Abstract: Cortisol awakening response (CAR) is closely associated with individual mental health. Traditional studies have used the mean value of CAR over several days to explore this relationship. However, their conclusions have been highly inconsistent due to the influence of state-dependent factors, such as sleep. Therefore, this study introduces CAR variability across multiple days as a novel measure to quantify CAR and investigates its relationship with mental health while regarding trait anxiety and psychological resilience as key variables under natural sleep observation and sleep deprivation manipulation scenarios. This study hypothesized that under the natural sleep condition, a smaller CAR variability reflects positive mental health characteristics, specifically higher psychological resilience scores and lower trait anxiety scores, while under the sleep deprivation condition, a greater CAR variability reflects positive mental health characteristics.
Two experiments were conducted to test these hypotheses. In Experiment 1, the CAR variability of 28 participants was reduced by decreasing sleep efficiency variability under natural sleep conditions. In Experiment 2, 40 participants experienced a full night of sleep deprivation following 2 natural sleep days to increase their CAR variability. All participants wore a sleep wristwatch to assess their sleep efficiency during the experimental period and provided 4 saliva samples at awakening and at 30, 45, and 60 minutes post-awakening to assess their CAR. The trait anxiety inventory and the Connor-Davidson resilience scale were used as key variables reflecting individual mental health. The relationship between CAR variability over multiple days and trait anxiety/psychological resilience was then examined.
Experiment 1 found a significant positive correlation between the variability of participants' subjective sleep efficiency over three days and their CAR variability. Additionally, a significant positive correlation was observed between CAR variability and trait anxiety scores over the three days, indicating that a smaller CAR variability in a stable environment is associated with lower levels of trait anxiety. Experiment 2 found no significant differences in CAR during the first two natural sleep days. However, on the third day, following a full night of sleep deprivation, CAR showed a marked blunting, with post-deprivation CAR levels being significantly lower than those recorded on the preceding two natural sleep days. Experiment 2 also found a significant positive correlation between CAR variability following sleep deprivation and psychological resilience, and this significant correlation was independent of daily stress factors. Therefore, a higher CAR variability in a changing environment is associated with higher levels of psychological resilience. When using the traditional CAR mean as a measure to explore its relationship with trait anxiety and psychological resilience, no significant correlations were found in either experiment.
These results suggest that in addition to mean CAR, CAR variability can be regarded as an effective physiological indicator for measuring mental health. Smaller CAR variability in stable environments is associated with lower levels of trait anxiety, while greater CAR variability in more variable environments is associated with higher levels of psychological resilience. This study emphasizes the importance of considering CAR variability over multiple days to understand how individuals adapt to daily stressors and challenges, thus providing new perspectives and potential approaches for promoting mental health and designing effective intervention strategies in the future.

Key words: cortisol awakening response (CAR), variability, sleep efficiency, trait anxiety, psychological resilience