research essay on alzheimer's disease

Advertisement

Diagnosis of Early Alzheimer’s Disease: Clinical Practice in 2021

• Open access
• Published: 09 June 2021
• Volume 8 , pages 371–386, ( 2021 )

Cite this article

You have full access to this open access article

• A. P. Porsteinsson 1 ,
• R. S. Isaacson 2 ,
• Sean Knox 3 ,
• M. N. Sabbagh 4 &
• I. Rubino 5  

82k Accesses

183 Citations

90 Altmetric

10 Mentions

Explore all metrics

Alzheimer’s disease is a progressive, irreversible neurodegenerative disease impacting cognition, function, and behavior. Alzheimer’s disease progresses along a continuum from preclinical disease, to mild cognitive and/or behavioral impairment and then Alzheimer’s disease dementia. Recently, clinicians have been encouraged to diagnose Alzheimer’s earlier, before patients have progressed to Alzheimer’s disease dementia. The early and accurate detection of Alzheimer’s disease-associated symptoms and underlying disease pathology by clinicians is fundamental for the screening, diagnosis, and subsequent management of Alzheimer’s disease patients. It also enables patients and their caregivers to plan for the future and make appropriate lifestyle changes that could help maintain their quality of life for longer. Unfortunately, detecting early-stage Alzheimer’s disease in clinical practice can be challenging and is hindered by several barriers including constraints on clinicians’ time, difficulty accurately diagnosing Alzheimer’s pathology, and that patients and healthcare providers often dismiss symptoms as part of the normal aging process. As the prevalence of this disease continues to grow, the current model for Alzheimer’s disease diagnosis and patient management will need to evolve to integrate care across clinical disciplines and the disease continuum, beginning with primary care. This review summarizes the importance of establishing an early diagnosis of Alzheimer’s disease, related practical ‘how-to’ guidance and considerations, and tools that can be used by healthcare providers throughout the diagnostic journey.

Similar content being viewed by others

Early Detection and Treatment of Patients with Alzheimer’s Disease: Future Perspectives

On the path to 2025: understanding the alzheimer’s disease continuum.

Alzheimer’s Disease: Clinical Aspects and Treatments

Avoid common mistakes on your manuscript.

Introduction

D ementia is among the greatest global health crises of the 21st century. Currently, more than 50 million people are living with dementia worldwide ( 1 ), with this number estimated to triple to 152 million by 2050 as the world’s population grows older ( 2 ). Alzheimer’s disease (AD) is the most common cause of dementia and is thought to account for 60–80% of dementia cases ( 3 ). Currently, the total annual cost for AD and other dementias in the USA is $305 billion and is predicted to increase to more than $1.1 trillion by 2050 ( 3 ). This substantial economic burden includes not only healthcare and hospice support for patients with AD ( 3 ) but also lost productivity from patients and caregivers ( 4 ).

AD is a progressive, neurodegenerative disease associated with cognitive, functional, and behavioral impairments, and characterized by two underlying pathological hallmarks: the progressive accumulation of extracellular amyloid beta (Aβ) plaques and intracellular neurofibrillary tangles (NFTs) ( 3 ). In AD, aggregated Aβ plaques are deposited within the brain as a result of either reduced Aβ clearance or excessive production ( 5 ); plaque deposition typically occurs ∼20 years before the onset of cognitive impairment ( 6 , 7 ). NFTs are formed by the abnormal accumulation of hyperphosphorylated-tau protein ( 5 ); these can be detected 10–15 years before the onset of symptoms ( 6 , 7 ).

AD follows a progressive disease continuum that extends from an asymptomatic phase with biomarker evidence of AD (preclinical AD), through minor cognitive (mild cognitive impairment [MCI]) and/or neurobehavioral (mild behavioral impairment [MBI]) changes to, ultimately, AD dementia. A number of staging systems have been developed to categorize AD across this continuum ( 7 – 9 ). While these systems vary in terms of how each stage is defined, all encompass the presence/absence of pathologic Aβ and NFTs, as well as deficits in cognition, function, and behavior ( 7 – 9 ). As a result, subtle but important differences exist in the nomenclature for each stage of AD depending on the selected clinical and research classifications (Figure 1 ).

Stages within the Alzheimer’s disease continuum

The AD continuum can be classified into different stages from preclinical AD to severe AD dementia; the nomenclature associated with each stage varies between the different clinical and research classifications. This figure provides a summary of the different naming conventions that are used within the AD community and the symptoms associated with each stage of the continuum; *Mild behavioral impairment is a construct that describes the emergence of sustained and impactful neuropsychiatric symptoms that may occur in patients ≥50 years old prior to cognitive decline and dementia ( 112 ); Abbreviations: Aβ, amyloid beta. AD, Alzheimer’s disease. FDA, Food and Drug Administration. IWG, International Working Group. MCI, mild cognitive impairment. NIA-AA, National Institute on Aging—Alzheimer’s Association

Preclinical AD, as the earliest stage in the AD continuum, comprises a long asymptomatic phase, in which individuals have evidence of AD pathology but no evidence of cognitive or functional decline, and their daily life is unaffected ( 8 ) (Figure 1 ). The duration of preclinical AD can vary between individuals, but typically lasts 6–10 years depending on the age of onset ( 10 , 11 ). The risk of progression from preclinical AD to MCI due to AD (with/without MBI) depends on a number of factors, including age, sex, and apolipoprotein E (ApoE) status ( 11 , 12 ); however, not all individuals who have underlying AD pathology will go on to develop MCI or AD dementia ( 13 , 14 ). A recent meta-analysis of six longitudinal cohorts followed up for an average of 3.8 years found that 20% of patients with preclinical AD progressed to MCI due to AD ( 11 ). A further study by Cho et al., with an average follow-up rate of 4 years, found that 29.1% of patients with preclinical AD progressed to MCI due to AD ( 12 ).

For patients who do progress to MCI due to AD (with/without MBI), initial clinical symptoms typically include short-term memory impairment, followed by subsequent decline in additional cognitive domains ( 15 ) (Figure 1 ). On a day-to-day basis, an individual with MCI due to AD may struggle to find the right word (language), forget recent conversations (episodic memory), struggle with completing familiar tasks (executive function), or get lost in familiar surroundings (visuospatial function) ( 15 , 16 ). As individuals have varying coping mechanisms and levels of cognitive reserve, patients’ experiences and symptomology vary widely; however, patients tend to remain relatively independent at this stage, despite potential marginal deficits in function. The prognosis for patients with MCI due to AD can be uncertain; one study that followed up patients with MCI due to AD for an average of 4 years found that 43.4% progressed to AD dementia ( 12 ). Other studies reported 32.7% and 70.0% of individuals with MCI due to AD progress to AD dementia within 3.2 and 3.6 years of follow-up, respectively ( 17 , 18 ). Patients who do progress to AD dementia will develop severe cognitive deficits that interfere with social functioning and will require assistance with activities of daily living ( 7 ) (Figure 1 ). As the disease progresses further, increasingly severe behavioral symptoms will develop that significantly burden patients and their caregivers, and the disease ultimately results in severe loss of independence and the need for round-the-clock care ( 3 ).

An early diagnosis of AD can provide patients the opportunity to collaborate in the development of advanced care plans with their family, caregivers, clinicians, and other members of the wider support team. Importantly, it also enables patients to seek early intervention with symptomatic treatment, lifestyle changes to maintain quality of life, and risk-reduction strategies that can provide clinically meaningful reductions in cognitive, functional, and behavioral decline ( 19 – 22 ). It can also help reduce healthcare system costs and constraints: a study by the Alzheimer’s Association found that diagnosing AD in the early stages could save approximately $7 trillion. These savings were due to lower medical and long-term care costs for patients with managed MCI than for those with unmanaged MCI and dementia ( 3 ). Furthermore, an early diagnosis will be vital for patients when a therapy addressing the underlying pathology of AD becomes available; currently 19 biologic compounds are under Phase 2 or 3 investigation ( 23 ). Physicians will need to be prepared for the approval of these treatments, to optimize the potential benefit and prolong preservation of patients’ cognitive function and independence beyond that associated with current standard of care ( 19 ).

As the prevalence of AD continues to grow, the advancement of AD patient diagnosis will require an orchestrated effort, starting in the primary care setting and subsequently involving multiple healthcare provider (HCP) specialties (e.g., nurse practitioner [NP] or physician assistant [PA]) throughout the disease continuum. Galvin et al. recently highlighted the need for HCPs to work as an integrated, patient-centered care team to accommodate the growing and diverse population of patients with AD, beginning with diagnosis ( 24 ). For patients to receive a timely diagnosis, it is vital to implement an approach that minimizes the burden placed on the patient, clinician, and healthcare system ( 25 ). Here, we summarize the importance of establishing an early diagnosis of AD, related practical ‘how-to’ guidance and considerations, and tools that can be used by healthcare providers throughout the diagnostic journey.

The importance of an early diagnosis

Historically, a diagnosis of AD has been one of exclusion, and one only made in the latter stages of disease ( 26 ); however, the disease process can take years to play out, exacting a significant toll on the patient, caregiver, and healthcare system along the way ( 27 ).

To mitigate this burden, the early and accurate detection of AD-associated symptoms in clinical practice represents a critically needed but challenging advancement in AD care ( 19 , 28 – 30 ). Usually, a patient with early signs/symptoms of AD will initially present in a primary care setting ( 30 ). For some patients, minor changes in cognition and/or behavior may be detected during a routine wellness visit or an appointment to discuss other comorbidities ( 24 ). As the PCP is often the first to observe a patient’s initial symptomatology, it is vital they recognize the early signs and symptoms, and understand how to use the most appropriate assessment tools designed to detect these early clinical effects of the disease.

Because the neuropathologic hallmarks of AD (Aβ plaques and NFTs) can be detected decades prior to the onset of symptoms ( 6 , 7 ), biomarkers reflecting this underlying pathology represent an important opportunity for early identification of patients at greatest risk of developing MCI due to AD. Biomarkers support the diagnosis of AD (especially important early on when symptoms can be subtle), and the U.S. Food and Drug Administration (FDA) has recently published guidelines that endorse their use in this population ( 9 ). The National Institute on Aging—Alzheimer’s Association (NIA-AA) has recently created a research framework that acknowledges the use of biomarkers for diagnosing AD in vivo and monitoring disease progression ( 7 ).

Important biomarker information can be gathered from imaging modalities such as magnetic resonance imaging (MRI) and positive emission tomography (PET) that visualize early structural and molecular changes in the brain, respectively ( 25 , 30 ). Fluid biomarker testing, such as cerebrospinal fluid (CSF) can also be used; CSF biomarkers can directly reflect the presence of Aβ and aggregated tau within the brain ( 7 , 31 ). As will be discussed in more depth later in this article, a large number of clinical studies have shown that Aβ and tau biomarkers can contribute diagnostically important information in the early stages of disease ( 32 ). There is ongoing research to expand the current range of tests that can be used by clinicians as part of the multistage diagnostic process ( 25 ). For instance, once approved, blood-based biomarkers could be used to identify patients at risk of developing AD and for monitoring disease progression ( 33 , 34 ), which would also reduce the current capacity constraints associated with PET imaging ( 25 ).

Practical guide for an early diagnosis of Alzheimer’s disease in clinical practice

As already raised, recent recommendations for evolving AD care to a more patient-centric, transdisciplinary model include guidance on realizing an efficient diagnostic process—one in which HCPs, payers, and specialists are encouraged to combine their efforts to ensure the early warning signs of AD are not overlooked ( 24 ). The recommendations include dividing the diagnosis of AD into the following steps: detect, assess/differentiate, diagnose, and treat (Figure 2 ). We present here a practical guide for the early diagnosis of AD, based on this outlined approach, including a case study to highlight each of these key steps.

A stepwise infographic to highlight key stages within the diagnostic process, along with the recommended tests to support each step

The diagnostic process for AD can be divided into the following steps: detect, assess/differentiate, diagnose, and treat. It is important for clinicians to utilize appropriate tests when investigating a patient suspected of having AD in the early stages. Here, we highlight the most valuable tests for each step and which ones should be used in a primary care or specialist setting; *FDG-PET is usually considered after a diagnostic work-up; Abbreviations: A-IADL-Q, Amsterdam Instrumental Activities of Daily Living Questionnaire. Aβ, amyloid beta. Ach, acetylcholine. BG, blood glucose. CSF, cerebrospinal fluid. FAQ, Functional Activities Questionnaire. FAST, Functional Analysis Screening Tool. FDG-PET, fluorodeoxyglucose-PET. GDS, Geriatric Depression Scale. IQCODE, Informant Questionnaire on Cognitive Decline in the Elderly. Mini-Cog, Mini Cognitive Assessment Instrument. MMSE, Mini-Mental State Examination. MoCA, Montreal Cognitive Assessment. MRI, magnetic resonance imaging. NMDA, N-Methyl-D-aspartic acid. NPI-Q, Neuropsychiatric Inventory Questionnaire. PCP, primary care physician. PET, positive emission tomography. p-tau, phosphorylated tau. QDRS, Quick Dementia Rating System. TSH, thyroid-stimulating hormone. t-tau, total tau depressive symptoms and anxiety, as well as irritability. Based on the patient’s symptoms, the PCP felt his presentation warranted further clinical assessment.

Step 1: Detect

The role of primary care in the early detection of ad.

The insidious and variable emergence of symptoms associated with AD and other dementias can make recognition extremely challenging, particularly in a primary care setting ( 30 , 35 ). Clinicians often have limited time with patients, so it is vital that they are able to quickly and accurately recognize the early signs and symptoms associated with AD ( Table 2 ) ( 3 , 30 , 36 ), and training for nurses, NPs, and PAs who may have more time to observe patients should provide substantial benefits. Although extremely variable, initial symptoms may include short-term memory loss or psychological concerns, including depressive symptoms and a loss of purpose ( 36 ).

Patients, family members, and even HCPs themselves may present barriers to the diagnosis of early-stage AD. Patients may hide their symptoms or even avoid making an appointment until their symptoms significantly affect their day-to-day life due to fear of the stigma associated with a diagnosis of AD ( 19 ). Additionally, patients, family members, and PCPs/HCPs may dismiss or misinterpret symptoms as simply part of the normal aging process ( 30 ). Retrieving information from a trusted family member or informant/caregiver is essential when trying to assess a patient for suspected AD, as this perspective can provide a more objective understanding of the daily routine, mood, and behavior of the patient, and how this may have changed over time ( 30 ). For patients presenting with even subtle symptoms associated with AD, it is important that the PCP/HCP conducts an initial assessment to confirm the presence of symptoms using a validated assessment for early-stage AD detection (Figure 2 ; Step 2: Assess/Differentiate).

Case study: Presentation

A 63-year-old Caucasian male (J.K.) presented to his PCP with short-term memory loss over the last 2 years ( Table 1A ). Accompanied by his wife, he acknowledged his job had been affected by issues with his short-term memory; however, he considered his memory similar to that of his peers. His wife reported that people at work had started to notice him struggling to keep up, and also that family had to remind him of his upcoming appointments. He admitted to having intermittent

Step 2: Assess and differentiate

Primary care: initial assessment when a patient presents.

When a patient initially presents with symptoms consistent with early stages of AD, a clinician must first conduct a comprehensive clinical assessment to rule out other potential non-AD causes of cognitive impairment (Figure 2 ). PCPs are well placed to conduct these initial assessments, as they may not require specialist input or hospital tests. During the initial assessment, the primary objective of the clinician should be to exclude possible reversible causes of cognitive impairment, such as depression, or vitamin, hormone, and electrolyte deficiencies ( 37 ). The initial assessment should include a thorough history to identify potential risk factors associated with AD, including a family history of AD or related dementias in first-degree relatives ( 31 , 38 ). Other known risk factors for AD that should be identified include age, female sex, ApoE ε4 status, physical inactivity, low education, diabetes, and obesity ( 3 ). It is also important to review for pre-existing medical conditions or prescribed medications that could be a cause of the patient’s cognitive impairment ( 36 ). Additionally, when conducting a thorough history, open-ended, probing questions should be directed to both the patient and the informant to ascertain how the patient’s cognition has changed over time and how the cognitive deficits affect their everyday activities; example questions for the initial assessment are detailed in Table 3 ( 30 ). Engaging with informants/caregivers is key to capturing additional information to help support all assessments. A routine differential diagnosis of AD begins with a detailed history, physical and neurologic examinations, and bloodwork analyses, followed by cognitive assessments and functional evaluation (Figure 2 ).

Primary care: Physical examination and blood analyses

A physical examination and blood tests can identify comorbid contributory medical conditions and reversible causes of cognitive impairment. A physical examination, including a mental status and neurological assessment, should be conducted to detect conditions such as depression and, for example, to look for signs such as issues with speaking or hearing as well as signs that could indicate a stroke ( 37 ). As part of the physical exam, a physician may ask the patient about diet and nutrition, review all medications (to see if these are the cause of any cognitive impairment, e.g. anti-cholinergics, analgesics, or sleep aids and anxiolytics), check blood pressure, temperature and pulse, and listen to the heart and lungs ( 36 , 39 ).

Blood tests can rule out potentially treatable illnesses as a cause of cognitive impairment, such as vitamin B 12 deficiency or thyroid disease ( 37 ). Suggested blood analyses include: 1) complete blood cell count; 2) blood glucose; 3) thyroid-stimulating hormone; 4) serum B 12 and folate; 5) serum electrolytes; 6) liver function; and 7) renal function tests ( 30 ). Although not routinely used in clinical practice, clinicians may request ApoE genotyping, as this can help assess the genetic risk of developing AD. ApoE is the dominant cholesterol carrier within the brain that supports lipid transport and injury repair ( 40 , 41 ), and the APOE gene exists as three polymorphic alleles: APOE ε2, ε3, and ε4. The ε4 allele of ApoE is associated with increased AD risk, whereas the ε2 allele is protective ( 40 , 42 ). The number of ApoE ε4 alleles a person carries increases their risk of developing AD and the age of disease onset ( 43 ). Homozygous ε4 carriers (those with two copies of the ε4 allele) have the greatest risk of developing AD and the lowest average age of onset ( 43 ). In some practice settings, ApoE genotyping can only be conducted by a genetic counselor; a referral for more comprehensive genetic testing may be considered by the HCP if there is a family history of early-onset AD or dementia. Consumer tests are also becoming more readily available for patients wanting to determine their risk of developing diseases such as AD based on genetic risk factors ( 44 ).

Primary care: Cognitive, functional, and behavioral assessments

Cognitive assessments.

If a patient is suspected of having AD following an initial assessment in primary care, and they are <65 years old, or if the case is complex, a referral to a dementia specialist such as a neurologist, geriatrician, or geriatric psychiatrist may be required for further evaluation. The specialist would then use an appropriate battery of cognitive, functional, and behavioral tests to assess the different aspects of disease, and ultimately to confirm diagnosis. However, not all patients with suspected cognitive deficits are immediately referred to a dementia specialist at this stage, which is only partly due to limited numbers of specialists ( 25 ) (Figure 2 ). In clinical practice, a two-stage process is often employed. This involves an initial ‘triage’ step conducted by non-specialists to clinically assess and select those patients who require further evaluation by a dementia specialist ( 45 ). During this ‘triage’ step, there are several clinical assessments available to non-specialists for assessing the presence of cognitive and functional impairments and behavioral symptoms (Table 4 ) ( 28 , 35 , 46 – 55 ).

Previous research has shown that clinicians have a tendency to choose one assessment over another due to their familiarity with the assessment, time constraints, or specific resources available to them within their clinic ( 30 ), but clinicians need to be aware of, and prepared to use, the most patient-appropriate assessments: the cultural, educational, and linguistic needs of the patient are important considerations ( 30 , 36 , 56 – 58 ). Some assessments have been translated into different languages or shortened, or have education-adjusted scoring classifications, where required ( 56 – 58 ).

Cognitive assessments that can be conducted quickly (<10 minutes), such as the Mini-Mental State Examination (MMSE) or Montreal Cognitive Assessment (MoCA), can be used by non-specialists to identify the presence and severity of cognitive impairment in patients before referring to a dementia specialist ( Table 4 ) ( 36 ). Both the MMSE and MoCA are used globally in clinical practice, particularly in primary care, but vary in terms of their sensitivity to identify AD in the early stages ( 28 , 59 ). The MMSE is sensitive and reliable for identifying memory and language deficits in general but has limitations in identifying impairments in executive functioning ( 59 ). MoCA was originally developed to improve the detection of MCI ( 28 ) and is more sensitive than the MMSE in its assessment of memory, visuospatial, executive, and language function, and orientation to time and place ( 59 ). Both tests are relatively easy to administer and take around 10 minutes to complete. Neither assessment requires extensive training by the clinician, although MoCA users do need to undergo a 1-hour certification as mandated by the MoCA Clinic and Institute ( 28 , 60 ).

For time-constrained clinicians, the Mini Cognitive Assessment Instrument (Mini-Cog) may be an appropriate tool to assess cognitive deficits that focus on memory, and components of visuospatial and executive function ( Table 4 ). The assessment includes the individual learning three items from a list, drawing a clock, and then recalling the three-item list. The Mini-Cog can be useful for clinicians in primary care, as it requires no training and the results are easy to interpret. As an alternative to these tests, PCPs might also consider using an informant-based structured questionnaire such as the AD8 or Informant Questionnaire on Cognitive Decline in the Elderly to help guide discussions with the patient and caregiver ( Table 4 ) ( 28 ).

Functional assessments

Functional assessments are valuable in identifying changes in a patient’s day-to-day functioning through the evaluation of their instrumental activities of daily living (IADLs). IADLs are complex activities that are necessary for the individual to function independently (e.g., cooking, shopping, and managing finances) and can be impaired during the early stages of cognitive impairment. While it is possible that functional decline may occur as a part of normal aging, a decline in a person’s IADL performance is strongly associated with neurodegenerative diseases such as AD ( 61 ). In the early stages of AD, patients may be functionally independent, and any impairment in IADLs may be subtle, such as difficulties paying bills or driving to new places. A patient’s functional independence is essential for their well-being and mental health ( 62 ), particularly in the early stages of the disease when the individual may still be working and socializing relatively independently ( 3 ). Consequently, functional independence is one of the most important clinical features for patients with AD. As the disease progresses, and patients have increasing functional impairment, this significantly impacts on their independence, and subsequently their and their family/caregiver’s quality of life.

Functional assessment is, therefore, an integral part of the diagnostic process for AD. The Functional Activities Questionnaire (FAQ) is an informant questionnaire that assesses the patient’s performance over a 4-week period and may take only a few minutes to complete ( Table 4 ). The questionnaire is scored from ‘normal’ to ‘dependent’, using numerical values assigned to categories, with higher scores indicative of increasing impairment ( 47 ). Previous research has shown that the FAQ has high sensitivity and reliability for detecting mild functional impairment in patients with MCI ( 47 ).

Determining an individual’s functional independence can be challenging and the clinician may require additional input from an informant to determine a patient’s functional decline and their ongoing ability to conduct activities of daily living ( 37 ). The clinician can gain greater insight through the informant into the patient’s day-to-day life and any issues the patient is having at home. This type of information is vital to the clinician, and when combined with other assessment tools, can help to narrow the differential diagnosis.

Behavioral assessments

Patients with suspected AD may experience several behavioral symptoms such as anxiety, disinhibition, apathy, and depression ( Table 2 ). In the early stages of disease, such symptoms are generally associated with poor long-term outcomes and caregiver burden, and are particularly distressing to both patients and their families ( 63 ). It is important for clinicians to use appropriate assessments to identify behavioral and psychiatric symptoms that are caused by neurodegenerative diseases, such as AD, rather than by alternative causes, such as a mood disorder.

The Geriatric Depression Scale (GDS) and Neuropsychiatric Inventory Questionnaire (NPI-Q) can be used by clinicians to assess neuropsychiatric symptoms in patients for whom early-stage AD is suspected ( Table 4 ). The GDS is a 15-item (or longer 30-item) questionnaire that assesses mood, has good reliability in older populations for detecting depression, and can be completed by the patient within 5–10 minutes ( 63 ). The NPI-Q can be used in conjunction with or as an alternative to the GDS. The NPI-Q is completed by a knowledgeable informant or caregiver who can report on the patient’s neuropsychiatric symptoms. The NPI-Q can be conducted in around 5 minutes to determine both the presence and severity of symptoms across several neuropsychiatric domains including depression, apathy, irritability, and disinhibition ( 49 ). Consequently, as it assesses depression, it can be used as an alternative to GDS if time constraints do not allow for both to be completed.

Behavioral symptoms can be non-specific, so it is important for clinicians to consider and rule out other potentially treatable causes of impairment when assessing this domain. For example, depression is associated with concentration and memory issues ( 64 ); apathy can occur in non-depressed elderly individuals and can impact cognitive function ( 65 ). Signs/symptoms such as social withdrawal, feelings of helplessness, or loss of purpose should be investigated closely, as these could be indicative of depression alone. It is important for clinicians to recognize that if changes over time in cognitive symptoms and mood symptoms match, then depression is most likely to be the root cause of subtle cognitive decline, rather than AD ( 28 ).

Primary care clinician checklist

If AD is still suspected following clinical assessment, referral to a specialist for further diagnostic testing, including imaging and fluid biomarkers, may be required. It is important the clinician confirms the following checks/assessments before the patient undergoes further evaluation:

Confirm medical and family history

Review the patient’s medications for any that could cause cognitive impairment

Perform blood tests to eliminate potential reversible causes of cognitive impairment

Conduct a quick clinical assessment to confirm the presence of cognitive impairment

Specialist role in assessment

Following the initial assessment in primary care, further cognitive, behavioral, functional, and imaging assessments can be carried out in a specialist setting. With their additional AD experience, access to other specialties, and possibly fewer time constraints than the PCP, the specialist is able to conduct a more comprehensive testing battery, using additional clinical assessments and biomarkers to determine causes of impairment and confirm diagnosis (Figure 2 ).

Because the cognitive impacts of early-stage AD may vary from patient to patient, it is important to consider which cognitive domains are affected in these early stages when considering which assessments to use. Specialists are able to conduct a full neuropsychological test battery that covers the major cognitive domains (executive function, social cognition/emotions, language, attention/concentration, visuospatial and motor function, learning and memory); preferably, a battery should contain more than one test per domain to ensure adequate sensitivity in capturing cognitive impairment ( 66 ). This step can help with obtaining an in-depth understanding of the subtle changes in cognition seen in the early stages of AD and enables the clinician to monitor subsequent changes over time.

Typically, episodic memory, executive function, visuospatial function, and language are the most affected cognitive domains in the early stages of AD ( 29 , 67 , 68 ). Currently, most cognitive assessment tools focus on a subset of the overall dimensions of cognition; it is therefore vital the clinician chooses the correct test to assess impairment in these specific cognitive domains that could be indicative of AD in the early stages. As cognitive impairment in the early stages of AD can be subtle and vary significantly between individuals ( 29 ), clinicians must choose appropriate, sensitive tests that can detect these changes and account for a patient’s level of activity and cognitive reserve ( 29 ). If there is large disparity in results across cognitive assessments, it is important for the clinician to shape their assessments based on the patient’s history. If the patient’s history is positive for neurodegenerative disease, but one assessment does not reflect this, it is important to conduct further tests to ascertain the cause of the cognitive impairment.

The Quick Dementia Rating System (QDRS) can be used by specialists to assess cognitive impairment ( Table 4 ). This short questionnaire (<5 minutes) is completed by a caregiver/informant and requires no training. The QDRS assesses several cognitive domains known to be affected by AD, including memory, language and communication abilities, and attention. The questionnaire can reliably discriminate between individuals with and without cognitive impairment and provides accurate staging for disease severity ( 28 ).

The Amsterdam IADL Questionnaire (A-IADL-Q) and Functional Assessment Screening Tool (FAST) can both be used to assess a patient’s functional ability ( Table 4 ) ( 53 ). The A-IADL-Q is a reliable computerized questionnaire that monitors a patient’s cognition, memory, and executive functioning over time. This questionnaire is completed by an informant of the patient and takes 10 minutes to complete ( 53 ). For patients with suspected early stage AD, the A-IADL-Q is a useful tool to monitor subtle changes in IADL independence over time and is less influenced by education, gender, and age than other functional assessments ( 53 ). The FAST is a useful assessment for clinicians to identify the occurrence of functional and behavioral problems in patients with suspected AD. The questionnaire is completed by informants who interact with the patient regularly; informants are required to answer Yes/No to a number of questions focusing on social and non-social scenarios ( 55 ).

Structural imaging

Structural imaging, such as MRI, provides clinically useful information when investigating causes of cognitive impairment ( 69 ) (Figure 2 ). MRI is routinely conducted to exclude alternative causes of cognitive impairment, rather than support a diagnosis of AD ( 37 , 70 ). It is well known that medial temporal lobe atrophy is the best MRI marker for identifying patients in the earliest stages of AD ( 70 , 71 ); however, specific patterns of atrophy may also be indicative of other neurodegenerative diseases. Atrophy alone is rarely sufficient to make a diagnosis. MRI findings can help to narrow the differential diagnosis, and the results should be considered in the context of the patient’s age and clinical examination ( 69 – 71 ).

Clinicians are advised to take a stepwise approach when reviewing structural imaging reports of a patient with suspected AD. These steps include: 1) excluding brain pathology that may be amenable to surgical intervention (e.g., the scan will show regions of hyper- or hypointensity rather than a uniform signal); 2) assessing for brain microbleeds (e.g., looking at signal changes within different areas of the brain can identify vascular comorbidities); and 3) assessing atrophy (e.g., medial temporal lobe atrophy is characteristic of AD) ( 69 ). Radiologists can conduct a quick and easy visual rating of any medial temporal lobe atrophy; these results can then be utilized by the specialist, in conjunction with a clinical assessment, to determine the likely cause of cognitive impairment. If the clinician is unable to determine a differential diagnosis, additional confirmatory tests can be requested.

Fluorodeoxyglucose-PET (FDG-PET) is a useful structural imaging biomarker that can support an early and differential diagnosis ( 72 ); however, specialists usually prefer to use this after their initial diagnostic work-up. As the brain relies almost exclusively on glucose as its source of energy, FDG (a glucose analog) can be combined with PET to identify regional patterns of reduced brain metabolism and neurodegeneration ( 70 , 72 ). FDG-PET is not recommended for diagnosing patients with preclinical AD, as there is no way to ascertain whether the hypometabolism is directly related to AD pathology ( 73 ); however, clinicians may refer patients with more established symptomatology for an FDG-PET scan to identify regions of glucose hypometabolism and neurodegeneration that could be indicative of AD ( 70 ).

Case study: Assess/differentiate

The initial assessment by the primary care clinician revealed that J.K.’s medical history was significant for hypertension, dyslipidemia, mild obesity, and glucose intolerance ( Table 1B ). There was no history of cerebrovascular events, significant head injuries, or focal findings on the neurologic exam. Besides the vascular risk factors, no medical conditions or current medications were found to be likely contributors to the cognitive deficit. The patient had a positive family history of dementia, where the onset typically occurred in the late 60s. Genotyping showed the patient to be a homozygous carrier of two ApoE ε4 alleles. Blood tests revealed elevated serum glucose and C-reactive protein but were otherwise normal. The patient had an unremarkable mental status examination, and his MoCA score was 21/30, with points lost on orientation, recall, and naming ( Table 1C ).

The patient was referred to a memory clinic for further assessment. The dementia specialist referred the patient for an MRI that predominantly showed mild small vessel disease and mild generalized atrophy with a significant reduction in hippocampal volume and ratio. Based on his medical and family history, cognitive assessments, and structural imaging results, the specialist deemed the severity of cognitive impairment to be in the mild range; consequently, the specialist referred the patient for biomarker assessment to determine the underlying cause.

Step 3: Diagnose

Historically, AD was only diagnosed postmortem until we developed the ability to ascertain the underlying pathology associated with the disease in new ways, namely imaging and fluid biomarkers. However, despite supportive results from single- and multicenter trials, the use and reimbursement of imaging and fluid biomarkers to support the diagnosis of AD still vary considerably between countries ( 70 ).

Imaging biomarkers

Recent advances have allowed physicians to visualize the proteins associated with AD, namely Aβ and tau, via PET scanning. Amyloid PET is currently the only imaging approach recommended by the Alzheimer’s Association and the Amyloid Imaging Task Force to support the diagnosis of AD ( 70 ). Amyloid PET utilizes tracers (florbetapir, flutemetamol, and florbetaben) that specifically bind to Aβ within amyloid plaques; a positive amyloid PET scan will show increased cortical retention of the tracer in regions of Aβ deposition within the brain ( 74 ), thus confirming the presence of Aβ plaques in the brain ( 74 , 75 ) and directly quantify brain amyloid pathology ( 76 ), thus making it a useful tool to supplement a clinical battery to diagnose AD ( 3 , 74 ). However, a positive amyloid PET scan alone does not definitively diagnose clinical AD, and these results must be combined with other clinical assessments, such as cognitive assessment, for an accurate diagnosis ( 74 ). It is also important to note that amyloid PET is expensive and not readily reimbursed by health insurance providers ( 70 ); if it is not possible to access amyloid PET, biomarker confirmation can be assessed using CSF.

Fluid biomarkers

An additional or alternative tool to amyloid PET is the collection and analysis of CSF for the presence of biomarkers associated with AD pathology. Patients who have symptoms suggestive of AD can be referred for a lumbar puncture to analyze their CSF for specific AD-associated biomarkers ( 3 ). CSF biomarkers are measures of the concentrations of proteins in CSF from the lumbar sac that reflect the rates of both protein production and clearance at a given timepoint ( 7 ). Lumbar punctures can be conducted safely and routinely in an outpatient setting or memory clinic ( 77 ). However, many patients still worry about the pain and possible side effects associated with the procedure and may require additional information and support from the clinician to undertake the procedure ( 77 ). Appropriate use criteria are available for HCPs to help identify suitable patients for lumbar puncture and CSF testing ( 78 ). For example, individuals presenting with persistent, progressing, and unexplained MCI, or those with symptoms suggestive of possible AD, should be referred for lumbar puncture and CSF testing ( 78 ). However, lumbar puncture and CSF testing are not recommended for determining disease severity in patients who have already received a diagnosis of AD or in lieu of genotyping for suspected autosomal dominant mutation carriers ( 78 ).

Because there is strong concordance between CSF biomarkers and amyloid PET, either can be used to confirm Aβ burden ( 79 ). As such, CSF biomarkers are widely accepted within the AD community to support a diagnosis ( 80 ). AD biomarkers from the brain can be detected in CSF well before the onset of overt clinical symptoms in early-stage AD ( 6 , 7 ). Core AD CSF biomarkers, such as Aβ42 (one of two main isoforms of Aβ and a major constituent of Aβ plaques) and phosphorylated tau (p-tau) and total tau (t-tau), can be measured to determine the presence of disease ( 80 ).

When interpreting CSF analyses for a patient with suspected AD, it is important to remember that AD is associated with decreased CSF Aβ42 and increased tau isoforms ( 32 ). Decreased CSF Aβ42 levels are a reflection of increased Aβ aggregation and deposition within the brain ( 32 ), and the concentration of CSF Aβ42 directly relates to the patient’s amyloid status (e.g., the presence or absence of significant amyloid pathology) and the total amount of Aβ peptides (e.g., Aβ42 and Aβ40) ( 32 ). Specialists’ use of ratios of these CSF biomarkers (e.g., Aβ42/40) rather than single CSF biomarkers alone has been shown to adjust for potential differences in Aβ production and provide a better index of the patient’s underlying amyloid-related pathology ( 81 ). The increase in CSF p-tau and t-tau associated with AD may directly reflect the aggregation of tau within the brain and neurodegeneration, respectively ( 32 ). P-tau in CSF provides a direct measure of the amount of hyperphosphorylated tau in the brain, which is strongly suggestive of the presence of NFTs, whereas CSF t-tau can predict the level of neurodegeneration in a patient with suspected AD; however, t-tau is also increased in other neurologic conditions ( 32 ).

Ultimately, the clinical decision to use amyloid PET or CSF to confirm amyloid and tau pathology can be affected by several practical factors (Table 5 ) ( 70 , 77 , 80 , 82 – 85 ).

