Abstract
Artificial intelligence (AI) is transforming the early detection, diagnosis, and care of neurodegenerative diseases in older adults. This article reviews current AI applications in imaging, digital biomarkers, and electronic health records for conditions like Alzheimer’s and Parkinson’s, highlighting achievements and limitations. It also explores the integration of AI into eldercare settings for personalized treatment, drug discovery, and monitoring. Challenges remain regarding bias, interpretability, and clinical adoption; however, AI holds significant promise for advancing equitable and effective care for aging populations.
Key Words
digital health, age-related diseases, Alzheimer’s disease, Parkinson’s disease, geriatric care, neurodegeneration, clinical adoption
Age-related neurodegenerative diseases such as Alzheimer’s and Parkinson’s are a growing public health challenge for older populations. The number of people living with dementia worldwide is projected to triple to more than 150 million by 2050 (R. Li et al., 2021). Alzheimer’s disease alone accounts for roughly 70% of dementia cases and has a 10- to 20-year preclinical phase before significant symptoms become visible (Li et al., 2021). This long silent period presents a critical window for intervention that depends upon the possibility of early, accurate detection.
Traditional diagnosis relies upon clinical exams and cognitive tests, which are time-consuming and often miss subtle early changes. Advanced biomarkers (i.e., PET scans or spinal fluid tests) can detect early pathology but can be too invasive or expensive for routine screening. Artificial Intelligence (AI) could be a promising tool to fill this gap. Developing novel tools and methods can lead to early detection, diagnosis, and even treatment for neurodegenerative diseases in older adults.
AI in Early Detection and Diagnosis
AI-driven techniques are enabling earlier and more accurate identification of neurodegenerative diseases than previously thought possible. As access to large collections of quality datasets becomes available, machine learned models are making significant strides in early detection of neurodegenerative diseases through multiple modalities. In imaging analysis, deep learning models—particularly Convolutional Neural Networks (CNNs)—have demonstrated high accuracy in identifying subtle brain changes associated with Alzheimer’s and Parkinson’s disease from MRI, CT, and PET scans (Yousefi et al., 2024). Some models exceed 90% accuracy in classifying Alzheimer’s, while others achieve near-perfect sensitivity in distinguishing early Parkinson’s. Such models can automatically recognize features (e.g., patterns of atrophy or dopamine neuron loss) that even experts might overlook, potentially flagging disease years before clinical diagnosis.
Beyond imaging, AI-based tools are helping uncover digital biomarkers in voice, movement, and sleep patterns (Ali et al., 2024). For example, vocal analysis can distinguish Parkinson’s patients with up to 99% accuracy, while wearable devices and smartphone apps track gait and tremors to identify motor-related dementias (Versel, 2013). Even nocturnal breathing patterns have proven effective for Parkinson’s detection with high accuracy (Yang et al., 2022).
‘In dementia care, AI-driven applications enhance
safety and engagement.’
Meanwhile, AI-driven analysis of electronic health records (EHRs) and cognitive data reveals hidden dementia risk. Advanced natural language processing (NLP) and use of large language models (LLMs) on clinical notes and patient history have identified early signs of dementia in individuals previously undiagnosed, flagging a significant proportion of high-risk cases and prompting earlier clinical intervention (Dorman, 2024; Zhu et al., 2024).
Collectively, these methods highlight AI’s growing role in detecting disease earlier, more accurately, and less invasively than had traditional approaches. Overall, AI-driven diagnostic tools are proving their worth in research: detecting Alzheimer’s or Parkinson’s earlier and with a high degree of accuracy across modalities (imaging, voice, text, etc.). It is important to note this is done in a non-invasive manner, such as using existing data or easy-to-obtain readings, and often with automated, rapid analysis. Such data augments clinicians’ ability to screen large, at-risk populations of older adults.
AI for Treatment and Care Management
Beyond diagnosis, AI-based tools increasingly support treatment and personalized care for neurodegenerative diseases, particularly among older adults. In Parkinson’s disease, AI helps fine-tune deep brain stimulation and adjust medication dosing by predicting symptom fluctuations and improving motor control (Kim, 2025).
In dementia care, AI-driven applications enhance safety and engagement—smart home systems monitor routines and flag deviations such as wandering or missed meals, enabling timely caregiver response. Medication management also benefits: AI-powered electronic records (eMAR) in nursing homes and assisted living facilities detect potential prescription errors, improving adherence and reducing risks associated with complex drug regimens (Nankee, 2025). Early pilots report increased safety and efficiency, especially in settings with staff shortages.
Additionally, AI enables personalized therapy by forecasting disease progression and helping clinicians tailor interventions. Machine models incorporating neuroimaging, genetics, and cognitive testing can predict rates of decline, inform treatment intensity, and identify patient subgroups more likely to respond to specific drugs or interventions. In sum, AI’s role in treatment is still emerging but spans everything from optimizing existing therapies to accelerating drug discovery.
