Artificial Intelligence and Digital Biomarkers: Advancing the Early  Diagnosis of Alzheimer’s Disease

Shanmuga, S

doi:10.65188/nurexus.1099

Shanmuga S

et al | DOI: 10.65188/nurexus.1099

Page 20

Journal of MedVerse Research & Practice

ISSN: 3107-4278

Artificial Intelligence and Digital Biomarkers: Advancing the Early

Diagnosis of Alzheimer’s Disease

S Shanmuga

1

Professor

Department of Pharmaceutics, Yong Loo Lin School of Medicine, Singapore

Email ID: hanmugaerpal@gmail.Com

Submission Date: 16.07.2026

Accepted Date:22.08.2026

Published Date: 31.08.2026

DOI: 10.65188/nurexus.1099

article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted

use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.

Abstract

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder characterized by the gradual deterioration of

brain cells, resulting in progressive impairment of cognitive and behavioral functions. The increasing burden of AD

highlights the need for innovative approaches that can facilitate early and accurate diagnosis. Artificial intelligence

(AI) has emerged as a promising tool in neurological and psychiatric care, with potential applications in disease

detection, prognosis assessment, treatment-response monitoring, and personalized patient management. AI-based

systems can analyze complex clinical and digital data to identify patterns that may not be readily recognized using

conventional diagnostic approaches.

This review examines the evolving role of AI and digital biomarkers in the diagnosis of Alzheimer’s disease, with

particular emphasis on recent technological innovations and AI-assisted analysis of individualized digital biomarkers.

It further discusses the potential of AI-based predictive models for early identification and progression assessment

of AD, as well as emerging approaches such as federated learning and privacy-preserving AI. The integration of

digital biomarkers with advanced AI algorithms may facilitate more efficient and personalized diagnostic pathways

while supporting continuous monitoring of disease progression and treatment response. However, the successful

clinical implementation of these technologies requires high-quality datasets, robust clinical validation, ethical

considerations, data privacy, and appropriate regulatory oversight. Overall, AI-enabled digital biomarkers represent

a promising direction for improving early detection, diagnostic accuracy, clinical decision-making, and personalized

care in patients with Alzheimer’s disease.

Keywords: Alzheimer’s disease; Artificial intelligence; Digital biomarkers; Neurodegeneration; Predictive models;

Federated learning; Early diagnosis.

Introduction

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder characterized by the gradual

deterioration of neuronal cells, resulting in progressive impairment of cognitive and behavioural functions.

These functions include memory, comprehension, communication, concentration, reasoning, and decision-

making. The increasing prevalence and mortality associated with AD have made it a major global public

health concern. According to the Centers for Disease Control and Prevention (CDC), Alzheimer’s disease

was the seventh leading cause of death in the United States in 2022. Before the COVID-19 pandemic,

Alzheimer’s disease ranked second only to stroke among the major causes of death [1,2].

Alzheimer’s disease is broadly classified according to the age at which symptoms begin. Late-onset

Alzheimer’s disease (LOAD) generally develops after 65 years of age, whereas early-onset Alzheimer’s

REVIEW ARTICLE

Shanmuga S

et al | DOI: 10.65188/nurexus.1099

Page 21

disease (EOAD) occurs before the age of 65 years. Although EOAD accounts for a smaller proportion of

cases, it represents an important clinical subgroup. The global burden of dementia has increased

considerably over recent decades. In 1990, approximately 20.3 million individuals worldwide were living

with dementia, which increased to approximately 43.8 million by 2016. Projections indicate that the number

of people affected by dementia may reach approximately 150 million by 2050, emphasizing the need for

improved approaches to early recognition and diagnosis [3,4].

Neurodegeneration is a defining pathological feature of Alzheimer’s disease, involving progressive loss

and dysfunction of neuronal cells. The disease process is frequently associated with early involvement of

the entorhinal cortex and hippocampal regions, followed by progressive involvement of other areas of the

brain. Both genetic and environmental factors contribute to the development of AD. Genetic influences are

particularly relevant in early-onset and late-onset disease, while conditions such as trisomy 21 are

associated with an increased risk of early-onset dementia [5]. Several additional factors have been

associated with an increased risk of Alzheimer’s disease, including previous traumatic brain injury,

depression, cardiovascular and cerebrovascular disorders, advanced parental age at birth, smoking, elevated

homocysteine levels, a family history of dementia, and the presence of the APOE ε4 allele [6-8].

