Peer-reviewed article

Chronic heat exposure and perceptual reasoning in preschool-aged children: Robust estimation, effect modification and period-specific analysis.

Guillaume Barbalat, Marie-Aline Charles, Ian Hough, Ludivine Launay, Itai Kloog et al. · 10 Jul 2026

Abstract

BACKGROUND: Emerging evidence links chronic heat exposure to impaired neurodevelopment, yet research on higher-order cognitive domains remains limited. Furthermore, the utility of existing evidence is often constrained by the use of traditional static exposure counterfactuals (e.g., "high" vs. "low" temperature throughout a long time-frame) that fail to reflect plausible, real-world temperature trajectories. In this study, we assessed the association between realistic scenarios of temperature increases during the prenatal and postnatal periods and infant perceptual reasoning scores. METHODS: Data were drawn from the French ELFE birth cohort (N = 8965). Perceptual reasoning was measured at age 3.5 using the British Ability Scale Picture Similarities subtest. Ambient heat was defined as the percentage of hot days where daily minimum temperature was higher than the 95th percentile across four temporal windows from conception to age 3.5. We estimated effects of incremental temperature shifts (+1°C to +3°C) using a sequentially doubly-robust estimator integrated with a SuperLearner machine learning ensemble. In addition, we assessed how these effects vary across a subset of covariates and explored critical windows of developmental vulnerability. RESULTS: While a +1°C increase showed no significant effect, population scores decreased by 0.7-0.8 points in the +2°C scenario and by 1.4-1.5 points in the +3°C scenario (all p < 0.001). For every unit increase in the European Deprivation Index, the negative effect of a +3°C shift deepened by 0.035 points (p < 0.04), while urban residence significantly intensified the heat-outcome relationship by 0.42 to 0.46 points (p < 0.015). Exploratory period-specific analyses identified the prenatal period and the third year of life as windows of highest sensitivity across all temperature shift scenarios. CONCLUSION: Elevations of 2-3°C in nocturnal temperatures throughout early development are significantly associated with reduced perceptual reasoning scores. Both urban residency and social deprivation may exacerbate these heat-related risks. High-sensitivity windows were predominantly identified during the prenatal period and third year of life.