Multi-modal sleep staging in the clinic for REM sleep behaviour disorder

Accurate REM identification is critical for diagnosing REM sleep behaviour disorder (RBD), yet many automated sleep-staging systems, especially single-channel EEG models trained on healthy cohorts, do not generalise well to real-life polysomnography (PSG) performed in patients. We compared a feature-based Random Forest (RF) model tuned for RBD with a state-of-the-art single-EEG deep architecture (AttnSleep), and assessed the impact of cohort adaptation and multimodal inputs (EEG, EOG, EMG, ECG). Experiments used 89 multi-site in-clinic PSGs (SleepWearables Phase-1) plus 53 MASS healthy controls ((mean age 63 ± 5 years), with 10-fold cross-validation and out-of-fold evaluation. When applied out-of-the-box after training on open-source healthy datasets, both models achieved moderate agreement overall (Cohen’s κ = 0.46), but performance declined in RBD, particularly for REM sleep (AttnSleep Cohen’s κ = 0.19 vs RF Cohen’s κ = 0.44), highlighting limited cross-cohort generalisation. The multi-modal model improved overall agreement (Cohen’s κ 0.59 versus 0.60) and performance in RBD (Cohen’s κ 0.45 versus 0.46), with gains most pronounced in REM (Cohen’s κ 0.45 versus 0.49). Attention-based modality analysis identified EEG as the dominant signal, increased EOG contribution during REM, and elevated ECG importance during N3. In RBD subjects, EOG weighting increased relative to non-RBD controls (Δ = +0.081), suggesting physiological relevance. Guided by these weights, a reduced four-channel EEG model matched full multimodal performance in non-RBD subjects, and adding EOG achieved the best overall configuration (Cohen’s κ = 0.61 overall; Cohen’s κ = 0.48 in RBD) with improved REM classification (53% compared to 45% recall). Inclusion of EOG also reduced inter-dataset variability in REM staging. Nonetheless, staging performance in RBD remained lower than in controls, particularly for REM. These results highlight (1) the limited generalisability of minimal-sensor models trained on healthy cohorts, (2) the value of mixed cohort-specific training, and (3) the benefit of multimodal integration and attention-guided channel selection, rather than minimal-sensor approaches alone for robust clinical sleep staging in pathological populations such as RBD.