The diagnosis of Alzheimer’s disease remains challenging because of overlapping clinical manifestations,

difficulties in establishing uniform diagnostic criteria, limited accessibility of specific biomarkers, and the

possibility of alternative neuropsychiatric and neurological diagnoses [9]. Over the past decade, substantial

progress has been made in the development of biomarkers that can support earlier and more accurate

identification of AD. These include neuroimaging biomarkers obtained through amyloid and tau positron

emission tomography (PET), as well as cerebrospinal fluid and plasma-based biomarkers.

Despite these advances, there is currently no definitive therapeutic approach capable of completely

preventing or reversing the progression of Alzheimer’s disease. Available treatments primarily aim to

alleviate symptoms and support functional independence. Consequently, there is increasing interest in

developing innovative technologies capable of identifying subtle disease-related changes at an early stage.

Artificial intelligence (AI), particularly when combined with digital biomarkers derived from behavioural,

cognitive, physiological, and neuroimaging data, has emerged as a promising approach. AI-based methods

can process large and complex datasets, recognize patterns, and potentially support earlier detection, risk

prediction, disease monitoring, and individualized clinical decision-making.

The integration of artificial intelligence with digital biomarkers therefore represents an important emerging

area in Alzheimer’s disease research. These technologies may complement established clinical,

biochemical, and neuroimaging approaches and potentially improve the efficiency and precision of

diagnosis. However, their successful application requires appropriate clinical validation, high-quality

datasets, consideration of ethical issues, and protection of patient privacy. The present review therefore

focuses on the evolving role of artificial intelligence and digital biomarkers in the diagnosis of Alzheimer’s

disease, including their potential applications, predictive capabilities, technological advances, and emerging

privacy-preserving approaches.

ROLE OF ARTIFICIAL INTELLIGENCE IN DIAGNOSING ALZHEIMER’S DISEASE

Artificial intelligence (AI) is increasingly being incorporated into the diagnosis and clinical management

of Alzheimer’s disease (AD). AI-based systems can assist psychiatrists, neurologists, and other healthcare

professionals by analyzing complex clinical, cognitive, imaging, and behavioral information to identify

disease-related patterns. These technologies can also support monitoring of disease progression and

assessment of treatment response. By presenting data-driven insights and individualized information, AI-

Shanmuga S

et al | DOI: 10.65188/nurexus.1099

Page 22

based platforms may improve clinical decision-making and encourage greater patient engagement with

treatment and follow-up [10].

The integration of AI with digital technologies has introduced new possibilities for identifying subtle

changes in cognition, behavior, and physiological functioning during everyday life. Machine learning (ML)

and deep learning (DL) algorithms are particularly useful for processing large and complex datasets and

identifying patterns associated with AD. These approaches may contribute to early disease detection,

prediction of disease progression, and development of individualized intervention strategies.

Digital Biomarkers and Passive Assessment

Digital biomarkers are measurable and quantifiable characteristics obtained from digitally recorded data

and can provide information regarding an individual’s physiological, behavioral, or cognitive status [11].

Smartphones, wearable devices, activity monitors, and other digital sensors can continuously collect

information that may be useful for monitoring patients with or at risk of AD.

Passive assessments represent an important development in this field because data can be collected

continuously without requiring active participation from the individual. Examples include sleep

characteristics, physical activity patterns, gait parameters, smartphone usage, and other measures of daily

functioning. These data can subsequently be analyzed using AI algorithms to identify subtle changes that

may precede clinically apparent cognitive impairment.

Active Digital Assessments

In addition to passive monitoring, active digital assessments can provide targeted information regarding

cognitive and behavioral functioning. Ecological momentary assessments (EMAs), smartphone-based

cognitive tests, and voice recordings can be used to obtain repeated measurements in real-world settings.

AI algorithms can combine information from multiple digital biomarkers to generate composite measures

of an individual's cognitive and functional status.

The integration of multiple digital indicators may improve the ability to identify early changes associated

with Alzheimer’s disease. Parameters such as typing performance, sleep efficiency, gait variability, speech

characteristics, and patterns of daily activity may provide complementary information when analyzed

collectively through AI-based systems.

