AI analysis finds mental health, not weather, drives cognitive difficulties
Psychosocial factors like loneliness 10x stronger than pollution or temperature...
A study using AI-assisted causal inference on longitudinal data from the All of Us Research Program (2018–2024) compared environmental exposures (weather, air pollution) and psychosocial factors (mental health, loneliness) on subjective cognitive difficulties. After accounting for within-location variation, environmental associations attenuated substantially, while psychosocial factors like mental-health burden and loneliness showed consistently larger effect sizes. These patterns were replicated in the 2024 BRFSS data.
- AI-assisted causal inference on All of Us data (n=large, longitudinal 2018–2024) found environmental effects on cognition vanish after controlling for location.
- Psychosocial factors (loneliness, mental health burden) showed up to 10x larger effect sizes than any environmental exposure.
- Findings replicated in 2024 BRFSS, confirming the pattern across two large U.S. datasets.
Why It Matters
Challenges focus on short-term environmental triggers; suggests interventions targeting loneliness and mental health could reduce cognitive decline risk.