Emerging diagnostic tools

Access constraints for amyloid PET have driven the need for alternative sensitive and specific CSF and blood-based biomarkers that can detect AD-associated pathology in the early stages ( 86 ). Significant efforts have been undertaken over the last decade to identify blood-based biomarkers to: 1) detect AD pathology; 2) identify those at risk of developing AD in the future; and 3) monitor disease progression ( 33 , 34 , 87 ). At present, only a limited number of approved blood-based assays are available to clinicians to detect AD pathology ( 88 ); however, several novel assays are currently under investigation, including those measuring various phosphorylated forms of tau, including p-tau181 and p-tau217 ( 89 ). Investigational use of plasma p-tau181 (an isoform of tau) has been shown to differentiate AD from other neurodegenerative diseases and predict cognitive decline in patients with AD ( 33 ). CSF p-tau217 (a different isoform of tau) is a promising biomarker under investigation for detecting preclinical and advanced AD ( 86 , 90 ). Given that blood testing is already a well-established part of clinical routines globally and can easily be performed in a variety of clinical settings, blood-based biomarkers could in future serve as the potential first step of a multistage diagnostic process. This would be a benefit to clinicians, particularly those in primary care, by helping to identify individuals requiring a referral to a specialist for diagnostic testing ( 87 ).

Case study: Diagnose

J.K. underwent a lumbar puncture for CSF analysis, which showed decreased Aβ42 and increased p-tau and t-tau protein ( Table 1D ). Based on the results from the genotyping, cognitive assessments, MRI, and CSF biomarkers, the clinician confirmed that the likely cause of the patient’s cognitive deficits was early-stage AD, especially in view of a positive family history of dementia with similar age of onset.

Step 4: Treat

The role of the clinician following a diagnosis of early-stage AD is to discuss the available management and treatment options while providing emotional and practical support to the patient, caregiver, and family where appropriate ( 37 ). Clinicians can also refer the patient and their caregiver(s) to social services for further support, as well as help connect them with reliable sources of information and even local research opportunities and clinical trials.

One important role for a clinician treating a patient diagnosed with early-stage AD is to closely monitor the patient’s disease progression through regular follow-up appointments (e.g., every 6–12 months); clinicians should encourage patients (and the caregiver) to make additional follow-up appointments, especially should symptoms worsen. Routine cognitive and functional assessments (Table 4 ) should be used to monitor disease progression; these tools can be used to identify unexpected trends, such as rapid decline, which could prompt the need for additional medical evaluation such as blood tests, imaging, or biomarker analyses. Results from such tests could help guide management and/or treatment decisions over the course of the patient’s disease.

Non-pharmacologic therapies (e.g., diet and exercise) may be employed for patients with early AD, with the goal to maintain or even improve cognitive function and retain their ability to perform activities of daily living. For patients in the early stages of disease, dietary changes (e.g., following a healthy diet high in green, leafy vegetables, fish, nuts, and berries), physical exercise, and cognitive training have demonstrated small but significant improvements in cognition ( 36 , 91 ). Nonpharmacologic therapies can have a positive impact on quality of life and are generally safe and inexpensive ( 36 ); however, compliance with these non-pharmacologic therapies should be monitored by the clinician. Research suggests that multimodal therapies, such as cognitive stimulation therapy, may also be more effective when used in combination with pharmacologic treatments ( 91 ).

Several pharmacologic treatments have received regulatory approval to treat the symptoms of mild to severe AD dementia. Acetylcholinesterase inhibitors (donepezil, rivastigmine, and galantamine) and N-methyl-D-aspartate receptor antagonists (memantine) can be prescribed to patients to temporarily ameliorate the symptoms of AD dementia such as cognitive and functional decline ( 92 – 96 ). Meta-analyses of donepezil, rivastigmine, and galantamine have shown that patients with mild-to-moderate AD dementia experience some benefits in cognitive function, activities of daily living, and clinician-rated global clinical state ( 93 , 94 , 97 ). Furthermore, treatment with acetylcholinesterase inhibitors and/or memantine has also been shown to modestly improve measures of global function and temporarily stabilize measures of activities of daily living ( 96 ). However, it is important to note that these drugs provide only temporary, symptomatic benefit and that not all patients respond to treatment ( 36 , 98 ). Critically, none of the current drugs available address the underlying pathophysiology or alter the ultimate disease course.

Following AD diagnosis, a comprehensive approach toward clinical care can be individualized based on the patient’s specific AD risk factors ( 20 , 21 ). Clinicians should consider managing uncontrolled vascular risk factors (e.g., hypertension, hyperlipidemia, diabetes) with antithrombotics, antihypertensives, lipid-lowering, and/or antidiabetic agents, respectively, to reduce the risk of cerebrovascular ischemia and stroke, and subsequent cognitive decline ( 36 , 99 ). They should also consider the management of the patient’s behavioral symptoms. For most patients in the early stages of disease, behavioral symptoms will be relatively mild, and no pharmacologic management is required; however, pharmacologic treatment, such as a low-dose selective serotonin reuptake inhibitor, can be prescribed for patients with AD-associated depression and anxiety ( 100 , 101 ).

Specialist clinician checklist

The specialist’s role is critical to further evaluating the initial checks/assessments, providing the diagnosis, and developing the individualized patient management plan:

Identify deficits to specific cognitive domains using appropriate tests

Confirm functional performance, using patient and caregiver assessments

Perform structural imaging to complete assessment of the patient

Confirm diagnosis with imaging or fluid biomarkers

Develop a personalized management and follow-up plan

Direct the patient to additional support resources such as the Alzheimer’s Association

Case study: Treat

Following diagnosis, J.K. was advised on the available management options and research opportunities ( Table 1E ). The specialist emphasized the need to control his vascular risk factors and suggested lifestyle modifications to optimize the management of his other medical problems. The patient’s neuropsychiatric symptoms were considered mild and did not require pharmacologic intervention. The patient was also provided with details for a local social worker and directed toward further disease-specific information from the Alzheimer’s Association related to his disease. The patient was encouraged to return for additional follow-up visits so that his disease and associated symptoms could be appropriately monitored and managed.

Future perspectives

An early diagnosis of AD will become increasingly important as treatments that alter the underlying disease pathology become available—particularly given the expectation that such treatments will be more effective in preserving cognitive function, and thus prolonging independence, when given early in the course of the disease ( 19 ). The approval of such treatments will likely lead to an increased awareness of cognitive impairment and other AD-associated symptoms among both the public and non-specialists, such as those in primary care settings. This may encourage more patients/family members to seek help at an earlier stage of disease than is currently seen in community practice. Increased use of sensitive screening measures to proactively assess for the presence of AD symptoms will help identify patients suspected of having early AD. Assessment of cognitive impairment during a Medicare Annual Wellness Visit is inconsistent; the U.S. Preventative Services Task Force, whilst recognizing the importance of MCI, has maintained its decision that there is insufficient evidence to support the mandate of cognitive screening. However, sensitive screening procedures, along with the availability of disease-modifying treatments, are likely to change their recommendations. There is also a need for a mandated, standardized screening approach internationally. Together, this will result in an increase in patients requiring diagnosis, increasing the demand for specialists to evaluate and diagnose, the need for amyloid confirmation, and wait times for patients, which will collectively put further pressure on an already-stretched healthcare infrastructure ( 25 ).

Nevertheless, efforts continue within the AD field to streamline the diagnostic process. Planning for and implementing change will not only improve patient management now but also help prepare healthcare systems for an approved disease-modifying treatment for AD. A flexible, multidisciplinary team approach is recommended to integrate the care needed to detect, assess, differentiate, diagnose, treat, and monitor a diverse AD population ( 24 ). The development of tests that could be carried out routinely in a primary care setting, such as blood-based AD biomarkers, would help PCPs and non-specialists identify which patients may need further evaluation or referral to a specialist ( 25 ). Interest also remains high in advancing imaging techniques, such as amyloid and tau PET, to support a diagnosis of AD. Although amyloid and tau PET are not currently readily available, they may be useful for specialists in the future to determine disease staging or track progression, or as a surrogate marker of cognitive status ( 74 ). The introduction of new screening and diagnostic tools could ultimately help lower the burden on specialists and ensure patients are diagnosed in a timely manner.

Conclusions

Consensus within the AD community has recently shifted to encourage the diagnosis of AD as early as possible. This shift will enable patients to plan their future and consider symptomatic therapies and lifestyle changes that could reduce cognitive deficits and ultimately help preserve their quality of life. Promisingly, new, potentially disease-modifying therapeutic candidates are on the horizon that could be effective in early AD by targeting and ameliorating the underlying biological mechanisms ( 92 , 102 ). This paper has outlined a menu of practical tools for clinicians to use in the real world to support an early diagnosis of AD and how they may best be incorporated into current clinical practice. Ultimately, a coordinated, multidisciplinary approach that encompasses primary care and specialist expertise is required to ensure timely detection, assessment and differentiation, diagnosis, and management of patients with AD.

Hughes J. This is one of the biggest global health crises of the 21st century. World Econ Forum 2017. https://www.weforum.org/agenda/2017/09/dementia-trillion-dollar-global-crisis/#:~:text=This%20World%20Alzheimer’s%20Day%2C%20the,crises%20of%20the%2021st%20century (accessed May 18, 2020).

Alzheimer’s Disease International. World Alzheimer Report 2019. Attitudes to dementia 2019. https://www.alz.co.uk/research/WorldAlzheimerReport2019.pdf (accessed February 13, 2020).

Alzheimer’s Association. 2020 Alzheimer’s disease facts and figures. Alzheimers Dement 2020;16:391–460.

Article   Google Scholar  

Deb A, Thornton JD, Sambamoorthi U, Innes K. Direct and indirect cost of managing alzheimer’s disease and related dementias in the United States. Expert Rev Pharmacoecon Outcomes Res 2017;17:189–202.

Article   PubMed   PubMed Central   Google Scholar  

Serrano-Pozo A, Frosch MP, Masliah E, Hyman BT. Neuropathological alterations in Alzheimer disease. Cold Spring Harb Perspect Med 2011;1:a006189.

Article   PubMed   PubMed Central   CAS   Google Scholar  

Bateman RJ, Xiong C, Benzinger TL, et al. Clinical and biomarker changes in dominantly inherited Alzheimer’s disease. N Engl J Med 2012;367:795–804.

Article   CAS   PubMed   PubMed Central   Google Scholar  

Jack CR, Bennett DA, Blennow K, et al. NIA-AA Research Framework: toward a biological definition of Alzheimer’s disease. Alzheimers Dement 2018;14:535–62.

Dubois B, Feldman HH, Jacova C, et al. Revising the definition of Alzheimer’s disease: a new lexicon. Lancet Neurol 2010;9:1118–27.

Article   PubMed   Google Scholar  

U.S. Food and Drug Administration (FDA). Early Alzheimer’s disease: developing drugs for treatment guidance for industry 2018. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/alzheimers-disease-developing-drugs-treatment-guidance-industy (accessed August 11, 2020).

Insel PS, Weiner M, Mackin RS, et al. Determining clinically meaningful decline in preclinical Alzheimer disease. Neurology 2019;93:e322–33.

Vermunt L, Sikkes SA, van den Hout A, et al. Duration of preclinical, prodromal, and dementia stages of Alzheimer’s disease in relation to age, sex, and APOE genotype. Alzheimers Dement 2019;15:888–98.

Cho SH, Woo S, Kim C, et al. Disease progression modelling from preclinical Alzheimer’s disease (AD) to AD dementia. Sci Rep 2021;11:4168.

Knopman DS, Parisi JE, Salviati A, et al. Neuropathology of cognitively normal elderly. J Neuropathol Exp Neurol 2003;62:1087–95.

Article   CAS   PubMed   Google Scholar  

Bennett DA, Schneider JA, Arvanitakis Z, et al. Neuropathology of older persons without cognitive impairment from two community-based studies. Neurology 2006;66:1837–44.

Kazim SF, Iqbal K. Neurotrophic factor small-molecule mimetics mediated neuroregeneration and synaptic repair: emerging therapeutic modality for Alzheimer’s disease. Mol Neurodegener 2016;11:50.

Tolbert S, Liu Y, Hellegers C, et al. Financial management skills in aging, MCI and dementia: cross sectional relationship to 18F-florbetapir PET cortical β-amyloid deposition. J Prev Alzheimers Dis 2019;6:274–82.

CAS   PubMed   Google Scholar  

Ye BS, Kim HJ, Kim YJ, et al. Longitudinal outcomes of amyloid positive versus negative amnestic mild cognitive impairments: a three-year longitudinal study. Sci Rep 2018;8:5557.

Roberts RO, Aakre JA, Kremers WK, et al. Prevalence and outcomes of amyloid positivity among persons without dementia in a longitudinal, population-based setting. JAMA Neurol 2018;75:970–9.

Dubois B, Padovani A, Scheltens P, Rossi A, Dell’Agnello G. Timely diagnosis for Alzheimer’s disease: a literature review on benefits and challenges. J Alzheimers Dis 2016;49:617–31.

Isaacson RS, Ganzer CA, Hristov H, et al. The clinical practice of risk reduction for Alzheimer’s disease: a precision medicine approach. Alzheimers Dement 2018;14:1663–73.

Isaacson RS, Hristov H, Saif N, et al. Individualized clinical management of patients at risk for Alzheimer’s dementia. Alzheimers Dement 2019;15:1588–602.

Gauthier SG. Alzheimer’s disease: the benefits of early treatment. Eur J Neurol 2005;12:11–6.

Cummings J, Lee G, Ritter A, Sabbagh M, Zhong K. Alzheimer’s disease drug development pipeline: 2020. Alzheimers Dement NY 2020;6:e12050.

Google Scholar  

Galvin JE, Aisen P, Langbaum JB, et al. Early stages of Alzheimer’s disease: evolving the care team for optimal patient management. Front Neurol 2021;11:592302.

Liu JL, Hlavka JP, Hillestad R, Mattke S. Assessing the preparedness of the U.S. Health Care System infrastructure for an Alzheimer’s treatment 2017. https://www.rand.org/pubs/research_reports/RR2272.html (accessed May 5, 2018).

Sabbagh MN, Lue L-F, Fayard D, Shi J. Increasing precision of clinical diagnosis of Alzheimer’s disease using a combined algorithm incorporating clinical and novel biomarker data. Neurol Ther 2017;6:83–95.

Balasa M, Gelpi E, Antonell A, et al. Clinical features and APOE genotype of pathologically proven early-onset Alzheimer disease. Neurology 2011;76:1720–5.

Galvin JE. Using informant and performance screening methods to detect mild cognitive impairment and dementia. Curr Geriatr Rep 2018;7:19–25.

Sabbagh MN, Boada M, Borson S, et al. Early detection of mild cognitive impairment (MCI) in primary care. J Prev Alzheimers Dis 2020;7:165–70.

Galvin JE, Sadowsky CH, NINCDS-ADRDA. Practical guidelines for the recognition and diagnosis of dementia. J Am Board Fam Med 2012;25:367–82.

Aisen PS, Cummings J, Jack CR, et al. On the path to 2025: understanding the Alzheimer’s disease continuum. Alzheimers Res Ther 2017;9:60.

Blennow K, Zetterberg H. Biomarkers for Alzheimer disease - current status and prospects for the future. J Intern Med 2018;284:643–63.

Karikari TK, Pascoal TA, Ashton NJ, et al. Blood phosphorylated tau 181 as a biomarker for Alzheimer’s disease: a diagnostic performance and prediction modelling study using data from four prospective cohorts. Lancet Neurol 2020;19:422–33.

Janelidze S, Mattsson N, Palmqvist S, et al. Plasma P-tau181 in Alzheimer’s disease: relationship to other biomarkers, differential diagnosis, neuropathology and longitudinal progression to Alzheimer’s dementia. Nat Med 2020;26:379–86.

Iliffe S, Robinson L, Brayne C, et al. Primary care and dementia: 1. diagnosis, screening and disclosure. Int J Geriatr Psychiatry 2009;24:895–901.

Arvanitakis Z, Shah RC, Bennett DA. Diagnosis and management of dementia: review. JAMA 2019;322:1589–99.

Robinson L, Tang E, Taylor J-P. Dementia: timely diagnosis and early intervention. BMJ 2015;350:h3029.

Zucchella C, Bartolo M, Pasotti C, Chiapella L, Sinforiani E. Caregiver burden and coping in early-stage Alzheimer disease. Alzheimer Dis Assoc Disord 2012;26:55–60.

Pfistermeister B, Tümena T, Gaßmann K-G, Maas R, Fromm MF. Anticholinergic burden and cognitive function in a large German cohort of hospitalized geriatric patients. PLoS One 2017;12:e0171353.

Liu C-C, Liu C-C, Kanekiyo T, Xu H, Bu G. Apolipoprotein E and Alzheimer disease: risk, mechanisms, and therapy. Nat Rev Neurol 2013;9:106–18.

Sadigh-Eteghad S, Sabermarouf B, Majdi A, Talebi M, Farhoudi M, Mahmoudi J. Amyloid-beta: a crucial factor in Alzheimer’s disease. Med Princ Pract 2015;24:1–10.

Karch CM, Goate AM. Alzheimer’s disease risk genes and mechanisms of disease pathogenesis. Biol Psychiatry 2015;77:43–51.

Ungar L, Altmann A, Greicius MD. Apolipoprotein E, gender, and Alzheimer’s disease: an overlooked, but potent and promising interaction. Brain Imaging Behav 2014;8:262–73.

23andMe. What to know about health report test result - 23andMe. 23&Me® 2020. https://www.23andme.com/test-info/ (accessed October 15, 2020).

Hendry K, Green C, McShane R, et al. AD-8 for detection of dementia across a variety of healthcare settings. Cochrane Database Syst Rev 2019;3:CD011121.

PubMed   Google Scholar  

Patnode CD, Perdue LA, Rossom RC, et al. Screening for cognitive impairment in older adults: an evidence update for the U.S. Preventive Services Task Force. Rockville (MD): Agency for Healthcare Research and Quality (US); Report No.: 19-05257-EF-1, 2020.

Teng E, Becker BW, Woo E, Cummings JL, Lu PH. Subtle deficits in instrumental activities of daily living in subtypes of mild cognitive impairment. Dement Geriatr Cogn Disord 2010;30:189–97.

Marshall GA, Amariglio RE, Sperling RA, Rentz DM. Activities of daily living: where do they fit in the diagnosis of Alzheimer’s disease? Neurodegener Dis Manag 2012;2:483–91.

Rosenberg PB, Mielke MM, Appleby BS, Oh ES, Geda YE, Lyketsos CG. The association of neuropsychiatric symptoms in MCI with incident dementia and Alzheimer disease. Am J Geriatr Psychiatry 2013;21:685–95.

Ismail Z, Emeremni CA, Houck PR, et al. A comparison of the E-BEHAVE-AD, NBRS and NPI in quantifying clinical improvement in the treatment of agitation and psychosis associated with dementia. Am J Geriatr Psychiatry 2013;21:78–87.

Bowden VM, Bowden CL. The Journal of Neuropsychiatry and Clinical Neurosciences. JAMA 1992;268:1473–4.

Galvin JE. The quick dementia rating system (QDRS): a rapid dementia staging tool. Alzheimers Dement Amst 2015;1:249–59.

Koster N, Knol DL, Uitdehaag BM, Scheltens P, Sikkes SA. The sensitivity to change over time of the Amsterdam IADL Questionnaire(©). Alzheimers Dement 2015;11:1231–40.

Sikkes SA, Pijnenburg YA, Knol DL, de Lange-de Klerk ES, Scheltens P, Uitdehaag BM. Assessment of instrumental activities of daily living in dementia: diagnostic value of the Amsterdam Instrumental Activities of Daily Living Questionnaire. J Geriatr Psychiatry Neurol 2013;26:244–50.

LaRue RH. Functional Assessment Screening Tool (FAST). In: Volkmar FR, editor. Encycl. Autism Spectr. Disord., New York, NY: Springer New York, 2018:1–2.

Lindgren N, Rinne JO, Palviainen T, Kaprio J, Vuoksimaa E. Prevalence and correlates of dementia and mild cognitive impairment classified with different versions of the modified Telephone Interview for Cognitive Status (TICS-m). Int J Geriatr Psychiatry 2019;34:1883–91.

Wang H, Fan Z, Shi C, et al. Consensus statement on the neurocognitive outcomes for early detection of mild cognitive impairment and Alzheimer dementia from the Chinese Neuropsychological Normative (CN-NORM) Project. J Glob Health 2019;9:020320.

Franzen S, van den Berg E, Goudsmit M, et al. A systematic review of neuropsychological tests for the assessment of dementia in non-western, low-educated or illiterate populations. J Int Neuropsychol Soc 2020;26:331–51.

Costa A, Bak T, Caffarra P, et al. The need for harmonisation and innovation of neuropsychological assessment in neurodegenerative dementias in Europe: consensus document of the Joint Program for Neurodegenerative Diseases Working Group. Alzheimers Res Ther 2017;9:27.

MoCA Test Inc. Upcoming mandatory training for MoCA testing. MoCA Montr - Cogn Assess 2021. https://www.mocatest.org/mandatory-moca-test-training/ (accessed February 3, 2021).

Tabira T, Hotta M, Murata M, et al. Age-related changes in instrumental and basic activities of daily living impairment in older adults with very mild Alzheimer’s disease. Dement Geriatr Cogn Disord Extra 2020;10:27–37.

Martyr A, Nelis SM, Quinn C, et al. The relationship between perceived functional difficulties and the ability to live well with mild-to-moderate dementia: findings from the IDEAL programme. Int J Geriatr Psychiatry 2019;34:1251–61.

Ismail Z, Smith EE, Geda Y, et al. Neuropsychiatric symptoms as early manifestations of emergent dementia: provisional diagnostic criteria for mild behavioral impairment. Alzheimers Dement 2016;12:195–202.

McAllister-Williams RH, Bones K, Goodwin GM, et al. Analysing UK clinicians’ understanding of cognitive symptoms in major depression: a survey of primary care physicians and psychiatrists. J Affect Disord 2017;207:346–52.

Richard E, Schmand B, Eikelenboom P, Yang SC, Ligthart SA. Symptoms of apathy are associated with progression from mild cognitive impairment to Alzheimers disease in non-depressed subjects. Dement Geriatr Cogn Disord 2012;33:204–9.

Weintraub S, Besser L, Dodge HH, et al. Version 3 of the Alzheimer Disease Centers’ Neuropsychological Test Battery in the Uniform Data Set (UDS). Alzheimer Dis Assoc Disord 2018;32:10–7.

Harrison JE, Hendrix S. Chapter 21 - The assessment of cognition in translational medicine: a contrast between the approaches used in Alzheimer’s disease and major depressive disorder. In: Nomikos GG, Feltner DE, editors. Handb. Behav. Neurosci., vol. 29, Elsevier, 2019:297–308.

Roberts R, Knopman DS. Classification and epidemiology of MCI. Clin Geriatr Med 2013;29:753–72.

Harper L, Barkhof F, Scheltens P, Schott JM, Fox NC. An algorithmic approach to structural imaging in dementia. J Neurol Neurosurg Psychiatry 2014;85:692–8.

Frisoni GB, Boccardi M, Barkhof F, et al. Strategic roadmap for an early diagnosis of Alzheimer’s disease based on biomarkers. Lancet Neurol 2017;16:661–76.

Dubois B, Feldman HH, Jacova C, et al. Advancing research diagnostic criteria for Alzheimer’s disease: the IWG-2 criteria. Lancet Neurol 2014;13:614–29.

Johnson KA, Fox NC, Sperling RA, Klunk WE. Brain imaging in Alzheimer disease. Cold Spring Harb Perspect Med 2012;2:a006213.

Dubois B, Hampel H, Feldman HH, et al. Preclinical Alzheimer’s disease: definition, natural history, and diagnostic criteria. Alzheimers Dement 2016;12:292–323.

Villemagne VL, Doré V, Burnham SC, Masters CL, Rowe CC. Imaging tau and amyloid-β proteinopathies in Alzheimer disease and other conditions. Nat Rev Neurol 2018;14:225–36.

Clark CM, Pontecorvo MJ, Beach TG, et al. Cerebral PET with florbetapir compared with neuropathology at autopsy for detection of neuritic amyloid-β plaques: a prospective cohort study. Lancet Neurol 2012;11:669–78.

Wong DF, Rosenberg PB, Zhou Y, et al. In vivo imaging of amyloid deposition in Alzheimer disease using the radioligand 18F-AV-45 (florbetapir F 18). J Nucl Med 2010;51:913–20.

Duits FH, Martinez-Lage P, Paquet C, et al. Performance and complications of lumbar puncture in memory clinics: results of the multicenter lumbar puncture feasibility study. Alzheimers Dement 2016;12:154–63.

Shaw LM, Arias J, Blennow K, et al. Appropriate use criteria for lumbar puncture and cerebrospinal fluid testing in the diagnosis of Alzheimer’s disease. Alzheimers Dement 2018;14:1505–21.

Hansson O, Seibyl J, Stomrud E, et al. CSF biomarkers of Alzheimer’s disease concord with amyloid-β PET and predict clinical progression: a study of fully automated immunoassays in BioFINDER and ADNI cohorts. Alzheimers Dement 2018;14:1470–81.

Blennow K, Dubois B, Fagan AM, Lewczuk P, de Leon MJ, Hampel H. Clinical utility of cerebrospinal fluid biomarkers in the diagnosis of early Alzheimer’s disease. Alzheimers Dement 2015;11:58–69.

Hansson O, Lehmann S, Otto M, Zetterberg H, Lewczuk P. Advantages and disadvantages of the use of the CSF Amyloid β (Aβ) 42/40 ratio in the diagnosis of Alzheimer’s Disease. Alzheimers Res Ther 2019;11:34.

Lee SA, Sposato LA, Hachinski V, Cipriano LE. Cost-effectiveness of cerebrospinal biomarkers for the diagnosis of Alzheimer’s disease. Alzheimers Res Ther 2017;9:18.

Dolgin E. Alzheimer’s disease is getting easier to spot. Nature 2018;559:S10–2.

Morris E, Chalkidou A, Hammers A, Peacock J, Summers J, Keevil S. Diagnostic accuracy of 18F amyloid PET tracers for the diagnosis of Alzheimer’s disease: a systematic review and meta-analysis. Eur J Nucl Med Mol Imaging 2016;43:374–85.

Engelborghs S, Niemantsverdriet E, Struyfs H, et al. Consensus guidelines for lumbar puncture in patients with neurological diseases. Alzheimers Dement Amst 2017;8:111–26.

Barthélemy NR, Bateman RJ, Hirtz C, et al. Cerebrospinal fluid phosphotau T217 outperforms T181 as a biomarker for the differential diagnosis of Alzheimer’s disease and PET amyloid-positive patient identification. Alzheimers Res Ther 2020;12:26.

O’Bryant SE, Mielke MM, Rissman RA, et al. Blood-based biomarkers in Alzheimer disease: current state of the science and a novel collaborative paradigm for advancing from discovery to clinic. Alzheimers Dement 2017;13:45–58.

C2N Diagnostics. Press release. Alzheimer’s breakthrough: C 2 N first to offer a widely accessible blood test. C2N Diagn 2021. https://www.c2ndiagnostics.com/press/press/2020/10/28/alzheimers-breakthrough-cn-first-to-offer-a-widely-accessible-blood-test (accessed January 25, 2021).

Barthélemy NR, Horie K, Sato C, Bateman RJ. Blood plasma phosphorylatedtau isoforms track CNS change in Alzheimer’s disease. J Exp Med 2020;217:e20200861.

Janelidze S, Stomrud E, Smith R, et al. Cerebrospinal fluid p-tau217 performs better than p-tau181 as a biomarker of Alzheimer’s disease. Nat Commun 2020;11:1683.

Chen J, Duan Y, Li H, Lu L, Liu J, Tang C. Different durations of cognitive stimulation therapy for Alzheimer’s disease: a systematic review and meta-analysis. Clin Interv Aging 2019;14:1243–54.

Cummings J, Fox N. Defining disease modifying therapy for Alzheimer’s disease. J Prev Alzheimers Dis 2017;4:109–15.

CAS   PubMed   PubMed Central   Google Scholar  

Birks JS, Harvey RJ. Donepezil for dementia due to Alzheimer’s disease. Cochrane Database Syst Rev 2018;6:CD001190.

Birks JS, Chong LY, Grimley Evans J. Rivastigmine for Alzheimer’s disease. Cochrane Database Syst Rev 2015;9:CD001191.

Tariot PN, Farlow MR, Grossberg GT, et al. Memantine treatment in patients with moderate to severe Alzheimer disease already receiving donepezil: a randomized controlled trial. JAMA 2004;291:317–24.

Cummings J. New approaches to symptomatic treatments for Alzheimer’s disease. Mol Neurodegener 2021;16:2.

Loy C, Schneider L. Galantamine for Alzheimer’s disease and mild cognitive impairment. Cochrane Database Syst Rev 2006;1:CD001747.

Atri A. The Alzheimer’s Disease Clinical Spectrum. Med Clin N Am 2019;103:263–93.

Berlowitz DR, Foy CG, Kazis LE, et al. Effect of intensive blood-pressure treatment on patient-reported outcomes. N Engl J Med 2017;377:733–44.

Porsteinsson AP, Drye LT, Pollock BG, et al. Effect of citalopram on agitation in Alzheimer’s disease - the citAD randomized controlled trial. JAMA 2014;311:682–91.

Sheline YI, Snider BJ, Beer JC, et al. Effect of escitalopram dose and treatment duration on CSF Aβ levels in healthy older adults: a controlled clinical trial. Neurology 2020;95:e2658–65.

Sheehan B. Assessment scales in dementia. Ther Adv Neurol Disord 2012;5:349–58.

Haubois G, Annweiler C, Launay C, et al. Development of a short form of Mini-Mental State Examination for the screening of dementia in older adults with a memory complaint: a case control study. BMC Geriatr 2011;11:59.

Horton DK, Hynan LS, Lacritz LH, Rossetti HC, Weiner MF, Cullum CM. An Abbreviated Montreal Cognitive Assessment (MoCA) for dementia screening. Clin Neuropsychol 2015;29:413–25.

Mini-Cog©. Mini-Cog© In other languages. Mini-Cog© Lang 2021. https://mini-cog.com/mini-cog-in-other-languages/ (accessed April 21, 2021).

Carnero Pardo C, de la Vega Cotarelo R, López Alcalde S, et al. Assessing the diagnostic accuracy (DA) of the Spanish version of the informant-based AD8 questionnaire. Neurologia 2013;28:88–94.

Harrison JK, Fearon P, Noel-Storr AH, McShane R, Stott DJ, Quinn TJ. Informant questionnaire on cognitive decline in the elderly (IQCODE) for the diagnosis of dementia within a secondary care setting. Cochrane Database Syst Rev 2015;3:CD010772.

Sanchez MA, Correa PC, Lourenço RA. Cross-cultural adaptation of the “Functional Activities Questionnaire - FAQ” for use in Brazil. Dement Neuropsychol 2011;5:322–7.

Kim G, DeCoster J, Huang C-H, Bryant AN. A meta-analysis of the factor structure of the Geriatric Depression Scale (GDS): the effects of language. Int Psychogeriatr 2013;25:71–81.

Mapi Research Trust. NPI - Officially distributed by Mapi Research Trust. Neuropsychiatr Inventory Quest NPI-Q 2021. https://eprovide.mapi-trust.org/instruments/neuropsychiatric-inventory-questionnaire (accessed April 21, 2021).

Mapi Research Trust. A-IADL-Q-SV - Amsterdam Instrumental Activity of Daily Living Questionnaire - Short version. Amst Instrum Act Dly Living Quest - Short Version -IADL-Q-SV n.d. https://eprovide.mapi-trust.org/instruments/amsterdam-instrumental-activity-of-daily-living-questionnaire-short-version (accessed April 21, 2021).

Ismail Z, Agüera-Ortiz L, Brodaty H, et al. The Mild Behavioral Impairment Checklist (MBI-C): a rating scale for neuropsychiatric symptoms in predementia populations. J Alzheimers Dis 2017;56:929–38.

Download references

Acknowledgements

The authors would like to acknowledge and thank Dr. Giovanni Frisoni, Geneva University Neurocenter, for his contribution towards the development of this manuscript.

Funding: The authors developed this manuscript concept during an assessment of Alzheimer’s disease educational needs. The development of this manuscript was funded by Biogen. Editorial support was provided by Jodie Penney, MSc, PhD, Helios Medical Communications, Cheshire, UK, which was funded by Biogen.

Author information

Authors and affiliations.

University of Rochester School of Medicine and Dentistry, Rochester, NY, USA

A. P. Porsteinsson

Weill Cornell Medical Center and New York-Presbyterian, New York, NY, USA

R. S. Isaacson

Biogen International GmbH, Neuhofstrasse 30, 6340, Baar, Switzerland

Cleveland Clinic Lou Ruvo Center for Brain Health s, Las Vegas, NV, USA

M. N. Sabbagh

Biogen Inc, Cambridge, MA, USA

You can also search for this author in PubMed   Google Scholar

Contributions

Authors’ contributions: All authors participated in the review of the literature and in the drafting and reviewing of the manuscript. All authors read and approved the final version of the manuscript for submission.

Corresponding author

Correspondence to Sean Knox .

Ethics declarations

Conflict of Interest: AP reports personal fees from Acadia Pharmaceuticals, Alzheon, Avanir, Biogen, Cadent Therapeutics, Eisai, Functional Neuromodulation, MapLight Therapeutics, Premier Healthcare Solutions, Sunovion, and Syneos; grants from Alector, Athira, Avanir, Biogen, Biohaven, Eisai, Eli Lilly, Genentech/Roche, and Vaccinex. RI has nothing to disclose. MS reports personal fees from Alzheon, Athira, Biogen, Cortexyme, Danone, Neurotrope, Regeneron, Roche-Genentech, and Stage 2 Innovations; stock options from Brain Health Inc, NeuroReserve, NeuroTau, Neurotrope, Optimal Cognitive Health Company, uMethod Health, and Versanum Inc. Additionally, he has intellectual property rights with Harper Collins. SK and IR report employment with Biogen.

Electronic supplementary material

Supplementary material, approximately 65.7 MB.

Rights and permissions

Open Access: This article is distributed under the terms of the Creative Commons Attribution 4.0 International License ( http://creativecommons.org/licenses/by/4.0/ ), which permits use, duplication, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Reprints and permissions

About this article

Porsteinsson, A.P., Isaacson, R.S., Knox, S. et al. Diagnosis of Early Alzheimer’s Disease: Clinical Practice in 2021. J Prev Alzheimers Dis 8 , 371–386 (2021). https://doi.org/10.14283/jpad.2021.23

Download citation

Received : 22 February 2021

Accepted : 28 April 2021

Published : 09 June 2021

Issue Date : July 2021

DOI : https://doi.org/10.14283/jpad.2021.23

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Alzheimer’s disease
• early diagnosis
• diagnostic work-up
• Find a journal
• Publish with us
• Track your research

Information

• Author Services

Initiatives

You are accessing a machine-readable page. In order to be human-readable, please install an RSS reader.

All articles published by MDPI are made immediately available worldwide under an open access license. No special permission is required to reuse all or part of the article published by MDPI, including figures and tables. For articles published under an open access Creative Common CC BY license, any part of the article may be reused without permission provided that the original article is clearly cited. For more information, please refer to https://www.mdpi.com/openaccess .

Feature papers represent the most advanced research with significant potential for high impact in the field. A Feature Paper should be a substantial original Article that involves several techniques or approaches, provides an outlook for future research directions and describes possible research applications.

Feature papers are submitted upon individual invitation or recommendation by the scientific editors and must receive positive feedback from the reviewers.

Editor’s Choice articles are based on recommendations by the scientific editors of MDPI journals from around the world. Editors select a small number of articles recently published in the journal that they believe will be particularly interesting to readers, or important in the respective research area. The aim is to provide a snapshot of some of the most exciting work published in the various research areas of the journal.

Original Submission Date Received: .

• Active Journals
• Find a Journal
• Proceedings Series
• For Authors
• For Reviewers
• For Editors
• For Librarians
• For Publishers
• For Societies
• For Conference Organizers
• Open Access Policy
• Institutional Open Access Program
• Special Issues Guidelines
• Editorial Process
• Research and Publication Ethics
• Article Processing Charges
• Testimonials
• Preprints.org
• SciProfiles
• Encyclopedia

Article Menu

• Subscribe SciFeed
• Recommended Articles
• PubMed/Medline
• Google Scholar
• on Google Scholar
• Table of Contents

Find support for a specific problem in the support section of our website.

