Hagit Cohen

Objective measurement of stress is critical in populations exposed to chronic adversity, yet existing tools fail to distinguish cognitive effort from physiological arousal at the individual level. We characterized neurophysiological stress signatures in a multimodal, ecologically valid naturalistic forest setting. One hundred and one healthy adults (87 women, 14 men) completed an auditory cognitive battery (resting state, working memory, emotional processing, acute startle) with simultaneous single-channel prefrontal EEG recording. A subset (N = 29) provided salivary cortisol, and heart rate variability (HRV) was assessed at baseline and 3 h post-session. Linear mixed models revealed differential EEG sensitivity profiles: an arousal intensity composite (A0), corroborated by conventional Beta and Alpha power, established a graded hierarchy (Startle ' Emotional ' Mental Load ' Rest), while low-frequency bands (Delta, Theta) and a neural complexity feature (VC9) showed non-specific elevation, peaking at acute startle. A cognitive engagement feature (ST4) increased uniformly across all active conditions regardless of demand type, and a regulatory control feature (T2) was selectively suppressed during mental load but not acute stress. In this exploratory cortisol subsample, higher baseline cortisol negatively predicted A0 and positively predicted ST4 during working memory, consistent with cortisol-mediated arousal attenuation and compensatory cognitive recruitment; these associations should be interpreted with caution given the limited sample size. VC9 and low-frequency spectral activity predicted reduced HRV 3 h post-session, and T2 predicted self-reported anger at the same time point. These results demonstrate dissociable EEG sensitivity profiles for physiological arousal and cognitive effort, supporting a scalable framework for personalized stress screening in populations exposed to prolonged instability.