REM sleep infraslow rhythms correlate with neuropsychiatric symptom severity

Rapid eye movement (REM) sleep is commonly divided into phasic and tonic substates based on higher and lower oculomotor activity, respectively. Our previous study described an alternative framework in which REM temporal structure forms a continuum of infraslow rhythms (∼0.01 Hz) across brain and body physiology. We now ask whether these infraslow rhythms vary with neuropsychiatric symptoms, given the long-standing but still inconsistent links between REM and mood- or trauma-related disorders.

Using a cross-sectional polysomnographic dataset from male veterans, we examined 59 participants for correlations between REM infraslow rhythms and neuropsychiatric symptom scores, excluding participants with moderate/severe apnea.

Oscillations in REM respiratory rate, but not cardiac rate or electroencephalogram beta power, showed higher infraslow frequencies in participants with multiple symptoms, including posttraumatic stress, pain, anxiety, depression, and cognitive difficulties. No significant correlations were found between these symptoms and conventional REM metrics (e.g., REM density, latency and duration). To further examine REM beyond phasic/tonic binarization, we analyzed the timing between oculomotor activity and infraslow oscillations. Eye movements occurred preferentially during periods of high respiratory rate, low beta power and intermediate cardiac rate, with no changes in these temporal relationships across symptom severity.

Neuropsychiatric associations with REM infraslow rhythms spanned multiple symptom domains but were physiologically specific, with respiratory rate oscillations representing the strongest readout of symptom burden. Furthermore, REM infraslow cycles across brain and body were differentially aligned with eye movements: a temporal pattern that may serve as a physiological target for neuropsychiatric and mechanistic research.

Rapid eye movement (REM) sleep has long been studied in relation to neuropsychiatric conditions, but traditional measures of the timing and duration of REM episodes have shown limited consistency for symptom tracking. This study addresses this gap by highlighting a less explored physiological feature within REM episodes: minute-scale, or infraslow, rhythms in respiratory and cardiac rate, as well as oculomotor and brain activity. We report that REM infraslow oscillations, particularly in respiratory rate, occur at higher frequency in participants with worse trauma/mood symptoms and cognitive difficulties. This work is evidence of a possible new REM sleep-based biomarker for multiple diseases.