Rapid eye movement (REM) sleep is commonly divided into phasic and tonic substates by oculomotor activity. The authors used an alternative framing in which REM exhibits a temporal continuum of infraslow rhythms (~0.01 Hz) across brain and body physiology. The primary objective was to test whether the characteristics of REM infraslow oscillations vary with neuropsychiatric symptom severity, given prior but inconsistent links between REM features and mood- or trauma-related disorders.
The analysis used a cross-sectional polysomnographic dataset collected from male veterans. The final analytic sample comprised 59 participants after excluding individuals with moderate or severe obstructive sleep apnea. The study correlated REM infraslow rhythm features with neuropsychiatric symptom scores spanning posttraumatic stress, pain, anxiety, depression, and cognitive difficulties. The report states that deidentified processed data will be made available upon reasonable request.
The investigators examined infraslow oscillations across multiple physiological measures recorded during REM sleep: respiratory rate, cardiac rate, and electroencephalogram (EEG) beta power. They contrasted these infraslow features with conventional REM metrics, including REM density, REM latency, and REM duration. In addition to frequency analyses, the team evaluated the temporal alignment between oculomotor (eye) activity and the different infraslow cycles.
Oscillations in REM respiratory rate—but not in cardiac rate or EEG beta power—showed higher infraslow frequencies in participants who presented with multiple neuropsychiatric symptoms. The symptom domains that tracked with higher respiratory infraslow frequency included posttraumatic stress, pain, anxiety, depression, and reported cognitive difficulties. The authors emphasize that this association was physiologically specific: respiratory rate oscillations provided the strongest readout of symptom burden compared with the other body and brain signals analyzed.
No significant correlations were observed between neuropsychiatric symptom severity and standard REM measures in this sample. Specifically, REM density, REM latency, and REM duration did not show meaningful relationships with the symptom scores examined. This contrast supports the idea that examining REM temporal structure at infraslow timescales can reveal associations not captured by traditional REM summary metrics.
To move beyond a binary phasic/tonic REM classification, the authors analyzed the timing between oculomotor events and the infraslow oscillations across physiological signals. Eye movements tended to occur preferentially during periods characterized by high respiratory rate, low EEG beta power, and intermediate cardiac rate. Importantly, these temporal relationships—the alignment of eye movements with particular phases of the infraslow cycles—did not change across levels of symptom severity in the cohort.
The study proposes that REM sleep’s infraslow temporal structure constitutes a physiologically informative continuum that links sleep physiology with multi-domain neuropsychiatric symptom burden. The strongest and most consistent association in this dataset was between symptom severity and infraslow respiratory rate oscillations during REM. Because conventional REM metrics did not track symptoms, the authors suggest that infraslow REM features could serve as novel physiological targets for mechanistic studies and for investigation in clinical populations with mood, trauma, pain, anxiety, or cognitive symptoms.
These findings are reported from a preprint and have not undergone peer review. The sample consisted of male veterans and the authors excluded moderate/severe sleep apnea, which may limit generalizability. The report does not provide additional mechanistic or causal inferences beyond the observed correlations.
The Institutional Review Boards of the VA Portland Health Care System and the University of Michigan approved the research. The authors declared no competing interests. Funding sources listed in the report include the Pritzker Neuropsychiatric Disorders Research Consortium, University of Michigan Neuroscience Scholars Fund, Hope for Depression Research Foundation, VA Merit Awards, and the National Institute on Aging. Deidentified, processed data will be made available upon reasonable request.