Please let us know what you think of our products and services.

Visit our dedicated information section to learn more about MDPI.

JSmol Viewer

Comprehensive review on alzheimer’s disease: causes and treatment.

1. Introduction

2. alzheimer’s disease diagnostic criteria, 3. alzheimer’s disease’s neuropathology, 3.1. senile plaques (sp), 3.2. neurofibrillary tangles (nfts), 3.3. synaptic loss, 4. the stages of alzheimer’s disease, 5. causes and risk factors of alzheimer’s disease, 5.1. alzheimer’s disease hypotheses, 5.1.1. cholinergic hypothesis, 5.1.2. amyloid hypothesis, 5.2. alzheimer’s disease risk factors, 5.2.1. aging, 5.2.2. genetics.

• Amyloid Precursor Protein (APP)
• Presenilin-1 (PSEN-1) and Presenilin-2 (PSEN-2)
• Apolipoprotein E (ApoE)
• ATP Binding Cassette Transporter A1 (ABCA1)
• Clusterin Gene (CLU) and Bridging Integrator 1 ( BIN1 )
• Evolutionarily Conserved Signaling Intermediate in Toll pathway (ECSIT)
• Estrogen Receptor Gene (ESR)
• Other Genes

5.2.3. Environmental Factors

• Air Pollution

5.2.4. Medical Factors

• Cardiovascular Disease (CVDs)
• Obesity and Diabetes

6. Treatment

6.1. symptomatic treatment of ad, 6.1.1. cholinesterase inhibitors.

• Rivastigmine
• Galantamine (GAL)

6.1.2. N -methyl d -aspartate (NMDA) Antagonists

6.2. promising future therapies, 6.2.1. disease-modifying therapeutics (dmt), 6.2.2. chaperones.

• Heat Shock Proteins (Hsps)
• Vacuolar sorting protein 35 (VPS35)

6.2.3. Natural Extract

7. conclusions, author contributions, conflicts of interest.

• De-Paula, V.J.; Radanovic, M.; Diniz, B.S.; Forlenza, O.V. Alzheimer’s disease. Sub-Cell. Biochem. 2012 , 65 , 329–352. [ Google Scholar ] [ CrossRef ]
• Cipriani, G.; Dolciotti, C.; Picchi, L.; Bonuccelli, U. Alzheimer and his disease: A brief history. Neurol. Sci. Off. J. Ital. Neurol. Soc. Ital. Soc. Clin. Neurophysiol. 2011 , 32 , 275–279. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Blass, J.P. Alzheimer’s disease. Dis. A Mon. Dm 1985 , 31 , 1–69. [ Google Scholar ] [ CrossRef ]
• Terry, R.D.; Davies, P. Dementia of the Alzheimer type. Annu. Rev. Neurosci. 1980 , 3 , 77–95. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Rathmann, K.L.; Conner, C.S. Alzheimer’s disease: Clinical features, pathogenesis, and treatment. Drug Intell. Clin. Pharm. 1984 , 18 , 684–691. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Yiannopoulou, K.G.; Papageorgiou, S.G. Current and future treatments in alzheimer disease: An update. J. Cent. Nerv. Syst. Dis. 2020 , 12. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Livingston, G.; Huntley, J.; Sommerlad, A.; Ames, D.; Ballard, C.; Banerjee, S.; Brayne, C.; Burns, A.; Cohen-Mansfield, J.; Cooper, C.; et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet 2020 , 396 , 413–446. [ Google Scholar ] [ CrossRef ]
• Schachter, A.S.; Davis, K.L. Alzheimer’s disease. Dialogues Clin. Neurosci. 2000 , 2 , 91–100. [ Google Scholar ] [ CrossRef ]
• Jatoi, S.; Hafeez, A.; Riaz, S.U.; Ali, A.; Ghauri, M.I.; Zehra, M. Low Vitamin B12 levels: An underestimated cause of minimal cognitive impairment and dementia. Cureus 2020 , 12 , e6976. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Cho, H.S.; Huang, L.K.; Lee, Y.T.; Chan, L.; Hong, C.T. Suboptimal baseline serum Vitamin B12 is associated with cognitive decline in people with Alzheimer’s disease undergoing cholinesterase inhibitor treatment. Front. Neurol. 2018 , 9 , 325. [ Google Scholar ] [ CrossRef ]
• McKhann, G.; Drachman, D.; Folstein, M.; Katzman, R.; Price, D.; Stadlan, E.M. Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology 1984 , 34 , 939–944. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Neugroschl, J.; Wang, S. Alzheimer’s disease: Diagnosis and treatment across the spectrum of disease severity. Mt. Sinai J. Med. N. Y. 2011 , 78 , 596–612. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• McKhann, G.M.; Knopman, D.S.; Chertkow, H.; Hyman, B.T.; Jack, C.R., Jr.; Kawas, C.H.; Klunk, W.E.; Koroshetz, W.J.; Manly, J.J.; Mayeux, R.; et al. The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimer’s Dement. J. Alzheimer’s Assoc. 2011 , 7 , 263–269. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Mayeux, R.; Stern, Y. Epidemiology of Alzheimer disease. Cold Spring Harb. Perspect. Med. 2012 , 2 , a006239. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Yaari, R.; Fleisher, A.S.; Tariot, P.N. Updates to diagnostic guidelines for Alzheimer’s disease. Prim. Care Companion Cns Disord. 2011 , 13 , 11f01262. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Serrano-Pozo, A.; Frosch, M.P.; Masliah, E.; Hyman, B.T. Neuropathological alterations in Alzheimer disease. Cold Spring Harb. Perspect. Med. 2011 , 1 , a006189. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Spires-Jones, T.L.; Hyman, B.T. The intersection of amyloid beta and tau at synapses in Alzheimer’s disease. Neuron 2014 , 82 , 756–771. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Singh, S.K.; Srivastav, S.; Yadav, A.K.; Srikrishna, S.; Perry, G. Overview of Alzheimer’s disease and some therapeutic approaches targeting abeta by using several synthetic and herbal compounds. Oxidative Med. Cell. Longev. 2016 , 2016 , 7361613. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Cras, P.; Kawai, M.; Lowery, D.; Gonzalez-DeWhitt, P.; Greenberg, B.; Perry, G. Senile plaque neurites in Alzheimer disease accumulate amyloid precursor protein. Proc. Natl. Acad. Sci. USA 1991 , 88 , 7552–7556. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Perl, D.P. Neuropathology of Alzheimer’s disease. Mt. Sinai J. Med. N. Y. 2010 , 77 , 32–42. [ Google Scholar ] [ CrossRef ]
• Armstrong, R.A. The molecular biology of senile plaques and neurofibrillary tangles in Alzheimer’s disease. Folia Neuropathol. 2009 , 47 , 289–299. [ Google Scholar ] [ PubMed ]
• Chen, G.F.; Xu, T.H.; Yan, Y.; Zhou, Y.R.; Jiang, Y.; Melcher, K.; Xu, H.E. Amyloid beta: Structure, biology and structure-based therapeutic development. Acta Pharmacol. Sin. 2017 , 38 , 1205–1235. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Tabaton, M.; Piccini, A. Role of water-soluble amyloid-beta in the pathogenesis of Alzheimer’s disease. Int. J. Exp. Pathol. 2005 , 86 , 139–145. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Brion, J.P. Neurofibrillary tangles and Alzheimer’s disease. Eur. Neurol. 1998 , 40 , 130–140. [ Google Scholar ] [ CrossRef ]
• Metaxas, A.; Kempf, S.J. Neurofibrillary tangles in Alzheimer’s disease: Elucidation of the molecular mechanism by immunohistochemistry and tau protein phospho-proteomics. Neural Regen. Res. 2016 , 11 , 1579–1581. [ Google Scholar ] [ CrossRef ]
• Overk, C.R.; Masliah, E. Pathogenesis of synaptic degeneration in Alzheimer’s disease and Lewy body disease. Biochem Pharm. 2014 , 88 , 508–516. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Lleo, A.; Nunez-Llaves, R.; Alcolea, D.; Chiva, C.; Balateu-Panos, D.; Colom-Cadena, M.; Gomez-Giro, G.; Munoz, L.; Querol-Vilaseca, M.; Pegueroles, J.; et al. Changes in synaptic proteins precede neurodegeneration markers in preclinical Alzheimer’s disease cerebrospinal fluid. Mol. Cell. Proteom. Mcp 2019 , 18 , 546–560. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Tarawneh, R.; D’Angelo, G.; Crimmins, D.; Herries, E.; Griest, T.; Fagan, A.M.; Zipfel, G.J.; Ladenson, J.H.; Morris, J.C.; Holtzman, D.M. Diagnostic and prognostic utility of the synaptic marker neurogranin in Alzheimer Disease. JAMA Neurol. 2016 , 73 , 561–571. [ Google Scholar ] [ CrossRef ]
• Dubois, B.; Hampel, H.; Feldman, H.H.; Scheltens, P.; Aisen, P.; Andrieu, S.; Bakardjian, H.; Benali, H.; Bertram, L.; Blennow, K.; et al. Preclinical Alzheimer’s disease: Definition, natural history, and diagnostic criteria. Alzheimer’s Dement. J. Alzheimer’s Assoc. 2016 , 12 , 292–323. [ Google Scholar ] [ CrossRef ]
• Kumar, A.; Sidhu, J.; Goyal, A. Alzheimer Disease. In StatPearls ; StatPearls Publishing: Treasure Island, FL, USA, 2020; Available online: https://www.ncbi.nlm.nih.gov/books/NBK499922/ (accessed on 8 December 2020).
• Wattmo, C.; Minthon, L.; Wallin, A.K. Mild versus moderate stages of Alzheimer’s disease: Three-year outcomes in a routine clinical setting of cholinesterase inhibitor therapy. Alzheimer’s Res. Ther. 2016 , 8 , 7. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Apostolova, L.G. Alzheimer disease. Continuum 2016 , 22 , 419–434. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Armstrong, R.A. Risk factors for Alzheimer’s disease. Folia Neuropathol. 2019 , 57 , 87–105. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Anand, P.; Singh, B. A review on cholinesterase inhibitors for Alzheimer’s disease. Arch. Pharmacal Res. 2013 , 36 , 375–399. [ Google Scholar ] [ CrossRef ]
• Babic, T. The cholinergic hypothesis of Alzheimer’s disease: A review of progress. J. Neurol. Neurosurg. Psychiatry 1999 , 67 , 558. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Ferreira-Vieira, T.H.; Guimaraes, I.M.; Silva, F.R.; Ribeiro, F.M. Alzheimer’s disease: Targeting the Cholinergic System. Curr. Neuropharmacol. 2016 , 14 , 101–115. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Monczor, M. Diagnosis and treatment of Alzheimer’s disease. Curr. Med. Chem. Cent. Nerv. Syst. Agents 2005 , 5 , 5–13. [ Google Scholar ] [ CrossRef ]
• Hampel, H.; Mesulam, M.M.; Cuello, A.C.; Farlow, M.R.; Giacobini, E.; Grossberg, G.T.; Khachaturian, A.S.; Vergallo, A.; Cavedo, E.; Snyder, P.J.; et al. The cholinergic system in the pathophysiology and treatment of Alzheimer’s disease. Brain A J. Neurol. 2018 , 141 , 1917–1933. [ Google Scholar ] [ CrossRef ]
• Paroni, G.; Bisceglia, P.; Seripa, D. Understanding the amyloid hypothesis in Alzheimer’s disease. J. Alzheimer’s Dis. Jad 2019 , 68 , 493–510. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Kametani, F.; Hasegawa, M. Reconsideration of amyloid hypothesis and tau hypothesis in Alzheimer’s disease. Front. Neurosci. 2018 , 12 , 25. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Ricciarelli, R.; Fedele, E. The amyloid cascade hypothesis in Alzheimer’s disease: It’s time to change our mind. Curr. Neuropharmacol. 2017 , 15 , 926–935. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Guerreiro, R.; Bras, J. The age factor in Alzheimer’s disease. Genome Med. 2015 , 7 , 106. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Riedel, B.C.; Thompson, P.M.; Brinton, R.D. Age, APOE and sex: Triad of risk of Alzheimer’s disease. J. Steroid Biochem. Mol. Biol. 2016 , 160 , 134–147. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Hou, Y.; Dan, X.; Babbar, M.; Wei, Y.; Hasselbalch, S.G.; Croteau, D.L.; Bohr, V.A. Ageing as a risk factor for neurodegenerative disease. Nat. Rev. Neurol. 2019 , 15 , 565–581. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Bekris, L.M.; Yu, C.E.; Bird, T.D.; Tsuang, D.W. Genetics of Alzheimer disease. J. Geriatr. Psychiatry Neurol. 2010 , 23 , 213–227. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Van Cauwenberghe, C.; Van Broeckhoven, C.; Sleegers, K. The genetic landscape of Alzheimer disease: Clinical implications and perspectives. Genet. Med. Off. J. Am. Coll. Med Genet. 2016 , 18 , 421–430. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Khanahmadi, M.; Farhud, D.D.; Malmir, M. Genetic of Alzheimer’s disease: A narrative review article. Iran. J. Public Health 2015 , 44 , 892–901. [ Google Scholar ]
• Li, N.M.; Liu, K.F.; Qiu, Y.J.; Zhang, H.H.; Nakanishi, H.; Qing, H. Mutations of beta-amyloid precursor protein alter the consequence of Alzheimer’s disease pathogenesis. Neural Regen. Res. 2019 , 14 , 658–665. [ Google Scholar ] [ CrossRef ]
• Tcw, J.; Goate, A.M. Genetics of beta-Amyloid precursor protein in Alzheimer’s disease. Cold Spring Harb. Perspect. Med. 2017 , 7 , a024539. [ Google Scholar ] [ CrossRef ]
• Bi, C.; Bi, S.; Li, B. Processing of mutant beta-amyloid precursor protein and the clinicopathological features of familial Alzheimer’s disease. Aging Dis. 2019 , 10 , 383–403. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Dai, M.H.; Zheng, H.; Zeng, L.D.; Zhang, Y. The genes associated with early-onset Alzheimer’s disease. Oncotarget 2018 , 9 , 15132–15143. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Zhao, J.; Liu, X.; Xia, W.; Zhang, Y.; Wang, C. Targeting amyloidogenic processing of APP in Alzheimer’s disease. Front. Mol. Neurosci. 2020 , 13 , 137. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Cai, Y.; An, S.S.; Kim, S. Mutations in presenilin 2 and its implications in Alzheimer’s disease and other dementia-associated disorders. Clin. Interv. Aging 2015 , 10 , 1163–1172. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Lanoiselee, H.M.; Nicolas, G.; Wallon, D.; Rovelet-Lecrux, A.; Lacour, M.; Rousseau, S.; Richard, A.C.; Pasquier, F.; Rollin-Sillaire, A.; Martinaud, O.; et al. APP, PSEN1, and PSEN2 mutations in early-onset Alzheimer disease: A genetic screening study of familial and sporadic cases. PLoS Med. 2017 , 14 , e1002270. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• De Strooper, B. Loss-of-function presenilin mutations in Alzheimer disease. Talking Point on the role of presenilin mutations in Alzheimer disease. Embo Rep. 2007 , 8 , 141–146. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Kelleher, R.J., 3rd; Shen, J. Presenilin-1 mutations and Alzheimer’s disease. Proc. Natl. Acad. Sci. USA 2017 , 114 , 629–631. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Walker, E.S.; Martinez, M.; Brunkan, A.L.; Goate, A. Presenilin 2 familial Alzheimer’s disease mutations result in partial loss of function and dramatic changes in Abeta 42/40 ratios. J. Neurochem. 2005 , 92 , 294–301. [ Google Scholar ] [ CrossRef ]
• Kim, J.; Basak, J.M.; Holtzman, D.M. The role of apolipoprotein E in Alzheimer’s disease. Neuron 2009 , 63 , 287–303. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Liu, C.C.; Liu, C.C.; Kanekiyo, T.; Xu, H.; Bu, G. Apolipoprotein E and Alzheimer disease: Risk, mechanisms and therapy. Nat. Rev. Neurol. 2013 , 9 , 106–118. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Giau, V.V.; Bagyinszky, E.; An, S.S.; Kim, S.Y. Role of apolipoprotein E in neurodegenerative diseases. Neuropsychiatr. Dis. Treat. 2015 , 11 , 1723–1737. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Koldamova, R.; Fitz, N.F.; Lefterov, I. ATP-binding cassette transporter A1: From metabolism to neurodegeneration. Neurobiol. Dis. 2014 , 72 Pt A , 13–21. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Nordestgaard, L.T.; Tybjaerg-Hansen, A.; Nordestgaard, B.G.; Frikke-Schmidt, R. Loss-of-function mutation in ABCA1 and risk of Alzheimer’s disease and cerebrovascular disease. Alzheimer’s Dement. J. Alzheimer’s Assoc. 2015 , 11 , 1430–1438. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Foster, E.M.; Dangla-Valls, A.; Lovestone, S.; Ribe, E.M.; Buckley, N.J. Clusterin in Alzheimer’s disease: Mechanisms, genetics, and lessons from other pathologies. Front. Neurosci. 2019 , 13 , 164. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Holler, C.J.; Davis, P.R.; Beckett, T.L.; Platt, T.L.; Webb, R.L.; Head, E.; Murphy, M.P. Bridging integrator 1 (BIN1) protein expression increases in the Alzheimer’s disease brain and correlates with neurofibrillary tangle pathology. J. Alzheimer’s Dis. Jad 2014 , 42 , 1221–1227. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Andrew, R.J.; De Rossi, P.; Nguyen, P.; Kowalski, H.R.; Recupero, A.J.; Guerbette, T.; Krause, S.V.; Rice, R.C.; Laury-Kleintop, L.; Wagner, S.L.; et al. Reduction of the expression of the late-onset Alzheimer’s disease (AD) risk-factor BIN1 does not affect amyloid pathology in an AD mouse model. J. Biol. Chem. 2019 , 294 , 4477–4487. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Soler-Lopez, M.; Badiola, N.; Zanzoni, A.; Aloy, P. Towards Alzheimer’s root cause: ECSIT as an integrating hub between oxidative stress, inflammation and mitochondrial dysfunction. Hypothetical role of the adapter protein ECSIT in familial and sporadic Alzheimer’s disease pathogenesis. Bioessays News Rev. Mol. Cell. Dev. Biol. 2012 , 34 , 532–541. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Mi Wi, S.; Park, J.; Shim, J.H.; Chun, E.; Lee, K.Y. Ubiquitination of ECSIT is crucial for the activation of p65/p50 NF-kappaBs in Toll-like receptor 4 signaling. Mol. Biol. Cell 2015 , 26 , 151–160. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Soler-Lopez, M.; Zanzoni, A.; Lluis, R.; Stelzl, U.; Aloy, P. Interactome mapping suggests new mechanistic details underlying Alzheimer’s disease. Genome Res. 2011 , 21 , 364–376. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Zhao, L.; Woody, S.K.; Chhibber, A. Estrogen receptor beta in Alzheimer’s disease: From mechanisms to therapeutics. Ageing Res. Rev. 2015 , 24 , 178–190. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Sundermann, E.E.; Maki, P.M.; Bishop, J.R. A review of estrogen receptor alpha gene (ESR1) polymorphisms, mood, and cognition. Menopause 2010 , 17 , 874–886. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Yaffe, K.; Lindquist, K.; Sen, S.; Cauley, J.; Ferrell, R.; Penninx, B.; Harris, T.; Li, R.; Cummings, S.R. Estrogen receptor genotype and risk of cognitive impairment in elders: Findings from the Health ABC study. Neurobiol. Aging 2009 , 30 , 607–614. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Goumidi, L.; Dahlman-Wright, K.; Tapia-Paez, I.; Matsson, H.; Pasquier, F.; Amouyel, P.; Kere, J.; Lambert, J.C.; Meirhaeghe, A. Study of estrogen receptor-alpha and receptor-beta gene polymorphisms on Alzheimer’s disease. J. Alzheimer’s Dis. Jad 2011 , 26 , 431–439. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Khorram Khorshid, H.R.; Gozalpour, E.; Saliminejad, K.; Karimloo, M.; Ohadi, M.; Kamali, K. Vitamin D Receptor (VDR) polymorphisms and late-onset Alzheimer’s disease: An association study. Iran. J. Public Health 2013 , 42 , 1253–1258. [ Google Scholar ] [ PubMed ]
• Liu, X.; Jiao, B.; Shen, L. The epigenetics of Alzheimer’s Disease: Factors and therapeutic implications. Front. Genet. 2018 , 9 , 579. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Wainaina, M.N.; Chen, Z.; Zhong, C. Environmental factors in the development and progression of late-onset Alzheimer’s disease. Neurosci. Bull. 2014 , 30 , 253–270. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Grant, W.B.; Campbell, A.; Itzhaki, R.F.; Savory, J. The significance of environmental factors in the etiology of Alzheimer’s disease. J. Alzheimer’s Dis. Jad 2002 , 4 , 179–189. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Moulton, P.V.; Yang, W. Air pollution, oxidative stress, and Alzheimer’s disease. J. Environ. Public Health 2012 , 2012 , 472751. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Croze, M.L.; Zimmer, L. Ozone atmospheric pollution and Alzheimer’s disease: From epidemiological facts to molecular mechanisms. J. Alzheimer’s Dis. Jad 2018 , 62 , 503–522. [ Google Scholar ] [ CrossRef ]
• Hu, N.; Yu, J.T.; Tan, L.; Wang, Y.L.; Sun, L.; Tan, L. Nutrition and the risk of Alzheimer’s disease. Biomed. Res. Int. 2013 , 2013 , 524820. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Abate, G.; Marziano, M.; Rungratanawanich, W.; Memo, M.; Uberti, D. Nutrition and AGE-ing: Focusing on Alzheimer’s disease. Oxidative Med. Cell. Longev. 2017 , 2017 , 7039816. [ Google Scholar ] [ CrossRef ]
• Koyama, A.; Hashimoto, M.; Tanaka, H.; Fujise, N.; Matsushita, M.; Miyagawa, Y.; Hatada, Y.; Fukuhara, R.; Hasegawa, N.; Todani, S.; et al. Malnutrition in Alzheimer’s disease, dementia with lewy bodies, and frontotemporal lobar degeneration: Comparison using serum albumin, total protein, and hemoglobin level. PLoS ONE 2016 , 11 , e0157053. [ Google Scholar ] [ CrossRef ]
• Adlard, P.A.; Bush, A.I. Metals and Alzheimer’s disease. J. Alzheimer’s Dis. Jad 2006 , 10 , 145–163. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Colomina, M.T.; Peris-Sampedro, F. Aluminum and Alzheimer’s disease. Adv. Neurobiol. 2017 , 18 , 183–197. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Huat, T.J.; Camats-Perna, J.; Newcombe, E.A.; Valmas, N.; Kitazawa, M.; Medeiros, R. Metal toxicity links to Alzheimer’s disease and neuroinflammation. J. Mol. Biol. 2019 , 431 , 1843–1868. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Sochocka, M.; Zwolinska, K.; Leszek, J. The infectious etiology of Alzheimer’s disease. Curr. Neuropharmacol. 2017 , 15 , 996–1009. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Fulop, T.; Itzhaki, R.F.; Balin, B.J.; Miklossy, J.; Barron, A.E. Role of microbes in the development of Alzheimer’s disease: State of the art—An international symposium presented at the 2017 IAGG congress in San Francisco. Front. Genet. 2018 , 9 , 362. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Muzambi, R.; Bhaskaran, K.; Brayne, C.; Smeeth, L.; Warren-Gash, C. Common bacterial infections and risk of incident cognitive decline or dementia: A systematic review protocol. BMJ Open 2019 , 9 , e030874. [ Google Scholar ] [ CrossRef ]
• Stampfer, M.J. Cardiovascular disease and Alzheimer’s disease: Common links. J. Intern. Med. 2006 , 260 , 211–223. [ Google Scholar ] [ CrossRef ]
• Santos, C.Y.; Snyder, P.J.; Wu, W.C.; Zhang, M.; Echeverria, A.; Alber, J. Pathophysiologic relationship between Alzheimer’s disease, cerebrovascular disease, and cardiovascular risk: A review and synthesis. Alzheimer’s Dement. 2017 , 7 , 69–87. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• De Bruijn, R.F.; Ikram, M.A. Cardiovascular risk factors and future risk of Alzheimer’s disease. BMC Med. 2014 , 12 , 130. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Alford, S.; Patel, D.; Perakakis, N.; Mantzoros, C.S. Obesity as a risk factor for Alzheimer’s disease: Weighing the evidence. Obes. Rev. Off. J. Int. Assoc. Study Obes. 2018 , 19 , 269–280. [ Google Scholar ] [ CrossRef ]
• Pegueroles, J.; Jimenez, A.; Vilaplana, E.; Montal, V.; Carmona-Iragui, M.; Pane, A.; Alcolea, D.; Videla, L.; Casajoana, A.; Clarimon, J.; et al. Obesity and Alzheimer’s disease, does the obesity paradox really exist? A magnetic resonance imaging study. Oncotarget 2018 , 9 , 34691–34698. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Anjum, I.; Fayyaz, M.; Wajid, A.; Sohail, W.; Ali, A. Does obesity increase the risk of dementia: A literature review. Cureus 2018 , 10 , e2660. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Lee, H.J.; Seo, H.I.; Cha, H.Y.; Yang, Y.J.; Kwon, S.H.; Yang, S.J. Diabetes and Alzheimer’s disease: Mechanisms and nutritional aspects. Clin. Nutr. Res. 2018 , 7 , 229–240. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Singh, R.; Sadiq, N.M. Cholinesterase Inhibitors. In StatPearls ; StatPearls Publishing: Treasure Island, FL, USA, 2020; Available online: https://www.ncbi.nlm.nih.gov/books/NBK544336/ (accessed on 8 December 2020).
• Eldufani, J.; Blaise, G. The role of acetylcholinesterase inhibitors such as neostigmine and rivastigmine on chronic pain and cognitive function in aging: A review of recent clinical applications. Alzheimers Dement 2019 , 5 , 175–183. [ Google Scholar ] [ CrossRef ]
• Sharma, K. Cholinesterase inhibitors as Alzheimer’s therapeutics (Review). Mol. Med. Rep. 2019 , 20 , 1479–1487. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Wang, R.; Reddy, P.H. Role of glutamate and NMDA receptors in Alzheimer’s disease. J. Alzheimer’s Dis. Jad 2017 , 57 , 1041–1048. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Kuns, B.; Rosani, A.; Varghese, D. Memantine. In StatPearls ; StatPearls Publishing: Treasure Island, FL, USA, 2020; Available online: https://www.ncbi.nlm.nih.gov/books/NBK500025/ (accessed on 8 December 2020).
• Briggs, R.; Kennelly, S.P.; O’Neill, D. Drug treatments in Alzheimer’s disease. Clin. Med. 2016 , 16 , 247–253. [ Google Scholar ] [ CrossRef ]
• Crous-Bou, M.; Minguillon, C.; Gramunt, N.; Molinuevo, J.L. Alzheimer’s disease prevention: From risk factors to early intervention. Alzheimer’s Res. Ther. 2017 , 9 , 71. [ Google Scholar ] [ CrossRef ]
• Crismon, M.L. Tacrine: First drug approved for Alzheimer’s disease. Ann. Pharmacother. 1994 , 28 , 744–751. [ Google Scholar ] [ CrossRef ]
• Qizilbash, N.; Birks, J.; Lopez Arrieta, J.; Lewington, S.; Szeto, S. Tacrine for Alzheimer’s disease. Cochrane Database Syst. Rev. 2000 , CD000202. [ Google Scholar ] [ CrossRef ]
• Cacabelos, R. Donepezil in Alzheimer’s disease: From conventional trials to pharmacogenetics. Neuropsychiatr. Dis. Treat. 2007 , 3 , 303–333. [ Google Scholar ] [ PubMed ]
• Kumar, A.; Sharma, S. Donepezil. In StatPearls ; StatPearls Publishing: Treasure Island, FL, USA, 2020; Available online: https://www.ncbi.nlm.nih.gov/books/NBK513257/ (accessed on 8 December 2020).
• Dooley, M.; Lamb, H.M. Donepezil: A review of its use in Alzheimer’s disease. Drugs Aging 2000 , 16 , 199–226. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Annicchiarico, R.; Federici, A.; Pettenati, C.; Caltagirone, C. Rivastigmine in Alzheimer’s disease: Cognitive function and quality of life. Ther. Clin. Risk Manag. 2007 , 3 , 1113–1123. [ Google Scholar ] [ PubMed ]
• Muller, T. Rivastigmine in the treatment of patients with Alzheimer’s disease. Neuropsychiatr. Dis. Treat. 2007 , 3 , 211–218. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Khoury, R.; Rajamanickam, J.; Grossberg, G.T. An update on the safety of current therapies for Alzheimer’s disease: Focus on rivastigmine. Ther. Adv. Drug Saf. 2018 , 9 , 171–178. [ Google Scholar ] [ CrossRef ]
• Birks, J.; Grimley Evans, J.; Iakovidou, V.; Tsolaki, M.; Holt, F.E. Rivastigmine for Alzheimer’s disease. Cochrane Database Syst. Rev. 2009 , CD001191. [ Google Scholar ] [ CrossRef ]
• Scott, L.J.; Goa, K.L. Galantamine: A review of its use in Alzheimer’s disease. Drugs 2000 , 60 , 1095–1122. [ Google Scholar ] [ CrossRef ]
• Prvulovic, D.; Hampel, H.; Pantel, J. Galantamine for Alzheimer’s disease. Expert Opin. Drug Metab. Toxicol. 2010 , 6 , 345–354. [ Google Scholar ] [ CrossRef ]
• Kim, J.K.; Park, S.U. Pharmacological aspects of galantamine for the treatment of Alzheimer’s disease. Excli J. 2017 , 16 , 35–39. [ Google Scholar ] [ CrossRef ]
• Wahba, S.M.; Darwish, A.S.; Kamal, S.M. Ceria-containing uncoated and coated hydroxyapatite-based galantamine nanocomposites for formidable treatment of Alzheimer’s disease in ovariectomized albino-rat model. Mater. Sci. Eng. C Mater. Biol. Appl. 2016 , 65 , 151–163. [ Google Scholar ] [ CrossRef ]
• Liu, J.; Chang, L.; Song, Y.; Li, H.; Wu, Y. The role of NMDA receptors in Alzheimer’s disease. Front. Neurosci. 2019 , 13 , 43. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Huang, Y.J.; Lin, C.H.; Lane, H.Y.; Tsai, G.E. NMDA Neurotransmission dysfunction in behavioral and psychological symptoms of Alzheimer’s disease. Curr. Neuropharmacol. 2012 , 10 , 272–285. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Companys-Alemany, J.; Turcu, A.L.; Bellver-Sanchis, A.; Loza, M.I.; Brea, J.M.; Canudas, A.M.; Leiva, R.; Vazquez, S.; Pallas, M.; Grinan-Ferre, C. A novel NMDA receptor antagonist protects against cognitive decline presented by senescent mice. Pharmaceutics 2020 , 12 , 284. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Folch, J.; Busquets, O.; Ettcheto, M.; Sanchez-Lopez, E.; Castro-Torres, R.D.; Verdaguer, E.; Garcia, M.L.; Olloquequi, J.; Casadesus, G.; Beas-Zarate, C.; et al. Memantine for the treatment of dementia: A Review on its current and future applications. J. Alzheimer’s Dis. Jad 2018 , 62 , 1223–1240. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Cummings, J.; Fox, N. Defining disease modifying therapy for Alzheimer’s Disease. J. Prev. Alzheimer’s Dis. 2017 , 4 , 109–115. [ Google Scholar ] [ CrossRef ]
• Huang, L.K.; Chao, S.P.; Hu, C.J. Clinical trials of new drugs for Alzheimer disease. J. Biomed. Sci. 2020 , 27 , 18. [ Google Scholar ] [ CrossRef ]
• Neumann, U.; Ufer, M.; Jacobson, L.H.; Rouzade-Dominguez, M.L.; Huledal, G.; Kolly, C.; Luond, R.M.; Machauer, R.; Veenstra, S.J.; Hurth, K.; et al. The BACE-1 inhibitor CNP520 for prevention trials in Alzheimer’s disease. Embo Mol. Med. 2018 , 10 , e9316. [ Google Scholar ] [ CrossRef ]
• Vandenberghe, R.; Riviere, M.E.; Caputo, A.; Sovago, J.; Maguire, R.P.; Farlow, M.; Marotta, G.; Sanchez-Valle, R.; Scheltens, P.; Ryan, J.M.; et al. Active Abeta immunotherapy CAD106 in Alzheimer’s disease: A phase 2b study. Alzheimers Dement 2017 , 3 , 10–22. [ Google Scholar ] [ CrossRef ]
• Cummings, J.; Lee, G.; Ritter, A.; Sabbagh, M.; Zhong, K. Alzheimer’s disease drug development pipeline: 2020. Alzheimers Dement 2020 , 6 , e12050. [ Google Scholar ] [ CrossRef ]
• Tolar, M.; Abushakra, S.; Hey, J.A.; Porsteinsson, A.; Sabbagh, M. Aducanumab, gantenerumab, BAN2401, and ALZ-801-the first wave of amyloid-targeting drugs for Alzheimer’s disease with potential for near term approval. Alzheimer’s Res. Ther. 2020 , 12 , 95. [ Google Scholar ] [ CrossRef ]
• Galimberti, D.; Scarpini, E. Disease-modifying treatments for Alzheimer’s disease. Ther. Adv. Neurol. Disord. 2011 , 4 , 203–216. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Ghezzi, L.; Scarpini, E.; Galimberti, D. Disease-modifying drugs in Alzheimer’s disease. Drug Des. Dev. Ther. 2013 , 7 , 1471–1478. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Medina, M. An Overview on the clinical development of tau-based therapeutics. Int. J. Mol. Sci. 2018 , 19 , 1160. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Davidowitz, E.J.; Krishnamurthy, P.K.; Lopez, P.; Jimenez, H.; Adrien, L.; Davies, P.; Moe, J.G. In vivo validation of a small molecule inhibitor of tau self-association in htau mice. J. Alzheimer’s Dis. Jad 2020 , 73 , 147–161. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Cortez, L.; Sim, V. The therapeutic potential of chemical chaperones in protein folding diseases. Prion 2014 , 8 , 197–202. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Campanella, C.; Pace, A.; Caruso Bavisotto, C.; Marzullo, P.; Marino Gammazza, A.; Buscemi, S.; Palumbo Piccionello, A. Heat shock proteins in Alzheimer’s disease: Role and targeting. Int. J. Mol. Sci. 2018 , 19 , 2603. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Wilhelmus, M.M.; de Waal, R.M.; Verbeek, M.M. Heat shock proteins and amateur chaperones in amyloid-Beta accumulation and clearance in Alzheimer’s disease. Mol. Neurobiol. 2007 , 35 , 203–216. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Martin-Pena, A.; Rincon-Limas, D.E.; Fernandez-Funez, P. Engineered Hsp70 chaperones prevent Abeta42-induced memory impairments in a Drosophila model of Alzheimer’s disease. Sci. Rep. 2018 , 8 , 9915. [ Google Scholar ] [ CrossRef ]
• Calderwood, S.K.; Murshid, A. Molecular chaperone accumulation in cancer and decrease in Alzheimer’s disease: The potential roles of HSF1. Front. Neurosci. 2017 , 11 , 192. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Repalli, J.; Meruelo, D. Screening strategies to identify HSP70 modulators to treat Alzheimer’s disease. Drug Des. Dev. Ther. 2015 , 9 , 321–331. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Li, X.; Shao, H.; Taylor, I.R.; Gestwicki, J.E. Targeting allosteric control mechanisms in heat shock protein 70 (Hsp70). Curr. Top. Med. Chem. 2016 , 16 , 2729–2740. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Abisambra, J.; Jinwal, U.K.; Miyata, Y.; Rogers, J.; Blair, L.; Li, X.; Seguin, S.P.; Wang, L.; Jin, Y.; Bacon, J.; et al. Allosteric heat shock protein 70 inhibitors rapidly rescue synaptic plasticity deficits by reducing aberrant tau. Biol. Psychiatry 2013 , 74 , 367–374. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Bohush, A.; Bieganowski, P.; Filipek, A. Hsp90 and its co-chaperones in neurodegenerative diseases. Int. J. Mol. Sci. 2019 , 20 , 4976. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Ou, J.R.; Tan, M.S.; Xie, A.M.; Yu, J.T.; Tan, L. Heat shock protein 90 in Alzheimer’s disease. Biomed Res. Int. 2014 , 2014 , 796869. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Wang, B.; Liu, Y.; Huang, L.; Chen, J.; Li, J.J.; Wang, R.; Kim, E.; Chen, Y.; Justicia, C.; Sakata, K.; et al. A CNS-permeable Hsp90 inhibitor rescues synaptic dysfunction and memory loss in APP-overexpressing Alzheimer’s mouse model via an HSF1-mediated mechanism. Mol. Psychiatry 2017 , 22 , 990–1001. [ Google Scholar ] [ CrossRef ] [ Green Version ]
• Li, J.G.; Chiu, J.; Ramanjulu, M.; Blass, B.E.; Pratico, D. A pharmacological chaperone improves memory by reducing Abeta and tau neuropathology in a mouse model with plaques and tangles. Mol. Neurodegener. 2020 , 15 , 1. [ Google Scholar ] [ CrossRef ]
• Vagnozzi, A.N.; Li, J.G.; Chiu, J.; Razmpour, R.; Warfield, R.; Ramirez, S.H.; Pratico, D. VPS35 regulates tau phosphorylation and neuropathology in tauopathy. Mol. Psychiatry 2019 . [ Google Scholar ] [ CrossRef ]
• Li, J.G.; Chiu, J.; Pratico, D. Full recovery of the Alzheimer’s disease phenotype by gain of function of vacuolar protein sorting 35. Mol. Psychiatry 2020 , 25 , 2630–2640. [ Google Scholar ] [ CrossRef ]
• Mecozzi, V.J.; Berman, D.E.; Simoes, S.; Vetanovetz, C.; Awal, M.R.; Patel, V.M.; Schneider, R.T.; Petsko, G.A.; Ringe, D.; Small, S.A. Pharmacological chaperones stabilize retromer to limit APP processing. Nat. Chem. Biol. 2014 , 10 , 443–449. [ Google Scholar ] [ CrossRef ]
• Knight, E.M.; Williams, H.N.; Stevens, A.C.; Kim, S.H.; Kottwitz, J.C.; Morant, A.D.; Steele, J.W.; Klein, W.L.; Yanagisawa, K.; Boyd, R.E.; et al. Evidence that small molecule enhancement of beta-hexosaminidase activity corrects the behavioral phenotype in Dutch APP(E693Q) mice through reduction of ganglioside-bound Abeta. Mol. Psychiatry 2015 , 20 , 109–117. [ Google Scholar ] [ CrossRef ]
• Andrade, S.; Ramalho, M.J.; Loureiro, J.A.; Pereira, M.D.C. Natural compounds for Alzheimer’s disease therapy: A systematic review of preclinical and clinical studies. Int. J. Mol. Sci. 2019 , 20 , 2313. [ Google Scholar ] [ CrossRef ] [ PubMed ] [ Green Version ]
• Ma, Y.; Yang, M.W.; Li, X.W.; Yue, J.W.; Chen, J.Z.; Yang, M.W.; Huang, X.; Zhu, L.L.; Hong, F.F.; Yang, S.L. Therapeutic effects of natural drugs on Alzheimer’s disease. Front. Pharmacol. 2019 , 10 , 1355. [ Google Scholar ] [ CrossRef ] [ PubMed ]

Click here to enlarge figure

Share and Cite

Breijyeh, Z.; Karaman, R. Comprehensive Review on Alzheimer’s Disease: Causes and Treatment. Molecules 2020 , 25 , 5789. https://doi.org/10.3390/molecules25245789

Breijyeh Z, Karaman R. Comprehensive Review on Alzheimer’s Disease: Causes and Treatment. Molecules . 2020; 25(24):5789. https://doi.org/10.3390/molecules25245789

Breijyeh, Zeinab, and Rafik Karaman. 2020. "Comprehensive Review on Alzheimer’s Disease: Causes and Treatment" Molecules 25, no. 24: 5789. https://doi.org/10.3390/molecules25245789

Article Metrics

Article access statistics, further information, mdpi initiatives, follow mdpi.