Initiatives, Innovations and Technological Advancements

Several initiatives are exploring the application of AI and digital biomarkers for Alzheimer’s disease

diagnosis and monitoring. The Alzheimer’s Drug Discovery Foundation’s Diagnostics Accelerator (DxA)

has supported the development of speech-based biomarkers through the SpeechDx Dataset. The dataset

serves as a repository of speech-related information intended to facilitate the development and validation

of clinically useful biomarkers for AD. A partnership with Siemens Healthineers was established to further

explore the development of speech-based diagnostic approaches.

Another emerging technology is MoCA Solo, an AI-enabled digital cognitive assessment platform

supported by the Diagnostics Accelerator. The platform aims to improve the accessibility of cognitive

assessment and facilitate earlier identification of cognitive impairment, particularly in primary-care settings

where patients may initially present with memory-related concerns.

Ocular Biomarkers and Artificial Intelligence

Ocular biomarkers represent another emerging non-invasive approach for identifying individuals at risk of

Alzheimer’s disease. Recent research has investigated the possibility of using ocular characteristics as

indicators of underlying neurodegenerative changes. AI-assisted analysis of ocular data may potentially

provide an accessible alternative or complementary approach to established biomarkers, particularly when

combined with other clinical and digital information [12,13].

Shanmuga S

et al | DOI: 10.65188/nurexus.1099

Page 23

Smartphone-Derived Digital Biomarkers and Remote Monitoring

Smartphone-based digital biomarkers are increasingly being investigated for remote assessment and

longitudinal monitoring of individuals with neurological disorders. Smartphones can passively capture

information such as activity patterns obtained from accelerometers and gyroscopes, device-use behavior,

and sleep-related measurements. Active assessments may include brief cognitive tests, mood or symptom

assessments, and voice recordings for machine-learning analysis [14].

AI algorithms can integrate these different data streams to identify patterns associated with reduced activity,

cognitive deterioration, changes in gait, sleep disturbances, or alterations in communication. Such

composite digital measures may provide a more continuous representation of an individual's functional

status than conventional clinic-based assessments performed at isolated time points.

Overall, the combination of artificial intelligence and digital biomarkers has the potential to complement

conventional clinical, neuropsychological, biochemical, and neuroimaging approaches in Alzheimer’s

disease. However, widespread clinical implementation requires rigorous validation, standardized data

collection, protection of patient privacy, and careful evaluation of the accuracy and reliability of AI-based

diagnostic systems.

AI-AUGMENTED ANALYSIS OF INDIVIDUAL DIGITAL BIOMARKERS: NEW ADVANCES

Speech Analysis

Speech analysis has emerged as a promising approach for identifying subtle cognitive and neurological

changes associated with Alzheimer’s disease (AD). Alterations in connected speech may be reflected

through differences in lexical-semantic content, including word selection and vocabulary diversity,

acoustic-phonetic characteristics such as pitch, speaking rate, and pauses, as well as grammatical

complexity. These changes may occur before the development of clinically obvious symptoms. Ongoing

research supported by the National Institutes of Health (NIH) is investigating machine learning and natural

language processing (NLP) techniques applied to audio recordings to identify speech-based digital

biomarkers associated with AD biomarker status and to predict cognitive trajectories over time [15].

Artificial intelligence can analyze both speech characteristics and automatically generated speech

transcripts to identify patterns associated with cognitive impairment. Amini et al. developed an AI-based

approach to analyze speech transcripts obtained during cognitive assessments in participants from the

Framingham Heart Study. The model analyzed linguistic structures without relying on acoustic

characteristics and demonstrated the ability to identify individuals with mild cognitive impairment (MCI)

who were likely to progress to Alzheimer’s disease within six years, with a reported predictive performance

exceeding 78%. Universal Sentence Encoder was used to generate vector representations of speech

segments, followed by logistic regression for prediction [16].

Gait Analysis

Gait characteristics are increasingly being investigated as potential digital biomarkers of cognitive decline

and Alzheimer’s disease. Changes such as reduced stride length, decreased cadence, and increased gait

variability may occur with progressive cognitive impairment. Wearable technologies, including

accelerometers incorporated into smartwatches and other sensor-based devices, as well as motion-capture

systems, can provide detailed information regarding gait patterns.

AI-based methods can analyze these complex movement signals and identify subtle abnormalities that may

not be apparent during routine clinical examination. Convolutional neural networks (CNNs) can be applied

to identify spatial patterns within gait-related data, whereas recurrent neural networks (RNNs) can analyze

temporal sequences and changes occurring over time. These approaches may assist in differentiating

healthy individuals from those at increased risk of cognitive impairment or AD [17].

