Machine-Learning Models of Cognitive Test Performance Incorporating Exposomic Pesticide Biomarkers in Older U.S. Adults

Description

Carlos A. Toro and Giulio Maria Pasinetti use machine learning to model thinking test scores in older United States adults, folding in urine markers of pesticide exposure, in 2026 work. The data came from a national health survey run from 2011 to 2014 and covered adults aged sixty and over with full thinking tests and exposure measures. Models also took in background, lifestyle, and mood information, plus scores on a depression questionnaire.

Sex, racial and ethnic group, and schooling ranked among the strongest predictors of test performance. Two breakdown products of a common insect repellent kept negative links to a speeded symbol test in tightened models, most clearly among adults in their sixties. The authors say exposure markers may add useful context to aging models, yet only long-term studies can tell whether the links reflect cause, shared exposures, or leftover mixing of factors.

Catalog record

Source: https://europepmc.org/article/MED/PMC13510830