Network control dynamics and subjective reactivity in physiological responses to psychosocial stress

The ability to endure psychosocial stressors is critical for mental and physical well-being. Clarifying mechanisms that differentiate high- from low-tolerant individuals may inform resilience-oriented interventions. This exploratory study aimed to predict tolerance to the socially evaluated cold pressor test (SECPT) from a multimodal set of psychological ratings and physiological markers, quantifying how psychophysiological responses account for individual differences in acute psychosocial stress tolerance. Thirty healthy adults completed a 5-min baseline followed by the SECPT. Self-reported perceptual and affective responses, electrodermal activity, and electroencephalography (EEG; sensor-level Granger connectivity computed over a frontoparietal FPN scalp montage) were acquired throughout; a brief semi-structured interview complemented quantitative findings. Models were evaluated with stratified 5-fold cross-validation. A Random Forest regressor with a square-root-transformed duration target explained 23.5% of the variance. Two composite features emerged as primary, directionally opposite predictors: the Stress Response Index showed a positive effect; higher perceived stress, arousal, and pain were associated with longer tolerance, whereas FPN Causal Connectivity showed a negative effect; stronger directed influence predicted shorter tolerance. The SECPT manipulation produced a perceptual profile of higher stress, pain, and arousal with lower affective valence and perceived dominance. Sympathetic activity predominated, with an early peak and a trend toward habituation. Global FPN connectivity was attenuated, most notably over parietal, central-parietal, and frontal interhemispheric circuits. Together, these results indicate that tolerance reflects an interplay between subjective reactivity and network control dynamics. The findings provide initial, mechanistically informed markers of psychosocial stress tolerance and motivate larger studies to test generalizability and temporal dynamics.