Subscribe to receive issue release notifications and newsletters from MDPI journals

Enter a Search Term

Why apoe4 makes women more susceptible to alzheimer’s.

By: Caleigh Findley, PhD, BrightFocus Foundation

• Research News

Reviewed By: Sharyn Rossi, PhD, BrightFocus Foundation  

Key Takeaways: 

• Women with APOE4 show accelerated immune aging that predisposes them to Alzheimer’s. Early intervention could potentially prevent or decrease cognitive decline in these women. 
• The study, funded by BrightFocus Foundation’s Alzheimer’s Disease Research program, uncovered a novel blood-based biomarker for Alzheimer’s and a potential target for drug development. 
• The researchers encourage progress toward personalized medicine in Alzheimer’s that accounts for individual genetic risk and sex. To that end, they will make the genetic data from the study into a resource for other researchers—empowering future Alzheimer’s discoveries.  

Men and women experience Alzheimer’s differently—right down to the way their brains and bodies respond to the disease, according to a new study funded by BrightFocus Foundation’s Alzheimer’s Disease Research program.  

Postdoctoral researcher and lead study author Neta Rosenzweig, PhD, uncovered not only a unique biological underpinning to Alzheimer’s risk for women, but also a potential drug target for future therapies. The study , published in Nature Medicine , builds on decades of findings from Alzheimer’s Disease Research grant recipient and senior study author, Oleg Butovsky, PhD , distinguished neurology chair and associate professor at Harvard Medical School.   

Their work suggests that blood immune cells infiltrate the brain and prevent the removal of amyloid plaques , driving cognitive impairment for people with the Alzheimer’s risk gene APOE4 . This interaction appears earlier and stronger in women and is a likely contributor to the increased risk of Alzheimer’s for women. The study authors encourage progress toward a personalized medicine approach to treatment based on these findings and current knowledge of sex differences in Alzheimer’s .  

Alzheimer’s and Inflammation: A Journey to Discovery 

Nearly 20 years ago, Dr. Butovsky began investigating a potential new biomarker for Alzheimer’s during his postdoctoral fellowship at the Weizmann Institute of Science. His mentor, Michal Schwartz, PhD, pioneered a seemingly controversial approach to neurodegeneration. She believed that immune cells found in the blood could be directed to help fight against devastating diseases like Alzheimer’s.  

“No one really got the word ’neuroimmunology’ because the brain people were absolutely disconnected from immunology people,” said Dr. Butovsky. The study of connections between the brain and immune system that now dominates much of the Alzheimer’s research space was in its infancy.  

Dr. Butovsky set out on a four-year collaboration between research laboratories. Taking cells from 80 human brains in the Dutch Brain Bank, he generated a large genomic dataset that examined men and women who either carried a copy of APOE4 or did not. APOE4 is the strongest genetic risk factor for Alzheimer’s, with female carriers being at greater risk than male carriers.  

Dr. Butovsky investigated how brain-specific immune cells, called microglia, are affected by APOE4 . Such a question had not been successfully examined at this level before.   

Microglia act as sensors of brain health, he explained. Any minor changes that throw the brain out of balance are reflected in these immune cells. From this genetic dataset came a big surprise that would spark years of continued research.  

They found that microglia sensed changes happening in blood immune cells, called neutrophils, that are first responders to fight off bacteria and pathogens. This was particularly true for female brains with APOE4 before any signs of cognitive decline. Such findings suggested that APOE4 puts women at greater risk for Alzheimer’s because it triggers changes in neutrophils in the blood, which are then picked up by microglia in the brain.  

“That was a really big discovery,” said Dr. Butovsky. The findings of this “tour de force” collaboration were published this year in Nature Medicine .  

Dr. Butovsky’s research team had previously made headlines in 2017 with a major discovery on how microglia change during these diseases and what drives that change.  

Microglia are not just brain sensors, they are also the janitors responsible for clearing amyloid plaques, Dr. Rosenzweig explains. “Microglia that are healthy should [clean up plaques] and resolve the inflammation quickly. But what happens in Alzheimer's is this process is impaired, and microglia fail to respond to neurodegeneration and to those plaques.” 

The damage to brain cells caused by plaques instigates a change in microglia. These immune cells go from a calm, peace-keeping state to a disease-fighting inflammatory response. In their prolonged attempt to remove plaques, the microglia generate chronic, damaging inflammation that will eventually prevent them from clearing plaques altogether.   

Dr. Butovsky’s seminal publication found that APOE cell signaling drives the change from healthy to diseased microglia in Alzheimer’s and other similar diseases. By identifying a driver, the researchers had uncovered a potential avenue to restore microglia back to health.  

Dr. Rosenzweig and team put this theory to the test in 2023. They showed not only that APOE4 keeps microglia from responding to neurodegeneration—but that she could correct this by selectively deleting this Alzheimer’s risk gene in animal models.  

New Perspectives on Alzheimer’s in Women

It was at this moment that Dr. Rosenzweig had gained enough understanding to turn her sight back to the blood immune cells and Dr. Butovsky’s work with the Dutch Brain Bank. She had also trained with Dr. Michal Schwartz —15 years apart.  

“I'm passionate about how the immune system talks to the brain cells, and how we can harness the good interactions and inhibit the bad interactions,” said Dr. Rosenzweig.  

With support from Alzheimer’s Disease Research, she harnessed findings from the Dutch Brain Bank dataset to explore how sex and APOE4 affect the interaction between microglia and neutrophils.  

Dr. Rosenzweig explains that neutrophils respond to the inflammation caused by plaque buildup in the brain during Alzheimer’s. “That’s when neutrophils come to play—they will slowly infiltrate the brain.”  

“What we found is a mechanism by which those neutrophils make microglia worse,” she continued. The study findings illustrate a communication pathway in APOE4 carriers by which neutrophils release an inflammatory factor, called interleukin-17F, that interacts with a protein on the surface of microglia to suppress their ability to remove plaques. This negative interaction is unique to APOE4 carriers—and especially strong for women.  

The suppressive action of neutrophils on microglia appears in cognitively healthy women with APOE4 , suggesting the gene is driving up Alzheimer’s risk for women by speeding up immune aging. Researchers showed that this interaction also marked the transition to mild cognitive impairment—a stage before Alzheimer’s.     

Whether these findings in neutrophils are a symptom or a driver of Alzheimer’s remains uncertain. However, the early onset of this neutrophil-microglia interaction is important to predicting disease progression and initiation, Dr. Butovsky explained.  

“If what we’re finding is right, well, then it says that neutrophils may be primary to the disease. They’re not just a responder, but the initiator of disease,” he said.  

A Future of Personalized Medicine and Scientific Collaboration 

Personalized medicine is the way of the future. Sex and genetics play a consequential role in how a person will experience disease, including Alzheimer’s.  

These factors can also impact how they respond to treatment, which researchers have already observed with approved Alzheimer’s therapies . As the study authors state, “Female patients were less responsive to [Leqembi] compared to male patients. Moreover, APOE4 carriers showed reduced response to [Leqembi].” Differences in biology spell real-world implications for what will be the most effective approach to treating Alzheimer’s— an approach that is unlikely to be one-size-fits-all. 

Further, the journey to publication was one of extensive collaboration between scientists, physicians, and donors. “This work would not come without all these amazing collaborators around,” said Dr. Butovsky. He aims to keep science moving forward by sharing the data bank he created as a resource for the larger scientific community, so that they may use it to unlock new discoveries.   

“You can see the momentum…to go and really re-explore, re-discover, [and] re-investigate,” said Dr. Butovsky. “This is what we tried 20 years ago, but very limited. With technology today—we’re starting a completely different era in time.” 

Additional Resources:   

• Watch: The Role of Genetics in Early Diagnosis and Treatment of Dementia  
• Eye-Brain Research Uncovers Potential New Way to Treat Glaucoma  
• Alzheimer’s and Inflammation  

About BrightFocus Foundation         

BrightFocus Foundation is a premier global nonprofit funder of research to defeat Alzheimer’s, macular degeneration, and glaucoma. Through its flagship research programs — Alzheimer’s Disease Research, National Glaucoma Research, and Macular Degeneration Research — the Foundation has awarded nearly $300 million in groundbreaking research funding over the past 51 years and shares the latest research findings, expert information, and resources to empower the millions impacted by these devastating diseases. Learn more at  brightfocus.org .  

The information provided in this section is a public service of BrightFocus Foundation, should not in any way substitute for the advice of a qualified healthcare professional, and is not intended to constitute medical advice. Although we make efforts to keep the medical information on our website updated, we cannot guarantee that the information on our website reflects the most up-to-date research.       

Please consult your physician for personalized medical advice; all medications and supplements should only be taken under medical supervision. BrightFocus Foundation does not endorse any medical product or therapy.  

• Krasemann S, Madore C, Cialic R, et al. The TREM2-APOE Pathway Drives the Transcriptional Phenotype of Dysfunctional Microglia in Neurodegenerative Diseases . Immunity . 2017;47(3):566-581.e9. doi:10.1016/j.immuni.2017.08.008
• Yin Z, Rosenzweig N, Kleemann KL, et al. APOE4 impairs the microglial response in Alzheimer's disease by inducing TGFβ-mediated checkpoints. Nat Immunol .  2023;24(11):1839-1853. doi:10.1038/s41590-023-01627-6
• Rosenzweig N, Kleemann KL, Rust T, et al. Sex-dependent APOE4 neutrophil-microglia interactions drive cognitive impairment in Alzheimer's disease . Nat Med . July 3, 2024. doi:10.1038/s41591-024-03122-3

Don't miss out.

Receive Alzheimer's breakthrough news, research updates, and expert information.

Donate to Alzheimer’s Disease Research

Your gift can help lead to treatments and a cure to end Alzheimer’s. Fund the latest, promising research and help provide valuable information to families living with this disease.

I would like to donate

Skip to content

Explore CUIMC

Diversity, equity, and inclusion.

At CUIMC, we are committed to continuous improvement in providing culturally inclusive medical education and clinical care.

Columbia Vagelos College of Physicians and Surgeons is dedicated to developing the next generation of leaders in medicine

Patient Care

• Find a Doctor

Search for a provider by specialty, expertise, location and insurance. Schedule an appointment online.

Read the latest news stories about CUIMC faculty, research, and events

A Cellular Community in the Brain Drives Alzheimer’s Disease

Share this page.

• Share on Facebook
• Share on X (formerly Twitter)
• Share on LinkedIn
• Share by email

An analysis of more than 1.6 million brain cells from older adults has captured the cellular changes that occur in the early stages of Alzheimer’s disease, potentially revealing new routes for preventing the most common cause of dementia in older individuals.

The study also identified a second community of cells that drives the older brain down a different path that does not lead to Alzheimer’s disease.

“Our study highlights that Alzheimer’s is a disease of many cells and their interactions, not just a single type of dysfunctional cell,” says Columbia neurologist Philip De Jager , who led the study with Vilas Menon , assistant professor of neurological sciences at Columbia University Vagelos College of Physicians and Surgeons, and Naomi Habib of the Hebrew University of Jerusalem.

“We may need to modify cellular communities to preserve cognitive function, and our study reveals points along the sequence of events leading to Alzheimer’s where we may be able to intervene.”

Crunching data from 1.6 million brain cells

The study was a technical marvel, cleverly combining new molecular technologies, machine-learning techniques, and a large collection of brains donated by aging adults.

Though previous studies of brain samples from Alzheimer’s patients have provided insights into molecules involved in the disease, they have not revealed many details about where in the long sequence of events leading to Alzheimer’s those genes play a role and which cells are involved at each step of the process.

“Past studies have analyzed brain samples as a whole and they lose all cellular detail,” De Jager says. “We now have tools to look at the brain in finer resolution, at the level of individual cells. When we couple this with detailed information on the cognitive state of brain donors before death, we can reconstruct trajectories of brain aging from the earliest stages of the disease.”

The new analysis required over 400 brains, which were provided by the Religious Orders Study and the Memory & Aging Project based at Rush University in Chicago.

Within each brain, the researchers collected several thousand cells from a brain region impacted by Alzheimer’s and aging. Every cell was then run through a process—single-cell RNA sequencing—that gave a readout of the cell’s activity and which of its genes were active.

Data from all 1.6 million cells were then analyzed by algorithms and machine-learning techniques developed by Menon and Habib to identify the types of cells present in the sample and their interactions with other cells.

“These methods allowed us to gain new insights into potential sequences of molecular events that result in altered brain function and cognitive impairment,” Menon says. “This was only possible thanks to the large number of brain donors and cells the team was fortunate enough to generate data from.”

Aging vs. Alzheimer’s

Because the brains came from people at different points in the disease process, the researchers were able to solve a major challenge in Alzheimer’s research: identifying the sequence of changes in cells involved in Alzheimer’s and distinguishing these changes from those associated with normal brain aging.

“We propose that two different types of microglial cells—the immune cells of the brain—begin the process of amyloid and tau accumulation that define Alzheimer’s disease,” De Jager says.

Then after the pathology has accumulated, different cells called astrocytes play a key role in altering electrical connectivity in the brain that leads to cognitive impairment. The cells communicate with each other and bring in additional cell types that lead to a profound disruption in the way the human brain functions.

“These are exciting new insights that can guide innovative therapeutic development for Alzheimer’s and brain aging,” De Jager says.

“By understanding how individual cells contribute to the different stages of the disease, we will know the best approach with which to reduce the activity of the pathogenic cellular communities in each individual, returning brain cells to their healthy state,” De Jager says.

Additional information

Top image of astrocytes and neurons in a dish by Kevin Richetin, University of Lausanne, CC BY-NC-ND 2.0 .

Philip De Jager, MD, PhD, is the Weil-Granat Professor of Neurology at Columbia University Vagelos College of Physicians and Surgeons, where he is also director of the Center for Translational and Computational Neuroimmunology, director of the Multiple Sclerosis Center, and deputy director of the Taub Institute for Research on Alzheimer's Disease and the Aging Brain.

Vilas Menon, PhD, is assistant professor in the Department of Neurology and Taub Institute for Research on Alzheimer Disease and the Aging Brain at Columbia University Vagelos College of Physicians and Surgeons.

The paper, “ Cellular communities reveal trajectories of brain aging and Alzheimer’s disease ,” was published Aug. 28 in Nature.

All authors: Gilad Sahar Green (Hebrew University of Jerusalem), Masashi Fujita (Columbia), Hyun-Sik Yang (Harvard), Mariko Taga (Columbia), Anael Cain (Hebrew University of Jerusalem), Cristin McCabe (Broad Institute), Natacha Comandante-Lou (Columbia), Charles C. White (Broad Institute), Anna K. Schmidtner (Hebrew University of Jerusalem), Lu Zeng (Columbia), Alina Sigalov (Columbia), Yangling Wang (Rush University Medical Center), Aviv Regev (Genentech), Hans-Ulrich Klein (Columbia), Vilas Menon (Columbia), David A. Bennett (Rush University Medical Center), Naomi Habib, (Hebrew University of Jerusalem), and Philip L. De Jager (Columbia).

The work was supported by the NIH (grants RF1AG057473, U01AG061356, U01AG046152, R01AG070438, R01AG015819, U01AG072572, R01AG066831, K23AG062750, K23AG062750, and R01AG080667), the Chan Zuckerberg Initiative (CS-02018-191971), the Israel Science Foundation (1709/19), the European Research Council (853409), Chinese Ministry of Science and Technology (3-15687), the Myers Foundation, Alzheimer’s Association (ADSF-21-816675), and a Minerva Fellowship of the Minerva Stiftung Gesellschaft fuer die Forschung mbH.

The Columbia authors declare no competing interests.

• Search Menu
• Sign in through your institution
• Advance articles
• Editor's Choice
• Continuing Education
• Author Guidelines
• Submission Site
• Open Access
• Why publish with this journal?
• About Archives of Clinical Neuropsychology
• About the National Academy of Neuropsychology
• Journals Career Network
• Editorial Board
• Advertising and Corporate Services
• Self-Archiving Policy
• Dispatch Dates
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

Introduction, conflict of interest, acknowledgements, author contributions, data availability.

• < Previous

BDNF Val66Met moderates episodic memory decline and tau biomarker increases in early sporadic Alzheimer’s disease

Data used in preparation of this article were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database ( adni.loni.usc.edu ). As such, the investigators within the ADNI contributed to the design and implementation of ADNI and/or provided data but did not participate in analysis or writing of this report. A complete listing of ADNI investigators can be found at: http://adni.loni.usc.edu/wp-content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf

• Article contents
• Figures & tables
• Supplementary Data

Diny Thomson, Emily Rosenich, Paul Maruff, Yen Ying Lim, for the Alzheimer’s Disease Neuroimaging Initiative, BDNF Val66Met moderates episodic memory decline and tau biomarker increases in early sporadic Alzheimer’s disease, Archives of Clinical Neuropsychology , Volume 39, Issue 6, September 2024, Pages 683–691, https://doi.org/10.1093/arclin/acae014

• Permissions Icon Permissions

Allelic variation in the brain-derived neurotrophic factor ( BDNF ) Val66Met polymorphism has been shown to moderate rates of cognitive decline in preclinical sporadic Alzheimer’s disease (AD; i.e., Aβ + older adults), and pre-symptomatic autosomal dominant Alzheimer’s disease (ADAD). In ADAD, Met66 was also associated with greater increases in CSF levels of total-tau (t-tau) and phosphorylated tau (p-tau 181 ). This study sought to determine the extent to which BDNF Val66Met is associated with changes in episodic memory and CSF t-tau and p-tau 181 in Aβ + older adults in early-stage sporadic AD.

Aβ + Met66 carriers ( n  = 94) and Val66 homozygotes ( n  = 192) enrolled in the Alzheimer’s Disease Neuroimaging Initiative who did not meet criteria for AD dementia, and with at least one follow-up neuropsychological and CSF assessment, were included. A series of linear mixed models were conducted to investigate changes in each outcome over an average of 2.8 years, covarying for CSF Aβ 42 , APOE ε4 status, sex, age, baseline diagnosis, and years of education.

Aβ + Met66 carriers demonstrated significantly faster memory decline ( d  = 0.33) and significantly greater increases in CSF t-tau ( d  = 0.30) and p-tau 181 ( d  = 0.29) compared to Val66 homozygotes, despite showing equivalent changes in CSF Aβ 42 .

These findings suggest that reduced neurotrophic support, which is associated with Met66 carriage, may increase vulnerability to Aβ-related tau hyperphosphorylation, neuronal dysfunction, and cognitive decline even prior to the emergence of dementia. Additionally, these findings highlight the need for neuropsychological and clinicopathological models of AD to account for neurotrophic factors and the genes which moderate their expression.

The development and application of biomarkers of beta-amyloid (Aβ) and tau in natural history studies of aging and dementia shows that Alzheimer’s disease (AD) pathology can emerge up to 30 years prior to individuals meeting any clinical criteria for dementia ( Villemagne , et al. , 2013 ). Careful prospective neuropsychological studies demonstrate that the pre-dementia stage of AD is characterized by a subtle but relentless decline in episodic memory ( Collie , & , Maruff , 2000 ; Pike , et al. , 2007 ; Twamley , et al. , 2006 ). Greater precision in neuropsychological models of early AD could therefore provide a basis for the detection and management of the disease in older adults who are not demented but who have elevated levels of AD biomarkers. One method of improving brain-behavior models is to examine how variation in AD biology (e.g., levels of Aβ), demographic characteristics (e.g., sex) or variation in genes (e.g., carriage of the apolipoprotein e4 allele [ APOE ε4) can influence the memory decline that occurs in early AD ( Buckley , et al. , 2018 ; Lim , et al. , 2014b ; Lim , et al. , 2017b ). There is now strong and well-replicated evidence that allelic variation in the brain-derived neurotrophic factor ( BDNF) Val66Met (rs6265) polymorphism exerts substantial influence on rates of neurodegeneration and cognitive decline in adults with elevated levels of Aβ (Aβ+) ( Boots , et al. , 2017 ; Franzmeier , et al. , 2021 ; Lim , et al. , 2013 ; Lim , et al. , 2014a ; Lim , et al. , 2015 ; Lim , et al. , 2016 ; Lim , et al. , 2017a ; Lim , et al. , 2018 ; Lim , et al. , 2021 ; Lim , et al. , 2022 ; van , den , Bosch , et al. , 2021 ). Understanding of relationships between BDNF Val66Met and changes in episodic memory and AD biomarkers could therefore inform neuropsychological models of early AD.

BDNF is expressed widely in the central nervous system (CNS) and is important for long-term potentiation and synaptic plasticity. Approximately 30% of the population carry the BDNF Met66 allele ( Shen , et al. , 2018 ), which is associated with reduced activity-dependent secretion of BDNF and specific impairments in hippocampal-dependent encoding and retrieval processes, hippocampal volume, and episodic memory in both cognitively normal samples ( Egan , et al. , 2003 ; Erickson , et al. , 2010 ; Hariri , et al. , 2003 ) and across the spectrum of AD. For example, when compared to compared to matched Aβ + Val66 homozygotes, non-demented Aβ + older adults who carry the Met66 allele have shown an ~ 18% greater episodic memory decline over 3–10 years ( Boots , et al. , 2017 ; Lim , et al. , 2013 ; Lim , et al. , 2014a ; Lim , et al. , 2021 ; van , den , Bosch , et al. , 2021 ) and an ~ 12% greater decrease in hippocampal volume over 3 years ( Lim , et al. , 2013 ; Lim , et al. , 2014a ), despite showing equivalent rates of Aβ accumulation. Non-demented Aβ + older adult Met66 carriers have also shown faster declines in other neuropsychological domains (e.g., executive function, language) across 3 years compared to Val66 homozygotes ( Lim , et al. , 2013 ). In pre-symptomatic autosomal dominant Alzheimer’s disease (ADAD), a quantitatively similar acceleration of Aβ + related neurodegeneration and cognitive decline also occurred in those who also carried the Met66 allele ( Lim , et al. , 2018 ), again, with equivalent decreases in levels of soluble Aβ 42 in CSF ( Lim , et al. , 2018 ; Lim , et al. , 2022 ). Further exploration of these relationships in ADAD indicated that Met66 carriage was also associated with a ~ 41% increase in CSF total tau (t-tau) and tau phosphorylated at threonine 181 (p-tau 181 ) over three years ( Lim , et al. , 2018 ).

In humans and animals, carriage of a Met66 allele is associated with reduced CNS levels of BDNF ( Egan , et al. , 2003 ). Hence, in pre-symptomatic ADAD, reduced CNS BDNF may allow the faster accumulation of tau and a consequent acceleration of symptoms ( Lim , et al. , 2018 ; Lim , et al. , 2022 ). This hypothesis accords with data from in vitro studies showing that reduction of BDNF in AD is specific to tangle-bearing neurons ( Ferrer , et al. , 1999 ; Murer , et al. , 1999 ), and with animal and post-mortem studies showing that the extent of BDNF reduction in the hippocampus was related to the magnitude of cognitive impairment ( Connor , et al. , 1997 ; Egan , et al. , 2003 ). However, it is possible that the relationship between Met66 carriage and higher tau levels observed in ADAD was a consequence of their younger age, more aggressive form of AD, or their carriage of mutations in PSEN -1, PSEN -2, or APP genes. Therefore, to understand the importance of BDNF in AD, it is necessary to determine whether BDNF Met66 carriage is related to increases in tau and memory decline in the pre-dementia stages of sporadic AD.

The aim of this study was to determine the extent to which Met66 carriage is related to rates of change in episodic memory, and in CSF t-tau, p-tau 181 , and Aβ 42 , in Aβ + older adults who do not meet clinical criteria for dementia. The hypothesis was that non-demented Aβ + Met66 carriers would show greater decline in episodic memory, and faster increases in CSF t-tau and p-tau 181 when compared to Val66 homozygotes, despite showing equivalent changes in CSF Aβ 42 .

Participants

Data used in the preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database ( adni.loni.usc.edu ). The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging, positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of mild cognitive impairment (MCI) and early AD. The ADNI study has consisted of four phases: ADNI-1, ADNI-GO, ADNI-2, and, most recently, ADNI-3. For up-to-date information, see www.adni-info.org .

Recruitment processes and inclusion/exclusion criteria for ADNI have been described in detail previously ( Petersen , et al. , 2010 ). Broadly, participants were included in ADNI if they were aged 55–90 years and did not have any significant physical, psychiatric or neurological disorders other than AD. At study entry, participants could be cognitively normal or meet clinical criteria for MCI or AD. Cognitive normality was determined by the absence of cognitive complaints, a Mini-Mental State Examination (MMSE) score of 24–30, a CDR score of 0, and a score of ≥9 (i.e., 16 years of education) or ≥ 5 (e.g., 8–15 years of education) on the Logical Memory II (Delayed Recall) subscale. MCI was distinguished by an MMSE score of 24–30 and a CDR score of 0.5 with the memory box score being ≥0.5. AD was further distinguished by a CDR of 0.5 or 1. Diagnoses of MCI and AD required scores of ≤8 (i.e., 16 years of education), ≤4 (8–15 years of education), or ≤ 2 (i.e., 0–7 years of education) on the Logical Memory II (Delayed Recall) subscale ( Petersen , et al. , 2010 ).

ADNI participants were included in the current study if they had provided CSF samples and completed neuropsychological assessments on at least two assessment timepoints, were classified as CSF Aβ+, and did not meet clinical criteria for AD dementia during the visit closest to their first lumbar puncture. Only ADNI-1, ADNI-2, and ADNI-GO participants were included in this study.

Genotype data for the BDNF Val66Met (rs6265) polymorphism was extracted using PLINK, an open-source program for analysing whole genome data ( Purcell , et al. , 2007 ). Genetic polymorphisms were not used diagnostically. Genotype data were cleaned by applying a minimum call rate for single-nucleotide variations (SNVs, formerly SNPs) and individuals (98%); SNVs not in Hardy–Weinberg equilibrium ( p  < 1 × 10–6) were excluded. No SNVs were removed because of low minor allele frequency. BDNF genotype was blind to all neuropsychological raters.

CSF biomarkers

Fasted CSF samples were collected via lumbar puncture on the morning of assessments, following procedures described previously ( Shaw , et al. , 2009 ). In summary, CSF concentrations CSF t-tau, CSF p-tau 181 and Aβ 42 were measured using Luminex bead-based multiplexed xMAP technology immunoassay (INNO-BIA Alzbio3; Innogenetics). All samples were shipped on dry ice to the ADNI Biomarker Core laboratory at the University of Pennsylvania Medical Center. Aβ + was classified when CSF Aβ 42 levels were below the previously validated cutpoint value of 980 pg/mL ( Hansson , et al. , 2018 ).

ADNI memory composite

Detailed methods for deriving the ADNI memory (ADNI-Mem) composite have been described previously ( Crane , et al. , 2012 ). Briefly, the ADNI-Mem composite is derived from a longitudinal single factor model that incorporates existing verbal episodic memory measures from the ADNI neuropsychological battery: The Rey Auditory Verbal Learning Test (RAVLT; Schmidt , 1996 ), the Alzheimer’s Disease Assessment Scale-Cognitive Subscale (ADAS-Cog; Rosen , et al. , 1984 ), three-word recall from the MMSE ( Folstein , et al. , 1975 ), and the Logical Memory subtest of the Wechsler Memory Scale. The composite score was originally standardized based on a sample of 803 eligible ADNI participants with longitudinal cognitive data available. It has a mean of 0 and variance of 1 ( Crane , et al. , 2012 ). Use of the composite score, as opposed to individual test scores, can circumvent challenges associated with variability in ADNI’s neuropsychological test battery (e.g., throughout the course of ADNI, two versions of the RAVLT and three versions of the ADAS-Cog were used), changes in tests used over time, and in handling missing data ( Crane , et al. , 2012 ).

Standard protocol approvals, registrations, and patient consent

Institutional review boards of all participating institutions of ADNI provided approval for the study. All participants at each site provided written informed consent prior to the commencement of any study procedures.

Data analysis

Analyses were performed in R v.4.1.3. using the following packages: “ggplot2”, “psych”, “gmodels”, “lme4”, “lmerTest”, “emmeans”, “effects”, “dplyr”, “multcomp”, “Hmisc”, “cowplot”, “tidyr”, “car”, “stringr”, “reshape2” and “lubridate”.

Baseline demographic differences between Met66 carriers and Val66 homozygotes were assessed using linear regressions for continuous variables and chi-squared tests for categorical variables. “Baseline” was defined as the first visit at which participants’ CSF samples were collected. An independent samples t -test was also conducted to assess differences in baseline CSF Aβ 42 levels. For longitudinal outcomes of interest (episodic memory, CSF t-tau, CSF p-tau 181 ,) a series of analyses of covariance (ANCOVA) were conducted with APOE ε4 status, sex, age, baseline diagnosis, years of education and CSF Aβ 42 levels were entered as covariates. To further explore episodic memory performance on the ADNI-Mem, an ANCOVA was also conducted using data derived from participants’ final visit. A change score (i.e., between baseline visit and final visit) was extracted by subtracting scores at baseline from scores at follow-up.

To test the hypothesis that Aβ + Met66 carriers would show faster declines in episodic memory and greater increases in CSF t-tau and p-tau 181 compared to Val66 homozygotes, despite showing equivalent changes in CSF Aβ 42 , a series of linear mixed effects models (LMM) were conducted for each outcome variable (i.e., episodic memory, CSF t-tau, CSF p-tau 181 , CSF Aβ 42 ,). LMMs were chosen due to their robustness to missing data in a longitudinal design, ability to account for individual variability in change across time, and as they account for fixed and random effects ( Verbeke , 1997 ).

Firstly, to investigate the extent to which Met66 carriage influenced change in Aβ over time compared to Val66 homozygosity, CSF Aβ 42 data was included as the dependent variable in an LMM including APOE e4 status, sex, age, baseline diagnosis and years of education as covariates. “Participant” was entered as a random effect. Then, to examine the extent to which Met66 carriage influenced change in episodic memory, CSF t-tau, CSF p-tau 181 and CSF Aβ 42 over time compared to Val66 homozygosity, data for each of these outcomes were included in a series of equivalent LMMs, with baseline CSF Aβ 42 as an additional covariate. Means (and standard errors, SE) of slopes were extracted to calculate the magnitude of difference in the rate of change in each biomarker and cognitive outcome between Met66 carriers and Val66 homozygotes, measured using Cohen’s d. Small (<0.20), medium (0.30–0.70) and large (>0.80) effect sizes were interpreted according to established convention ( Cohen , 2016 ).

Demographic and clinical characteristics

A total of 286 participants (94 Met66 carriers, 192 Val66 homozygotes) from ADNI-1, ADNI-GO and ADNI-2 met criteria for inclusion in the analysis. Table 1 summarizes the demographic and clinical characteristics of Aβ + Met66 carriers and Val66 homozygotes at baseline. Groups did not differ significantly on any clinical or demographic measure, except for their baseline scores on the ADNI-Mem, where Met66 carriers showed significantly poorer performance, of a moderate magnitude, compared to Val66 homozygotes ( d  = 0.30). On average, participants were followed for 2.8 years ( SD  = 1.82, range  = 0.70–10.21), including an average of 2.71 assessments ( SD  = 1.13, range  = 2–7).

Baseline demographic, clinical and biological characteristics of Aβ + non-demented Met66 carriers and Val66 homozygotes

Note : Bold values emphasize statistical significance at p  < 0.05.

Abbreviations. ADNI-Mem = ADNI memory composite; APOE ε4 = apolipoprotein E; MCI = mild cognitive impairment; SD = standard deviation; GDS = Geriatric Depression Scale; CDR = Clinical Dementia Rating Scale; CDR-SOB = Clinical Dementia Rating Scale Sum of Boxes; MMSE = Mini-Mental State Examination; CSF = cerebrospinal fluid; t-tau = total tau; p-tau 181  = phosphorylated tau at threonine 181; ADNI = Alzheimer’s Disease Neuroimaging Initiative.

