Nightmare disorder shows reduced slow oscillation-dominant spindle coupling in NREM sleep
I. Papp, Zoran Cvetković, Nazanin Biabani, Jan Rosenzweig, Camilla Speicher, Alessio Delogu, Alexander D. Nesbitt, Panagis Drakatos, Ines R. Violante, Dijana Vilić, Clive Ballard, David O’Regan, Peter J. Goadsby, Ivana Rosenzweig
npj Biological Timing and Sleep June 22, 2026 DOI: 10.1038/s44323-026-00094-0 (opens in new tab) via OpenAlex
Summary
AI-generated from the abstractNightmare disorder involves not only REM sleep but also altered NREM sleep physiology. In a retrospective case-control study of 26 adults with nightmare disorder and 32 controls, those with nightmare disorder showed reduced frontal slow-oscillation–spindle coupling, while delta–spindle coupling remained normal, resulting in lower coupling dominance. Stage-adjusted K-complex density across the first 6 hours of sleep was also lower in the nightmare disorder group. These findings suggest that nightmare disorder involves changes in NREM sleep microarchitecture, not just REM sleep, and warrant prospective confirmation.
Study at a glance
| Characteristics | Retrospective case-control study Peer reviewed |
|---|---|
| Sample size | 58 |
| Population | Adults with nightmare disorder and controls |
| Topics | Dreaming |
| Keywords | Polysomnography Eye movement Sleep system call Sleep stages |
| Key finding | Nightmare disorder showed reduced frontal slow-oscillation–spindle coupling and lower K-complex density, indicating altered NREM sleep microarchitecture. |
Abstract
Abstract Nightmare disorder is historically conceptualised as a rapid eye movement (REM) sleep parasomnia, but evidence also points to altered non-rapid eye movement (NREM) sleep physiology. Here, in a retrospective case-control study, we analysed overnight polysomnography from 26 adults with nightmare disorder and 32 controls using a harmonised event-based EEG pipeline. Nightmare disorder showed reduced frontal slow-oscillation–spindle coupling, whereas delta–spindle coupling was preserved, yielding lower coupling dominance than in controls (0.137 ± 0.083 versus 0.238 ± 0.096; p = 6.8 × 10 −5 ). Stage-adjusted K-complex density across the first 6 h from sleep onset was also lower in nightmare disorder (0.463 ± 0.345 versus 0.723 ± 0.475 events per minute of available N2/N3 sleep; p = 0.019), whereas peak-timing differences were not robust. These findings suggest altered NREM sleep microarchitecture in nightmare disorder rather than a phenotype confined to REM sleep, and warrant prospective confirmation in harmonised, clinically phenotyped cohorts.