PaperarXiv:2610.36583

MM-Net: A Multimodal Multi-Task Network for Sleep Staging and Sleep Apnea Assessment After Ischemic Stroke

PreprintDated Oct 27, 2026Added Oct 29, 2026

Authors: Md Wahiduzzaman Suva, Esm E Moula Chowdhury Abha, Md Imtiaj Alam Sajin, Md Iftekharul Mobin, Jia Uddin

arXiv 2610.36583 (cs.LG), 2026-10-27

TL;DR

MM-Net is a multimodal multi-task network for joint sleep staging and respiratory-event detection in polysomnography after ischemic stroke. It fuses sleep-adapted EEG-EOG representations and engineered neural descriptors with learned cardiorespiratory features through cross-attention, and uses cross-fitting with ridge-based embedding alignment so pretrained-encoder embeddings stay consistent between training and unseen patients. On iSLEEPS (99 patients with subacute ischemic stroke), ten-fold patient-independent evaluation reached 76.3% staging accuracy (Cohen's kappa 0.656) and 0.825 respiratory-event AUC, with further results on Sleep-EDF, ISRUC-I and UCD.

Caveats

Ten-fold patient-independent; ISRUC-I 0.889 retrained; not a wearable input.

Abstract

Automated polysomnography after ischemic stroke is challenging because of altered sleep-related neural activity and concurrent respiratory disturbances. Existing stroke-specific approaches mainly address sleep-stage classification, leaving joint analysis of sleep architecture and respiratory events underexplored. Moreover, adapting pretrained EEG encoders introduces a representation inconsistency when downstream models are trained on embeddings from previously seen patients but evaluated on unseen individuals. We propose MM-Net, a multimodal multi-task framework that resolves this inconsistency through cross-fitting and ridge-based embedding alignment. Sleep-adapted EEG-EOG representations and engineered neural descriptors are fused with learned cardiorespiratory features through cross-attention and shared temporal modeling for simultaneous sleep staging and respiratory-event detection. On iSLEEPS (99 patients with subacute ischemic stroke), ten-fold patient-independent evaluation achieved 76.3% staging accuracy (Cohen's = 0.656) and 0.825 respiratory-event AUC, outperforming the evaluated baselines under matched protocols. The sleep-adapted encoder improved the downstream model when its embeddings were aligned, raising staging accuracy by 1.4 percentage points over the frozen encoder (p = 0.020) versus 0.2 points without alignment. MM-Net also achieved 86.4% staging accuracy on Sleep-EDF, above every single-EEG architecture reimplemented on the same recordings, and 80.3% staging accuracy with 0.889 respiratory AUC on ISRUC-I, exceeding its retrained staging baselines. Without retraining, the iSLEEPS-trained model retained respiratory AUCs of 0.833 on ISRUC-I and 0.841 on UCD. These findings support representation-consistent neural adaptation and joint physiological modeling across stroke, healthy-sleep, and sleep-disorder cohorts.

Verbatim from arXiv (opens in a new tab) · arXiv metadata (abstracts are released under CC0)

Results on ChatPPG leaderboards · 1

All leaderboards

Leaderboard	Method	Result	Rank
iSLEEPS event AUROCSleep apnea (OSA) screening	MM-NetCaveat: Ten-fold patient-independent; ISRUC-I 0.889 retrained; not a wearable input.	0.825event AUROC, higher is better	Unranked

1 row is credited to this paper. Ranks apply within each leaderboard only, as reported by the source.

Tasks & signals

Tasks

Sleep apnea (OSA) screeningSleep stagingSleep apnea

Signals

ECG SpO₂ Resp EEG EOG

Keywords

Sleep StagingSleep ApneaStrokePSGMultimodalMulti TaskCross AttentionEEG Encoder Adaptation

Datasets used

Sleep-EDF Database Expanded (PhysioNet)Resp · Temp · EEG

How it works

Cite this paper

Cite this paper with the fields recorded here. Check it against the publisher before submitting.

Suva, M. W., Abha, E. E. M. C., Sajin, M. I. A., Mobin, M. I., & Uddin, J. (2026). MM-Net: A Multimodal Multi-Task Network for Sleep Staging and Sleep Apnea Assessment After Ischemic Stroke. arXiv 2610.36583. https://arxiv.org/abs/2610.36583

Suva, Md Wahiduzzaman, Esm E Moula Chowdhury Abha, Md Imtiaj Alam Sajin, Md Iftekharul Mobin, and Jia Uddin. 2026. "MM-Net: A Multimodal Multi-Task Network for Sleep Staging and Sleep Apnea Assessment After Ischemic Stroke." arXiv 2610.36583. https://arxiv.org/abs/2610.36583.

Suva, M. W., Abha, E. E. M. C., Sajin, M. I. A., Mobin, M. I. and Uddin, J. (2026) 'MM-Net: A Multimodal Multi-Task Network for Sleep Staging and Sleep Apnea Assessment After Ischemic Stroke', arXiv 2610.36583. Available at: https://arxiv.org/abs/2610.36583

Suva, Md Wahiduzzaman, et al. "MM-Net: A Multimodal Multi-Task Network for Sleep Staging and Sleep Apnea Assessment After Ischemic Stroke." arXiv 2610.36583, 2026. https://arxiv.org/abs/2610.36583.

M. W. Suva, E. E. M. C. Abha, M. I. A. Sajin, M. I. Mobin, and J. Uddin, "MM-Net: A Multimodal Multi-Task Network for Sleep Staging and Sleep Apnea Assessment After Ischemic Stroke," arXiv 2610.36583, 2026, [Online]. Available: https://arxiv.org/abs/2610.36583.

@misc{suva2026mmnet,
  title         = {{MM-Net: A Multimodal Multi-Task Network for Sleep Staging and Sleep Apnea Assessment After Ischemic Stroke}},
  author        = {Suva, Md Wahiduzzaman and Abha, Esm E. Moula Chowdhury and Sajin, Md Imtiaj Alam and Mobin, Md Iftekharul and Uddin, Jia},
  year          = {2026},
  howpublished  = {arXiv 2610.36583},
  eprint        = {2610.36583},
  archivePrefix = {arXiv}
}

Oct 29, 2026Added to the ChatPPG research hub