Abstract Introduction During laboratory-based, voluntary exposure to total sleep deprivation (TSD), relatively large decreases in positive mood and relatively small increases in negative mood have been observed, with the magnitudes of change varying across individuals. People differ in their trait emotion reactivity, i.e., their sensitivity to emotion-evoking stimuli or events. As higher reactivity is associated with greater difficulty regulating emotion, we hypothesized this trait could moderate mood changes during TSD. Methods N=96 healthy adults (ages 21-38, 47% female) participated in one of three 4-day/3-night in-laboratory TSD studies. In each study, after 10h baseline sleep, participants were exposed to 38h TSD, followed by 10h recovery sleep. Participants completed the Emotion Reactivity Scale (ERS), and their self-reported affect was assessed every 2-4h during wake using the Positive and Negative Affect Schedule (PANAS). The 11 test bouts during TSD shared across studies were included in analyses (day 2: 09:00, 13:00, 21:00, 23:00; day 3: 01:00, 03:00, 05:00, 07:00, 09:00, 13:00, 21:00). Positive and negative affect ratings were analyzed with linear mixed-effects regression with fixed effects of test bout (categorical), ERS score (continuous), and their interaction, covariates for study, sex, and age, and a random intercept over participants. Results Participants showed the expected decrease in positive affect as a function of test bout (time awake and time of day, p< 0.001), and a smaller, non-significant increase in negative affect (p=0.17). Higher ERS scores were associated with increased negative affect during TSD (p=0.002), but this effect was small, and negative affect showed a floor effect. Emotion reactivity was not a significant predictor of decreased positive affect during TSD (p=0.97). Conclusion As hypothesized, emotion reactivity predicted individual variability in negative affect during TSD, but unexpectedly it did not predict the decrease in positive affect associated with TSD. This suggests that the trait measured by the ERS may be biased toward negative affect, and/or that positive affective changes may be more difficult to predict during the low arousal state induced by TSD. Studies inducing greater variability in negative affect during TSD (e.g., through exposure to a stressor) are needed to confirm our findings. Support (if any) NIH CA167691, ONR N00014-13-1-0302, and CDMRP W81XWH-16-1-0319 and W81XWH-20-1-0442.