Shanmuga S

et al | DOI: 10.65188/nurexus.1099

Page 24

Ocular Biomarkers

The eye, particularly the retina, has been described as a potential “window to the brain” because of its

shared embryological origin and vascular characteristics with the central nervous system. Artificial

intelligence is being increasingly applied to ophthalmic imaging modalities to investigate retinal features

associated with Alzheimer’s disease.

AI algorithms can analyze retinal photographs, including fundus images, and optical coherence tomography

angiography (OCT-A) scans to identify subtle abnormalities such as changes in retinal vascular density and

retinal nerve fiber layer thickness. These imaging characteristics may be associated with AD pathology and

cognitive performance. AI-assisted ocular imaging therefore represents a potentially non-invasive and

accessible screening approach for identifying individuals who may be at increased risk of Alzheimer’s

disease [18].

Multimodal Integration of Digital Biomarkers

Although individual digital biomarkers demonstrate considerable potential, the major advantage of AI in

Alzheimer’s disease may lie in its ability to integrate heterogeneous data streams. Multimodal digital

phenotyping can incorporate information from ocular imaging, sleep patterns, physical activity, smartphone

interactions, movement characteristics, and speech analysis. AI algorithms can combine these different

sources to generate composite digital measures that may provide a more comprehensive assessment of

neurological and cognitive health than individual biomarkers alone.

The use of composite scores may be particularly relevant because Alzheimer’s disease is a complex and

heterogeneous disorder with effects across multiple cognitive, behavioral, and functional domains.

However, further research is required to establish appropriate data-fusion approaches and to validate AI-

generated composite biomarkers against established clinical and pathological standards. Clinical validation

against recognized biomarkers, including cerebrospinal fluid and positron emission tomography measures,

remains essential before widespread clinical adoption [19].

AI-Based Predictive Models for Alzheimer’s Disease

AI-based prediction of Alzheimer’s disease has progressed from relatively simple models using limited

demographic and clinical information to sophisticated systems capable of integrating large-scale

multimodal datasets. These developments have created opportunities for more individualized assessment

of disease risk, prediction of disease progression, and identification of subtle symptoms at an earlier stage.

Early Predictive Models and Risk-Factor Identification

Early applications of AI for Alzheimer’s disease prediction primarily relied on traditional machine-learning

methods applied to demographic characteristics, clinical records, and established risk factors. Logistic

regression and support vector machine (SVM) approaches were commonly used with structured datasets.

These models incorporated variables such as age, sex, socioeconomic characteristics, general health status,

lifestyle factors including smoking, and genetic risk factors.

Although these approaches provided an initial framework for risk prediction, their ability to identify

complex nonlinear relationships and interactions among multiple variables was relatively limited.

Consequently, subsequent research has focused on more sophisticated machine-learning and deep-learning

techniques capable of processing larger and more diverse datasets [20].

Next-Generation Predictive Models Using Clinical, Psychosocial and Genetic Data

The subsequent development of AI-based prediction models incorporated advanced techniques such as

random forests, classification algorithms, extreme gradient boosting (XGBoost), and deep-learning

Shanmuga S

et al | DOI: 10.65188/nurexus.1099

Page 25

methods. These approaches enabled researchers to incorporate a broader range of predictors, including

detailed clinical information, psychosocial characteristics, genetic factors, and personal health-related

variables.

For example, an XGBoost model developed using National Alzheimer’s Coordinating Center (NACC) data

incorporated 11 significant variables, including education level, depression, insomnia, age, body mass

index, number of medications, sex, stenting history, systolic blood pressure, neuroticism, and REM sleep

abnormalities. Model-interpretability approaches were also applied to investigate the contribution of

individual variables to the predicted risk. Higher educational attainment was associated with a protective

effect, whereas advanced age and polypharmacy were associated with increased risk [21].

The availability of large research databases, including the NACC Uniform Data Set (NACC-UDS) and

Alzheimer’s Disease Neuroimaging Initiative (ADNI), has further facilitated the development of

sophisticated predictive models by providing access to extensive clinical, cognitive, imaging, and

biomarker datasets.