Limitations and Challenges
Despite its promise, AI faces notable challenges in neurodegenerative disease care. First, accuracy in research settings doesn’t always translate to clinical practice. Models trained on controlled datasets can produce false positives when applied to broader populations, leading to overdiagnosis and unnecessary stress or interventions—particularly in diseases like Alzheimer’s (Dang et al., 2025).
Moreover, distinguishing between similar conditions with overlapping symptoms remains difficult. For example, disorders like Alzheimer’s, Lewy body dementia, and Parkinson’s can confuse AI systems, especially if training data lacks diagnostic and demographic variety. Reliable differentiation often requires multimodal data and clinician oversight and feedback (Xue et al., 2024).
Additionally, data bias and quality limit generalizability. Many training datasets underrepresent diverse populations or rare diseases such as amyotrophic lateral sclerosis (ALS), reducing model accuracy, especially for those outside typical demographics. Machine learned models require further development to incorporate data samples such as these instead of treating them as outliers. Working with real-life data (EHR records, sensors, etc.) has its share of problems due to noise or inconsistency, making it more difficult for AI to detect patterns reliably (Dang et al., 2025).
Finally, interpretability remains a major barrier. This opaqueness is problematic in medicine—geriatricians and neurologists are understandably hesitant to act on an AI’s judgment unless they understand why a patient was flagged. Such lack of interpretability hampers clinician trust and adoption of AI systems. To gain trust and regulatory approval, significant efforts are being developed in Explainable AI (XAI), which highlights relevant features influencing predictions. Addressing these challenges is essential to ensure AI tools are effective, equitable, fair, and clinically integrated (Dang et al., 2025).
Integration Into Healthcare Systems
The incorporation of AI-driven systems into U.S. elder care and clinical practice is progressing, with hospitals and long-term care facilities beginning to pilot systems for dementia and Parkinson’s care, but widespread implementation remains to be seen. A notable example is Cedars–Sinai Medical Center’s (Los Angeles, CA) recent initiative: Its team used AI to analyze EHR records for signs of undiagnosed dementia in admitted older patients. Because many older adults with mild or early onset dementia may be mis- or undiagnosed, the fact that Cedars-Sinai’s AI analysis of EHR revealed up to 10% of EHR records showing signs of altered mental status or cognitive dysfunction, means it successfully flagged cognitively impaired individuals who had no previous formal dementia diagnosis (Dorman, 2024).
‘Technology alone isn’t enough; healthcare
workers must be equipped
to interpret and act upon AI insights.’
Following this study, Cedars–Sinai instituted staff training so nurses and physicians would know how to respond when AI raises a dementia alert. This underscores a key aspect of integration: Technology alone isn’t enough; healthcare workers must be equipped to interpret and act upon AI insights.
The practical and ethical concerns remain relevant in the context of successful implementation and require careful attention. Integrating AI into healthcare IT systems poses questions on interoperability and cybersecurity, especially with sensitive patient data, along with governance and access control required in AI systems. Policymakers are cautiously supportive, with federal initiatives funding AI research and regulatory pathways emerging. However, widespread adoption hinges on concerns over approval to ensure compliance with patient privacy regulations, such as HIPAA.
Ultimately, a collaborative approach involving technologists, clinicians, and policymakers will not only be essential, but necessary to realize the potential of AI to improve the safety, quality of life, and equitable access to care for older adults.
Future Directions
The coming years promise substantial growth in AI’s role fighting neurodegenerative diseases. Research is moving toward multimodal AI systems that combine various data sources—neuroimaging, genetic profiles, speech and handwriting samples, lifestyle factors—to paint a more holistic and accurate picture of an individual’s brain health. These integrated models could further improve early detection and extend into differentiating the complexities of overlapping conditions by learning to recognize underlying data signatures for each disease. In conjunction with XAI, model transparency, along with clinician oversight, can not only show why a prediction was made but also enhance decision-making post-prediction.
Regarding treatment, AI-driven drug discovery is accelerating. AI-based algorithms can identify novel compounds and repurpose existing ones for neurodegenerative targets (Li et al., 2025). It remains to be seen which early candidates will reach the market as currently there are no AI-derived Alzheimer’s medications. In clinical care, AI may enable precision medicine by predicting which patients will benefit from specific therapies. Assistive technologies, including intelligent robots and voice interfaces, also could play a growing role, supporting daily routines, monitoring health changes, and delivering personalized cognitive stimulation.
It is necessary for policy and ethical frameworks to evolve alongside AI-driven growth and innovation. We can expect updated guidelines on AI in geriatric care from professional communities to ensure AI tools are accurate, unbiased, and safely implemented. Keeping eyes on patient data and machine model explainability, security, and privacy protections will continue to be the critical path of success, especially to expand data-sharing efforts and the inclusion of diverse patient populations.
AI demonstrates capability in research and treatment of neurogenerative disease, and as such, is positioned to become a powerful ally in addressing age-related neurodegenerative conditions. From advances in early detection, to personalized treatment, to enhancements in overall elder care, AI offers tools to enhance outcomes for older adults (Adam, 2024).
