Artificial intelligence offers new insights into neurodegenerative diseases

The relationship between neurodegenerative diseases, brain tissue abnormalities and associated symptoms is often complicated and unclear. Symptoms can overlap between different conditions and the clinical picture varies greatly from patient to patient, frequently leading to misdiagnoses. Researchers and doctors need more clarity, and artificial intelligence (AI) seems to be able to play an important role in this.

AI as a diagnostic tool

Dutch researchers have deployed AI to better understand these connections. In their publication in Nature Medicine, they describe how they used data from the Dutch Brain Bank. This unique database contains detailed descriptions of neuropathological diagnoses and symptoms from thousands of donors, combined with post-mortem brain tissue.
By ‘feeding’ AI with this information, they created a model that identified 90 different symptoms, divided into five domains: psychiatric, cognitive, motor and sensory symptoms. In addition, they developed a second AI model that could make diagnoses based on the clinical picture. Although the model generally worked accurately, it struggled with rare disorders.

Wrong diagnoses detected

A surprising outcome of the study was the discovery of a group of donors who had been misdiagnosed while alive. For example, some patients with Alzheimer's disease showed symptoms more consistent with Parkinson's disease, while in others frontotemporal dementia was mistaken for Alzheimer's. This highlights how complex the relationship between symptoms and brain abnormalities can be, and how AI can help detect such misdiagnoses.

A unique resource for future research

In doing so, the research team has created a valuable resource for future research. This data can help researchers better understand the development of symptoms, support molecular biologists in discovering the underlying mechanisms of neurodegenerative diseases, and provide computational scientists with material to build predictive models for better diagnosis and prognosis of dementia and other brain disorders.

The dementia risk self-test forms the basis of your personalised prevention strategy to reduce your risk.