Differences over time between Aβ + Met66 carriers and Val66 homozygotes on episodic memory and AD biomarkers

The results of the LMM indicated a significant BDNF × time interaction for episodic memory, whereby Aβ + Met66 carriers showed significantly greater decline in episodic memory, measured by the ADNI-Mem composite, compared to Met66 carriers over an average of 2.8 years ( Table 2 ; Figure 1 ). The magnitude of this difference was, by convention, moderate ( d  = 0.33). Further examination of this difference indicated that although Met66 carriers performed worse than Val66 homozygotes on the ADNI-Mem composite at baseline ( Table 1 ), Met66 carriers also performed significantly worse than Val66 homozygotes ( Table 3 ) at the final visit, even after accounting for the number of visits. When performance on the ADNI-Mem composite was considered as a change from baseline to final visit, Met66 carriers also showed greater decline than Val66 homozygotes ( Table 3 ). The magnitude of this difference in the rate of episodic decline is equivalent to that in the difference in slopes derived from the LMM.

Summary of results from the LMMs exploring two-way interactions between BDNF Val66Met × time on episodic memory (measured by ADNI-Mem), CSF t-tau, p-tau 181 and Aβ 42 , outcomes over 10 years in Aβ + Met66 carriers and Val66 homozygotes

Note: Bold values emphasize statistical significance at p  < 0.05; Covariates: CSF Aβ 42 (except for CSF Aβ 42 outcome), APOE ε4 status, sex, age, baseline diagnosis, years of education.

Abbreviations. ADNI-Mem = ADNI memory composite; BDNF  = brain-derived neurotrophic factor; Est. = estimate, SE = standard error; CSF = cerebrospinal fluid; t-tau = total tau; p-tau 181  = phosphorylated tau at threonine 181; Aβ 42  = amyloid beta peptide 42; ADNI = Alzheimer’s Disease Neuroimaging Initiative; pg/mL = picograms per milliliter.

Rates of Change in (A) CSF t-tau, (B) ptau181, (C) Aβ 42 , and (D) Episodic Memory (Measured by the ADNI Memory Composite) Over 10 Years in Non-Demented Older Adult Aβ+ Met66 Carriers and Val66 Homozygotes, Adjusted for CSF Aβ 42 (Except for the CSF Aβ 42 Outcome), APOE ε4 Status, Sex, Age, Baseline Diagnosis, and Years of Education.

Summary of Met66 carriers’ and Val66 homozygotes’ memory performance on the ADNI-Mem at the final visit, including change between baseline and final visit and the modelled slope derived from the linear mixed model

Note : SD = standard deviation; CI = confidence interval; LMM = linear mixed model

The LMMs also indicated a significant BDNF × time interaction for CSF t-tau and p-tau 181 , whereby Aβ + Met66 carriers showed significantly greater increases in CSF t-tau ( d  = 0.30) and p-tau 181 ( d  = 0.29) compared to Val66 homozygotes over an average of 2.8 years ( Table 2 ; Figure 1 ). The magnitude of these differences in rates of change were, by convention, moderate. Met66 carriers did not show a statistically greater decrease in CSF Aβ 42 compared to Val66 homozygotes, with the magnitude of this difference small ( d  = 0.23).

Supporting our hypothesis, the BDNF Val66Met polymorphism exerted a substantial influence on the rate of decline in episodic memory and increases in CSF levels of t-tau and p-tau 181 over an average of 2.8 years. Despite showing equivalent changes in CSF Aβ 42 , non-demented Aβ + Met66 carriers showed not only poorer memory performance at baseline compared to Val66 homozygotes ( d  = 0.30), but also showed faster memory decline compared to Val66 homozygotes (~15%). Similarly, Aβ + Met66 carriers showed faster increases in CSF t-tau (~12%) and p-tau 181 (~15%) compared to Val66 homozygotes. These findings are consistent with observations made in ADAD over the same time period, where Met66 was also associated with an ~ 29% greater decline in episodic memory, and ~ 41% greater increase in CSF t-tau and p-tau 181 when compared to ADAD Val66 homozygotes ( Lim , et al. , 2018 ). These results further support the hypothesis that BDNF Val66Met moderates the effects of Aβ on AD clinical disease progression via its effect on tau ( Lim , et al. , 2018 ; Lim , et al. , 2022 ) by showing this effect on tau in the sporadic form of AD. Despite Aβ + being the hallmark characteristic of AD, younger Aβ- Met66 carriers do not show any change in CSF Aβ 42 , CSF tau, brain volume or cognition ( Lim , et al. , 2018 ). Therefore, these findings show that in non-demented Aβ + older adults, variation in the BDNF Val66Met polymorphism has a clinically important influence on increases in tau, neurodegeneration, and memory decline.

The magnitude of difference in the rate of episodic memory decline between groups is consistent with previous observations in cross-sectional and longitudinal studies in the preclinical ( Boots , et al. , 2017 ; Lim , et al. , 2013 ) and prodromal ( Lim , et al. , 2014a ; Lim , et al. , 2018 ) stages of sporadic AD, as well as in pre-symptomatic ADAD ( Lim , et al. , 2016 ; Lim , et al. , 2021 ; Lim , et al. , 2022 ). Although hippocampal volume was not measured in this study, previous studies show that hippocampal atrophy is~14% faster in Aβ + Met66 carriers compared to Val66 homozygotes across preclinical and prodromal sporadic AD ( Lim , et al. , 2013 ; Lim , et al. , 2014a ) and pre-symptomatic ADAD ( Lim , et al. , 2018 ). A recent study conducted in preclinical, prodromal, and clinical stages of AD dementia, conducted over 12.5 years, suggested that the nature of memory decline in Met66 carriers may change with disease severity. Specifically, although Met66 carriage was associated with faster memory decline in preclinical and prodromal AD, in the clinical stages of AD dementia, memory decline is only evident in Val66 homozygotes, as memory performance in Met66 carriers had reached the lowest possible values (i.e., a floor effect; Lim , et al. , 2021 ). This likely reflects that any protective effects conferred by Val66 homozygosity on synaptic plasticity and neuronal survival become limited once individuals progress to AD dementia. The sample of Aβ + individuals with clinical AD dementia, BDNF genotyping, and CSF biomarker outcomes in ADNI was not sufficiently large ( n  = 64; 36% Met carriers) to test this interaction. As such, future studies using larger samples of individuals with AD dementia are necessary to clarify how neurotrophic factors influence tau levels in advanced AD. Additionally, although memory differences in Aβ + Met66 carriers and Val66 homozygotes have been consistently observed, the magnitude of difference in memory decline between groups remains moderate in the magnitude, suggesting that their utility in clinical practice will be limited. However, the results of this study do inform current neuropsychological models of AD by confirming that subtle memory declines are related to accumulation of tau, and that this is hastened when there is reduced expression of growth factors such as BDNF.

Recent advances in understanding of tau biology have allowed the development of models of tau kinetics according to site-specific levels in mass-spectrometry measured tau phosphorylation. It is now agreed that site-specific tau phosphorylation may reflect different clinical stages of AD. For example, CSF tau phosphorylation occupancy at threonine 181 and 217 (p-tau 217 ) increases with initial Aβ accumulation, although phosphorylation occupancy at threonine 205 (p-tau 205 ) increases only when brain atrophy and clinical symptoms emerge ( Barthélemy , et al. , 2020 ). Recently, we showed that these different levels of tau phosphorylation were also influenced by Met66 carriage. In pre-symptomatic ADAD mutation carriers, Met66 carriers showed greater CSF p-tau 181 and p-tau 217 phosphorylation compared to Val66 homozygotes, but equivalent levels of p-tau 205 phosphorylation ( Lim , et al. , 2022 ). Conversely, in symptomatic mutation carriers, Met66 carriers showed greater levels of CSF t-tau and p-tau 205 phosphorylation compared to Val66 homozygotes, but equivalent phosphorylation levels of CSF p-tau 181 and p-tau 217 . The association between Met66 carriage and site-specific tau phosphorylation suggests that Met66 carriers experience greater disease progression in early stages of AD. This is reflected by greater initial increases in p-tau 181 and p-tau 217 phosphorylation relative to Val66 homozygotes, and followed by greater increases in t-tau and p-tau 205 phosphorylation as neuronal dysfunction increases ( Barthélemy , et al. , 2020 ). Measures of p-tau 217 and p-tau 205 are not yet available for the ADNI sample. Future studies are needed to confirm the hypothesis that changes in levels of p-tau 217 and p-tau 205 isoforms reflect different stages of clinical disease severity in sporadic AD.

The processes by which BDNF Val66Met influences tau hyperphosphorylation and episodic memory decline, even prior neurofibrillary tangle formation, has not been fully elucidated. Initially, research in animal and cellular models postulated that tau was responsible for downregulating BDNF expression ( Rosa , et al. , 2016 ). However, if downregulation of BDNF by tau was the driver of cognitive decline and neurodegeneration observed in clinical studies, human Met66 carriers and Val66 homozygotes should show equivalent increases in CSF tau in early AD. This was not observed in the current or previous studies ( Lim , et al. , 2018 ). On the contrary, the clinical data suggest that as BDNF Met66 reduces BDNF availability in the CNS, this reduction in neurotrophic factors may allow faster Aβ + related tau hyperphosphorylation, subsequent neurodegeneration and cognitive decline. Animal studies do show that loss of BDNF may mediate neurotoxicity of tau downstream of abnormal increases in Aβ ( Rosa , et al. , 2016 ). Lower levels of circulating BDNF can also antagonize the major receptor site for BDNF, tropomyosin receptor kinase B (TrkB), which can precipitate rapid tau hyperphosphorylation and subsequent synaptic dysfunction and neuronal degeneration ( Elliott , et al. , 2005 ; Xiang , et al. , 2019 ). Additionally, BDNF Val66Met may moderate neuroinflammatory responses that affect vulnerability to the downstream effects of Aβ. In vitro studies show that astrocytes can quickly increase expression of BDNF in response to increasing levels of Aβ ( Kimura , et al. , 2006 ), which may be an attempt to protect neurons against AD pathogenesis ( Faria , et al. , 2014 ). Thus, one possible integration of these observations is that lower BDNF availability associated with Met66 carriage could reduce resilience to inflammatory processes in early disease stages.

The current finding that BDNF Val66Met influences increases in tau in Aβ + older adults, which is consistent with previous findings across different cohorts with varying disease severity and Aβ aetiology, have important implications for the field. First, the effect of the Met66 allele on cognitive decline, at least in the pre-dementia stages of the disease, is substantial [~8%–~40% in preclinical sporadic AD studies ( Boots , et al. , 2017 ; Lim , et al. , 2013 ; Lim , et al. , 2021 ; van , den , Bosch , et al. , 2021 ); ~6%–21% in prodromal sporadic AD studies ( Lim , et al. , 2014a ; Lim , et al. , 2021 ); ~29% in pre-symptomatic ADAD ( Lim , et al. , 2018 )], and, therefore, understanding of this effect should inform AD clinicopathological models. Second, the Met66 allele is very common (e.g., 33% of the white population), and given the centrality of Aβ and tau models in AD and in the development of pharmacotherapeutics, the potential influence of variation in BDNF Val66Met on study outcomes should be considered. Third, given the frequency of the polymorphism, the magnitude of its effects on tau and cognitive outcomes, and the absence of effects on Aβ, variation on the BDNF Val66Met polymorphism could provide a useful clinical tool for challenging knowledge about the extent to which other fluid or imaging biomarkers are associated with Aβ or tau accumulation. For example, examination of the effect of BDNF Val66Met on markers of neuroinflammation or synaptic function may further clarify its role in AD clinical progression. Finally, these findings suggest that loss of BDNF that occurs downstream of abnormal increases in Aβ may increase vulnerability to neuronal dysfunction via tau and precipitate accelerated clinical disease progression in pre-dementia stages of AD. Conversely, these findings also suggest that greater BDNF availability can protect neurons, as well as the cognitive functions that depend on them, from Aβ-related cell death. Thus, pharmacologically increasing neurotrophic support in the early stages of AD may be a potential therapeutic target to delay the clinical manifestation of AD dementia ( Lu , et al. , 2013 ). For example, in vitro studies suggest that agonist antibodies (e.g., AS86) that target TrkB promote synaptic growth and repair ( Wang , et al. , 2020 ). Studies in rodents also suggest that TrkB receptor agonists such as 7,8-dihydroxyflavone (7,8-DHF) can block delta-secretase activation, attenuate AD pathology, and reduces cognitive dysfunction ( Chen , et al. , 2021 ). It is important to note, however, that despite the promising therapeutic potential of these TrkB agonists, their safety, and efficacy in humans remains unclear.

Some limitations are noted which restrict the generalizability of the current results. First, although the maximum follow-up time was 10 years, the average follow-up time was only 2.8 years. This limited the ability to investigate the role of BDNF Val66Met on clinical disease progression. Second, cognitively normal and MCI participants were combined to enable examination across a more well-powered sample of Aβ + older adults. Although baseline diagnosis was controlled statistically in the statistical modelling, it will be important for future studies with larger sample sizes to clarify the role of BDNF Val66Met on changes in tau and memory across the preclinical, prodromal, and dementia stages of AD. Alternatively, harmonization of multiple prospective cohort studies may enable a sufficiently powered study of the effects of BDNF Val66Met on disease progression at each stage of AD. This would allow for more detailed investigations into whether BDNF Val66Met’s effects on CSF tau biomarkers differ across populations and whether the pattern of memory decline shown in Met66 carriers can be characterized by impairments in specific components of memory across the disease course of AD (e.g., delayed recall, recognition). Given that the Met66 allele is present in ~30% of the white population ( Shen , et al. , 2018 ), a further benefit of expanding the sample to include cohorts is the ability to explore the role of BDNF Val66Met across various racial and ethnic groups. This would be informative as the prevalence of the Met66 allele varies across populations (e.g., approximately 70% of individuals from Asian populations carry the Met66 allele; Shen , et al. , 2018 ). A third limitation of the study is that CSF measures of Aβ 42 can be susceptible to variability in pre-analytical handling of CSF samples ( Fourier , et al. , 2015 ). There is also variability in handling of CSF samples across ADNI cohorts. As such, there has been movement towards using a ratio of Aβ 42 /Aβ 40 for a more sensitive classification of Aβ positivity. Although this ratio was not available for all participants in the current study, it will be important for studies to test this in the future. Finally, although CSF markers of tau provide a single value reflecting neuronal death and neurofibrillary tangle formation, they cannot determine the topography of tau pathology in the brain ( Jack , et al. , 2016 ). It will thus be important for future studies to examine whether changes in tau PET are similar to the changes in CSF tau and p-tau 181 observed in this study.

These limitations notwithstanding, the results of this study provide support for the hypothesis that BDNF Val66Met moderates downstream effects of Aβ + on tau and memory decline in preclinical and prodromal sporadic AD. This study also illuminates that brain areas necessary for episodic memory are dependent on growth factors, such as BDNF, and that factors which compromise the availability of BDNF, such as BDNF Met66 carriage, give way to increases in CSF tau and accelerated episodic memory decline. Consistent with other studies in Aβ + older adults at risk of sporadic AD and adults with pre-symptomatic ADAD, subtle episodic memory declines can be detected in the long pre-dementia phase of AD, even prior to meeting clinical criteria for dementia. Taken together, these data show that memory loss in Aβ + adults is related to loss of neuronal function, and that sufficient availability of neuronal growth factors may forestall tau hyperphosphorylation, accelerated neuronal dysfunction and memory decline in the pre-dementia stages of sporadic AD. Given the long pre-dementia stage of AD, these findings highlight the opportunity for more precise detection and management of memory decline in older adults who present with abnormal AD biomarkers and risk factors for reduced growth factor availability.

Data collection and sharing for this project was funded by the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH-12-2-0012). ADNI is funded by the National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering, and through generous contributions from the following: AbbVie, Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private sector contributions are facilitated by the Foundation for the National Institutes of Health ( www.fnih.org ). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer’s Therapeutic Research Institute at the University of Southern California. ADNI data are disseminated by the Laboratory for Neuro Imaging at the University of Southern California.

DT is supported by an Australian Government Research Training Program Scholarship. ER is supported by an Alzheimer’s Association Research Fellowship (23AARF-1025519). YYL is supported by an NHMRC Career Development Fellowship (GNT1162645), and Emerging Leadership Grant (GNT2009550).

None declared.

We thank all ADNI participants for their commitment and dedication to helping advance research into the early detection and causation of AD.

Diny Thomson (Conceptualization, Formal analysis, Methodology, Writing—original draft, Writing—review & editing), Emily Rosenich (Writing—review & editing), Paul Maruff (Conceptualization, Writing—review & editing), Yen Ying Lim (Conceptualization, Formal analysis, Methodology, Writing—review & editing).

Genetic, biomarker and clinical data from ADNI are publicly accessible and are available through a formal application on http://adni.loni.usc.edu .

Barthélemy , N. R. , Li , Y. , Joseph-Mathurin , N. , Gordon , B. A. , Hassenstab , J. , Benzinger , T. L. S. , et al.  ( 2020 ). A soluble phosphorylated tau signature links tau, amyloid and the evolution of stages of dominantly inherited Alzheimer's disease . Nature Medicine , 26 ( 3 ), 398 – 407 . https://doi.org/10.1038/s41591-020-0781-z .

Google Scholar

Boots , E. A. , Schultz , S. A. , Clark , L. R. , Racine , A. M. , Darst , B. F. , Koscik , R. L. , et al.  ( 2017 ). BDNF Val66Met predicts cognitive decline in the Wisconsin Registry for Alzheimer's Prevention . Neurology , 88 ( 22 ), 2098 – 2106 . https://doi.org/10.1212/WNL.0000000000003980 .

van den Bosch , K. A. , Verberk , I. M. W. , Ebenau , J. L. , van der Lee , S. J. , Jansen , I. E. , Prins , N. D. , et al.  ( 2021 ). BDNF-Met polymorphism and amyloid-beta in relation to cognitive decline in cognitively normal elderly: The SCIENCe project . Neurobiology of Aging , 108 , 146 – 154 . https://doi.org/10.1016/j.neurobiolaging.2021.08.018 .

Buckley , R. F. , Mormino , E. C. , Amariglio , R. E. , Properzi , M. J. , Rabin , J. S. , Lim , Y. Y. , et al.  ( 2018 ). Sex, amyloid, and APOE ε4 and risk of cognitive decline in preclinical Alzheimer's disease: Findings from three well-characterized cohorts . Alzheimer's & Dementia , 14 ( 9 ), 1193 – 1203 . https://doi.org/10.1016/j.jalz.2018.04.010 .

Chen , C. , Ahn , E. H. , Liu , X. , Wang , Z.-H. , Luo , S. , Liao , J. , et al.  ( 2021 ). Optimized TrkB agonist ameliorates Alzheimer’s disease pathologies and improves cognitive functions via inhibiting delta-secretase . ACS Chemical Neuroscience , 12 ( 13 ), 2448 – 2461 . https://doi.org/10.1021/acschemneuro.1c00181 .

Cohen , J. ( 2016 ). A power primer. In Kazdin , A. E. (Ed.), Methodological issues and strategies in clinical research (pp. 279 – 284 ). Washington, DC: American Psychological Association .

Google Preview

Collie , A. , & Maruff , P. ( 2000 ). The neuropsychology of preclinical Alzheimer's disease and mild cognitive impairment . Neuroscience & Biobehavioral Reviews , 24 ( 3 ), 365 – 374 . https://doi.org/10.1016/S0149-7634(00)00012-9 .

Connor , B. , Young , D. , Yan , Q. , Faull , R. , Synek , B. , & Dragunow , M. ( 1997 ). Brain-derived neurotrophic factor is reduced in Alzheimer's disease . Molecular Brain Research , 49 ( 1–2 ), 71 – 81 . https://doi.org/10.1016/s0169-328x(97)00125-3 .

Crane , P. K. , Carle , A. , Gibbons , L. E. , Insel , P. , Mackin , R. S. , Gross , A. , et al.  ( 2012 ). Development and assessment of a composite score for memory in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) . Brain Imaging and Behavior , 6 ( 4 ), 502 – 516 . https://doi.org/10.1007/s11682-012-9186-z .

Egan , M. F. , Kojima , M. , Callicott , J. H. , Goldberg , T. E. , Kolachana , B. S. , Bertolino , A. , et al.  ( 2003 ). The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function . Cell , 112 ( 2 ), 257 – 269 . https://doi.org/10.1016/s0092-8674(03)00035-7 .

Elliott , E. , Atlas , R. , Lange , A. , & Ginzburg , I. ( 2005 ). Brain-derived neurotrophic factor induces a rapid dephosphorylation of tau protein through a PI-3Kinase signalling mechanism . European Journal of Neuroscience , 22 ( 5 ), 1081 – 1089 . https://doi.org/10.1111/j.1460-9568.2005.04290.x .

Erickson , K. I. , Prakash , R. S. , Voss , M. W. , Chaddock , L. , Heo , S. , McLaren , M. , & Kramer, A. F. (2010) . Brain-derived neurotrophic factor is associated with age-related decline in hippocampal volume. Journal of Neuroscience , 30 ( 15 ), 5368 – 5375 . https://doi.org/10.1523/JNEUROSCI.6251-09.2010 .

Faria , M. C. , Gonçalves , G. S. , Rocha , N. P. , Moraes , E. N. , Bicalho , M. A. , Cintra , M. T. G. , et al.  ( 2014 ). Increased plasma levels of BDNF and inflammatory markers in Alzheimer's disease . Journal of Psychiatric Research , 53 , 166 – 172 . https://doi.org/10.1016/j.jpsychires.2014.01.019 .

Ferrer , I. , Marín , C. , Rey , M. J. , Ribalta , T. , Goutan , E. , Blanco , R. , et al.  ( 1999 ). BDNF and full-length and truncated TrkB expression in Alzheimer disease. Implications in therapeutic strategies . Journal of Neuropathology and Experimental Neurology , 58 ( 7 ), 729 – 739 . https://doi.org/10.1097/00005072-199907000-00007 .

Folstein , M. F. , Folstein , S. E. , & McHugh , P. R. ( 1975 ). “Mini-mental state”: A practical method for grading the cognitive state of patients for the clinician . Journal of Psychiatric Research , 12 ( 3 ), 189 – 198 . https://doi.org/10.1016/0022-3956(75)90026-6 .

Fourier , A. , Portelius , E. , Zetterberg , H. , Blennow , K. , Quadrio , I. , & Perret-Liaudet , A. (2015) . Pre-analytical and analytical factors influencing Alzheimer's disease cerebrospinal fluid biomarker variability . Clinica Chimica Acta , 449 , 9 – 15 . https://doi.org/10.1016/j.cca.2015.05.024 .

Franzmeier , N. , Ren , J. , Damm , A. , Monté-Rubio , G. , Boada , M. , Ruiz , A. , et al.  ( 2021 ). The BDNFVal66Met SNP modulates the association between beta-amyloid and hippocampal disconnection in Alzheimer’s disease . Molecular Psychiatry , 26 ( 2 ), 614 – 628 . https://doi.org/10.1038/s41380-019-0404-6 .

Hansson , O. , Seibyl , J. , Stomrud , E. , Zetterberg , H. , Trojanowski , J. Q. , Bittner , T. , et al.  ( 2018 ). CSF biomarkers of Alzheimer's disease concord with amyloid-β PET and predict clinical progression: A study of fully automated immunoassays in BioFINDER and ADNI cohorts . Alzheimer's & Dementia , 14 ( 11 ), 1470 – 1481 . https://doi.org/10.1016/j.jalz.2018.01.010 .

Hariri , A. R. , Goldberg , T. E. , Mattay , V. S. , Kolachana , B. S. , Callicott , J. H. , Egan , M. F. , et al.  ( 2003 ). Brain-derived neurotrophic factor val66met polymorphism affects human memory-related hippocampal activity and predicts memory performance . Journal of Neuroscience , 23 ( 17 ), 6690 – 6694 . https://doi.org/10.1523/JNEUROSCI.23-17-06690.2003 .

Jack , C. R. , Bennett , D. A. , Blennow , K. , Carrillo , M. C. , Feldman , H. H. , Frisoni , G. B. , et al.  ( 2016 ). A/T/N: An unbiased descriptive classification scheme for Alzheimer disease biomarkers . Neurology , 87 ( 5 ), 539 – 547 . https://doi.org/10.1212/WNL.0000000000002923 .

Kimura , N. , Takahashi , M. , Tashiro , T. , & Terao , K. ( 2006 ). Amyloid β up-regulates brain-derived neurotrophic factor production from astrocytes: Rescue from amyloid β-related neuritic degeneration . Journal of Neuroscience Research , 84 ( 4 ), 782 – 789 . https://doi.org/10.1002/jnr.20984 .

Lim , Y. Y. , Williamson , R. , Laws , S. M. , Villemagne , V. L. , Bourgeat , P. , Fowler , C. , et al.  ( 2017a ). Effect of APOE genotype on amyloid deposition, brain volume, and memory in cognitively normal older individuals . Journal of Alzheimer's Disease , 58 ( 4 ), 1293 – 1302 . https://doi.org/10.3233/JAD-170072 .

Lim , Y. Y. , Hassenstab , J. , Cruchaga , C. , Goate , A. , Fagan , A. M. , Benzinger , T. L. S. , et al.  ( 2016 ). BDNF Val66Met moderates memory impairment, hippocampal function and tau in preclinical autosomal dominant Alzheimer's disease . Brain , 139 ( 10 ), 2766 – 2777 . https://doi.org/10.1093/brain/aww200 .

Lim , Y. Y. , Hassenstab , J. , Goate , A. , Fagan , A. M. , Benzinger , T. L. S. , Cruchaga , C. , et al.  ( 2018 ). Effect of BDNFVal66Met on disease markers in dominantly inherited Alzheimer's disease . Annals of Neurology , 84 ( 3 ), 424 – 435 . https://doi.org/10.1002/ana.25299 .

Lim , Y. Y. , Laws , S. M. , Perin , S. , Pietrzak , R. H. , Fowler , C. , Masters , C. L. , et al.  ( 2021 ). BDNF VAL66MET polymorphism and memory decline across the spectrum of Alzheimer's disease . Genes, Brain and Behavior , 20 ( 5 ), e12724 . https://doi.org/10.1111/gbb.12724 .

Lim , Y. Y. , Maruff , P. , Barthélemy , N. R. , Goate , A. , Hassenstab , J. , Sato , C. , et al.  ( 2022 ). Association of BDNF Val66Met with tau hyperphosphorylation and cognition in dominantly inherited Alzheimer disease . JAMA Neurology , 79 ( 3 ), 261 – 270 . https://doi.org/10.1001/jamaneurol.2021.5181 .

Lim , Y. Y. , Maruff , P. , Pietrzak , R. H. , Ames , D. , Ellis , K. A. , Harrington , K. , et al.  ( 2014a ). Effect of amyloid on memory and non-memory decline from preclinical to clinical Alzheimer’s disease . Brain , 137 ( 1 ), 221 – 231 . https://doi.org/10.1093/brain/awt286 .

Lim , Y. Y. , Rainey-Smith , S. , Lim , Y. , Laws , S. M. , Gupta , V. , Porter , T. , et al.  ( 2017b ). BDNF Val66Met in preclinical Alzheimer's disease is associated with short-term changes in episodic memory and hippocampal volume but not serum mBDNF . International Psychogeriatrics , 29 ( 11 ), 1825 – 1834 . https://doi.org/10.1017/S1041610217001284 .

Lim , Y. Y. , Villemagne , V. L. , Laws , S. M. , Ames , D. , Pietrzak , R. H. , Ellis , K. A. , et al.  ( 2014b ). Effect of BDNF Val66Met on memory decline and hippocampal atrophy in prodromal Alzheimer's disease: A preliminary study . PLoS One , 9 ( 1 ), e86498. https://doi.org/10.1371/journal.pone.0086498 .

Lim , Y. Y. , Villemagne , V. L. , Laws , S. M. , Ames , D. , Pietrzak , R. H. , Ellis , K. A. , et al.  ( 2013 ). BDNF Val66Met, Aβ amyloid, and cognitive decline in preclinical Alzheimer's disease . Neurobiology of Aging , 34 ( 11 ), 2457 – 2464 . https://doi.org/10.1016/j.neurobiolaging.2013.05.006 .

Lim , Y. Y. , Villemagne , V. L. , Laws , S. M. , Pietrzak , R. H. , Snyder , P. J. , Ames , D. , et al.  ( 2015 ). APOE and BDNF polymorphisms moderate amyloid β-related cognitive decline in preclinical Alzheimer's disease . Molecular Psychiatry , 20 ( 11 ), 1322 – 1328 . https://doi.org/10.1038/mp.2014.123 .

Lu , B. , Nagappan , G. , Guan , X. , Nathan , P. J. , & Wren , P. ( 2013 ). BDNF-based synaptic repair as a disease-modifying strategy for neurodegenerative diseases . Nature Reviews Neuroscience , 14 ( 6 ), 401 – 416 . https://doi.org/10.1038/nrn3505 .

Murer , M. , Boissiere , F. , Yan , Q. , Hunot , S. , Villares , J. , Faucheux , B. , et al.  ( 1999 ). An immunohistochemical study of the distribution of brain-derived neurotrophic factor in the adult human brain, with particular reference to Alzheimer's disease . Neuroscience , 88 ( 4 ), 1015 – 1032 . https://doi.org/10.1016/s0306-4522(98)00219-x .

Petersen , R. C. , Aisen , P. , Beckett , L. A. , Donohue , M. , Gamst , A. , Harvey , D. J. , et al.  ( 2010 ). Alzheimer's Disease Neuroimaging Initiative (ADNI): clinical characterization . Neurology , 74 ( 3 ), 201 – 209 . https://doi.org/10.1212/WNL.0b013e3181cb3e25 .

Pike , K. E. , Savage , G. , Villemagne , V. L. , Ng , S. , Moss , S. A. , Maruff , P. , et al.  ( 2007 ). β-amyloid imaging and memory in non-demented individuals: evidence for preclinical Alzheimer's disease . Brain , 130 ( 11 ), 2837 – 2844 . https://doi.org/10.1093/brain/awm238 .

Purcell , S. , Neale , B. , Todd-Brown , K. , Thomas , L. , Ferreira , M. A. , Bender , D. , et al.  ( 2007 ). PLINK: a tool set for whole-genome association and population-based linkage analyses . The American Journal of Human Genetics , 81 ( 3 ), 559 – 575 . https://doi.org/10.1086/519795 .

Rosa , E. , Mahendram , S. , Ke , Y. D. , Ittner , L. M. , Ginsberg , S. D. , & Fahnestock , M. ( 2016 ). Tau downregulates BDNF expression in animal and cellular models of Alzheimer's disease . Neurobiology of Aging , 48 , 135 – 142 . https://doi.org/10.1016/j.neurobiolaging.2016.08.020 .

Rosen , W. G. , Mohs , R. C. , & Davis , K. L. ( 1984 ). A new rating scale for Alzheimer's disease . The American Journal of Psychiatry. , 141 ( 11 ), 1356 – 1364 . https://doi.org/10.1176/ajp.141.11.1356 .

Schmidt , M. ( 1996 ). Rey auditory verbal learning test: A handbook (Vol. 17) . Los Angeles, CA : Western Psychological Services Los Angeles .

Shaw , L. M. , Vanderstichele , H. , Knapik-Czajka , M. , Clark , C. M. , Aisen , P. S. , Petersen , R. C. , et al.  ( 2009 ). Cerebrospinal fluid biomarker signature in Alzheimer's disease neuroimaging initiative subjects . Annals of Neurology , 65 ( 4 ), 403 – 413 . https://doi.org/10.1002/ana.21610 .

Shen , T. , You , Y. , Joseph , C. , Mirzaei , M. , Klistorner , A. , Graham , S. L. , et al.  ( 2018 ). BDNF polymorphism: A review of its diagnostic and clinical relevance in neurodegenerative disorders . Aging and Disease , 9 ( 3 ), 523 – 536 . https://doi.org/10.14336/AD.2017.0717 .

Twamley , E. W. , Ropacki , S. A. L. , & Bondi , M. W. ( 2006 ). Neuropsychological and neuroimaging changes in preclinical Alzheimer's disease . Journal of the International Neuropsychological Society , 12 ( 5 ), 707 – 735 . https://doi.org/10.1017/S1355617706060863 .

Verbeke , G. ( 1997 ). Linear mixed models for longitudinal data. In Linear mixed models in practice (pp. 63 – 153 ). New York, NY: Springer .

Villemagne , V. L. , Burnham , S. , Bourgeat , P. , Brown , B. , Ellis , K. A. , Salvado , O. , et al.  ( 2013 ). Amyloid β deposition, neurodegeneration, and cognitive decline in sporadic Alzheimer's disease: A prospective cohort study . The Lancet Neurology , 12 ( 4 ), 357 – 367 . https://doi.org/10.1016/S1474-4422(13)70044-9 .

Wang , S. , Yao , H. , Xu , Y. , Hao , R. , Zhang , W. , Liu , H. , et al.  ( 2020 ). Therapeutic potential of a TrkB agonistic antibody for Alzheimer's disease . Theranostics , 10 ( 15 ), 6854 – 6874 . https://doi.org/10.7150/thno.44165 .

Xiang , J. , Wang , Z.-H. , Ahn , E. H. , Liu , X. , Yu , S.-P. , Manfredsson , F. P. , et al.  ( 2019 ). Delta-secretase-cleaved Tau antagonizes TrkB neurotrophic signalings, mediating Alzheimer’s disease pathologies . Proceedings of the National Academy of Sciences , 116 ( 18 ), 9094 – 9102 . https://doi.org/10.1073/pnas.1901348116 .

Author notes

• alzheimer's disease
• biological markers
• brain-derived neurotrophic factor
• disease transmission
• tau proteins
• episodic memory
• older adult
• val66met polymorphism

Email alerts

Citing articles via.

• Recommend to your Library

Affiliations

• Online ISSN 1873-5843
• Copyright © 2024 Oxford University Press
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

• History and Mission
• Structure and Cores
• Annual Report
• Media Guide
• Guiding Principles for Diversity
• Land Acknowledgement
• Volunteer for a study
• Open Studies
• Brain Donation
• Classes and Series
• Precious Memories Choir
• Mind Readers Book Club
• What is Alzheimer's disease?
• COVID-19 Brain Health Resource Guide
• Resources for People with Dementia and Care Partners
• How to get a diagnosis
• Raising awareness of Alzheimer’s disease in Wisconsin’s Indian Country
• Dementia Matters
• Dementia Matters COVID-19 Special Series
• Biomarker videos
• Request a Talk
• Professional Education
• Black Leaders for Brain Health
• Research Services and Resources
• Apply for Resources
• Data Dictionary
• Developmental Projects
• Acknowledging the Center
• IEA Innovation Fund
• Funded Pilot Grants, 2006-2019
• Internal Center Links
• Referring Patients to Research
• Updates in Dementia Care E-newsletter
• Dr. Daniel I. Kaufer Lecture
• How to join a research lab
• REC Scholar Program
• Research Day 2024
• Junior Fellowship
• Professional Development Award
• Lightning Funding Award

Gut Feelings: The Links Between Gut Health and Alzheimer’s Disease

An expanding field of research is looking at how the gut affects different parts of people’s health, but how does it affect brain health? Drs. Barb Bendlin and Tyler Ulland join the podcast to talk about their 2023 study, which suggests a link between gut health, aging and changes related to Alzheimer’s disease. They discuss their findings on how gut inflammation could impact brain health, as well as explain what it means to have good gut health and how food, medications, where one lives and other factors can impact the gut microbiome.

Guests: Barbara Bendlin, PhD, professor, Division of Geriatrics and Gerontology, deputy director, University of Wisconsin (UW) Center for Health Disparities Research, and Tyler Ulland, PhD, associate professor of pathology and laboratory medicine, UW School of Medicine and Public Health, leader, Ulland Lab

Learn more about the 2024 Fall Community Conversation: The Impact of Social Connections on Brain Health and register to attend in person on our website.

Read Dr. Bendlin and Dr. Ulland’s study, “ Gut inflammation associated with age and Alzheimer’s disease pathology: a human cohort study ,” on the National Library of Medicine website.

Learn more about Dr. Bendlin and Dr. Uland’s study in the article, “ Gut inflammation linked to aging and Alzheimer’s disease ,” on the UW School of Medicine and Public Health website.

Learn more about Dr. Ulland from his profile on the Ulland lab webpage .

Learn more about Dr. Bendlin from her profile on the UW Center for Health Disparities Research website .

Connect with us

Find transcripts and more at our website .

Email Dementia Matters : [email protected]

Follow us on Facebook and Twitter .

Subscribe to the Wisconsin Alzheimer’s Disease Research Center’s e-newsletter .