Prognosis Using AI and Digital Biomarkers

The most advanced predictive approaches combine digital biomarkers with sophisticated AI techniques to

provide more individualized estimates of Alzheimer’s disease risk and progression. These systems can

integrate digital biomarker streams with conventional clinical, genetic, and neuroimaging information. Such

multimodal approaches may enable continuous monitoring and provide a more detailed representation of

individual disease trajectories [22].

ADVANCED ARTIFICIAL INTELLIGENCE METHODS FOR PERSONALIZED

FORECASTING

State-Space Models

State-space models (SSMs) are useful for analyzing longitudinal data obtained from repeated cognitive

assessments or measurements of biomarkers over multiple time points. These models can characterize

changes occurring over time and may help estimate underlying disease states from observed clinical and

digital measurements.

Deep Learning Temporal Models

Deep-learning temporal models, including recurrent neural networks (RNNs) and long short-term memory

(LSTM) networks, can be applied to multivariate time-series data. By analyzing clinical, cognitive,

imaging, and digital information collected at different time points, these models can identify complex

temporal relationships and potentially improve prediction of individual disease progression.

Graph Neural Networks

Graph neural networks (GNNs) provide an approach for representing relationships among interconnected

data elements. In Alzheimer’s disease research, GNNs may be applied to structural or functional brain-

connectivity networks obtained from neuroimaging or to patient-similarity networks constructed from

multimodal clinical features. Their ability to model relationships between interconnected variables may

provide additional information regarding disease patterns and progression.

Deep Generative Models and AI-Driven Digital Twins

Deep generative models, including conditional restricted Boltzmann machines (CRBMs), represent another

emerging approach for personalized disease modelling. These techniques may facilitate the development

of patient-specific “digital twins” capable of simulating potential disease trajectories under different

hypothetical conditions. Such approaches could support in-silico experimentation and assessment of

potential treatment effects.

Shanmuga S

et al | DOI: 10.65188/nurexus.1099

Page 26

Comprehensive predictive systems may integrate multiple data modalities, including structural MRI, PET,

diffusion tensor imaging (DTI), resting-state functional MRI (rs-fMRI), cerebrospinal fluid biomarkers

such as Aβ42, total tau and phosphorylated tau, genetic information including APOE genotype and single-

nucleotide polymorphisms, clinical and cognitive assessments, demographic characteristics, and newly

developed digital biomarkers [23-25].

Explainable Artificial Intelligence for Prediction Models

As AI models become increasingly complex, explainable artificial intelligence (XAI) is important for

improving clinical interpretability and confidence in model-generated predictions. Methods such as SHAP

(SHapley Additive exPlanations), LIME (Local Interpretable Model-agnostic Explanations), and Grad-

CAM (Gradient-weighted Class Activation Mapping) can help identify the contribution of individual

variables or regions to AI predictions. XAI approaches can potentially demonstrate which brain regions,

clinical characteristics, or digital biomarkers contribute most strongly to predictions of Alzheimer’s disease

onset or progression. Improved interpretability may help clinicians understand AI-generated predictions

rather than treating them as unexplained “black-box” outputs. Furthermore, XAI can assist in evaluating

model fairness by examining performance across different patient populations and identifying potential

sources of bias [26].

FEDERATED LEARNING AND PRIVACY-PRESERVING AI

The development of reliable artificial intelligence (AI) models for Alzheimer’s disease requires access to

large, diverse, and high-quality datasets. However, the use of healthcare data is restricted by strict privacy

and data-protection requirements. Federated learning (FL) and other privacy-preserving AI approaches

have therefore gained increasing importance because they enable collaborative model development while

reducing the need to directly share sensitive patient information [36-47].

One emerging approach is the use of secure federated learning architectures to facilitate analysis across

geographically distributed biomedical databases. Such frameworks can enable AI models to be trained

using information from multiple institutions and populations while keeping the original patient-level data

within the respective institutions. For example, an NIH-supported secure federated learning initiative has

proposed analysis of distributed biobank resources containing more than 100,000 neuroimaging scans and

genomic datasets from institutions across the United States, Europe, India, and Japan. Federated deep-

learning approaches, data-harmonization methods, and secure homomorphic encryption can be

incorporated to generate site-independent predictive models while maintaining patient confidentiality.

The ultimate objective of these approaches extends beyond simple classification of disease risk at the

population level. Advanced AI systems aim to generate individualized predictions of cognitive and

pathological trajectories, thereby supporting personalized assessment of Alzheimer’s disease progression.

