Bottling up emotions tracks with worse sleep, but the link is small

Expressive suppression — keeping a lid on the outward signs of emotion — is one of the most commonly used emotion regulation strategies, and it has long been suspected of carrying a cost. Individual studies linking it to sleep in healthy adults have produced mixed and sometimes contradictory results, which leaves it unclear whether there is any reliable association at all. It is also unclear whether the picture depends on culture, given that norms around emotional display vary, or on the level of ambient stress people are living under. Synthesising the correlational literature is the way to test both questions.

The authors searched PubMed, EBSCO and Web of Science from database inception to May 2025 for studies of emotion regulation and sleep quality. To be included, a study had to use non-clinical adult samples, measure expressive suppression with the Emotion Regulation Questionnaire, assess sleep quality with a standardised self-report instrument such as the PSQI or ISI, and report zero-order Pearson correlations. Twenty-three independent samples totalling 13,636 participants met these criteria. A random-effects model estimated the pooled correlation, and pre-specified subgroup and moderation analyses tested cultural background, pandemic context, sleep measure, study design and publication period.

Expressive suppression was significantly associated with poorer sleep quality, but the effect was small: a pooled correlation of r = .14 (95% CI 0.12 to 0.16). Heterogeneity between studies was low, at I2 = 28.21%, suggesting the estimate was fairly consistent across samples. The association was stronger in studies explicitly focused on COVID-19-related stress than in those conducted in general contexts. Notably, several moderators did not matter: cultural background, the type of sleep measure used, study design and publication period showed no significant moderating effect on the association.

This settles part of the inconsistency in earlier single studies by showing that a small but statistically reliable association does exist across a large pooled sample. The null result for cultural background is interesting in itself, since it undercuts the idea that suppression is benign for sleep in cultures where restraint is more normative. The pandemic finding raises the possibility that suppression's costs surface most under major psychosocial stress, though that comparison is between study contexts rather than a controlled manipulation. Because everything rests on cross-sectional correlations, the direction of the relationship remains unresolved.

This is a reason to think how you handle emotion and how you sleep are related, not evidence that suppressing feelings ruins your sleep. The size of the association is modest, so it sits alongside far stronger sleep influences rather than displacing them. It is also worth remembering that poor sleep can plausibly make people more likely to clamp down on emotion, which correlational data cannot separate from the reverse. If anything, the pandemic subgroup hints the relationship is most visible when life is already stressful.