Enjoy Dementia Matters ? Consider making a gift to the Dementia Matters fund through the UW Initiative to End Alzheimer’s . All donations go toward outreach and production.

Information For

Participants & Families Brain Donors & Families Caregivers Researchers Providers Media Employees

Contact Us Join a Research Lab Subscribe to E-news Give

This website may not display properly if you are using any version of Internet Explorer Feedback, questions or accessibility issues: [email protected]

Privacy Notice | © 2024 Board of Regents of the University of Wisconsin System

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 20 June 2022

Multimodal deep learning for Alzheimer’s disease dementia assessment

• Shangran Qiu 1 , 2   na1 ,
• Matthew I. Miller 1   na1 ,
• Prajakta S. Joshi 3 , 4 , 5 ,
• Joyce C. Lee 1 ,
• Chonghua Xue 1 , 3 ,
• Yunruo Ni 1 ,
• Yuwei Wang 1 ,
• Ileana De Anda-Duran   ORCID: orcid.org/0000-0003-3629-9465 6 ,
• Phillip H. Hwang   ORCID: orcid.org/0000-0001-6780-6808 3 ,
• Justin A. Cramer 7 ,
• Brigid C. Dwyer 8 ,
• Honglin Hao 9 ,
• Michelle C. Kaku 8 ,
• Sachin Kedar 10 , 11 , 12 ,
• Peter H. Lee 13 ,
• Asim Z. Mian 14 ,
• Daniel L. Murman 10 ,
• Sarah O’Shea 8 ,
• Aaron B. Paul 13 ,
• Marie-Helene Saint-Hilaire 8 ,
• E. Alton Sartor 8 ,
• Aneeta R. Saxena 8 ,
• Ludy C. Shih   ORCID: orcid.org/0000-0002-6590-8365 8 ,
• Juan E. Small   ORCID: orcid.org/0000-0002-4931-3564 13 ,
• Maximilian J. Smith 13 ,
• Arun Swaminathan 10 ,
• Courtney E. Takahashi 8 ,
• Olga Taraschenko 10 ,
• Hui You 15 ,
• Jing Yuan   ORCID: orcid.org/0000-0002-3923-925X 9 ,
• Yan Zhou 9 ,
• Shuhan Zhu 8 ,
• Michael L. Alosco 8 , 16 ,
• Jesse Mez 5 , 8 , 16 ,
• Thor D. Stein 16 , 17 , 18 , 19 ,
• Kathleen L. Poston   ORCID: orcid.org/0000-0003-3424-7143 20 ,
• Rhoda Au 3 , 5 , 8 , 16 , 21 &
• Vijaya B. Kolachalama   ORCID: orcid.org/0000-0002-5312-8644 1 , 16 , 22 , 23  

Nature Communications volume  13 , Article number:  3404 ( 2022 ) Cite this article

44k Accesses

89 Citations

106 Altmetric

Metrics details

• Alzheimer's disease
• Translational research

Worldwide, there are nearly 10 million new cases of dementia annually, of which Alzheimer’s disease (AD) is the most common. New measures are needed to improve the diagnosis of individuals with cognitive impairment due to various etiologies. Here, we report a deep learning framework that accomplishes multiple diagnostic steps in successive fashion to identify persons with normal cognition (NC), mild cognitive impairment (MCI), AD, and non-AD dementias (nADD). We demonstrate a range of models capable of accepting flexible combinations of routinely collected clinical information, including demographics, medical history, neuropsychological testing, neuroimaging, and functional assessments. We then show that these frameworks compare favorably with the diagnostic accuracy of practicing neurologists and neuroradiologists. Lastly, we apply interpretability methods in computer vision to show that disease-specific patterns detected by our models track distinct patterns of degenerative changes throughout the brain and correspond closely with the presence of neuropathological lesions on autopsy. Our work demonstrates methodologies for validating computational predictions with established standards of medical diagnosis.

Similar content being viewed by others

Generalizable deep learning model for early Alzheimer’s disease detection from structural MRIs

A novel CNN architecture for accurate early detection and classification of Alzheimer’s disease using MRI data

Multimodal deep learning models for early detection of Alzheimer’s disease stage

Introduction.

Alzheimer’s disease (AD) is the most common cause of dementia worldwide 1 , and future expansions in caseload due to an aging population are likely to accentuate existing needs for health services 2 . This increase in clinical demand will likely contribute to an already considerable burden of morbidity and mortality among the elderly 3 , thus requiring improvements in the treatment and timely identification of AD. Significant efforts have been made in recent years towards the development of cerebrospinal fluid (CSF) biomarkers 4 , as well as advanced imaging modalities such as amyloid and tau positron emission tomography (PET) 5 , 6 , 7 , 8 . Furthermore, novel generations of disease-modifying therapies for AD are now coming into clinical purview 9 , though their efficacy remains controversial. Despite this progress, many emerging diagnostic and treatment modalities remain limited to research contexts, and the backbone of antemortem diagnosis remains traditional clinical assessment, neuropsychological testing 10 , and magnetic resonance imaging (MRI) 11 . Mild cognitive impairment (MCI), a prodromal stage of dementia, may also be a subtle early presentation of AD whose diagnosis similarly requires significant clinical acumen from qualified specialists. Complicating matters is the presence of a multitude of other non-Alzheimer’s disease dementia (nADD) syndromes whose clinical presentations often overlap with AD. Thus, common causes of dementias outside of AD such as vascular dementia (VD), Lewy body dementia (LBD), and frontotemporal dementia (FTD) widen the differential diagnosis of neurodegenerative conditions and contribute to variability in diagnostic sensitivity and specificity 12 .

Reliably differentiating between normal cognitive aging, MCI, AD, and other dementia etiologies requires significant clinical acumen from qualified specialists treating memory disorders, yet timely access to memory clinics is often limited for patients and families. This is a major problem in remote, rural regions within developed countries and in still economically developing nations, where there is a dearth of specialized practitioners. Furthermore, the need for skilled clinicians is rising, yet the United States is facing a projected shortage of qualified clinicians, such as neurologists, in coming decades 13 , 14 . As increasing clinical demand intersects with a diminishing supply of medical expertise, machine learning methods for aiding neurologic diagnoses have begun to attract interest 15 . Complementing the high diagnostic accuracy reported by other groups 16 , we have previously reported interpretable deep learning approaches capable of distinguishing participants with age-appropriate normal cognition (NC) from those with AD using magnetic resonance imaging (MRI) scans, age, sex, and mini-mental state examination (MMSE) 17 . Others have also demonstrated the efficacy of deep learning in discriminating AD from specific types of nADD 18 , 19 , 20 . However, clinical evaluation of persons presenting in a memory clinic involves consideration of multiple etiologies of cognitive impairment. Therefore, the ability to successfully differentiate between NC, MCI, AD, and nADD across diverse study cohorts in a unified framework remains to be developed.

In this study, we report the development and validation of a deep learning framework capable of accurately classifying subjects with NC, MCI, AD, and nADD in multiple cohorts of participants with different etiologies of dementia and varying levels of cognitive function (Table  1 , Fig.  1 ). Using data from the National Alzheimer’s Coordinating Center (NACC) 21 , 22 , we developed and externally validated models capable of classifying cognitive status using MRI, non-imaging variables, and combinations thereof. To validate our approach, we demonstrated comparability of our model’s accuracy to the diagnostic performance of a team of practicing neurologists and neuroradiologists. We then leveraged SHapley Additive exPlanations (SHAP) 23 , to link computational predictions with well-known anatomical and pathological markers of neurodegeneration. Our strategy provides evidence that automated methods driven by deep learning may approach clinical standards of accurate diagnosis even amidst heterogeneous datasets.

Multimodal data including MRI scans, demographics, medical history, functional assessments, and neuropsychological test results were used to develop deep learning models on various classification tasks. Eight independent datasets were used for this study, including NACC, ADNI, AIBL, FHS, LBDSU, NIFD, OASIS, and PPMI. We selected the NACC dataset to develop three separate models: (i) an MRI-only CNN model (ii) non-imaging models in the form of traditional machine learning classifiers, which did not use any MRI data (iii) a fusion model combining imaging and non-imaging data within a hybrid architecture joining a CNN to a CatBoost model. The MRI-only model was validated across all eight cohorts, whereas external validation of non-imaging and fusion models was performed only on OASIS. First, T1-weighted MRI scans were input to a CNN to calculate a continuous DEmentia MOdel (DEMO) score to assess cognitive status on a 0 to 2 scale, where “0” indicated NC “1” indicated MCI and “2” indicated DE. DEMO scores were converted to class labels using an optimal thresholding algorithm, with these assignments constituting the COG task. For individuals with DE diagnosis, the multi-task CNN model simultaneously discriminated their risk of having AD versus nADD, a classification that we refer to as the ADD task. We denoted the probability of AD diagnosis as the ALZheimer (ALZ) score. Both MRI-derived DEMO scores and ALZ scores were then input alongside non-imaging variables to various machine learning classifiers to form fusion models, which then predicted outcomes on the COG and ADD tasks, respectively. A portion of cases with confirmed dementia ( n  = 50) from the NACC testing cohort was randomly selected for direct comparison of the fusion model with an international team of practicing neuroradiologists. Both the model and neuroradiologists completed the ADD task using available MRI scans, age, and gender. Additionally, a portion of NACC cases ( n  = 100) was randomly selected to compare the fusion model performance to practicing neurologists, with both the model and clinicians having access to a common set of multimodal data. Lastly, model predictions were compared with neuropathology grades from NACC, ADNI and FHS cohorts ( n  = 110).

We divided the process of differential diagnosis into staged tasks. The first, which we refer to as the COG task, labeling persons as having either NC, MCI, or dementia (DE) due to any cause. Of note, the COG task may be seen as comprising three separate binary classification subtasks: (i) COG NC task: Separation of NC and MCI/DE cases (ii) COG MCI task: Separation of MCI from NC/DE cases, and (iii) COG DE task: Separation of DE from NC/MCI cases. After completion of the overall COG task, we next formulated the ADD task, in which we assigned all persons labeled as DE to a diagnosis of either AD or nADD. Successive completion of the COG and the ADD tasks allowed execution of an overarching 4-way classification that fully delineated NC, MCI, AD, and nADD cases (See  Supplementary Information : Glossary of Tasks, Models, and Metrics).

We also created three separate models: (i) MRI-only model: A convolutional neural network (CNN) that internally computed a continuous DEmentia MOdel (DEMO) score to complete the COG task, as well as an ALZheimer’s (ALZ) score to complete the ADD task. (ii) Non-imaging model: A traditional machine learning classifier that took as input only scalar-valued clinical variables from demographics, past medical history, neuropsychological testing, and functional assessments. As in the MRI-only model, the non-imaging model also computed the DEMO and the ALZ scores from which the COG and the ADD tasks could be completed. We tested multiple machine learning architectures for these purposes and ultimately selected a CatBoost model as our final non-imaging model architecture. (iii) Fusion model: This framework linked a CNN to a CatBoost model. With this approach, the DEMO and the ALZ scores computed by the CNN were recycled and used alongside available clinical variables. The CatBoost model then recalculated these scores in the context of the additional non-imaging information. We provide definitions of our various prediction tasks, cognitive metrics, and model types within the  Supplementary Information . Further details of model design may be found within the Methods.

Assessment for confounding

We used two-dimensional t-distributed stochastic neighbor embedding (tSNE) to assess for the presence of confounding relationships between disease status and certain forms of metadata. Using this approach, we observed no obvious clustering of post-processed MRI embeddings among the eight cohorts used for testing of MRI-only models (Fig.  2a, b ). Within the NACC cohort, we also observed no appreciable clustering based on individual Alzheimer’s Disease Research Centers (ADRCs, Fig.  2c, d ) or scanner manufacturer (Fig.  2e, f ). Relatedly, although tSNE analysis of CNN hidden layer activations did yield clustering of NACC data points (Fig.  2b ), this was an expected phenomenon given the selection of NACC as our cohort for model training. Otherwise, we appreciated no obvious conglomeration of embeddings from hidden layer activations due to specific ADRCs (Fig.  2d ) or scanner manufacturers (Fig.  2f ). Lastly, Mutual Information Scores (MIS) computed from the NACC cohort indicated negligible correlation of diagnostic labels (NC, MCI, AD, and nADD) between specific scanner manufacturers (MIS = 0.010, Fig.  2g ) and ADRCs (MIS = 0.065, Fig.  2h ).

Unsupervised clustering of post-processed MRIs and hidden layer activations assessed for systematic biases in input data and model predictions, respectively. a Two-dimensional (2D) t-distributed stochastic neighbor embedding (tSNE) embeddings of downsampled MRI scans are shown. The downsampling was performed on the post-processed MRI scans using spline interpolation with a downsampling factor of 8 on each axis. Individual points represent MRIs from a single subject and are colored according to their original cohort (either NACC, ADNI, AIBL, FHS, LBDSU, NIFD, OASIS, or PPMI). b We demonstrate 2D tSNEs of hidden-layer activations from the penultimate CNN hidden layer. Individual points correspond to internal representations of MRI scans during testing and are colored by cohort label. c Plot of 2D tSNE embeddings of downsampled MRI scans from the NACC dataset is shown. Individual points representing MRI scans are colored by the unique identifier of one of twenty-one Alzheimer Disease Research Centers (ADRCs) that participate in the NACC collaboration. d tSNE embeddings for penultimate layer activations colored by ADRC ID are shown. e Plot of 2D tSNE embeddings of downsampled MRI scans from the NACC dataset is shown. Embeddings in this plot are the same as those in c but colored according to the manufacturer of the scanner used to acquire each MRI, either General Electric (GE), Siemens, or Philips. f Plot of 2D tSNE of penultimate layer activations is shown for cases in the NACC dataset. Embeddings are equivalent to those visualized in d but are now colored by the manufacturer of the scanner used for image acquisition. g A tabular representation of disease category counts by manufacturer is presented. Only cases from the NACC dataset are included. We provide the Mutual Information Score (MIS) to quantify the correlation between disease type and scanner manufacturer. h We also provided a tabular representation of disease category counts stratified by ADRC ID in the NACC dataset. MIS is once again shown to quantify the degree of correlation between diagnostic labels and individual centers participating in the NACC study. Source data are provided as a Source Data file.

Deep learning model performance

We observed that our fusion model provided the most accurate classification of cognitive status for NC, MCI, AD and nADD across a range of clinical diagnosis tasks (Table  2 ). We found strong model performance on the COG NC task between both the NACC test set (Fig.  3a , Row 1) and an external validation set (OASIS; Fig.  3b , Row 1) as indicated by area under the receiver operating characteristic (AUC) curve values of 0.945 [95% confidence interval (CI): 0.939, 0.951] and 0.959 [CI: 0.955, 0.963], respectively. Similar values for area under precision-recall (AP) curves were also observed, yielding 0.946 [CI: 0.940, 0.952] and 0.969 [CI: 0.964, 0.974], respectively. Such correspondence between AUC and AP performance supports robustness to class imbalance across datasets. In the COG DE task, comparable results were also seen, as the fusion model yielded respective AUC and AP scores of 0.971 [CI: 0.966, 0.976]/0.917 [CI: 0.906, 0.928] (Fig.  3a , Row 2) in the NACC dataset and 0.971 [CI: 0.969, 0.973]/0.959 [CI: 0.957, 0.961] in the OASIS dataset (Fig.  3b , Row 2). Conversely, classification performance dropped slightly for the ADD task, with respective AUC/AP values of 0.773 [CI: 0.712, 0.834]/0.938 [CI: 0.918, 0.958] in the NACC dataset (Fig.  3a , Row 3) and 0.773 [CI: 0.732, 0.814]/0.965 [CI: 0.956, 0.974] in the OASIS dataset (Fig.  3b , Row 3).

a , b ROC curves showing true positive rate versus false positive rate and PR curves showing the positive predictive value versus sensitivity on the a NACC test set and b OASIS dataset. The first row in a and b denotes the performance of the MRI-only model, the non-imaging model, and the fusion model (CNN + CatBoost) trained to classify cases with NC from those without NC (COG NC task). The second row shows ROC and PR curves of the MRI-only model, the non-imaging model, and the fusion model for the COG DE task aimed at distinguishing cases with DE from those who do not have DE. The third row illustrates performance of the MRI-only model, the non-imaging model, and the fusion model focused on discriminating AD from nADD. For each curve, mean AUC was computed. In each plot, the mean ROC/PR curve and standard deviation are shown as bolded lines and shaded regions, respectively. The dotted lines in each plot indicate the classifier with the random performance level. c , d Fifteen features with highest mean absolute SHAP values from the fusion model are shown for the COG and ADD tasks, respectively across cross-validation rounds ( n  = 5). Error bars overlaid on bar plots are centered at the mean of the data and extend + /− one standard deviation. For each task, the MRI scans, demographic information, medical history, functional assessments, and neuropsychological test results were used as inputs to the deep learning model. The left plots in c and d illustrate the distribution of SHAP values and the right plots show the mean absolute SHAP values. All the plots in c and d are organized in decreasing order of mean absolute SHAP values. e , f For comparison, we also constructed traditional machine learning models to predict cognitive status and AD status using the same set of features used for the deep learning model, and the results are presented in e and f , respectively. The heat maps show fifteen features with the highest mean absolute SHAP values obtained for each model. Source data are provided as a Source Data file.

Relative to the fusion model, we observed moderate performance reductions across classifications in our MRI-only model. For the COG NC task, the MRI-only framework yielded AUC and AP scores of 0.844 [CI: 0.832, 0.856]/0.830 [CI: 0.810, 0.850] (NACC) and 0.846 [CI: 0.840, 0.852]/0.890 [CI: 0.884, 0.896] (OASIS). Model results were comparable on the COG DE task, in which the MRI-only model achieved respective AUC and AP scores of 0.869 [CI: 0.850, 0.888]/0.712 [CI: 0.672, 0.752] (NACC) and 0.858 [CI: 0.854, 0.862]/0.772 [CI: 0.763, 0.781] (OASIS). For the ADD task as well, the results of the MRI-only model were approximately on par with those of the fusion model, giving respective AUC and AP scores of 0.766 [CI: 0.734, 0.798]/0.934 [CI: 0.917, 0.951] (NACC) and 0.694 [CI: 0.659,0.729]/0.942 [CI: 0.931, 0.953] (OASIS). For both fusion and MRI-only models, we also reported ROC and PR curves for the ADD task stratified by nADD subtypes in the  Supplementary Information (Figs.  S1 and S2 ).

Interestingly, we note that a non-imaging model generally yielded similar results to those of both the fusion and MRI-only models. Specifically, a CatBoost model trained for the COG NC task gave AUC and AP values 0.936 [CI: 0.929, 0.943] /0.936 [CI: 0.930, 0.942] (NACC), as well as 0.959 [CI: 0.957, 0.961]/0.972 [CI: 0.970, 0.974] (OASIS). Results remained strong for the COG DE task, with AUC/PR pairs of 0.962 [CI: 0.957, 0.967]/0.907 [0.893, 0.921] (NACC) and 0.971 [CI: 0.970, 0.972]/0.955 [CI: 0.953, 0.957] (OASIS). For the ADD task, the non-imaging model resulted in respective AUC/PR scores of 0.749 [CI: 0.691, 0.807]/0.935 [CI: 0.919, 0.951] (NACC) and 0.689 [CI: 0.663, 0.715]/0.947 [CI: 0.940, 0.954] (OASIS). A full survey of model performance metrics across all classification tasks may be found in the  Supplementary Information (Tables  S1 – S4 ). Performance of the MRI-only model across all external datasets is demonstrated via ROC and PR curves (Fig.  S3 ).

To assess the contribution of various imaging and non-imaging features to classification outcomes, we calculated fifteen features with highest mean absolute SHAP values for the COG (Fig.  3c ) and the ADD prediction tasks using the fusion model (Fig.  3d ). Though MMSE score was the primary discriminative feature for the COG task, the DEMO score derived from the CNN portion of the model ranked third in predicting cognitive status. Analogously, the ALZ score derived from the CNN was the most salient feature in solving the ADD task. Interestingly, the relative importance of features remained largely unchanged when a variety of other machine learning classifiers were substituted to the fusion model in lieu of the CatBoost model (Fig.  3e, f ). This consistency indicated that our prediction framework was robust to the specific choice of model architecture, and instead relied on a consistent set of clinical features to achieve discrimination between NC, MCI, AD, and nADD classes. Relatedly, we also observed that non-imaging and fusion models retained predictive performance across a variety of input feature combinations, showing flexibility to operate across differences in information availability. Importantly, however, the addition of MRI-derived DEMO and ALZ scores improved 4-way classification performance across all combinations of non-imaging variables (Figs.  S4 and S5 ).

Neuroimaging signatures of AD and non-AD dementia

The provenance of model predictions was visualized by pixel-wise SHAP mapping of hidden layers within the CNN model. The SHAP matrices were then correlated to physical locations within each subject’s MRI to visualize conspicuous brain regions implicated in each stage of cognitive decline from NC to dementia (Fig.  4a ). This approach allowed neuroanatomical risk mapping to distinguish regions associated with AD from those with nADD (Fig.  4b ). Indeed, the direct overlay of color maps representing disease risk on an anatomical atlas derived from traditional MRI scans facilitates interpretability of the deep learning model. Also, the uniqueness of the SHAP-derived representation allows us to observe disease suggestive regions that are specific to each outcome of interest (Table  S5 and Fig.  S6 ).

a , b SHAP value-based illustration of brain regions that are most associated with the outcomes. The first columns in both a and b show a template MRI oriented in axial, coronal, and sagittal planes. In a , the second, third and fourth columns show SHAP values from the input features of the second convolutional block of the CNN averaged across all NACC test subjects with NC, MCI, and dementia, respectively. In b , the second and third columns show SHAP values averaged across all NACC test subjects with AD and nADD, respectively. c Brain region-specific SHAP values for both AD and nADD cases obtained from the NACC testing data are shown. The violin plots are organized per lobe and in decreasing order of mean absolute SHAP values. d , e Network of brain regions implicated in the classification of AD and nADD, respectively. We selected 33 representative brain regions for graph analysis and visualization of sagittal regions, as well as 57 regions for axial analyses. Nodes representing brain regions are overlaid on a two-dimensional brain template and sized according to weighted degree. The color of the segments connecting different nodes indicates the sign of correlation and the thickness of the segments indicates the magnitude of the correlation. It must be noted that not all nodes can be seen either from the sagittal or the axial planes. Source data are provided as a Source Data file.

A key feature of SHAP is that a single voxel or a sub-region within the brain can contribute to accurate prediction of one or more class labels. For example, the SHAP values were negative in the hippocampal region in NC participants, but they were positive in participants with dementia, underscoring the well-recognized role of the hippocampus in memory function. Furthermore, positive SHAP values were observed within the hippocampal region for AD and negative SHAP values for the nADD cases, indicating that hippocampal atrophy has direct proportionality with AD-related etiology. The SHAP values sorted according to their importance on the parcellated brain regions also further confirm the role of hippocampus and its relationship with dementia prediction, particularly in the setting of AD (Fig.  4c ), as well as nADD cases (Fig.  S7 ). In the case of nADD, the role of other brain regions such as the lateral ventricles and frontal lobes was also evident. Evidently, SHAP-based network analysis revealed pairwise relationships between brain regions that simultaneously contribute to patterns indicative of AD (Fig.  4d ). The set of brain networks evinced by this analysis also demonstrate marked differences in structural changes between AD and nADD (Fig.  4e ).

Neuropathologic validation

In addition to mapping hidden layer SHAP values to original neuroimaging, correlation of deep learning predictions with neuropathology data provided further validation of our modeling approach. Qualitatively, we observed that areas of high SHAP scores for the COG task correlated with region-specific neuropathological scores obtained from autopsy (Fig.  5a ). Similarly, the severity of regional neuropathologic changes in these persons demonstrated a moderate to high degree of concordance with the regional cognitive risk scores derived from our CNN using the Spearman’s rank correlation test. Of note, the strongest correlations appeared to occur within areas affected by AD pathology such as the temporal lobe, amygdala, hippocampus, and parahippocampal gyrus (Fig.  5b ). Using the one-way ANOVA test, we also rejected a null hypothesis of there being no significant differences in DEMO scores between semi-quantitative neuropathological score groups (0–3) with a confidence level of 0.95, including for the global ABC severity scores of Thal phase for Aβ (A score F-test: F(3, 51) = 3.665, p  = 1.813e-2), Braak & Braak for neurofibrillary tangles (NFTs) (B score F-test: F(3, 102) = 11.528, p  = 1.432e-6), and CERAD neuritic plaque scores (C score F-test: F(3, 103) = 4.924, p  = 3.088e-3) (Fig.  5c ). We further performed post hoc testing using Tukey’s procedure to compare pairwise group means of DEMO scores, observing consistently significant differences between individuals with the highest and lowest burdens of neurodegenerative findings, respectively (Fig.  S8 ). Of note, we also observed an increasing trend of ALZ score with the semi-quantitative neuropathological scores (Fig.  5d ).

We correlated model findings with regional ABC scores of neuropathologic severity obtained autopsied participants in NACC, ADNI, and FHS cohorts ( n  = 110). a An example case from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset is displayed in sagittal, axial, and coronal views. The SHAP values derived from the second convolutional block and neuropathologic ABC scores are mapped to brain regions where they were measured at the time of autopsy. Visually, high concordance is observed between anatomically mapped SHAP values regardless of the hidden layer from which they are derived. Concordance is observed between the SHAP values and neurofibrillary tangles (NFT) scores within the temporal lobe. b A heatmap is shown demonstrating Spearman correlations between population-averaged SHAP values from the input features of the second convolutional layer and stain-specific ABC scores at various regions of the brain. A strong positive correlation is observed between the SHAP values and neuropathologic changes within several areas well-known to be affected in AD such as the hippocampus/parahippocampus, amygdala and temporal gyrus. c Beeswarm plots with overlying box-and-whisker diagrams are shown to denote the distribution of ABC system sub-scores (horizontal axis) versus model-predicted cognitive scores (vertical axis). The displayed data points represent a pooled set of participants from ADNI, NACC, and FHS for whom neuropathology reports were available from autopsy. Each symbol represents a study participant, boxes are centered at the median and extend over the interquartile range (IQR), while bottom and top whiskers represent 1st and 3rd quartiles −/+ 1.5 x IQR, respectively. We denote p  < 0.05 as *; p  < 0.001 as **, and p  < 0.0001 as *** based on post-hoc Tukey testing. d A heatmap demonstrating the distribution of neuropathology scores versus model predicted AD probabilities. Herein, each column within the map represents a unique individual whose position along the horizontal axis is a descending function of AD risk according to the deep learning model. The overlying hatching pattern represents the dataset (ADNI, NACC, and FHS), from which everyone is drawn. Source data are provided as a Source Data file.

Expert-level validation

Lastly, to provide clinical benchmarking of our modeling approach, both neurologists and neuroradiologists were recruited to perform diagnostic tasks on a subset of NACC cases. The approach and performance of the neurologists and the neuroradiologists indicated variability across different clinical practices (See  Supplementary Information : Neurologist and Neuroradiologist Accounts), with a moderate inter-rater agreement as evaluated using pairwise kappa (κ) scoring for all the tasks. Among neurologists specifically, we observed average κ = 0.600 for the COG NC task (Fig.  6a , Row 1) and average κ = 0.601 for the COG DE task (Fig.  6a , Row 2). Among neuroradiologists performing the ADD task, we found average κ = 0.292 (Fig.  6b ). In the overall 4-way classification of NC, MCI, AD, and nADD, we observed that the accuracy of our fusion model (mean: 0.558, 95% CI: [0.482,0.634]) reached that of neurologists (mean: 0.565, 95% CI: [0.529,0.601]). Interestingly, a similar level of 4-way accuracy was achieved by a non-imaging CatBoost model (mean: 0.544, 95% CI: [0.517,0.571]), though not on an MRI-only model (mean: 0.412, 95% CI: [0.380,0.444]). However, an MRI-only model did yield a moderate improvement in diagnostic accuracy (mean: 0.692, 95% CI: [0.649,0.735]) over neuroradiologists (mean: 0.566, 95% CI: [0.516,0.616]) in the ADD task (Fig.  6b ). Full performance metrics (including accuracy, sensitivity, specificity, F-1 score, and Matthews Correlation Coefficient) may be found in Tables  S6 and S7 for respective comparison of machine learning models to neurologists and neuroradiologists in diagnostic simulations. Performance metrics for simple thresholding of various neuropsychologic test scores can be found in Table  S8 . We also sought to correlate region-specific SHAP values with structural changes observed by the neuroradiologists throughout the brain, with particular attention towards limbic and temporal lobe structures. Statistically significant correlations between regional SHAP averages and clinically graded atrophy severity suggested a connection between CNN features and widely known markers of dementia (Fig.  6c ).

a For the COG NC task (Row 1), the diagnostic accuracy of board-certified neurologists ( n  = 17) is compared to the performance of our deep learning model using a random subset of cases from the NACC dataset ( n  = 100). Metrics from individual clinicians are plotted in relation to the ROC and PR curves from the trained model. Individual clinician performance is indicated by the blue plus symbol and averaged clinician performance along with error bars is indicated by the green plus symbol on both the ROC and PR curves. The mean ROC/PR curve and the standard deviation are shown as the bold line and shaded region, respectively. A heatmap of pairwise Cohen’s kappa statistic is also displayed to demonstrate inter-rater agreement across the clinician cohort. For the COG DE task (Row 2), ROC, PR, and interrater agreement graphics are illustrated with comparison to board-certified neurologists in identical fashion. For these tasks, all neurologists were granted access to multimodal patient data, including MRIs, demographics, medical history, functional assessments, and neuropsychological testing. The same data was used as input to train the deep learning model. b For validation of our ADD task, a random subset ( n  = 50) of cases with dementia from the NACC cohort was provided to the team of neuroradiologists ( n  = 7), who classified AD versus those with dementia due to other etiologies (nADD). As above, the diagnostic accuracy of the physician cohort is compared to model performance using ROC and PR curves. Graphical conventions for visualizing model and clinician performance are as described above in a and, once more, pairwise Cohen’s kappa values are shown to demonstrate inter-rater agreement. c SHAP values from the second convolutional layer averaged from selected brain regions are shown plotted against atrophy scores assigned by neuroradiologists. Orange and blue points (and along with regression lines and 95% confidence intervals) represent left and right hemispheres, respectively. Spearman correlation coefficients and corresponding two-tailed p values are also shown and demonstrate a statistically significant proportionality between SHAP scores, and the severity of regional atrophy assigned by clinicians. Source data are provided as a Source Data file.

In this work, we presented a range of machine learning models that can process multimodal clinical data to accurately perform a differential diagnosis of AD. These frameworks can achieve multiple diagnostic steps in succession, first delineating persons based on overall cognitive status (NC, MCI, and DE) and then separating likely cases of AD from those with nADD. Importantly, our models are capable of functioning with flexible combinations of imaging and non-imaging data, and their performance generalized well across multiple datasets featuring a diverse range of cognitive statuses and dementia subtypes.

Our fusion model demonstrated the highest overall classification accuracy across diagnostic tasks, achieving results on par with neurologists recruited from multiple institutions to complete clinical simulations. Notably, similar levels of performance were observed both in the NACC testing set, and in the OASIS external validation set. Our MRI-only model also surpassed the average diagnostic accuracy of practicing neuroradiologists and maintained a similar level of performance in 6 additional external cohorts (ADNI, AIBL, FHS, NIFD, PPMI, and LBDSU), thereby suggesting that diagnostic capability was not biased to any single data source. It is also worth noting that the DEMO and the ALZ scores bore strong analytic importance like that of traditional information used for dementia diagnosis. For instance, in the ADD task, the ALZ score was shown by SHAP analysis to have a greater impact in accurately predicting disease status than key demographic and neuropsychological test variables used in standard clinical practice such as age, sex, and MMSE score. These CNN-derived scores maintained equal levels of importance when used in other machine learning classifiers, suggesting wide utility for digital health workflows.

Furthermore, post-hoc analyses demonstrated that the performance of our machine learning models was grounded in well-established patterns of dementia-related neurodegeneration. Network analyses evinced differing regional distributions of SHAP values between AD and nADD populations, which were most pronounced in areas such as the hippocampus, amygdala, and temporal lobes. The SHAP values in these regions also exhibited a strong correlation with atrophy ratings from neuroradiologists. Although recent work has shown that explainable machine learning methods may identify spurious correlations in imaging data 24 , we feel that our ability to link regional SHAP distributions to both anatomic atrophy and also semi-quantitative scores of Aβ amyloid, neurofibrillary tangles, and neuritic plaques links our modeling results to a gold standard of postmortem diagnosis. More generally, our approach demonstrates a means by which to assimilate deep learning methodologies with validated clinical evidence in health care.

Our work builds on prior efforts to construct automated systems for the diagnosis of dementia. Previously, we developed and externally validated an interpretable deep learning approach to classify AD using multimodal inputs of MRI and clinical variables 17 . Although this approach provided a novel framework, it relied on a contrived scenario of discriminating individuals into binary outcomes, which simplified the complexity of a real-world setting. Our current work extends this framework by mimicking a memory clinic setting and accounting for cases along the entire cognitive spectrum. Though numerous groups have taken on the challenge of nADD diagnosis using deep learning 18 , 19 , 20 , 25 , 26 , even these tasks were constructed as simple binary classifications between disease subtypes. Given that the practice of medicine rarely reduces to a choice between two pathologies, integrated models with the capability to replicate the differential diagnosis process of experts more fully are needed before deep learning models can be touted as assistive tools for clinical-decision support. Our results demonstrate a strategy for expanding the scope of diagnostic tasks using deep learning, while also ensuring that the predictions of automated systems remain grounded in established medical knowledge.

Interestingly, it should be noted that the performance of a non-imaging model alone approached that of the fusion model. However, the inclusion of neuroimaging data was critical to enable verification of our modeling results by clinical criteria (e.g., cross-correlation with post-mortem neuropathology reports). Such confirmatory data sources cannot be readily assimilated to non-imaging models, thus limiting the ability to independently ground their performance in non-computational standards. Therefore, rather than viewing the modest contribution of neuroimaging to diagnostic accuracy as a drawback, we argue that our results suggest a path towards balancing demands for transparency with the need to build models using routinely collected clinical data. Models such as ours may be validated in high-resource areas where the availability of advanced neuroimaging aids interpretability. As physicians may have difficulty entrusting medical decision-making to black box model in artificial intelligence 27 , grounding our machine learning results in the established neuroscience of dementia may help to facilitate clinical uptake. Nevertheless, we note that our non-imaging model may be best suited for deployment among general practitioners (GPs) and in low-resource settings.

Functionally, we also contend that the flexibility of inputs afforded by our approach is a necessary precursor to clinical adoption at multiple stages of dementia. Given that sub-group analyses suggested significant 4-way diagnostic capacity on multiple combinations of training data (i.e., demographics, clinical variables, and neuropsychological tests), our overall framework is likely adaptable to many variations of clinical practice without requiring providers to significantly alter their typical workflows. For example, GPs frequently perform cognitive screening with or without directly ordering MRI tests 28 , 29 , 30 , whereas memory specialists typically expand testing batteries to include imaging and advanced neuropsychological testing. This ability to integrate along the clinical care continuum, from primary to tertiary care allows our deep learning solution to address a two-tiered problem within integrated dementia care by providing a tool for both screening and downstream diagnosis.

Our study has several limitations. To begin, in cases of mixed dementia, the present models default to a diagnosis of AD whenever this condition is present, thus attributing a single diagnosis to participants with multiple comorbidities. Given the considerable prevalence of mixed dementias 31 , future work may include the possibility of a multi-label classification that may allow for the identification of co-occurring dementing conditions (e.g., LBD and AD, VD and AD) within the same individual. Our cohorts also did not contain any confirmed cases of atypical AD, which is estimated to affect approximately 6% of elderly-onset cases and one-third of patients with early-onset disease 32 . We must also note that MCI is a broad category by itself that includes persons who may or may not progress to dementia. When relevant data becomes available across many cohorts, future investigations could include MCI subjects who are amnestic and non-amnestic, to understand distinct signatures of those who have prodromal AD. We also acknowledge that our study data is predominantly obtained from epidemiologic studies which primarily focus on AD and that variables that optimize the identification of this illness may in fact detract from the accurate diagnosis of certain nADDs. For instance, we noted that the performance of our fusion models was slightly lower than that of the MRI-only model for distinguishing AD from non-parkinsonian dementias such as FTD and VD. We speculate that certain forms of neuropsychological testing such as the MMSE, which have well-known limitations in specificity 33 , may bias predictions towards more common forms of dementia such as AD. Although we validated the various models using data from a population-based cohort (i.e., FHS), it is possible that multimodal analysis frameworks have the potential to decrease diagnostic accuracy for less common dementias. Future modeling efforts may optimize for the identification of these diseases by including additional clinical data tailored to their diagnosis: for instance, the inclusion of motor examination to assess for parkinsonism, FLAIR images for vascular injury, or cognitive fluctuations and sleep behavior abnormalities for LBD. Lastly, although we have compared our model to the performance of individual neurologists and neuroradiologists, future studies may consider comparison to consensus reviews by teams of collaborating clinicians.

