한국정보처리학회:학술대회논문집 (Annual Conference of KIPS)

한국정보처리학회 (Korea Information Processing Society)
권호 표지

Multivariate Masked Autoencoders for Obstructive Sleep Apnea Diagnosis

  • Thanh-Cong Do (Dept. of Artificial Intelligence Convergence, Chonnam National University) ;
  • Hyung-Jeong Yang (Dept. of Artificial Intelligence Convergence, Chonnam National University) ;
  • Hyeong-Chae Yang (Otolaryngology Department, Chonnam National University)
  • 발행 : 2024.10.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

Obstructive sleep apnea (OSA) is a common sleep disorder, causing disrupted sleep and reduced oxygen levels in the blood. Full-night polysomnography (PSG) and home sleep apnea tests (HSAT) may offer acceptable diagnosis results but have shown some limitations. In recent years, deep learning and data analysis methods are progressively employed on electronic health records, and various methods have been developed for OSA event detection. Self-supervised learning has shown some advantages over supervised training methods, by learning more generalized feature representation of data. Unlike text or image processing, the high information density in multivariate time-series data makes it more challenging to utilize self-supervised learning. In this research, based on the characteristics of sleep dataset, we propose a self-supervised approach with Masked Autoencoder, which masks portions of the input and attempting to reconstruct them. This enables the model to learn more generalized features from unlabeled data. We evaluate our proposed framework with three independent sleep datasets, which have shown significant improvement compared to supervised learning models.

키워드

과제정보

This work was supported by Institute of Information & communications Technology Planning & Evaluation (IITP) under the Artificial Intelligence Convergence Innovation Human Resources Development (IITP-2023-RS-2023-00256629) grant funded by the Korea government(MSIT) This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT). (RS-2023-00208397) This research was supported by the MSIT(Ministry of Science and ICT), Korea, under the ITRC(Information Technology Research Center) support program(IITP-2024-RS-2024-00437718) supervised by the IITP(Institute for Information & Communications Technology Planning & Evaluation)

참고문헌

  1. Wang, E., Koprinska, I. , Jeffries, B. Sleep Apnea Prediction Using Deep Learning. IEEE Journal of Biomedical and Health Informatics, vol. 27, no. 11, pp. 5644-5654, Nov. 2023, doi: 10.1109/JBHI.2023.3305980.
  2. Benjafield, A. V. et al. Estimation of the global prevalence and burden of obstructive sleep apnoea: a literature-based analysis. Lancet Respir. Med. 7, 687-698 (2019).
  3. Gottlieb, D. J. & Punjabi, N. M. Diagnosis and management of obstructive sleep apnea: a review. JAMA 323, 1389-1400 (2020).
  4. Levy, J., Alvarez, D., Del Campo, F. et al. Deep learning for obstructive sleep apnea diagnosis based on single channel oximetry. Nat Commun 14, 4881 (2023).
  5. Massie, F., Van Pee, B. & Bergmann, J. Correlations between home sleep apnea tests and polysomnography outcomes do not fully reflect the diagnostic accuracy of these tests. J. Clin. Sleep Med. 18, 871-876 (2022).
  6. Deviaene, M. et al. Automatic screening of sleep apnea patients based on the spo 2 signal. IEEE J. Biomed. Health Inform. 23, 607-617 (2018).
  7. Behar, J. A. et al. Single-channel oximetry monitor versus in-lab polysomnography oximetry analysis: does it make a difference? Physiol. Meas. 41, 044007 (2020).
  8. He, Kaiming, et al. "Masked Autoencoders Are Scalable Vision Learners." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (2022): 16000-16009.
  9. Alshaer, H.; Hummel, R.; Mendelson, M.; Marshal, T.; Bradley, T.D. Objective Relationship between Sleep Apnea and Frequency of Snoring Assessed by Machine Learning. J. Clin. Sleep Med. 2019, 15, 463-470.
  10. Singh, S.A.; Majumder, S. A novel approach osa detection using single-lead ECG scalogram based on deep neural network. J. Mech. Med. Biol. 2019, 19, 1950026.
  11. Song, C.; Liu, K.; Zhang, X.; Chen, L.; Xian, X. An obstructive sleep apnea detection approach using a discriminative hidden markov model from ECG signals. IEEE Trans. Biomed. Eng. 2015, 63, 1532-1542.
  12. Zarei, A.; Beheshti, H.; Asl, B.M. Detection of sleep apnea using deep neural networks and single-lead ECG signals. Biomed. Signal Process. Control 2022, 71, 103125.
  13. Hochreiter, Sepp, and Jurgen Schmidhuber. "Long short-term memory." Neural computation 9.8 (1997): 1735-1780.
  14. Ashish Vaswani, Noam M. Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. Attention is all you need. In Advances in Neural Information Processing Systems (NeurIPS), 2017.
  15. Thornton, A. T., Singh, P., Ruehland, W. R. & Rochford, P. D. Aasm criteria for scoring respiratory events: interaction between apnea sensor and hypopnea definition. Sleep 35, 425-432 (2012).