Abstract Introduction Nightmares are emotionally intense sleep disturbances linked to impaired emotion regulation and physiological arousal. Although stress sensitivity is theorized to increase vulnerability to nightmares, little research has examined whether specific HPA-axis indices, particularly the Cortisol Awakening Response (CAR), predict nightmare frequency in non-clinical young adults. This study evaluated whether CAR, perceived stress, and sleep quality were associated with nightmare occurrence in college students. Methods Participants were female college students (N=41; Mage = 22.7). Participants completed the PSQI, including the nightmare item (5h), and provided eight saliva samples across two days for cortisol assessment. CAR was computed from the first two daily samples at 30-min and 60-min post-waking. Perceived stress was measured with the Perceived Stress Scale. Nightmare frequency was coded dichotomously (any nightmare or none). Logistic regression with 1,000-sample bootstrapping tested whether CAR, perceived stress, and non-nightmare sleep disturbances predicted nightmare presence. Additional correlations and exploratory reverse models were examined. Results Nightmares were reported by 14.6% of participants (n=6). CAR values ranged from -1.00 to 0.65 μg/dL, with several participants showing blunted responses. CAR was significantly correlated with nightmare presence (r = .35, p = .03) and with global sleep disturbance (r = .34, p = .04). In the logistic regression model, CAR significantly predicted the presence of nightmares (B = 4.62, bootstrapped p = .023, 95% CI 0.09, 6.81). Perceived stress and general sleep disturbances were not significant predictors. The overall model did not reach statistical significance, reflecting limited power and restricted nightmare distribution. An exploratory reverse model indicated that nightmares modestly predicted higher CAR (β = .33, p = .055; R² = .164) Conclusion A higher cortisol awakening response was associated with greater likelihood of reporting nightmares, suggesting that morning HPA-axis reactivity may reflect individual differences relevant to disturbed dreaming. Self-reported stress did not predict nightmares once CAR was included, indicating that biological stress reactivity may be more sensitive than subjective stress in identifying nightmare risk. Findings emphasize the value of incorporating physiological markers into nightmare research and highlight the need for longitudinal, multi-day sampling to clarify directionality and underlying mechanisms in the stress-nightmare association. Support (if any)
Nichols et al. (Fri,) studied this question.