In conclusion, our interpretable, multimodal deep learning framework was able to obtain high accuracy signatures of dementia status from routinely collected clinical data, which was validated against data from independent cohorts, neuropathological findings, and expert-driven assessment. Moreover, our approach provides a solution that may be utilized across different practice types, from GPs to specialized memory clinics at tertiary care centers. We envision performing a prospective observational study in memory clinics to confirm our model’s ability to assess dementia status at the same level as the expert clinician involved in dementia care. If confirmed in such a head-to-head comparison, our approach has the potential to expand the scope of machine learning for AD detection and management, and ultimately serve as an assistive screening tool for healthcare practitioners.

Study population

This study was exempted from local institutional review board approval, as all neuroimaging and clinical data were obtained in deidentified format upon request from external study centers, who ensured compliance with ethical guidelines and informed consent for all participants. No compensation was provided to participants.

We collected demographics, medical history, neuropsychological tests, and functional assessments as well as magnetic resonance imaging (MRI) scans from 8 cohorts (Table  1 ), totaling 8916 participants after assessing for inclusion criteria. There were 4550 participants with normal cognition (NC), 2412 participants with mild cognitive impairment (MCI), 1606 participants with Alzheimer’s disease dementia (AD) and 348 participants with dementia due to other causes. The eight cohorts include the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset ( n  = 1821) 34 , 35 , 36 , the National Alzheimer’s Coordinating Center (NACC) dataset ( n  = 4822) 21 , 22 , the frontotemporal lobar degeneration neuroimaging initiative (NIFD) dataset ( n  = 253) 37 , the Parkinson’s Progression Marker Initiative (PPMI) dataset ( n  = 198) 38 , the Australian Imaging, Biomarker and Lifestyle Flagship Study of Ageing (AIBL) dataset ( n  = 661) 39 , 40 , 41 , the Open Access Series of Imaging Studies-3 (OASIS) dataset ( n  = 666) 42 , the Framingham Heart Study (FHS) dataset ( n  = 313) 43 , 44 , and in-house data maintained by the Lewy Body Dementia Center for Excellence at Stanford University (LBDSU) ( n  = 182) 45 .

We labeled the participants according to the clinical diagnosis (See  Supplementary Information : Data to clinicians and diagnostic criterion). Subjects were labeled according to the clinical diagnoses provided by each study cohort. We kept MCI diagnoses without further consideration of underlying etiology to simulate a realistic spectrum of MCI presentations. For any subjects with documented dementia and primary diagnosis of Alzheimer’s disease dementia, an AD label was assigned regardless of the presence of additional dementing comorbidities. Subjects with dementia but without confirmed AD diagnosis were labeled as nADD. Notably, we elected to conglomerate all nADD subtypes into a singular label given that subdividing model training across an arbitrary number of prediction tasks ran the risk of diluting overall diagnostic accuracy. The ensemble of these 8 cohorts provided us a considerable number of participants with various forms of dementias as their primary diagnosis, including Alzheimer’s disease dementia (AD, n  = 1606), Lewy body dementia (LBD, n  = 63), frontotemporal dementia (FTD, n  = 193), vascular dementia (VD, n  = 21), and other causes of dementia ( n  = 237). We provided a full survey of nADD dementias by cohort in the  Supplementary Information (Table  S9 ).

Data inclusion criterion

Subjects from each cohort were eligible for study inclusion if they had at least one T1-weighted volumetric MRI scan within 6 months of an officially documented diagnosis. We additionally excluded all MRI scans with fewer than 60 slices. For subjects with multiple MRIs and diagnosis records within a 6-month period, we selected the closest pairing of neuroimaging and diagnostic label. Therefore, only one MRI per subject was used. For the NACC and the OASIS cohorts, we further queried all available variables relating to demographics, past medical history, neuropsychological testing, and functional assessments. We did not use the availability of non-imaging features to exclude individuals in these cohorts and used K-nearest neighbor imputation for any missing data fields. Our overall data inclusion workflow may be found in Fig.  S9 , where we reported the total number of subjects from each cohort before and after application of the inclusion criterion. See Information Availability by Cohort in the Supplementary Information .

MRI harmonization and preprocessing

To harmonize neuroimaging data between cohorts, we developed a pipeline of preprocessing operations (Fig.  S10 ) that was applied in identical fashion to all MRIs used in our study. This pipeline broadly consisted of two phases of registration to a standard MNI-152 template. We describe Phase 1 as follows:

Scan axes were reconfigured to match the standard orientation of MNI-152 space.

Using an automated thresholding technique, a 3D volume-of-interest within the original MRI was identified containing only areas with brain tissue.

The volume-of-interest was skull-stripped to isolate brain pixels.

A preliminary linear registration of the skull-stripped brain to a standard MNI-152 template was performed. This step approximated a linear transformation matrix from the original MRI space to the MNI-152 space.

Phase 2 was designed to fine-tune the quality of linear registration and parcellate the brain into discrete regions. These goals were accomplished by the following steps:

The transformation matrix computed from linear registration in Phase 1 was applied to the original MRI scan.

Skull stripping was once again performed after applying the linear registration computed from the initial volume of interest to isolate brain tissue from the full registered MRI scan.

Linear registration was applied again to alleviate any misalignments to MNI-152 space.

Bias field correction was applied to account for magnetic field inhomogeneities.

The brain was parcellated by applying a nonlinear warp of the Hammersmith Adult brain atlas to the post-processed MRI.

All steps of our MRI-processing pipeline were conducted using FMRIB Software Library v6.0 (FSL) (Analysis Group, Oxford University). The overall preprocessing workflow was inspired by the harmonization protocols of the UK Biobank ( https://git.fmrib.ox.ac.uk/falmagro/UK_biobank_pipeline_v_1 ). We manually inspected the outcome of the MRI pipeline on each scan to filter out cases with poor quality or significant processing artifacts.

Evaluation of MRI harmonization

We further assessed our image harmonization pipeline by clustering the data using the t-distributed stochastic neighbor embedding (tSNE) algorithm 46 . We performed this procedure in order to ensure that (i) input data for all models was free of site-, scanner-, and cohort-specific biases and (ii) such biases could not be learned by a predictive model. To accomplish (i), we performed tSNE using pixel values from post-processed, 8x-downsampled MRI scans. For (ii), we performed tSNE using hidden-layer activations derived from the penultimate layer of a convolutional neural network (CNN) developed for our prediction tasks (see “Model Development” below). For the NACC dataset, we assessed clustering of downsampled MRIs and hidden layer activations based on specific Alzheimer’s Disease Research Centers (ADRCs) and scanner manufacturers (i.e., Siemens, Philips, and General Electric). We also repeated tSNE analysis based on specific cohorts (i.e., NACC, ADNI, FHS, etc.) using all available MRIs across our datasets. We also calculated mutual information scores (MIS) between ADRC ID, scanner brand, and diagnostic labels (NC, MCI, AD, and nADD) in the NACC dataset. This metric calculates the degree of similarity between two sets of labels on a common set of data. As with the tSNE analysis, the MIS calculation helped us to exclude the presence of confounding site- and scanner-specific biases on MRI data.

Harmonization of non-imaging data

To harmonize the non-imaging variables across datasets, we first surveyed the available clinical data in all eight cohorts (See Information Availability by Cohort and Non-Imaging Features Used in Model Development  in the  Supplementary Information ). We specifically examined information related to demographics, past medical history, neuropsychological test results, and functional assessments. Across a range of clinical features, we found the greatest availability of information in the NACC and the OASIS datasets. Additionally, given that the NACC and the OASIS cohorts follow Uniform Data Set (UDS) guidelines, we were able to make use of validated conversion scales between UDS versions 2.0 and 3.0 to align all cognitive measurements onto a common scale. We supply a full listing of clinical variables along with missing information rates per cohort in Fig.  S11 .

Overview of the prediction framework

We developed predictive models to meet two main objectives. The first, which we designated the COG task, was to predict the overall degree of cognitive impairment (either NC, MCI, or dementia [DE]) in each participant based on neuroimaging. To meet this goal, we predicted a continuous 0–2 score (NC: 0, MCI: 1, DE: 2), which we denote as the DEmentia MOdel (DEMO) score. Of note, the COG task may also be regarded as consisting of three separate subtasks: (i) separation of NC from MCI and DE (COG NC task), (ii) separation of MCI from NC and DE (COG MCI task), and (iii) separation of DE from NC and MCI (COG DE task). The second objective, which we designated the ADD task, was to predict whether a participant held a diagnosis of AD or nADD given that they were already predicted as DE in the COG task. For ease of reference, we denoted the probability of a person holding an AD diagnosis as the ALZheimer (ALZ) score. Following the sequential completion of the COG and ADD tasks, we were able to successfully separate AD participants from NC, MCI, and nADD subjects.

MRI-only model

We used post-processed volumetric MRIs as inputs and trained a CNN model. To transfer information between the COG and ADD tasks, we trained a common set of convolutional blocks to act as general-purpose feature extractors. The DEMO and the ALZ scores were then calculated separately by appending respective fully connected layers to the shared convolutional backbone. We conducted the COG task as a regression problem using mean square error loss between the DEMO score and available cognitive labels. We performed the ADD task as a classification problem using binary cross entropy loss between the reference AD label and the ALZ score. The MRI-only model was trained using the NACC dataset and validated on all the other cohorts. To facilitate presentation of results, we pooled data from all the external cohorts (ADNI, AIBL, FHS, LBDSU, NIFD, OASIS, and PPMI), and computed all the model performance metrics.

Non-imaging model

In addition to an MRI-only model, we developed a range of traditional machine learning classifiers using all available non-imaging variables shared between the NACC and the OASIS datasets. We first compiled vectors of demographics, past medical history, neuropsychological test results, and functional assessments. We scaled continuous variables by their mean and standard deviations and one-hot encoded categorical variables. These non-imaging data vectors were then passed as input to CatBoost, XGBoost, random forest, decision tree, multi-layer perceptron, support vector machine and K-nearest neighbor algorithms. Like the MRI-only model, each non-imaging model was sequentially trained to complete the COG and the ADD tasks by calculating the DEMO and the ALZ scores, respectively. We ultimately found that a CatBoost model yielded the best overall performance per area-under-receiver-operating-characteristic curve (AUC) and area-under-precision-recall curve (AP) metrics. We, therefore, selected this algorithm as the basis for follow-up analyses.

To mimic a clinical neurology setting, we developed a non-imaging model using data that is routinely collected for dementia diagnosis. A full listing of the variables used as input may be found in our  Supplementary Information . While some features such as genetic status (APOE ε4 allele) 47 , or cerebrospinal fluid measures 10 have great predictive value, we have purposefully not included them for model development because they are not part of the standard clinical work-up of dementia.

To infer the extent to which completeness of non-imaging datasets influenced model performance, we conducted multiple experiments using different combinations of clinical data variables. The following combinations were input to the CatBoost algorithm for comparison: (1) demographic characteristics alone, (2) demographic characteristics and neuropsychological tests, (3) demographic characteristics and functional assessments, (4) demographic characteristics and past medical history, (5) demographic characteristics, neuropsychological tests and functional assessments, (6) demographic characteristics, neuropsychological tests and past medical history, and (7) demographic characteristics, neuropsychological tests, past medical history, and functional assessments.

Fusion model

To best leverage every aspect of the available data, we combined both MRI and non-imaging features into a common “fusion” model for the COG and the ADD tasks. The combination of data sources was accomplished by concatenating the DEMO and the ALZ scores derived from the MRI-only model to lists of clinical variables. The resultant vectors were then given as input to traditional machine learning classifiers as described above. Based on the AUC and the AP metrics, we ultimately found that a CNN linked with CatBoost model yielded the highest performance in discriminating different cognitive categories; the combination of CNN and CatBoost models was thus used as the final fusion model for all further experiments. Similarly, to our procedure with the non-imaging model, we studied how MRI features interacted with different subsets of demographic, past medical history, neuropsychological, and functional assessment variables. As with our non-imaging model, development and validation of fusion models was limited to NACC and OASIS only given limited availability of non-imaging data in other cohorts.

Training strategy and data splitting

We trained all models on the NACC dataset using cross validation. NACC was randomly divided into 5 folds of equal size with constant ratios of NC, MCI, AD, and nADD cases. We trained the model on 3 of the 5 folds and used the remaining two folds for validation and testing, respectively. Each tuned model was also tested on the full set of available cases from external datasets. Performance metrics for all models were reported as a mean across five folds of cross validation along with standard deviations and 95% confident intervals. A graphical summary of our cross-validation strategy may be found within Fig.  S12 . Prior to training, we also set aside two specialized cohorts within NACC for neuropathologic validation and head-to-head comparison with clinicians. In the former case, we identified 74 subjects from whom post-mortem neuropathological data was available within 2 years of an MRI scan. In the latter, we randomly selected 100 age- and sex-matched groups of patients (25 per diagnostic category) to provide simulated cases to expert clinicians.

SHAP analysis

SHAP is a unified framework for interpreting machine learning models which estimates the contribution of each feature by averaging over all possible marginal contributions to a prediction task 23 . Though initially developed for game theory applications 48 , this approach may be used in deep learning-based computer vision by considering each image voxel or a network node as a unique feature. By assigning SHAP values to specific voxels or by mapping internal network nodes back to the native imaging space, heatmaps may be constructed over input MRIs.

Though a variety of methods exist for estimating SHAP values, we implemented a modified version of the DeepLIFT algorithm 49 , which computes SHAP by estimating differences in model activations during backpropagation relative to a standard reference. We established this reference by integrating over a “background” of training MRIs to estimate a dataset-wide expected value. For each testing example, we then calculated SHAP values for the overall CNN model as well as for specific internal layers. Two sets of SHAP values were estimated for the COG and ADD tasks, respectively. SHAP values calculated over the full model were directly mapped back to native MRI pixels whereas those derived for internal layers were translated to the native imaging space via nearest neighbor interpolation.

Network analysis

We sought to perform a region-by-region graph analysis of SHAP values to determine whether consistent differences in ADD and nADD populations could be demonstrated. To visualize the relationship of SHAP scores across various brain regions, we created graphical representations of inter-region SHAP correlations within the brain. We derived region-specific scores by averaging voxel-wise SHAP values according to their location within the registered MRI. Subsequently, we constructed acyclic graphs in which nodes were defined as specific brain regions and edges as inter-regional correlations measured by Spearman’s rank correlation and Pearson correlation coefficient, separately. To facilitate visualization and convey structural information, we manually aligned the nodes to a radiographic projection of the brain.

Once correlation values were calculated between every pair of nodes, we filtered out the edges with p value larger than 0.05 and ranked the remaining edges according to the absolute correlation value. We used only the top N edges ( N  = 100 for sagittal view, N  = 200 for axial view) for the graph. We used color to indicate the sign of correlation and thickness to represent the magnitude of correlation. We used the following formula to derive the thickness:

where the threshold is defined as the minimum of the absolute value of all selected edges’ correlation value. The radius of nodes represents the weighted degree of the node which is defined as the sum of the edge weights for edges incident to that node. More specifically, we calculated the radius using the following equation:

In the above equation, we used 20 as a bias term to ensure that every node has at-least a minimal size to be visible on the graph. Note as well that the digit inside each node represents the index of the region name. Derivation of axial and sagittal nodes from the Hammersmith atlas is elaborated in Table  S5 .

Neuropathologic validation

Neuropathologic evaluations are considered to be the gold standard for confirming the presence and severity of neurodegenerative diseases 50 . We validated our model’s ability to identify regions of high risk of dementia by comparing the spatial distribution of the model-derived scores with post-mortem neuropathological data from NACC, FHS, and ADNI study cohorts, derived from the National Institute on Aging Alzheimer’s Association guidelines for the neuropathologic assessment of AD 51 . Hundred and ten participants from NACC ( n  = 74), ADNI ( n  = 25) and FHS ( n  = 11) who met the study inclusion criteria, had MRI scans taken within 2 years of death and with neuropathologic data were included in the neuropathologic validation. The data was harmonized in the format of the Neuropathology Data Form Version 10 of the NACC established by the National Institute on Aging. The neuropathological lesions of AD (i.e., amyloid β deposits (Aβ), neurofibrillary tangles (NFTs), and neuritic plaques (NPs)) were assessed in the entorhinal, hippocampal, frontal, temporal, parietal, and occipital cortices. The regions were based on those proposed for standardized neuropathological assessment of AD and the severity of the various pathologies were classified into four semi-quantitative score categories (0 = None, 1 = Mild, 2 = Moderate, 3 = Severe) 52 . Based on the NIA-AA protocol, the severity of neuropathologic changes were evaluated using a global “ABC” score which incorporates histopathologic assessments of amyloid β deposits by the method of Thal phases: 53 (A), staging of neurofibrillary tangles (B) silver-based histochemistry 54 , or phospho-tau immunohistochemistry 55 , and scoring of neuritic plaques (C). Spearman’s rank correlation was used to correlate the DEMO score predictions with the A, B, C scores, and ANOVA and Tukey’s tests were used to assess the differences in the mean DEMO scores across the different levels of the scoring categories. Lastly, a subset of the participants from ADNI ( n  = 25) and FHS ( n  = 11) had regional semi-quantitative Aβ, NFT, and NP scores, which was also used to validate the model predictions.

We sought to test our model’s predictions against the diagnostic acumen of clinicians who are involved in care of patients with dementia. We recruited an international cohort of practicing neurologists and neuroradiologists to participate in simulated diagnostic tasks using a subset of NACC cases (see “Training strategy and data splitting” above). Neurologists were provided with 100 cases that included imaging data (T1-weighted brain MRI scans) and non-imaging data (demographics, medical history, neuropsychological tests, and functional assessments) and asked to provide diagnostic impressions of NC, MCI, AD, and nADD. Notably, the model was not directly compared to neurologists for the ADD task given that our framework only performs this prediction on patients internally identified as demented. Due to this computational pre-selection, it was not feasible to consistently compare a common cohort of persons with neurologists who also must perform a differential diagnosis of NC, MCI, AD, and nADD. Neuroradiologists were provided with imaging data (T1-weighted brain MRI scans), age, and gender from 50 known DE cases and then asked to provide diagnostic impressions of AD or nADD. For each case, the neuroradiologists also answered a questionnaire to grade the extent of atrophy in each sub-region of the brain on a scale of 0 to 4, where higher values indicate greater atrophy. A case sample and example questionnaires provided to neurologists and neuroradiologists, respectively, may be found within the  Supplementary Information (Data to clinicians and diagnostic criterion). For both groups of clinicians, we also calculated inter-annotator agreement using Cohen’s kappa (κ). Additionally, to compare our machine learning models to neuropsychological assessments, we performed the COG NC , COG DE , and ADD tasks using all possible whole number cutoffs of neuropsychiatric test scores available in the NACC dataset. Following this approach, we performed simple thresholding for binary classifications.

Performance metrics

We presented the performance by computing the mean and the standard deviation over the model runs. We generated receiver operating characteristic (ROC) and precision-recall (PR) curves based on model predictions on the NACC test data as well as on the other datasets. For each ROC and PR curve, we also computed the area under curve (AUC & AP) values. Additionally, we computed sensitivity, specificity, F1-score and Matthews correlation coefficient on each set of model predictions. The F1-score considers both precision and recall of a test whereas the MCC is a balanced measure of quality for dataset classes of different sizes of a binary classifier. We also calculated inter-annotator agreement using Cohen’s kappa (κ), as the ratio of the number of times two experts agreed on a diagnosis. We computed average pairwise κ for each sub-group task that provided an overall measure of agreement between the neurologists and the neuroradiologists, respectively.

Statistical analysis

We used one-way ANOVA test and the χ2 test for continuous and categorical variables, respectively to assess the overall levels of differences in the population characteristics between NC, MCI, AD, and nADD groups across the study cohorts. To validate our CNN model, we evaluated whether the presence and severity of the semi-quantitative neuropathology scores across the neuropathological lesions of AD (i.e., amyloid β deposits (Aβ), neurofibrillary tangles (NFTs), and neuritic plaques (NPs)) reflected the DEMO score predicted by the CNN model. We stratified the lesions based on A, B, and C scores and used Spearman’s rank correlation to assess their relationship with the DEMO scores. Next using one-way ANOVA analysis, we evaluated the differences in the mean DEMO scores across the different levels of the scoring categories for the A, B, and C scores. We used the Tukey-Kramer test to identify the pairwise statistically significant differences in the mean DEMO score between the levels of scoring categories (0–3). Similarly, to analyze the correspondence between SHAP values and a known marker of neurodegenerative disease, we correlated SHAPs with the radiologist impressions of atrophy. Utilizing the segmentation maps derived from each participant, we calculated regional SHAP averages on each of the 50 test cases given to neuroradiologists with the 0–4 regional atrophy scales assigned by the clinicians. We calculated Pearson’s correlation coefficients with two-tailed p values that indicates the probability that an uncorrelated system producing Pearson’s correlation coefficient as extreme as the observed value in the neuroanatomic regions known to be implicated in AD pathology. All statistical analyses were conducted at a significance level of 0.05. Confidence intervals for model performance were calculated by assuming a normal distribution of AUC and AP values across cross-validation experiments using t-student distribution with 4 degrees of freedom.

Computational hardware and software

We processed all MRIs and non-imaging data on a computing workstation with Intel i9 14-core 3.3 GHz processor, and 4 NVIDIA RTX 2080Ti GPUs. Python (version 3.7.7) was used for software development. Each deep learning model was developed using PyTorch (version 1.5.1), and plots were generated using the Python library matplotlib (version 3.1.1) and numpy (version 1.18.1) was used for vectorized numerical computation. Other Python libraries used to support data analysis include pandas (version 1.0.3), scipy (version 1.3.1), tensorflow (version 1.14.0), tensorboardX (version 1.9), torchvision (version 0.6) and scikit-learn (version 0.22.1). Using a single 2080Ti GPU, the average run time for training the deep learning model was 10 h, and inference task took less than a minute. All clinicians reviewed MRIs using 3D Slicer (version 4.10.2) ( https://www.slicer.org/ ) and recorded impressions in REDCap (version 11.1.3). Additionally, statistics for neuropathology analysis were completed using SAS (version 9.4).

Reporting summary

Further information on research design is available in the  Nature Research Reporting Summary linked to this article.

Data availability

Data from ADNI, AIBL, NACC, NIFD, OASIS and PPMI can be downloaded from publicly available resources. Data from FHS and LBDSU are available upon request and will be subjected to institutional approval. Source data for figures are provided with this paper.  Source data are provided with this paper.

Code availability

Python scripts are made available on GitHub ( https://github.com/vkola-lab/ncomms2022 ).

Nichols, E. et al. Global, regional, and national burden of Alzheimer’s disease and other dementias, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol. 18 , 88–106 (2019).

Article   Google Scholar  

Mehta, K. M. & Yeo, G. W. Systematic review of dementia prevalence and incidence in United States race/ethnic populations. Alzheimer’s Dement. 13 , 72–83 (2017).

James, B. D. et al. Contribution of Alzheimer disease to mortality in the United States. Neurology 82 , 1045–1050 (2014).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Palmqvist, S. et al. Discriminative accuracy of plasma phospho-tau217 for Alzheimer disease vs other neurodegenerative disorders. Jama 324 , 772–781 (2020).

Article   CAS   PubMed   Google Scholar  

Nordberg, A. PET imaging of amyloid in Alzheimer’s disease. lancet Neurol. 3 , 519–527 (2004).

Bohnen, N. I., Djang, D. S., Herholz, K., Anzai, Y. & Minoshima, S. Effectiveness and safety of 18F-FDG PET in the evaluation of dementia: a review of the recent literature. J. Nucl. Med. 53 , 59–71 (2012).

Mattsson, N. et al. Predicting diagnosis and cognition with 18F-AV-1451 tau PET and structural MRI in Alzheimer’s disease. Alzheimer’s Dement. 15 , 570–580 (2019).

Ossenkoppele, R. et al. Associations between tau, Aβ, and cortical thickness with cognition in Alzheimer disease. Neurology 92 , e601–e612 (2019).

Sevigny, J. et al. The antibody aducanumab reduces Aβ plaques in Alzheimer’s disease. Nature 537 , 50–56 (2016).

Article   ADS   CAS   PubMed   Google Scholar  

McKhann, G. M. et al. The diagnosis of dementia due to Alzheimer’s disease: recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimer’s Dement. 7 , 263–269 (2011).

Knopman, D. et al. Practice parameter: diagnosis of dementia (an evidence-based review): report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurology 56 , 1143–1153 (2001).

Beach, T. G., Monsell, S. E., Phillips, L. E. & Kukull, W. Accuracy of the clinical diagnosis of Alzheimer disease at National Institute on Aging Alzheimer Disease Centers, 2005–2010. J. Neuropathol. Exp. Neurol. 71 , 266–273 (2012).

Article   PubMed   Google Scholar  

Dall, T. M. et al. Supply and demand analysis of the current and future US neurology workforce. Neurology 81 , 470–478 (2013).

Article   PubMed   PubMed Central   Google Scholar  

Dall, T. M. Physician workforce shortages: what do the data really say? Acad. Med 90 , 1581–1582 (2015).

Pedersen, M. et al. Artificial intelligence for clinical decision support in neurology. Brain Commun. 2 , fcaa096 (2020).

Lu, D., Popuri, K., Ding, G. W., Balachandar, R. & Beg, M. F. Multimodal and multiscale deep neural networks for the early diagnosis of Alzheimer’s disease using structural MR and FDG-PET images. Sci. Rep. 8 , 1–13 (2018).

Google Scholar  

Qiu, S. et al. Development and validation of an interpretable deep learning framework for Alzheimer’s disease classification. Brain 143 , 1920–1933 (2020).

Wada, A. et al. Differentiating Alzheimer’s disease from dementia with Lewy bodies using a deep learning technique based on structural brain connectivity. Magn. Reson. Med. Sci. 18 , 219 (2019).

Nemoto, K. et al. Differentiating dementia with lewy bodies and Alzheimer’s disease by deep learning to structural MRI. J. Neuroimaging 31 , 579–587 (2021).

Ma, D. et al. Differential diagnosis of frontotemporal dementia, Alzheimer’s disease, and normal aging using a multi-scale multi-type feature generative adversarial deep neural network on structural magnetic resonance images. Front. Neurosci. 14 , 853 (2020).

Beekly, D. L. et al. The National Alzheimer’s Coordinating Center (NACC) database: the uniform data set. Alzheimer Dis. Associated Disord. 21 , 249–258 (2007).

Beekly, D. L. et al. The national Alzheimer’s coordinating center (NACC) database: an Alzheimer disease database. Alzheimer Dis. Associated Disord. 18 , 270–277 (2004).

Lundberg, S. M. & Lee, S.-I. A Unified Approach to Interpreting Model Predictions: in Proceedings of the 31st International Conference on Neural Information Processing Systems . 4768–4777 (2017).

DeGrave, A. J., Janizek, J. D. & Lee, S.-I. AI for radiographic COVID-19 detection selects shortcuts over signal. Nat. Mach. Intell. 3 , 610–619 (2021).

Wang, Y. et al. Classification of subcortical vascular cognitive impairment using single MRI sequence and deep learning convolutional neural networks. Front. Neurosci. 13 , 627 (2019).

Castellazzi, G. et al. A machine learning approach for the differential diagnosis of Alzheimer and vascular dementia fed by MRI selected features. Front. Neuroinform. 14 , 25 (2020).

Wang, F., Kaushal, R. & Khullar, D. Should health care demand interpretable artificial intelligence or accept “black box” medicine? Ann. Intern. Med. 172 , 59–60 (2020).

Weiner, M. & Khachaturian, Z. The use of MRI and PET for clinical diagnosis of dementia and investigation of cognitive impairment: a consensus report. Alzheimer’s Assoc. Chic., IL 1 , 1–15 (2005).

Kara, S. et al. Guidelines, training and quality assurance: influence on general practitioner MRI referral quality. J. Prim. Health Care 11 , 235–242 (2019).

Article   ADS   PubMed   Google Scholar  

Bernstein, A. et al. Dementia assessment and management in primary care settings: a survey of current provider practices in the United States. BMC Health Serv. Res . 19 , 919 (2019).

Zekry, D., Hauw, J. J. & Gold, G. Mixed dementia: epidemiology, diagnosis, and treatment. J. Am. Geriatrics Soc. 50 , 1431–1438 (2002).

Graff-Radford, J. et al. New insights into atypical Alzheimer’s disease in the era of biomarkers. Lancet Neurol. 20 , 222–234 (2021).

Wind, A. W. et al. Limitations of the Mini‐Mental State Examination in diagnosing dementia in general practice. Int. J. Geriatr. psychiatry 12 , 101–108 (1997).

Jack, C. R. Jr et al. The Alzheimer’s disease neuroimaging initiative (ADNI): MRI methods. J. Magn. Reson. Imaging. 27 , 685–691 (2008).

Mueller, S. G. et al. Ways toward an early diagnosis in Alzheimer’s disease: the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Alzheimer’s Dement. 1 , 55–66 (2005).

Petersen, R. C. et al. Alzheimer’s Disease Neuroimaging Initiative (ADNI): clinical characterization. Neurology 74 , 201–209 (2010).

Boxer, A. L. et al. Frontotemporal degeneration, the next therapeutic frontier: molecules and animal models for frontotemporal degeneration drug development. Alzheimers Dement 9 , 176–188 (2013).

Marek, K. et al. The parkinson progression marker initiative (PPMI). Prog. Neurobiol. 95 , 629–635 (2011).

Article   PubMed Central   Google Scholar  

Ellis, K., Ames, D., Martins, R., Hudson, P. & Masters, C. The Australian biomarkers lifestyle and imaging flagship study of ageing. Acta Neuropsychiatr. 18 , 285 (2006).

PubMed   Google Scholar  

Ellis, K. A. et al. The Australian Imaging, Biomarkers and Lifestyle (AIBL) study of aging: methodology and baseline characteristics of 1112 individuals recruited for a longitudinal study of Alzheimer’s disease. Int Psychogeriatr. 21 , 672–687 (2009).

Ellis, K. A. et al. Addressing population aging and Alzheimer’s disease through the Australian imaging biomarkers and lifestyle study: collaboration with the Alzheimer’s Disease Neuroimaging Initiative. Alzheimers Dement 6 , 291–296 (2010).

LaMontagne, P. J. et al. OASIS-3: Longitudinal neuroimaging, clinical, and cognitive dataset for normal aging and Alzheimer disease. medRxiv , https://doi.org/10.1101/2019.12.13.19014902 (2019).

Mahmood, S. S., Levy, D., Vasan, R. S. & Wang, T. J. The Framingham Heart Study and the epidemiology of cardiovascular disease: a historical perspective. lancet 383 , 999–1008 (2014).

Massaro, J. M. et al. Managing and analysing data from a large-scale study on Framingham Offspring relating brain structure to cognitive function. Stat. Med. 23 , 351–367 (2004).

Linortner, P. et al. White matter hyperintensities related to Parkinson’s disease executive function. Mov. Disord. Clin. Pr. 7 , 629–638 (2020).

Van der Maaten, L. & Hinton, G. Visualizing data using t-SNE. J. Mach. Learn. Res. 9 (2008).

Choudhury, P., Ramanan, V. K. & Boeve, B. F. APOE ɛ4 allele testing and risk of Alzheimer disease. JAMA 325 , 484–485 (2021).

Shapley, L. S. 17. A value for n-person games . (Princeton University Press, 2016).

Shrikumar, A., Greenside, P. & Kundaje, A. Learning important features through propagating activation differences: in. Int. Conf. Mach. Learn. 70 , 3145–3153 (2017).

Besser, L. M. et al. The revised national Alzheimer’s coordinating center’s neuropathology form-available data and new analyses. J. Neuropathol. Exp. Neurol. 77 , 717–726 (2018).

Montine, T. J. et al. National Institute on Aging–Alzheimer’s Association guidelines for the neuropathologic assessment of Alzheimer’s disease: a practical approach. Acta Neuropathologica 123 , 1–11 (2012).

Hyman, B. T. et al. National Institute on Aging–Alzheimer’s Association guidelines for the neuropathologic assessment of Alzheimer’s disease. Alzheimer’s Dement. 8 , 1–13 (2012).

Thal, D. R., Rub, U., Orantes, M. & Braak, H. Phases of A beta-deposition in the human brain and its relevance for the development of AD. Neurology 58 , 1791–1800 (2002).

Braak, H. & Braak, E. Neuropathological stageing of Alzheimer-related changes. Acta Neuropathologica 82 , 239–259 (1991).

Braak, H., Alafuzoff, I., Arzberger, T., Kretzschmar, H. & Del Tredici, K. Staging of Alzheimer disease-associated neurofibrillary pathology using paraffin sections and immunocytochemistry. Acta Neuropathol. 112 , 389–404 (2006).

Download references

Acknowledgements

This project was supported by grants from the Karen Toffler Charitable Trust (VBK), the Michael J. Fox Foundation (KLP), the Lewy Body Dementia Association (KLP), the Alzheimer’s Drug Discovery Foundation (RA), the American Heart Association (20SFRN35460031, VBK), and the National Institutes of Health (R01-HL159620 [VBK], R21-CA253498 [VBK], RF1-AG062109 [RA], RF1-AG072654 [RA], U19-AG065156 [KLP], P30-AG066515 [KLP], R01-NS115114 [KLP], K23-NS075097 [KLP], U19-AG068753 [RA] and P30-AG013846 [RA, VBK]). We acknowledge the efforts of several investigators from the ADNI, AIBL, FHS, LBDSU, NACC, NIFD, OASIS, and PPMI studies for providing access to data.

Author information

These authors contributed equally: Shangran Qiu, Matthew I. Miller.

