Related Experiment Video
Updated: Jul 10, 2026
Human Circadian Phenotyping and Diurnal Performance Testing in the Real World
Published on: April 7, 2020
Sleep-timing mismatch (MSFsc-lights-off interval) and its association with polysomnographic sleep parameters in a
Kultida Wongprakarnsanti1, Jarturong Siritienthong1, Tantawan Awirutworakul2
1Department of Psychiatry, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok, Thailand.
Objective:
To determine whether the difference between individual sleep midpoint (MSFsc) and a fixed laboratory bedtime predicts polysomnographic (PSG) sleep parameters under clinical conditions enriched for obstructive sleep apnea (OSA) evaluation.
Methods:
Overnight PSG data from 645 clinically referred adults were analyzed under a standardized schedule (lights-off 21:30; lights-on 06:00). Each participant's sleep midpoint was indexed by mid-sleep on free days corrected for sleep debt (MSFsc). The sleep-timing mismatch (Δt) was defined as the difference between MSFsc and each participant's individually recorded lights-off time. Associations between Δt and PSG parameters were examined using correlation analyses, MSFsc terciles, and multivariable regression models adjusted for age, sex, BMI, AHI, and insomnia severity, with Δt as the primary exposure. Social jetlag, MSW, and ΔTIB were examined in sensitivity analyses. Categorical MSFsc (terciles) and interaction terms were evaluated in secondary analyses.
Results:
Greater Δt was independently associated with longer sleep onset latency, lower sleep efficiency, higher N1 sleep, and lower REM percentage. These associations were unchanged in sensitivity analyses additionally adjusting for social jetlag, MSW, and ΔTIB. Categorical MSFsc by tercile showed a graded pattern, with later MSFsc associated with poorer sleep consolidation and lower REM percentage. Categorical MSFsc alone did not independently predict PSG parameters once Δt was included. In interaction analyses, Δt × late MSFsc tercile (T3 vs T1) interactions were significant across SOL, SE, N1%, and REM% (all p ≤ 0.005), indicating that the Δt-sleep architecture association was most pronounced in patients with later MSFsc (later sleep midsleep).
Conclusions:
The sleep-timing mismatch (MSFsc-lights-off interval) relative to a fixed laboratory bedtime is a robust predictor of objective sleep initiation, sleep continuity, and REM architecture, exceeding the explanatory value of categorical MSFsc alone. Incorporating Δt may strengthen PSG interpretation in clinical samples with high OSA prevalence.
