2026 ECNP Congress

PS02-2163

Eyes-closed episodes during eyes-open resting-state functional MRI: prevalence and temporal characteristics of unintended sleep onset in healthy participants

Elena Gerber — V. Matev, K. Al Barede, P. Stöhrmann, P. Falb, S.M. Tüchler, I. Pörnbacher, A. Sahl, J. Lenz, S. Graf, M. Klöbl, R. Lanzenberger, G. Dörl · Medical University of Vienna, Department of Psychiatry and Psychotherapy, Vienna, Austria

2026 ECNP Congress

Introduction

Resting-state functional magnetic resonance imaging (rs-fMRI) is frequently used in neuroimaging studies to investigate pathology- or drug-associated network change [1]. Fluctuations in vigilance up to unintended sleep during rs-fMRI pose a methodological challenge that remains largely unaddressed. A seminal paper showed that approximately one third of participants lost wakefulness within the first minutes of an eyes-closed rs-fMRI scan [2]. Sleep-onset episodes at the wake-sleep boundary are typically characterized by electroencephalography (EEG) changes and may be as brief as 15 seconds [3]. Furthermore, eyes typically fall shut during these episodes, which is evident in EEG signal close to the eyes. Here, we annotate episodes of prolonged eyes-closed episodes (EC), which may reflect sustained vigilance lapses. We hypothesize that EC of ≥15 seconds occur at substantial prevalence during 20-minute eyes-open resting-state scans in healthy, non-sleep-deprived participants.

Methods

Thirty-one healthy participants (mean age 25.58 ± 3.61 years, 13 females, 18 males) underwent repeated 20-minute EO resting-state scans with simultaneous EEG-fMRI acquisition using an MR-compatible EEG system (Brain Products GmbH) and a 3 T Siemens Prisma scanner. Participants were explicitly instructed to keep their eyes open throughout scanning. EC episodes were identified via visual inspection of EEG recordings, specifically evaluating the absence of ocular blinking artefacts in fronto-polar electrodes as an indicator of shut eyes. EC episodes with a duration of ≥15 seconds were included in the analysis [5]. After processing, our final sample consisted of 31 participants with 131 measurements. EC prevalence, cumulative duration, and temporal onset were quantified at measurement and subject level. Additional sensitivity analyses were conducted across different EC episode duration cut-offs of 20, 30, 45, and 60 seconds.

Results

Twenty-eight of 31 healthy participants (90.3%) showed EC episodes during a 20-minute EO resting-state scan, occurring in 99 of 131 measurements (75.6%). Across participants, EC episodes (15 seconds) accounted for a mean of 21.56% ± 19.86% of total scan time (median 14.23%, IQR 30.26%). Per individual scan, mean EC duration was 4.74min ± 5.26 min (median 2.22 min, IQR 8.76). The first EC episode occurred at a median of 2.54 minutes after scan onset, with the majority of episodes occurring within the first five minutes across all measurement time points. Sensitivity analysis showed that EC-episodes were detectable across stricter cut-offs, though prevalence naturally decreased with increasing duration thresholds: for a 60 seconds cut-off, still 54.8% of participants and 32.8% of measurements contained instances classified as EC episodes.

Conclusions

EC episodes occur for the majority of healthy participants during standard 20-minute EO resting-state scans, although with considerable variation in duration. This heterogeneity suggests that the potential impact of EC episodes on resting-state data may differ across participants. These findings extend prior evidence of sleep-related vigilance lapses in rs-fMRI [2] to an EO protocol with explicit wakefulness instructions. Our results problematize unmonitored EC episodes during common rs-fMRI scans, given the bias in Blood-Oxygen-Level-Dependent (BOLD) signal due to fluctuating vigilance [4]. Investigations on psychiatric disorders, medication, and pharmaco-fMRI studies might also suffer from the resulting bias, given they often modulate vigilance.

References

  1. Lynch CJ, Elbau IG, Ng T, Ayaz A, Zhu S, Wolk D, et al. Frontostriatal salience network expansion in individuals in depression. Nature. 2024;633:624–633.
  2. Tagliazucchi E, Laufs H. Decoding wakefulness levels from typical fMRI resting-state data reveals reliable drifts between wakefulness and sleep. Neuron. 2014;82(3):695–708.
  3. Hertig-Godeschalk A, Skorucak J, Malafeev A, Achermann P, Mathis J, Schreier DR. Microsleep episodes in the borderland between wakefulness and sleep. Sleep. 2020;43(1):zsz163. https://doi.org/10.1093/sleep/zsz163
  4. Gaez I, et al. Dynamic functional MRI markers of drowsiness during sleep onset period. iScience. 2025;28(8):113088.
  5. Berry RB, et al. The AASM Manual for the Scoring of Sleep and Associated Events. Version 2.5. American Academy of Sleep Medicine; 2018.