Estradiol and non-REM sleep attenuate physiological and emotional responses to social-evaluative stress in healthy women

Abstract

Background: This study examined whether individual differences in estradiol and sleep predict autonomic and cognitive responses to social-evaluative stress in healthy young women.

Methods: Forty-two healthy women underwent overnight in-laboratory sleep monitoring, followed by a social-evaluative stress task the next morning, which involved listening to a playback of their own karaoke singing. Pre-task estradiol levels were measured via blood sampling. Autonomic responses were assessed using pupil dilation during the playback, while cognitive responses were indexed by subjective stress ratings collected immediately afterward.

Results: Higher estradiol concentrations (range: 40-1129 pmol/L) were associated with reduced peak pupil dilation (range: 0.21-1.70 mm above the 95% confidence interval upper bound of baseline) and cumulative pupil dilation (total pupil expansion exceeding the individual baseline over 45 s; range: 64-3139 mm, p < 0.001), indicating lower autonomic arousal. In contrast, estradiol was not associated with subjective stress ratings (range: 10-98 mm on a 100-mm scale). Greater N3 sleep duration (range: 27-172 min), but not N1 or N2, was associated with lower subjective stress (p = 0.032), whereas REM sleep duration (range: 53-152 min) showed no association with either outcome. None of the sleep measures was significantly related to pupillometry outcomes, and no interaction effects were observed between estradiol and sleep measures.

Conclusions: Estradiol and N3 sleep appear to independently modulate distinct facets of the response to social-evaluative stress: higher estradiol levels were associated with reduced autonomic arousal, while greater N3 sleep duration was linked to lower subjective stress. By revealing parallel pathways through which estradiol and sleep shape physiological and cognitive reactions to stress, this work contributes to a more nuanced understanding of the biological factors that may influence mental health and stress resilience.

Keywords: Autonomic arousal; Cognitive appraisal; Emotion regulation; Estradiol; Non-REM sleep; Social stress.

Fig. 1

Overview of study protocol. Forty-two women of reproductive age who were naturally cycling (i.e., not using hormonal contraceptives) participated in the study. Each underwent an 8-h in-laboratory sleep opportunity (23:00–07:00), during which a headband was used to assess sleep. The following morning, venous blood samples were collected approximately 30 min after awakening to quantify estradiol concentrations. Participants then completed an eye-tracking task comprising three phases: a 30-s resting baseline, 45 s of karaoke playback (used as a social-evaluative stressor), and a 30-s post-task period. Pupil dilation was continuously recorded throughout the task. Immediately following the playback, participants rated their perceived stress using a visual analog scale (VAS). For illustrative purposes, the pupil dilation response of one randomly selected participant across the task timeline is shown

Fig. 2

Elevated morning estradiol concentrations were significantly associated with reduced peak pupil dilation during the karaoke replay. The blue line represents the predicted association, and the estimates shown are derived from a generalized linear model, with morning blood estradiol concentrations measured after sleep entered as a continuous predictor and peak pupil response (expressed as a percentage from baseline) as the outcome variable. Each gray “x” represents an individual participant’s observed data point. For adjusted estimates of baseline-corrected peak pupil diameter, see the Results section

Fig. 3

Elevated morning estradiol concentrations were significantly associated with a reduced cumulative peak pupil response during the karaoke replay. The blue line represents the predicted association, and the estimates shown are derived from a generalized linear model, with morning blood estradiol concentrations measured after sleep entered as a continuous predictor and cumulative pupil dilation as the outcome variable. Each gray “x” represents an individual participant’s observed data point. For adjusted estimates, see the Results section

Fig. 4

More time spent in non-REM sleep during the night was significantly associated with lower subjective stress during the karaoke task. The blue line represents the predicted association, and the estimates shown are derived from a generalized linear model, with time spent in non-REM sleep on the pre-test night entered as a continuous predictor and subjective stress ratings (range: 0–100 mm) as the outcome variable. Each gray “x” represents an individual participant’s observed data point. For adjusted estimates, see the Results section