Predicting Cognitive Outcome Through Nutrition and Health Markers Using Supervised Machine Learning.

Background: Machine learning (ML) use in health research is growing, yet its application to predict cognitive outcomes using diverse health indicators is under-investigated.

Objective: We utilized ML models to predict cognitive performance based on a set of health and behavioral factors, aiming to identify key contributors to cognitive function for insights into potential personalized interventions.

Methods: Data from 374 adults aged 19-82 years (227 females) were used to develop ML models predicting cognitive performance (reaction time in milliseconds [ms]) on a modified Eriksen flanker task. Features included demographics, anthropometric measures, dietary indices (healthy eating index [HEI], Dietary Approaches to Stop Hypertension [DASH], Mediterranean, and Mediterranean-DASH Intervention for Neurodegenerative Delay [MIND]), self-reported physical activity, systolic (SBP) and diastolic blood pressure (DBP). The dataset was split (80:20) for training and testing. Predictive models (Decision Trees, Random Forest, AdaBoost, XGBoost, Gradient Boosting, Linear, Ridge, and Lasso Regression) were used with hyperparameter tuning and cross-validation. Feature importance was calculated using Permutation Importance, while Performance using Mean Absolute Error (MAE) and Mean Squared Error (MSE).

Results: Random Forest Regressor exhibited the best performance, with the lowest MAE (ms) (training: 0.66, testing: 0.78) and MSE (ms²) (training: 0.70, testing:1.05). Age was the most significant feature (score: 0.208), followed by DBP (0.169), BMI (0.079), SBP (0.069), and HEI (0.048). Ethnicity (0.005) and sex (0.003) had minimal predictive effect.

Conclusion: Age, blood pressure, and BMI showed strong associations with cognitive performance, while diet quality had a subtler effect. These findings highlight the potential of ML models for developing personalized interventions and preventive strategies for cognitive decline.