However, large-scale biological datasets present several challenges, including high dimensionality, sparse

observations, irregular sampling intervals, missing information, and imbalanced datasets. These limitations

can affect the reliability and generalizability of predictive models.

Several computational approaches can be used to address these challenges. Transfer learning may allow

models trained on one dataset or population to be adapted to another setting, while advanced data-

imputation techniques can reduce the impact of missing observations. Synthetic data-generation approaches

based on variational autoencoders (VAEs) and generative adversarial networks (GANs) may also help

overcome data sparsity and imbalance while supporting the development of more robust AI models. These

approaches can be combined with multimodal digital biomarkers to improve individualized prediction and

longitudinal monitoring of Alzheimer’s disease.

Shanmuga S

et al | DOI: 10.65188/nurexus.1099

Page 27

Overall, federated learning and privacy-preserving AI provide an important framework for overcoming the

tension between the need for large-scale data and the requirement to protect sensitive health information.

Their integration with multimodal clinical, imaging, genetic, and digital biomarker data may facilitate the

development of more generalizable and personalized AI systems for Alzheimer’s disease. Nevertheless,

continued evaluation of model security, clinical validity, fairness, and real-world performance is necessary

before these technologies can be widely incorporated into routine clinical practice.

Summary

Artificial intelligence and digital biomarkers are emerging as promising tools for improving the early

detection, diagnosis, and monitoring of Alzheimer’s disease. AI can analyze diverse information from

speech, gait, ocular imaging, cognitive assessments, wearable devices, and smartphones to identify

subtle disease-related changes that may occur before obvious clinical symptoms. Advanced machine-

learning and deep-learning models can further integrate clinical, genetic, neuroimaging, and digital data

to improve individualized prediction of disease onset and progression. Multimodal approaches,

explainable AI, and personalized forecasting models may enhance clinical decision-making while

improving understanding of disease trajectories. Federated learning and privacy-preserving AI offer

additional opportunities to develop robust models using distributed healthcare datasets without directly

sharing sensitive patient information. However, successful clinical implementation requires high-quality

data, rigorous clinical validation, ethical oversight, transparency, fairness, and strong privacy protection.

Overall, the integration of AI with digital biomarkers has considerable potential to complement

conventional diagnostic methods and support earlier, more personalized, and efficient care for

individuals with Alzheimer’s disease.

Conclusion

Artificial intelligence and digital biomarkers offer promising opportunities for the early diagnosis and

monitoring of Alzheimer’s disease. AI-based analysis of speech, gait, ocular imaging, cognitive

performance, and smartphone data may detect subtle changes and support personalized disease prediction.

Advanced approaches such as deep learning, explainable AI, and privacy-preserving methods can further

improve their clinical potential. However, high-quality data, clinical validation, ethical safeguards, and

patient privacy remain essential for reliable implementation. Overall, AI-augmented digital biomarkers may

complement conventional diagnostic methods and contribute to earlier and more personalized Alzheimer’s

disease care.

Declaration

Conflict of Interest: The authors declare that they have no competing interests or conflicts of interest

related to this work.

Funding: Nil

Reference

1. Ahmad FB, Cisewski JA, Xu J, Anderson RN. Provisional mortality data—United States, 2022. MMWR

Morb Mortal Wkly Rep. 2023 May 5;72(18):488–492.

2. Mendez MF. Early-Onset Alzheimer Disease. Neurol Clin. 2017 May;35(2):263–281.

3. Li X, Feng X, Sun X, Hou N, Han F, Liu Y. Global, regional, and national burden of Alzheimer’s disease

and other dementias, 1990–2019. Front Aging Neurosci. 2022;14:937486.

4. Santos CY, Snyder PJ, Wu WC, Zhang M, Echeverria A, Alber J. Pathophysiologic relationship between

Alzheimer’s disease, cerebrovascular disease, and cardiovascular risk: A review and synthesis. Alzheimers

Dement (Amst). 2017;7:69–87.

5. Nicolas G, Acuña-Hidalgo R, Keogh MJ, et al. Somatic variants in autosomal dominant genes are a rare

Shanmuga S

et al | DOI: 10.65188/nurexus.1099

Page 28

cause of sporadic Alzheimer’s disease. Alzheimers Dement. 2018 Dec;14(12):1632–1639.

