Alzheimer’s Dement
AI model links diet, medical history to Alzheimer’s risk
Clinical takeaway: AI tools built from routine clinical and lifestyle information may eventually help identify patients at higher Alzheimer’s risk and flag who may need closer evaluation.
Alzheimer’s risk is usually discussed in terms of brain biomarkers, but this analysis points to a broader clinical picture. Investigators used an artificial intelligence model to test whether factors such as diet, medical history, and lifestyle could help distinguish people with Alzheimer’s disease from controls, and whether gut microbiome patterns might help explain some of those associations.
The study found that overall dietary patterns mattered more than nutrient-by-nutrient intake, and that medical history was similarly informative. Vascular conditions and depression were among the medical factors most closely tied to higher risk. Processed and refined foods were associated with higher-risk patterns, while plant-based and omega-3-rich diets were more protective.
Exploratory microbiome analyses suggested reduced microbial diversity and fewer short-chain fatty acid-producing bacteria in higher-risk groups, consistent with the idea that diet-related gut dysbiosis may contribute to neuroinflammation and Alzheimer’s risk. These findings help frame a possible mechanism.
Investigators analyzed a public dataset of 9,832 participants with questionnaire-based diet, lifestyle, and medical history data. They grouped 120 variables into five domains, used statistical screening to select the most informative features, and trained several machine learning models with validation in an independent test set. They also analyzed microbiome sequencing data from 2,000 samples to explore whether gut bacterial patterns might help explain the clinical associations.
“Beneficial bacteria responsible for producing short-chain fatty acids, compounds that maintain the gut’s protective barrier and actively suppress neuroinflammation, were significantly depleted,” said Faezeh Karimi, PhD, project lead and senior lecturer at the University of Technology Sydney School of Computer Science.
“Microbial diversity was reduced,” she continued. “In their place, a more inflammatory microbial environment had taken hold, one that appears capable of sending damaging signals through the gut-brain axis directly to the brain.”
A questionnaire-based AI approach could eventually help flag higher-risk older patients before more specialized evaluation, though the researchers say the model still needs further validation.
Source: Jabeen T. Alzheimer’s Dement. 2026 Apr 19. Multi-modal machine learning and gut microbiome pathway analysis for Alzheimer’s risk prediction