Despite remaining challenges in validation, integration, and ethical concerns, AI is showing promise and is likely to become a more common tool to help clinicians in their work. Through thoughtful consideration of patient-centered design and integration, AI-based tools can bridge the gap to meet the needs of a rapidly aging population and transform how we understand and manage brain health.
Faizan Wajid, PhD, is a senior data and research scientist at Explore Digits, where he is developing AI-driven tools to support healthcare providers—from policy makers to caregivers. His research interests include sensors and signal analysis using wearable technologies, as well as LLM.
Photo credit: meeboonstudio
References
Adam, J. (2024, November 1). Can AI cure Alzheimer’s disease? Labiotech. https://www.labiotech.eu/in-depth/ai-alzheimer/
Ali, M. H., Mohammed, S. L., & Al-Naji, A. (2024). Parkinson’s disease detection from voice using artificial intelligence techniques: A review. AIP Conference Proceedings. https://pubs.aip.org/aip/acp/article-abstract/3232/1/040010/3316679/Parkinson-s-disease-detection-from-voice-using?redirectedFrom=PDF
Dang, V. N., Casamitjana, A., Hernández-González, J., Lekadir, K., & the Disease Neuroimaging Initiative (2025). Mitigating overdiagnosis bias in CNN-Based Alzheimer’s disease diagnosis for the elderly. Universitat de Barcelona. https://diposit.ub.edu/dspace/bitstream/2445/220330/1/Mitigating%20Overdiagnosis%20Bias.pdf
Dorman, A. (2024). Hospital study on AI identifies patients with undiagnosed dementia coming from LTC. McKnights Senior Living. https://www.mcknightsseniorliving.com/news/hospital-study-on-ai-identifies-undiagnosed-dementia-patients-coming-from-ltc
Kim, M. (2025, April 3). An AI-enhanced treatment has reduced Parkinson’s symptoms for some people. Washington Post. https://www.washingtonpost.com/wellness/2025/04/03/parkinsons-adaptive-dbs-deep-brain-stimulation/
Li, R., Wang, X., Lawler, K., Garg, S., Bai, Q. & Alty, J. (2021). Applications of artificial intelligence to aid detection of dementia: A narrative review on current capabilities and future directions. ArXiv: Artificial Intelligence. https://arxiv.org/abs/2104.14073
Li, V. O. K., Han, Y., Kaistha, T. Zhang, Q., Downey, J., Gozes, I., & Lam, J. C. K. (2025). DeepDrug as an expert guided and AI driven drug repurposing methodology for selecting the lead combination of drugs for Alzheimer’s disease. Scientific Reports, 15, 2093.
Nankee, B. (2025). How AI is enhancing medication management in senior living. ecp. https://www.ecp123.com/blog/how-ai-is-enhancing-medication-management-in-senior-living
Versel, N. (2013). Phone detection of Parkinson’s approaches 99 percent accuracy. mobilehelathnews. https://www.mobihealthnews.com/20068/phone-detection-of-parkinsons-approaches-99-percent-accuracy
Xue, C., Kowshik, S. S., Lteif, D., Puducheri, S., Jasodanand, V. H., Zhou, O. T., Walia, A. S., Guney, O. B., Zhang, J. D., Poésy, S., Kaliaev, V., Andreu-Arasa, C., Dwyer, B. C., Farris, C. W., Hao, H., Kedar, S., Mian, A. Z., Murman, D. L., O’Shea, S. A., Paul, A. B., Rohatgi, S., Saint-Hilaire, M.-H., Sartor, E. A., Setty, B. N., & Kolachalama, V. B. (2024). AI-based differential diagnosis of dementia etiologies on multimodal data. Nature Medicine, 30, 2977–2989. https://www.nature.com/articles/s41591-024-03118-z
Yang, Y., Yuan, Y., Zhang, G., et al. (2022). Artificial intelligence-enabled detection and assessment of Parkinson’s disease using nocturnal breathing signals. Nature Medicine, 28, 2207–2215. https://www.nature.com/articles/s41591-022-01932-x
Yousefi, M., Matin, A., Mohamadi, Z., Shaghayegh, K., Dasoomi, H., Atabi, A., Sarkeshikian, S. A., Dehaki, M. A., and Bayati, H., Mashayekhi, N., Varmazyar, S., Rahimian, Z., Asadi Anar, M., Shafiei, D., & Mohebbi, A. (2024). Machine learning-based algorithms for virtual early detection and screening of neurodegenerative and neurocognitive disorders: A systematic review. Frontiers in Neurology, 15. https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2024.1413071/full
Zhu, W., Tang, H., Zhang, H., Rajamohan, H. R., Huang, S. L., Ma, X., Chaudhari, A., Madaan, D., Almahmoud, E., Chopra, S., Dodson, J. A., Brody, A. A., Masurkar, A. V., & Razavian, N. (2024). Predicting risk of Alzheimer’s diseases and related dementias with AI foundation model on electronic health records. medRxiv [Preprint].
