ABSTRACT This study examined associations between trait anxiety and relative sleep‐stage proportions, with a focus on REM sleep, and explored differences across sex and age groups. A total of 128 non‐clinical adults (75 women, 53 men; age range 18–82 years, M = 35.5, SD = 18.4) were recruited from a university setting. Sleep stages were recorded using Fitbit Charge 5 devices, and a complementary verification substudy ( n = 6) was conducted to confirm whether the observed correspondence between wearable‐derived and polysomnographic estimates replicated trends in validation studies. Higher trait anxiety (STAI‐T) correlated positively with REM sleep proportion ( r = 0.34, p < 0.005) and negatively with light sleep ( r = −0.20, p = 0.02), while deep sleep was unrelated ( p = 0.49). Women exhibited higher STAI‐T scores ( t (126) = 3.86, p < 0.005) and greater REM proportions ( t (126) = 2.65, p = 0.01) than men. Participants aged ≥ 50 years showed lower anxiety ( t (126) = 3.79, p < 0.005) and reduced REM sleep ( t (126) = 4.27, p < 0.005) than younger adults. Multivariate analyses revealed differences between sex–age groups in anxiety measures and sleep architecture ( p < 0.005). Additionally, sleep‐related anxiety (DSEA‐S) was higher in men from the ≥ 50 group, while stable across age groups in women. Overall, these results indicate that REM sleep and trait anxiety are linked in healthy adults and that patterns of anxiety and sleep vary according to sex and age group. Findings support the ecological validity of wearables for assessing sleep architecture.
Mediano et al. (2026) studied this question.