INDIVIDUAL DIFFERENCES IN AUTONOMIC NERVOUS SYSTEM REACTIVITY PREDICT LANGUAGE LEARNING UNDER ACUTE PSYCHOSOCIAL STRESS

Learning a new language in adulthood can be socially stressful, yet how acute psychosocial stress affects the ability to learn linguistic patterns and regularities remains unclear. Here, we investigated (1) whether acute psychosocial stress affects statistical language learning (SLL)—i.e., implicit learning of probabilistic patterns from continuous speech streams—and (2) whether individual variability in autonomic nervous system (ANS) reactivity is associated with SLL ability. Sixty-five adults participated in a six-hour experiment including stress and recovery periods, mimicking everyday life. Participants were exposed to novel speech streams immediately after a Virtual Reality Trier Social Stress Test (VR-TSST) and during a less stressful period. Learning was tested immediately after exposure and approximately four hours later. Heart-rate variability (HRV) and electrodermal activity were recorded. Principal component analyses yielded composite indices of sympathetic (SNS) and parasympathetic (PNS) reactivity, quantifying individual responses to stress and recovery. Acute stress did not influence SLL at the group level. However, individual differences in SLL performance were associated with congruence of ANS reactivity: participants whose SNS and PNS branches were congruently reactive—both high or both low—showed higher SLL performance than those with incongruent reactivity across all conditions. Learning also showed phase-dependent associations with ANS reactivity, with SNS reactivity supporting encoding and PNS reactivity supporting consolidation. Stable learning across phases was observed only in individuals with congruent ANS reactivity. These findings link ANS reactivity to SLL, underscoring how individual differences in physiological regulation affect learning under stress and motivating physiologically informed, personalized approaches to language learning.