Temporal changes in the discrepancy between subjective and objective sleep: an in-home electroencephalography study

Sleep misperception has traditionally been attributed to neurophysiological abnormalities or psychological traits; however, most evidence is based on single-night or laboratory-based assessments. This raises the possibility that misperception may partly reflect measurement limitations rather than stable individual pathology. This study aimed to examine whether sleep misperception changes across consecutive nights and to identify variability-based phenotypes using repeated in-home electroencephalography (EEG).

Eighty-three adults (40-64 years) underwent 7 consecutive nights of in-home EEG monitoring and completed daily sleep questionnaires (OSA-MA). Sleep misperception was quantified using the Sleep Perception Index (SPI), and participants were classified as underestimators (n = 15), concordant estimators (n = 56), or overestimators (n = 12). Night-to-night variability in sleep parameters was assessed using coefficients of variation (CV).

Sleep misperception decreased over time, particularly among overestimators (β = -4.05, p < 0.001). Underestimators also showed an early correction by Night 2, although this change did not reach statistical significance. Compared with concordant estimators, overestimators exhibited greater variability in total sleep time (CV: 0.18 vs. 0.12, p = 0.033), whereas underestimators showed greater variability in sleep efficiency (CV: 0.11 vs. 0.06, p = 0.024) and N3 delta power (CV: 0.37 vs. 0.18, p = 0.003). Daily diary-based assessments (OSA-MA) demonstrated superior performance in identifying sleep-related discrepancies compared to retrospective measures.

These findings suggest that sleep misperception is not a fixed trait but a dynamic phenomenon that evolves with repeated measurement. Distinct variability patterns across sleep domains further indicate that misperception reflects instability in sleep quantity and quality rather than a single underlying mechanism. Multi-night, ethologically valid monitoring may therefore provide a more accurate framework for understanding and assessing sleep perception.