Authors and Affiliations

Department of Medicine, Boston University School of Medicine, Boston, MA, USA

Shangran Qiu, Matthew I. Miller, Joyce C. Lee, Chonghua Xue, Yunruo Ni, Yuwei Wang & Vijaya B. Kolachalama

Department of Physics, College of Arts & Sciences, Boston University, Boston, MA, USA

Shangran Qiu

Department of Anatomy and Neurobiology, Boston University School of Medicine, Boston, MA, USA

Prajakta S. Joshi, Chonghua Xue, Phillip H. Hwang & Rhoda Au

Department of General Dentistry, Boston University School of Dental Medicine, Boston, MA, USA

Prajakta S. Joshi

The Framingham Heart Study, Boston University School of Medicine, Boston, MA, USA

Prajakta S. Joshi, Jesse Mez & Rhoda Au

School of Public Health and Tropical Medicine, Tulane University, New Orleans, LA, USA

Ileana De Anda-Duran

Department of Radiology, College of Medicine, University of Nebraska Medical Center, Omaha, NE, USA

Justin A. Cramer

Department of Neurology, Boston University School of Medicine, Boston, MA, USA

Brigid C. Dwyer, Michelle C. Kaku, Sarah O’Shea, Marie-Helene Saint-Hilaire, E. Alton Sartor, Aneeta R. Saxena, Ludy C. Shih, Courtney E. Takahashi, Shuhan Zhu, Michael L. Alosco, Jesse Mez & Rhoda Au

Department of Neurology, Peking Union Medical College Hospital, Peking Union Medical College, Chinese Academy of Medical Sciences, Beijing, China

Honglin Hao, Jing Yuan & Yan Zhou

Department of Neurological Sciences, College of Medicine, University of Nebraska Medical Center, Omaha, NE, USA

Sachin Kedar, Daniel L. Murman, Arun Swaminathan & Olga Taraschenko

Department Neurology, Emory University School of Medicine, Atlanta, GA, USA

Sachin Kedar

Department Ophthalmology, Emory University School of Medicine, Atlanta, GA, USA

Department of Radiology, Lahey Hospital & Medical Center, Burlington, MA, USA

Peter H. Lee, Aaron B. Paul, Juan E. Small & Maximilian J. Smith

Department of Radiology, Boston University School of Medicine, Boston, MA, USA

Asim Z. Mian

Department of Radiology, Peking Union Medical College Hospital, Peking Union Medical College, Chinese Academy of Medical Sciences, Beijing, China

Boston University Alzheimer’s Disease Research Center, Boston, MA, USA

Michael L. Alosco, Jesse Mez, Thor D. Stein, Rhoda Au & Vijaya B. Kolachalama

Department of Pathology and Laboratory Medicine, Boston University School of Medicine, Boston, MA, USA

Thor D. Stein

Boston VA Healthcare System, Boston, MA, USA

Bedford VA Healthcare System, Bedford, MA, USA

Department of Neurology, Stanford University, Palo Alto, CA, USA

Kathleen L. Poston

Department of Epidemiology, Boston University School of Public Health, Boston, MA, USA

Department of Computer Science, Boston University, Boston, MA, USA

Vijaya B. Kolachalama

Faculty of Computing & Data Sciences, Boston University, Boston, MA, USA

You can also search for this author in PubMed   Google Scholar

Contributions

S.Q. and M.I.M. contributed equally to this work. S.Q. and M.I.M. performed literature search. S.Q. designed and developed the deep learning framework, performed image processing, and constructed the models. S.Q., P.S.J., and J.C.L., performed data collection, data harmonization and statistical analysis. C.X., Y.N., Y.W., I.D.A., and P.H.H. assisted with data collection and performed sub-group analyses. S.Q., M.I.M., P.S.J., J.C.L., and C.X. generated the figures and tables. P.S.J. and S.Q. performed the analysis on the neuropathology data. J.A.C., B.C.D., H.H., M.C.K., S.K., P.H.L., A.Z.M., D.L.M., S.O., A.B.P., M-H.S-H., E.A.S., A.R.S., L.C.S., J.E.S., M.J.S., A.S., C.E.T., O.T., H.Y., J.Y., Y.Z., S.Z. and K.L.P. are the practicing clinicians who reviewed the cases. M.L.A., J.M., T.D.S, K.L.P., and R.A. provided the clinical relevance context. K.L.P. and R.A. provided access to data. M.I.M. and V.B.K. wrote the manuscript. All authors reviewed and approved the manuscript. V.B.K. conceived, designed, and directed the entire study.

Corresponding author

Correspondence to Vijaya B. Kolachalama .

Ethics declarations

Competing interests.

V.B.K. reports honoraria from invited scientific presentations to industry not exceeding $5000/year. He also serves as a consultant to Davos Alzheimer’s Collaborative. R.A. is a scientific advisor to Signant Health and consultant to Biogen. K.L.P. reports honoraria from invited scientific presentations to universities and professional societies not exceeding $5,000/year and has received consulting fees from Curasen. The remaining authors declare no competing interests.

Peer review

Peer review information.

Nature Communications thanks Arman Eshaghi and the other anonymous reviewer(s) for their contribution to the peer review of this work. Peer review reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary information, peer review file, reporting summary, source data, source data, rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Qiu, S., Miller, M.I., Joshi, P.S. et al. Multimodal deep learning for Alzheimer’s disease dementia assessment. Nat Commun 13 , 3404 (2022). https://doi.org/10.1038/s41467-022-31037-5

Download citation

Received : 21 August 2021

Accepted : 06 May 2022

Published : 20 June 2022

DOI : https://doi.org/10.1038/s41467-022-31037-5

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

Multimodal diagnosis model of alzheimer’s disease based on improved transformer.

BioMedical Engineering OnLine (2024)

Comparing a pre-defined versus deep learning approach for extracting brain atrophy patterns to predict cognitive decline due to Alzheimer’s disease in patients with mild cognitive symptoms

• Ida Arvidsson
• Olof Strandberg
• Niklas Mattsson-Carlgren

Alzheimer's Research & Therapy (2024)

Recent advances in Alzheimer’s disease: Mechanisms, clinical trials and new drug development strategies

• Yinglu Zhang

Signal Transduction and Targeted Therapy (2024)

Identifying autism spectrum disorder from multi-modal data with privacy-preserving

• Haishuai Wang

npj Mental Health Research (2024)

Leveraging electronic health records and knowledge networks for Alzheimer’s disease prediction and sex-specific biological insights

• Alice S. Tang
• Katherine P. Rankin
• Marina Sirota

Nature Aging (2024)

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: Translational Research newsletter — top stories in biotechnology, drug discovery and pharma.

An official website of the United States government

Here's how you know

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

APOE Genetics as a Major Determinant of Alzheimer’s Disease Pathobiology

On this page:

Purpose and Background

Registration, contact information.

Alzheimer's Disease and Related Dementia (AD/ADRD) physicians and biologists, and scientists interested in brain aging, neurodegenerative diseases, genetics, whole genome sequencing, global and local genetic ancestry analyses, single-cell genomic analyses, neurons-astrocytes-microglia interactions, and genetically driven therapeutic targets.

September 5, 2024 | 8:00 a.m. – 5:00 p.m. ET September 6, 2024 | 8:00 a.m. – 3:00 p.m. ET

This is a hybrid workshop. Participants can attend virtually via Zoom, or In Person on NIH Main Campus:

John Edwards Porter Neuroscience Research Center, Building 35 9000 Rockville Pike, Bethesda, MD 20892

This workshop will bring together investigators who have intensely investigated the effects of APOE on the brain, and the mechanism of disease to engage with the audience on new technologies centered on APOE for the successful identification of genetically driven therapeutic approaches. The expertise of investigators ranging from genetics, molecular biology, functional genomics, and AI/ML fields will also present the latest findings across diverse populations both in terms of global and local genetic ancestries. Discussion of supported research will for example examine why APOE4 risk differs dramatically among populations and how these differences can be leveraged to better understand AD pathobiology. During the two-day meeting participants will explore the best next steps to develop study approaches that will pinpoint the molecular, genetic, and epigenetic factors associated with APOE risk for AD. Information provided at this workshop is expected to generate new mechanistic insights into APOE impact in the brain and its multivariate cell types. 

Pease register in advance for this webinar.

Register for this workshop

All times are in Eastern Daylight Time.

Day 1 | Thursday, September 5

8:00 a.m. Welcome

8:10 a.m. NIA Opening Remarks, Eliezer Masliah, M.D., National Institute on Aging (NIA)

8:30 a.m. Keynote Lecture, APOE4 as a toxic gain of function molecule, David M. Holtzman, M.D., Washington University 

9:00 a.m. Session 1 | APOE genetics (Part 1) Session Moderator: Jeffery (Jeff) Vance, M.D., Ph.D., University of Miami 

• The APOE4 Story: From Discovery to Diversity, Peggy Pericak-Vance, Ph.D., University of Miami
• Local versus global ancestry in APOE: African and African American, Hispanic, and Asian populations, Farid Rajabli, Ph.D., University of Miami
• Differential APOE effects on gene expression in different human brain cell types, Anthony (Tony) Griswold, Ph.D., University of Miami  

10:20 a.m.   Break

10:35 a.m. Session 1 | APOE genetics (Part 1 Continued)

• Dissecting APOE biology by CRISPR-based functional genomics, Martin Kampmann, Ph.D., University of California, San Francisco
• South Asians in India (LASI-DAD): Impact of APOE4 with Social Determinants of Health (SDOH), Sharon Kardia, Ph.D., University of Michigan
• A Genetic Modifier of ε4/AD Association and APOE Expression, Lindsay Farrer, Ph.D., Boston University 

11:35 a.m. Session 2 | APOE genetics (Part 2) Moderator: Jeffery (Jeff) Vance, M.D., Ph.D., University of Miami 

• PSG haplotype is protective for APOE4, Jeffery (Jeff) Vance, M.D., Ph.D., University of Miami
• Fibronectin 1 and APOE- ε4, Richard Mayeux, M.D., Columbia University 
• Interaction of Haptoglobin and APOE in Alzheimer Disease, Jonathan Haines, Ph.D., Case Western Reserve University 

12:30 p.m. Lunch

1:15p.m. Session 2 | APOE genetics (Part 2 Continued)

• APOE ε2 Allele and Protective Variants in APOE ε4/ε4 Carriers on Alzheimer’s Disease Risk, Gyungah Jun, Ph.D., Boston University
• Rare Protective APOE Variants, Michael Greicius, M.D., MPH., Stanford University

1:55 p.m. Session 3 | Mechanisms of Disease-APOE Moderator: Takahisa Kanekiyo, M.D., Ph.D., Mayo Clinic, Jacksonville, FL 

• APOE-Genotype Dependent Single-Cell Transcriptomics of Alzheimer's Disease, Li Hui Tsai, Ph.D., Massachusetts Institute of Technology 
• The role of APOE genotype in microglial behavior and gene expression, Alison Goate, D.Phil., Mount Sinai 
• The role of APOE genetics in glial lipid metabolism and inflammation, Julia TCW, Ph.D., Boston University 
• The role of genetic variants and their influence on the immune response in myeloid cells/microglia, Christopher (Chris) Glass, M.D., Ph.D., University of California, San Diego 

3:15 p.m. Break

3:30 p.m. Session 3 | Mechanisms of Disease-APOE (Continued)

• An allelic series of lipidated ApoE drives CNS lipofuscinosis , Gilbert (Gil) Di Paolo, Ph.D., Denali Therapeutics 
• APOE4/4 is linked to damaging lipid droplets in Alzheimer's disease microglia, Michael Haney, Ph.D., University of Pennsylvania
• Biology of APOE Protective Variants, Yadong Huang, M.D., Ph.D., Gladstone Institute/ University of California, San Francisco 
• Biological effects of APOE3ch on amyloid-induced tau seeding/spreading, Yun Chen, Washington University 

4:50 p.m. Wrap Up

5:00 p.m. Adjourn

Day 2 | Friday, September 6

8:00 a.m. Session 4 | New technologies and moving towards therapeutics of APOE4 Moderator:  Julia TCW, Ph.D., Boston University  

• (Machine) Learning Features of the Alzheimer’s Disease Landscape, Olivier Lichtarge, M.D., Ph.D., Baylor College of Medicine
• Leveraging deep molecular profiling to understand APOE dependent and independent pathology, Carlos Cruchage, Ph.D., Washington University in St. Louis
• APOE4 impact on vasculature, Sally Temple, Ph.D., Neural Stem Cell Institute
• ApoeE2 and its role in plaque deposition, neuroinflammation, and neurodegeneration, Bradley (Brad) Hyman, M.D., Ph.D., Massachusetts General Hospital
• Clinical trial of APOE2 gene therapy, Ronald Crystal, M.D., Weill Cornell Medical College

9:50 a.m. Break

10:10 a.m. Session 4 | New technologies and moving towards therapeutics of APOE4 (Continued)

• Potential therapeutic role for peripheral APOE, Guojun Bu, Ph.D., Hong Kong University of Science and Technology  
• Antisense oligonucleotides for Alzheimer’s disease – a focus on APOE, Hien Zhao, Ph.D., Ionis Pharmaceuticals
• RNAi Modulation of ApoE: Delicate Balance between Plague Clearance and Glia Activation, Anastasia Khvorova, Ph.D., University of Massachusetts
• Combination therapy in NACC and ADNI Alzheimer’s participants: Impact of APOE genotype and Sex, Francesca Vitali, Ph.D., University of Arizona 
• Using biomarkers in persons with different APOE variants to inform the study, treatment and prevention of AD, Eric Reiman, M.D., Banner Health
• Therapeutic Correction of ApoE4-Mediated Endolysosomal Dysfunction in Alzheimer's Disease, Joachim Herz, M.D., University of Texas Southwestern Medical Center 

12:00 p.m. Lunch

1:00 p.m. Session 5 | Brainstorm Moderator: David M. Holtzman, M.D., Washington University and Jeffery (Jeff) Vance, M.D., Ph.D., University of Miami 

• What are the therapeutic implications of lowering or raising APOE variants?
• What more do we need to understand from a mechanistic standpoint?
• What else needs to be understood about APOE variants and the effect in different ancestries?
• How does APOE variant impact Aβ immunotherapy and other diseases?

2:45 p.m. Wrap Up: Discussion Summary and Meeting Outcomes

3:00 p.m. Adjourn

Please contact Marilyn Miller at [email protected] , Michael Bennani at [email protected] , and Tiffany Rolle at  [email protected] for questions you may have about the workshop.

Reasonable Accommodations: If you need reasonable accommodation to participate in this event, please contact the meeting organizer listed under Contact information. Please make your request no later than 1 week before the event.

nia.nih.gov

An official website of the National Institutes of Health

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List

Current understanding of Alzheimer’s disease diagnosis and treatment

Jason weller.

1 Department of Neurology, Boston VA Hospital, 150 South Huntington Street, Jamaica Plain, MA, 02130, USA

2 Department of Neurology, Boston University School of Medicine, 72 East Concord Street C-309, Boston, MA, USA

Andrew Budson

Alzheimer’s disease is the most common cause of dementia worldwide, with the prevalence continuing to grow in part because of the aging world population. This neurodegenerative disease process is characterized classically by two hallmark pathologies: β-amyloid plaque deposition and neurofibrillary tangles of hyperphosphorylated tau. Diagnosis is based upon clinical presentation fulfilling several criteria as well as fluid and imaging biomarkers. Treatment is currently targeted toward symptomatic therapy, although trials are underway that aim to reduce the production and overall burden of pathology within the brain. Here, we discuss recent advances in our understanding of the clinical evaluation and treatment of Alzheimer’s disease, with updates regarding clinical trials still in progress.

Dementia is a clinical syndrome characterized by progressive decline in two or more cognitive domains, including memory, language, executive and visuospatial function, personality, and behavior, which causes loss of abilities to perform instrumental and/or basic activities of daily living. Alzheimer’s disease (AD) is by far the most common cause of dementia and accounts for up to 80% of all dementia diagnoses 1 . Although the overall death rate in the United States from stroke and cardiovascular disease is decreasing, the proportion of deaths related to AD is going up, increasing by 89% between 2000 and 2014 2 . Direct and indirect costs for healthcare related to AD are estimated at nearly $500 billion annually 3 . The definitive diagnosis of AD requires post-mortem evaluation of brain tissue, though cerebrospinal fluid (CSF) and positron emission tomography (PET) biomarkers combined with several relatively new clinical criteria can aid diagnosis in living patients 4 . Current treatments available include cholinesterase inhibitors for patients with any stage of AD dementia and memantine for people with moderate-to-severe AD dementia. These medications have been shown to enhance the quality of life for both patient and caregiver when prescribed at the appropriate time during the course of illness; however, they do not change the course of illness or the rate of decline 5 .

Clinical research is advancing toward more definitive treatment of the hallmark pathology in AD with the expectation that these therapies will attenuate the progressive cognitive decline associated with this illness ( Figure 1 ). This review will attempt to summarize the accepted evaluation methods and describe current and future therapies for patients with suspected AD.

Rate of decline of memory (M) over time (t, months to years). Memory declines slowly in normal aging (1). Alzheimer’s disease is marked by more rapid cognitive decline, often starting earlier in life (2). Current therapies enhance cognition without changing the rate of decline in AD (3). The anticipated effect of novel therapies is reduction in the rate of decline (4).

Building upon the original 1984 diagnostic criteria, the National Institute on Aging–Alzheimer’s Association (NIA–AA) revised the clinical criteria for the diagnosis of mild cognitive impairment (MCI) and the different stages of dementia due to AD in 2011 6 – 8 . The use of supportive biomarker evidence (imaging, serum, and CSF) of AD pathology were included to aid in the delineation of AD from other forms of dementia as well as in the diagnosis of MCI due to AD. The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) re-classified delirium, dementia, amnestic and other geriatric cognitive disorders into the more encompassing neurocognitive disorders 9 . This change was made to better discriminate between different neurodegenerative diseases, such as AD, dementia with Lewy bodies, and frontotemporal dementia, as well as to include both major neurocognitive disorder (equivalent to dementia) and mild neurocognitive disorder (equivalent to MCI) 4 . Finally, the newer criteria allow for the use of current and future biomarkers in the diagnosis of degenerative brain disease.

The development of non-invasive diagnostic imaging recently resulted in a test which increases the diagnostic accuracy in AD 10 . After injection of a radiolabeled tracer agent, patients undergo a specialized PET scan that detects the deposition of amyloid-β (Aβ) peptides into plaques in the living brain. In 2012, clinicians were able to accurately diagnose the disease (later autopsy proven) using this method with up to 96% sensitivity and 100% specificity. Over the next year, this same test demonstrated similar results in patients with milder disease 11 . Nearly a decade after researchers at the University of Pittsburgh created the first tracer, the US Food and Drug Administration approved the use of florbetapir for the detection of AD pathology. Now, the list of amyloid-specific PET ligands includes florbetaben and flutemetamol in addition to florbetapir, all of which have a similar profile 12 , 13 . However, the use of amyloid PET imaging in practice is still limited owing to its cost for most patients, as it is not covered by most insurance carriers. Currently, the majority of patients who undergo amyloid PET imaging do so as part of participation in clinical trials.

A more-invasive but less-costly evaluation involves examination of CSF for Aβ42, hyperphosphorylated tau peptide (p-tau), and total tau protein content 14 . This method has slightly less diagnostic accuracy (85–90%), carries the risks and inconveniences involved with a lumbar puncture procedure, and often takes weeks to obtain results because of the dearth of laboratory facilities which perform the fluid analysis. However, a head-to-head comparison showed no difference in diagnostic accuracy between CSF Aβ42:p-tau ratio and amyloid PET imaging biomarkers, suggesting that the best test for individual patients depends upon availability, cost, and patient/provider preference 15 . Less-invasive serum assays designed to detect the quantity of circulating proteins implicated in AD are currently in development and show promise. In 2017, one test discriminated among normal cognition, MCI, and dementia due to AD in a small number of patients with sensitivities and specificities of 84% and 88%, respectively 16 . Another blood test that shows promise is the serum microRNA profile screen that demonstrated validity and reproducibility in smaller trials 17 . With validation by future larger-scale studies, the hope is that a simple blood test may aid in the diagnosis of AD 18 .

Current treatment

At present, only two classes of pharmacologic therapy are available for patients with AD. The cholinesterase inhibitors donepezil, rivastigmine, and galantamine are recommended therapy for patients with mild, moderate, or severe AD dementia as well as Parkinson’s disease dementia 19 . Memantine, which has activity as both a non-competitive N-methyl-D-aspartate receptor antagonist and a dopamine agonist, is approved for use in patients with moderate-to-severe AD (mini-mental state examination [MMSE] <15) who show difficulty with attention and alertness 20 . For patients who choose alternative therapy, the nutraceutical huperzine A has shown benefit in both memory function and activities of daily living 21 . However, while huperzine A is a government-approved medication outside of the US, it is not regulated by the US Food and Drug Administration and may be subject to fluctuations in potency and purity. Vitamin D deficiency was also identified as an independent risk factor for the development of dementia of any cause, and supplementation is recommended for patients in whom deficiency is diagnosed 22 . Although many retrospective, observational studies alluded to the role of inflammation in the development of AD by showing a reduced risk of AD with the use of non-steroidal anti-inflammatory drugs, a more-thorough investigation failed to note any significant difference in cognitive performance in patients who took these medications 23 . In the past decade, omega-3 fatty acid supplements including fish oil have received much attention owing to their cardiovascular benefits. Two recent randomized, controlled, double-blinded studies showed improvement in thinking and memory in patients with MCI who took fish oil supplements, though these studies were limited by small sample size 24 , 25 .

Finally, the management of cardiovascular risk factors contributes to overall brain health in both cerebrovascular disease and neurodegenerative disease 26 . Recent systematic reviews found that people who adhere to the Mediterranean diet (meals consisting of fresh produce, wholegrains, olive oil, legumes, and seafood while limiting dairy and poultry products and avoiding red meat, sweets, and processed foods) have reduced risk of developing cognitive decline and AD 27 , 28 . Regular aerobic exercise, long known to prevent metabolic conditions such as diabetes mellitus and coronary artery disease, also shows preservation of function and reduces caregiver burden in patients with AD 29 . Not only does physical exercise prevent loss of strength and agility as patients age but it also reduces neuropsychiatric symptoms and the increased care requirements associated with these issues. Recreational physical activity increases cognitive function later in life, with benefit noted regardless of age at the initiation of exercise 30 . Less atrophy was observed in the brains of patients with genetic risk factors for AD who exercised regularly compared with those who did not, suggesting that aerobic activity prevents neurodegeneration 31 . Although larger controlled studies are still needed to examine the long-term effects of physical activity in patients with biomarker-proven AD pathology, the inherent systemic benefits and lack of health risks should lead all healthcare providers to recommend regular exercise for their patients, regardless of cognitive function.

Future treatment

Research into future treatments of AD involve targeting of the etiologic pathologies: neurofibrillary tangles (composed of p-tau) and senile plaques (Aβ). However, there remains debate as to which abnormality is the best target to slow or halt neurologic decline as well as how soon treatment should be initiated 32 , 33 . Another approach aims to fortify transcortical networks and enhance inter-neuronal connections in order to enhance cognitive function 34 . From previous studies, we learned that early identification of an at-risk population and subsequent treatment in the pre-clinical stage is the approach most likely to slow or halt the progression of AD 35 . Clinical trials are underway that aim to recruit asymptomatic patients with a genetic predisposition or biomarkers suggestive of higher risk of developing Alzheimer’s dementia, with results expected early in the next decade. The EU/US/Clinical Trials in AD Task Force in 2016 examined many of these trials in an attempt to identify the most effective measures of patient recruitment and retention, infrastructure development, and patient assessment including biomarkers and objective testing for clinical outcomes 35 . Some of the persistent challenges identified include the timeline of recruitment and recruitment failures, difficulty in predicting success based upon prior studies for certain drugs, and the overall costs for such large-scale clinical trials. With a more cooperative effort between researchers, private and public funding, and screening of at-risk populations, a better predictor of successful clinical trials can be created.

Anti-amyloid

According to the amyloid cascade hypothesis, toxic plaques are the earliest manifestation of disease, a statement supported by evidence of Aβ up to 20 years prior to the onset of symptoms 36 . Researchers found in 2013 that this abnormal amyloid plaque induces the phosphorylation of tau protein, which then spreads almost infectiously via microtubule transport to neighboring neurons, leading to neuronal death 37 . One class of medications developed using this evidence is the monoclonal antibodies (passive immunotherapy). This type of treatment involves injection of an antibody that targets abnormal Aβ and facilitates its removal from the brain. Two such monoclonal antibodies were initially developed in 2014 to remove these plaques from the brains of people with AD 38 , 39 . Neither medication improved cognitive scores in patients with mild-to-moderate disease (MMSE 16–26), leading researchers to conclude that these medications may show benefit only when administered in the early stages of MCI and mild dementia. However, a new study regarding the effect of this class of medication in patients with few to no symptoms (MMSE 20–26) but a positive amyloid PET imaging result also failed to show a significant difference in cognitive outcomes between the study group and asymptomatic controls 40 . Studies involving similar drugs in this class are ongoing, with the goal of improving or preserving cognition in patients with MCI due to AD.

Another approach to decreasing Aβ plaque burden in the brain is the inhibition of the enzymes that produce the Aβ peptide from its precursor, amyloid precursor protein (APP). Currently, multiple drugs are in development which target β-site APP cleaving enzyme 1 (BACE1), which is thought to be essential for the production of Aβ peptides 41 . Though previous studies of BACE1 inhibitors failed to yield meaningful results in human subjects, the novel agent verubecestat recently achieved a more than 40-fold reduction in Aβ levels in the brains of rodents and primates, and it has shown a good safety profile in early human trials 42 . Currently, another drug is under investigation for its effect on memory and cognitive function in older patients with positive biomarkers or family history of AD, known as the EARLY study.

Researchers showed in 2014 that combination therapy with a monoclonal antibody and a BACE1 inhibitor significantly reduced the amount of Aβ in amyloid-producing mice 43 . While there are no current trials underway utilizing this approach in humans, many experts believe that combination therapy employing both approaches to eliminate Aβ will ultimately lead to success in AD treatment 44 .

Since p-tau appears to be the downstream pathology and is likely the direct cause of symptoms in AD, drugs to reduce the burden of this protein are also in development 45 . Many different tau vaccines have shown both safety and efficacy in animal models 46 , and, in one recent small study, an anti-tau drug demonstrated a good safety profile and even stimulated a positive immune response in human patients 47 . Several other early phase trials of drugs which target the tau protein are currently underway, though results are yet to be published 48 . Table 1 outlines the treatments and targets currently under investigation.

Potential treatments currently undergoing clinical investigation. APP, amyloid precursor protein; BACE1, β-site amyloid precursor protein cleaving enzyme 1; p-tau, hyperphosphorylated tau peptide; RAGE, receptor for advanced glycation end products.

*Medications under investigation as combination therapy. Source: www.clinicaltrials.gov .

Neural circuitry

The failure of some targeted therapies toward Aβ in large-scale clinical trials has led to the hypothesis that, although the abnormal protein is implicated at the onset of AD, the progression of clinical symptoms is due to more global neural network dysfunction 49 . Gamma oscillation, a high-frequency brainwave rhythm, is associated with inter-neuronal communication in virtually all brain networks 50 and may help to distinguish between true and false memories 51 . Recently, researchers at the Massachusetts Institute of Technology found that induction of gamma-frequency oscillations led to reduced Aβ deposition and improved cognitive outcomes in an AD mouse model 52 . This was done by using a non-invasive 40 Hz photic stimulator to entrain the desired frequency in the mouse cortex. This method is also currently in early phase trials in humans, utilizing both visual and auditory stimulation.

As recently as 2010, the diagnosis and management of AD relied upon clinical symptom reporting that fit the pattern of memory dysfunction and loss of functional independence in multiple cognitive domains. With the reclassification system devised by the NIA–AA and DSM-5, the spectrum of AD has grown to include pre-clinical disease and MCI, helping to lay the foundation for early identification of at-risk patients. There are now a few widely available diagnostic studies that augment the clinical evaluation for a more accurate diagnosis of AD pathology, including bodily fluids and imaging studies, with good specificity.

However, the treatment options for AD remain supportive and symptomatic without attenuation of the ultimate prognosis. Medications such as cholinesterase inhibitors and memantine improve memory and alertness, respectively, without changing the life expectancy or overall progression of AD dementia. Lifestyle modifications including diet and exercise remain the only interventions with evidence showing lower AD risk and possible prevention of overall cognitive decline, and these interventions are first-line recommendations for all patients regardless of cognitive function. The pathological features associated with AD, Aβ and p-tau, are the current targets for potential treatments; however, early success in comparative studies and smaller clinical trials are thus far not reproducible in larger-scale administrations. Although limited evidence suggests that earlier identification of AD pathology will lead to better and more-definitive treatment, the results of larger-scale interventions are not yet available for review. Given the rising prevalence and mortality of AD coupled with the growing total healthcare costs, there continues to be a sense of urgency in the medical community to develop effective means for the early diagnosis and successful treatment of this progressive neurodegenerative disease.

Abbreviations

Aβ, amyloid β; AD, Alzheimer’s disease; APP, amyloid precursor protein; BACE1, β-site amyloid precursor protein cleaving enzyme 1; CSF, cerebrospinal fluid; DSM-5, Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition; MCI, mild cognitive impairment; MMSE, mini-mental state examination; NIA–AA, National Institute on Aging–Alzheimer’s Association; p-tau, hyperphosphorylated tau peptide; PET, positron emission tomography

[version 1; referees: 2 approved]

Funding Statement

The author(s) declared that no grants were involved in supporting this work.

Editorial Note on the Review Process

F1000 Faculty Reviews are commissioned from members of the prestigious F1000 Faculty and are edited as a service to readers. In order to make these reviews as comprehensive and accessible as possible, the referees provide input before publication and only the final, revised version is published. The referees who approved the final version are listed with their names and affiliations but without their reports on earlier versions (any comments will already have been addressed in the published version).

The referees who approved this article are:

• Hemachandra Reddy , Garrison Institute on Aging, Texas Tech University Health Sciences Center, Lubbock, TX, USA No competing interests were disclosed.
• Erik Portelius , Institute of Neuroscience and Physiology, Department of Psychiatry and Neurochemistry, The Sahlgrenska Academy at the University of Gothenburg, Mölndal, Sweden No competing interests were disclosed.

• Alzheimer's disease & dementia
• Arthritis & Rheumatism
• Attention deficit disorders
• Autism spectrum disorders
• Biomedical technology
• Diseases, Conditions, Syndromes
• Endocrinology & Metabolism
• Gastroenterology
• Gerontology & Geriatrics
• Health informatics
• Inflammatory disorders
• Medical economics
• Medical research
• Medications
• Neuroscience
• Obstetrics & gynaecology
• Oncology & Cancer
• Ophthalmology
• Overweight & Obesity
• Parkinson's & Movement disorders
• Psychology & Psychiatry
• Radiology & Imaging
• Sleep disorders
• Sports medicine & Kinesiology
• Vaccination
• Breast cancer
• Cardiovascular disease
• Chronic obstructive pulmonary disease
• Colon cancer
• Coronary artery disease
• Heart attack
• Heart disease
• High blood pressure
• Kidney disease
• Lung cancer
• Multiple sclerosis
• Myocardial infarction
• Ovarian cancer
• Post traumatic stress disorder
• Rheumatoid arthritis
• Schizophrenia
• Skin cancer
• Type 2 diabetes
• Full List »

share this!

August 28, 2024

This article has been reviewed according to Science X's editorial process and policies . Editors have highlighted the following attributes while ensuring the content's credibility:

fact-checked

peer-reviewed publication

trusted source

Epigenetics blood markers can help understand dementia risk

by University of Exeter

Recent research suggests that epigenetic markers in the blood could be useful for understanding dementia risk. Two linked papers from the University of Exeter and Maastricht University have together progressed research to show the potential for DNA methylation, an epigenetic marker, in understanding how genetics and lifestyle factors influence dementia risk.

DNA methylation is a chemical tag added to DNA, which can turn genes on and off. Genetic and lifestyle factors can alter the levels of the DNA methylation tag on genes, with some of these factors already known to increase the risk of developing dementia. By assessing DNA methylation, this can help scientists understand the extent to which these different factors influence risk of dementia and the mechanisms by which they bring about disease.

In the largest study of its kind, researchers assessed DNA methylation at 800,000 sites in the genome in blood samples collected from 900 people in the European Medical Information Framework for Alzheimer's disease Multimodal Biomarker Discovery (EMIF-AD MBD) study.

The study includes extensive clinical information on participants, who all provided spinal fluid samples, which have been used for diagnosis and monitoring of Alzheimer's disease, because it is in direct contact with the brain. However, collecting the fluid is an invasive procedure, so the team investigated whether they could instead use blood samples , through analyzing blood epigenetic signatures that are associated with Alzheimer's disease biomarkers, as this would be cheaper and easier to collect in practice.

In the first of the two papers, led by Professor Katie Lunnon at the University of Exeter Medical School, the team showed that DNA methylation signatures in blood can mirror some protein biomarker levels in spinal fluid samples, which are used for assessing dementia. The team explored these signatures in conjunction with 15 different spinal fluid biomarkers that are used for diagnosing dementia and showed changes in the methylation status of key genes for a number of these biomarkers.

The first paper is titled "Blood DNA methylomic signatures associated with CSF biomarkers of Alzheimer's disease in the EMIF-AD study. Alzheimer's and Dementia." The findings are published in the journal Alzheimer's & Dementia .

In a second paper in the same journal, led by Dr. Ehsan Pishva at Maastricht University in the Netherlands, the team generated epigenetic risk scores using blood DNA methylation signatures as a proxy for 14 known dementia risk factors. Some of these were modifiable lifestyle risks, including physical activity, diet and some were non-modifiable, such as age and having heart disease.

The second paper is titled "Blood-based multivariate methylation risk score for cognitive impairment and dementia. Alzheimer's and Dementia."

They showed that their epigenetic risk scores can improve the prediction of the risk of cognitive decline and dementia onset, even at early stages. Early detection is crucial to better lifestyle management, and to accessing potential new treatments. The paper highlights how genetic, lifestyle, and environmental factors are contributing to the development and progression of dementia through epigenetic mechanisms.

Professor Katie Lunnon, at the University of Exeter Medical School, is lead author on one of the studies, and leads the Dementia Genomics Team who have previously published a number of pioneering papers exploring epigenetics in the brain and blood in different dementias. She said, "We know that a number of genetic and lifestyle factors can increase the risk of developing Alzheimer's disease and other dementias. Epigenetics is a particularly exciting research field because it can mediate the interaction between our genetic makeup, which is fixed at conception, and environmental risks, which we can potentially modify."

Dr. Ehsan Pishva, at Maastricht University, who led the other paper and leads the Dementia Systems Biology team, said, "Our epigenetic risk score can improve the prediction of risk of cognitive impairment in different populations, marking a significant advancement in dementia research. The study, which involved advanced analysis of large epigenetic datasets from multiple independent dementia cohorts, found that the epigenetic risk score was a predictor of future cognitive decline in Alzheimer's disease and Parkinson's disease cohorts.

"Our findings highlight the potential of using blood-derived epigenetic measurements as a non-invasive approach to assess dementia risk, paving the way for future studies to explore more personalized and preventive health care strategies in tackling cognitive impairment."

Jarno Koetsier et al, Blood‐based multivariate methylation risk score for cognitive impairment and dementia, Alzheimer's & Dementia (2024). DOI: 10.1002/alz.14061

Explore further

Feedback to editors

Managing early stages of abortion care at home after 12 weeks is safe and reduces time spent in hospital, study finds

9 hours ago

Billions worldwide consume inadequate levels of micronutrients critical to human health, new study finds

Blocking the longevity gene S6K1 extends lifespan in mice by reducing inflammation

11 hours ago

Machine learning predicts which patients will continue taking opioids after hand surgery

Study describes a new molecular pathway involved in the control of reproduction

12 hours ago

Machine learning helps identify rheumatoid arthritis subtypes

Immune protection against tuberculosis reinfection driven by cells that dampen lung inflammation, shows study

Girls with mental health conditions have lower HPV vaccination coverage

13 hours ago

Mankai plant found to reduce post-meal sugar levels in diabetics

Unique chicken line advances research on autoimmune disease that affects humans

Related stories.

Largescale brain epigenetics study provides new insights into dementia

Jun 10, 2021

Older Native Americans may experience higher levels of cognitive impairment than previously thought

May 15, 2024

Midlife blood test may predict cognitive decline, Alzheimer's in later life

Dec 21, 2023

Epigenetic 'age' predicts cognitive function

Apr 20, 2022

Study identifies 10 risk factors for Alzheimer's disease

Jul 21, 2020

Study is first to show role of genomic changes in specific brain cells in Alzheimer's disease

Oct 3, 2022

Recommended for you

Navigating the digestive tract: Study offers first detailed map of the small intestine

High-resolution brain tumor mapping reveals possible reason why some patients don't respond to new drug

16 hours ago

Gene therapy gets a turbo boost from researchers

17 hours ago

Scientists aim to understand why men with prostate cancer are at higher risk of Alzheimer's

18 hours ago

Oncoprotein activity increases prostate cancer progression, finds study

Crucial role of JUN protein in restraining liver cancer growth discovered

Let us know if there is a problem with our content.

Use this form if you have come across a typo, inaccuracy or would like to send an edit request for the content on this page. For general inquiries, please use our contact form . For general feedback, use the public comments section below (please adhere to guidelines ).

Please select the most appropriate category to facilitate processing of your request

Thank you for taking time to provide your feedback to the editors.

Your feedback is important to us. However, we do not guarantee individual replies due to the high volume of messages.

E-mail the story

Your email address is used only to let the recipient know who sent the email. Neither your address nor the recipient's address will be used for any other purpose. The information you enter will appear in your e-mail message and is not retained by Medical Xpress in any form.

Newsletter sign up

Get weekly and/or daily updates delivered to your inbox. You can unsubscribe at any time and we'll never share your details to third parties.

More information Privacy policy

Donate and enjoy an ad-free experience

We keep our content available to everyone. Consider supporting Science X's mission by getting a premium account.

E-mail newsletter