6. Liljegren M, Landqvist Waldö M, Rydbeck R, Englund E. Police interactions among neuropathologically

confirmed dementia patients: Prevalence and cause. Alzheimer Dis Assoc Disord. 2018 Oct-Dec;32(4):346–

350.

7. Tong BC, Wu AJ, Li M, Cheung KH. Calcium signaling in Alzheimer’s disease & therapies. Biochim

Biophys Acta Mol Cell Res. 2018 Nov;1865(11 Pt B):1745–1760.

8. Tahami Monfared AA, Byrnes MJ, White LA, Zhang Q. Alzheimer’s disease: Epidemiology and clinical

progression. Neurol Ther. 2022 Jun;11(2):553–569.

9. Zetterberg H, Bendlin BB. Biomarkers for Alzheimer’s disease—preparing for a new era of disease-

modifying therapies. Mol Psychiatry. 2021 Jan;26(1):296–308.

10. Zhang W, Li Y, Ren W, Liu B. Artificial intelligence technology in Alzheimer’s disease research.

Intractable Rare Dis Res. 2023 Nov;12(4):208–212.

11. Wang Y, Li Y, Wang S, Zhang Y, Chen K, Zhang J, et al. Development and validation of machine learning

models with blood digital biomarkers for Alzheimer’s disease and mild cognitive impairment. Lancet Digit

Health. 2025 Mar 5;7(3):e210–e221.

12. Gure TR, Liss J, Abraham J, et al. Cognitive digital biomarkers from automated transcription of speech for

early Alzheimer’s detection. Sensors (Basel). 2024;24(5):1572.

13. Livingston G, et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission.

Lancet. 2020;396(10248):413–446.

14. Oh HS, Urey DY, Karlsson L, Zhu Z, Shen Y, Farinas A, et al. A cerebrospinal fluid synaptic protein

biomarker for prediction of cognitive resilience versus decline in Alzheimer’s disease. Nat Med. 2025

May;31(5):1592–1603.

15. Guo T, Shaw LM, Li Y, et al. Advances in blood biomarkers for Alzheimer’s disease: Ultra-sensitive

detection technologies and clinical applications. Front Aging Neurosci. 2024 Jul 30;16:11297492.

16. Amini S, Hao B, Yang J, Karjadi C, Kolachalama VB, Au R, Paschalidis IC. Prediction of Alzheimer’s

disease progression within 6 years using speech: A novel approach leveraging language models. Alzheimers

Dement. 2024 Aug;20(8):5262–5270.

17. Nunez N, Sanz A, Garcia A, et al. Multidimensional digital biomarker phenotypes for mild cognitive

impairment and early Alzheimer’s disease. Front Digit Health. 2024;6:1265846.

18. Rabin LA, Smart CM, Amariglio RE. Subjective cognitive decline in preclinical Alzheimer’s disease. Annu

Rev Clin Psychol. 2017;13:369–396.

19. Sperling RA, et al. Toward defining the preclinical stages of Alzheimer’s disease: Recommendations from

the National Institute on Aging–Alzheimer’s Association workgroups. Alzheimers Dement. 2011;7(3):280–

292.

20. zdok D, Meyer-Lindenberg A. Machine learning for precision psychiatry: Opportunities and challenges.

Biol Psychiatry Cogn Neurosci Neuroimaging. 2018;3(3):223–230.

21. Ismail Z, Rajji TK, Shulman KI. Brief cognitive screening instruments: An update. Int J Geriatr Psychiatry.

2010 Feb;25(2):111–120.

22. Piau A, et al. A smartphone application for self-management of health and chronic conditions: A systematic

review. Health Informatics J. 2019;25(3):1060–1078.

23. Varatharajah Y, et al. Predicting short-term MCI-to-AD progression using multi-modal deep learning. Sci

Rep. 2019;9:2235.

24. Kourtis LC, Regele OB, Wright JM, Jones GB. Digital biomarkers for Alzheimer’s disease: The

mobile/wearable devices opportunity. NPJ Digit Med. 2019;2:9.

25. Insel TR. Digital phenotyping: Technology for a new science of behavior. JAMA. 2017;318(13):1215–

1216.

26. Eyigoz E, Mathur S, Santamaria M, Cecchi G, Naylor M. Linguistic markers predict onset of Alzheimer’s

disease. EClinicalMedicine. 2020;28:100583.

27. Velazquez M, Lee Y; for the Alzheimer’s Disease Neuroimaging Initiative. Random forest model for

feature-based Alzheimer’s disease conversion prediction from early mild cognitive impairment subjects.

