Researchers have developed an artificial intelligence model capable of identifying individuals at high risk for Alzheimer's disease using readily available medical and dietary information. The AI's findings also suggest a link between gut health and the neurodegenerative condition.
Traditional methods for early Alzheimer's detection, such as costly scans or invasive biomarkers, limit widespread access. This new study tested the efficacy of using questionnaire-based data, including medical history and dietary patterns, for risk prediction. The research also investigated if gut microbiome patterns could explain the connections between diet, health history, and neurodegeneration.
The machine learning study analyzed data from 9,832 participants, incorporating demographic, dietary, lifestyle, nutritional, and medical factors. Four algorithms were trained and externally validated. Microbiome sequencing data from 2,000 participants allowed for exploratory microbial composition analysis.
Results showed that medical history and dietary patterns were the strongest predictors of Alzheimer's risk, achieving an area under the curve (AUC) of 0.871 and 0.874, respectively. These significantly outperformed demographic, lifestyle, and isolated nutritional intake measures. Influential features identified by the model included vascular conditions, depression, and eating behaviors. Exploratory analysis of the gut microbiome revealed dysbiosis markers, suggesting potential gut-brain inflammatory pathways involved in disease development.
These findings indicate that Alzheimer's risk prediction may become feasible at a community level through low-cost questionnaires, potentially guiding earlier monitoring or preventive strategies. Further longitudinal studies are needed to validate causality and refine accuracy, exploring the impact of microbiome-targeted or dietary interventions on reducing future Alzheimer's incidence.