PaperarXiv:2502.17486

Multimodal Sleep Stage and Sleep Apnea Classification Using Vision Transformer: A Multitask Explainable Learning Approach

Preprint2025Added Sep 26, 2026

Authors: Kianoosh Kazemi, Iman Azimi, Michelle Khine, Rami N. Khayat, Amir M. Rahmani, Pasi Liljeberg (6 authors)

arXiv preprint (eess.SP, cs.LG)

Kazemi, Kianoosh et al., "Multimodal Sleep Stage and Sleep Apnea Classification Using Vision Transformer: A Multitask Explainable Learning Approach", arXiv:2502.17486, CC BY 4.0

TL;DR

1D-Vision Transformer multitask model jointly classifying sleep stages and sleep apnea from multimodal PPG + respiratory flow + respiratory effort. Reported overall 5-stage accuracy 78% (κ 0.66) and apnea classification accuracy 74% (κ 0.58); attention-weight analysis highlights respiratory troughs/peaks. Included because it jointly does staging (not apnea-only) with PPG as a primary multimodal channel.

Form factor: other

Abstract

Sleep is an essential component of human physiology, contributing significantly to overall health and quality of life. Accurate sleep staging and disorder detection are crucial for assessing sleep quality. Studies in the literature have proposed PSG-based approaches and machine-learning methods utilizing single-modality signals. However, existing methods often lack multimodal, multilabel frameworks and address sleep stages and disorders classification separately. In this paper, we propose a 1D-Vision Transformer for simultaneous classification of sleep stages and sleep disorders. Our method exploits the sleep disorders' correlation with specific sleep stage patterns and performs a simultaneous identification of a sleep stage and sleep disorder. The model is trained and tested using multimodal-multilabel sensory data (including photoplethysmogram, respiratory flow, and respiratory effort signals). The proposed method shows an overall accuracy (cohen's Kappa) of 78% (0.66) for five-stage sleep classification and 74% (0.58) for sleep apnea classification. Moreover, we analyzed the encoder attention weights to clarify our models' predictions and investigate the influence different features have on the models' outputs. The result shows that identified patterns, such as respiratory troughs and peaks, make a higher contribution to the final classification process.

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

Results on ChatPPG leaderboards

All leaderboards

No results from this paper on ChatPPG leaderboards yet. Browse the leaderboards.

Tasks & signals

Tasks

Sleep stagingSleep apnea

Signals

PPG Resp

Keywords

sleepstagingapneamultitaskViTexplainable

How it works

Cite this paper

APA

Kazemi, K., Azimi, I., Khine, M., Khayat, R. N., Rahmani, A. M., & Liljeberg, P. (2025). Multimodal Sleep Stage and Sleep Apnea Classification Using Vision Transformer: A Multitask Explainable Learning Approach. arXiv preprint. https://doi.org/10.48550/arXiv.2502.17486

BibTeX

@misc{kazemi2025multimodal,
  title         = {{Multimodal Sleep Stage and Sleep Apnea Classification Using Vision Transformer: A Multitask Explainable Learning Approach}},
  author        = {Kazemi, Kianoosh and Azimi, Iman and Khine, Michelle and Khayat, Rami N. and Rahmani, Amir M. and Liljeberg, Pasi},
  year          = {2025},
  howpublished  = {arXiv preprint},
  doi           = {10.48550/arXiv.2502.17486},
  eprint        = {2502.17486},
  archivePrefix = {arXiv}
}

Built from the hub record. Check it against the publisher before submitting.

Sep 26, 2026Added to the ChatPPG research hub

Multimodal Sleep Stage and Sleep Apnea Classification Using Vision Transformer: A Multitask Explainable Learning Approach

TL;DR

Abstract

Results on ChatPPG leaderboards