PLoS One. 2021 Apr 29;16(4):e0244773.

28. Al-Adhaileh MH. A long short-term memory biomarker-based prediction framework for Alzheimer’s

disease. Sensors (Basel). 2022 Jan 31;22(4):1475.

Shanmuga S

et al | DOI: 10.65188/nurexus.1099

Page 29

29. Rahman AU, et al. A novel 3D-CNN with DAG framework for enhanced Alzheimer’s disease diagnosis

using structural MRI. IEEE Conference. 2025 Feb 26.

30. Mahmud T, Barua K, Habiba SU, Sharmen N, Hossain MS, Andersson K. An explainable AI paradigm for

Alzheimer’s diagnosis using deep transfer learning. Diagnostics (Basel). 2024 Feb 5;14(3):345.

31. AlexNet Alzheimer MRI model. Kaggle. 2022. Available from: Kaggle

32. Bahri A, et al. Alzheimer’s multiclass classification using explainable AI techniques. Appl Sci. 2024 Sep

13;14(18):8287.

33. Kavitha C, Mani V, Srividhya SR, Khalaf OI, Tavera Romero CA. Early-stage Alzheimer’s disease

prediction using machine learning models. Front Public Health. 2022 Mar 3;10:853294.

34. Khan MA, et al. HSI-LFS-BERT: Novel hybrid swarm intelligence based linguistic feature selection and

computational intelligent model for Alzheimer’s detection using audio transcript. IEEE Access.

2022;10:117717–117732.

35. Prajapati R, Khatri U, Kwon GR. An efficient deep neural network binary classifier for Alzheimer’s disease

classification. International Conference on Artificial Intelligence and Smart Systems (ICAIS). 2021:204–

209.

36. Mc Ardle R, Morris R, Hickey A, Del Din S, Koychev I, et al. Gait in mild Alzheimer’s disease: Feasibility

of multi-center measurement in the clinic and home with body-worn sensors: A pilot study. J Alzheimers

Dis. 2020;75(2):559–570.

37. Den Haan J, Verbraak FD, Visser PJ, Bouwman FH. Retinal thickness in Alzheimer’s disease: A systematic

review and meta-analysis. Alzheimers Dement (Amst). 2020;12(1):e12028.

38. Musiek ES, Holtzman DM. Mechanisms linking circadian clocks, sleep, and neurodegeneration. Science.

2016;354(6315):1004–1008.

39. Rentz DM, Dekhtyar M, Sherman J, Burnham S, Blacker D, et al. The feasibility of at-home iPad cognitive

testing for use in clinical trials. J Prev Alzheimers Dis. 2016;3(1):8–12.

40. Paschalidis IC, Tandon A, Sun J, et al. Predicting progression to Alzheimer’s disease using speech patterns:

A machine learning approach. Alzheimers Dement. 2024;20(3):512–521.

41. Lugito NPH, Djuwita R, Adisasmita A, Simadibrata M. Blood pressure lowering effect of Lactobacillus

containing probiotic. Int J Probiotics Prebiotics. 2022;17(1):1–13.

42. Sharun K, Banu SA, Mamachan M, Abualigah L, Pawde AM, Dhama K. Unleashing the future: Exploring

the transformative prospects of artificial intelligence in veterinary science. J Exp Biol Agric Sci.

2024;12(3):297–317.

43. Liu X, Wang Y, Zhang Y, et al. Gait analysis using wearable sensors for early detection of cognitive decline

in Alzheimer’s disease. Front Aging Neurosci. 2023;15:1182932.

44. Lim JK, Li QX, Villemagne VL, et al. Unlocking ocular biomarkers for early detection of Alzheimer’s

disease. Front Neurosci. 2024;18:11848398.

45. Coley N, Coste J, Andrieu S, et al. Sleep and activity patterns as digital biomarkers for early Alzheimer’s

disease: A prospective cohort study. Alzheimers Dement. 2022;18(9):1725–1737.

46. Alhasawi Y, Alghamdi S. Federated Learning for Decentralized DDoS Attack Detection in IoT Networks.

IEEE Access. 2024;12:36132–36145.

47. Zhu X, Suk HI, Lee SW, Shen D. Subtype and stage inference of Alzheimer’s disease by multi-task

learning. Sci Rep. 2021;11:21630.