Could AI help uncover hidden Parkinson’s subtypes years earlier? Multimodal means combining many different types of information such as genetics, brain scans, spinal fluid markers and cognitive testing into one analysis. Reyes and colleagues describe in a new paper in npj Artificial Intelligence how a new AI framework called SPARROW may help identify Parkinson’s disease subtypes using multimodal reasoning and interpretable AI. Key points: - SPARROW combined genetics, MRI brain scans, spinal fluid biomarkers and cognitive testing into a single reasoning framework for Parkinson’s disease subtyping. - The AI system achieved strong accuracy in distinguishing slow, moderate and fast progressing Parkinson’s disease subtypes using baseline patient data. - The system generated step by step reasoning traces, allowing health care providers to better understand how the AI arrived at each subtype decision. My take: Parkinson’s disease is not one disease. It is many diseases hiding under one name. This study is exciting because it moves beyond simply identifying Parkinson’s and instead asks whether AI can help us understand who may progress faster and why. The interpretability piece is critical. Black box AI will struggle in neurology clinics. Transparent reasoning may be the bridge that builds trust and utility. The performance of this approach may not be as good as a clinician. Also, the scientists turned the imaging data into text. Here are 5 points that resonated w/ me: 1- Parkinson’s disease likely contains multiple biological subtypes and these differences may explain why symptoms and progression vary so widely between folks. 2- Cognitive testing still emerged as one of the strongest signals in predicting subtype and progression. 3- MRI and genetic information alone were weak predictors, however when combined w/ other data they improved accuracy substantially. 4- The AI system was designed to tolerate missing data, which mirrors real world neurology practice where every test is not always available. 5- The future of Parkinson’s care may involve AI systems that help health care providers personalize prognosis, treatment selection and monitoring much earlier in the disease course. nature.com/articles/s44387-0…

Exposome and brain aging: your environment may be shaping your brain faster than you think. The exposome refers to the sum of all environmental exposures you experience across your life including air, water, social conditions and policies. Hernandez and colleagues describe in a new paper in Nature Medicine how the exposome across 34 countries influences brain aging and risk for cognitive decline. Key Points: - Combined environmental exposures explained far more about brain aging than any single factor alone. - Physical exposures like air pollution, temperature and limited green space were linked to structural brain aging. - Social exposures like inequality, reduced rights and poverty were linked to functional brain aging and brain network changes. My take: This is one of those papers that shifts how we think. Brain aging is not just about genes or disease, it is about the world we live in. The cumulative burden of exposures may quietly accelerate aging even before symptoms appear. If we want to change brain health, we must think beyond the clinic and into our communities. Here are 5 points that resonated w/ me: 1- Your environment is not background noise, it is an active driver of brain health across your lifetime. 2- Air quality, water, climate and access to green space may directly influence how fast your brain ages. 3- Social factors like inequality, education and community stability matter just as much as biology. 4- Multiple small exposures can combine and amplify risk more than any single factor alone. 5- Improving environments and policies may be one of the most powerful ways to protect brain health for future generations. nature.com/articles/s41591-0… #parkinson #alzheimer #dementia