Cargando…

Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea

Scoring polysomnography for obstructive sleep apnea diagnosis is a laborious, long, and costly process. Machine learning approaches, such as deep neural networks, can reduce scoring time and costs. However, most methods require prior filtering and preprocessing of the raw signal. Our work presents a...

Descripción completa

Detalles Bibliográficos
Autores principales: Almarshad, Malak Abdullah, Al-Ahmadi, Saad, Islam, Md Saiful, BaHammam, Ahmed S., Soudani, Adel
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10536445/
https://www.ncbi.nlm.nih.gov/pubmed/37765980
http://dx.doi.org/10.3390/s23187924
_version_ 1785112867332161536
author Almarshad, Malak Abdullah
Al-Ahmadi, Saad
Islam, Md Saiful
BaHammam, Ahmed S.
Soudani, Adel
author_facet Almarshad, Malak Abdullah
Al-Ahmadi, Saad
Islam, Md Saiful
BaHammam, Ahmed S.
Soudani, Adel
author_sort Almarshad, Malak Abdullah
collection PubMed
description Scoring polysomnography for obstructive sleep apnea diagnosis is a laborious, long, and costly process. Machine learning approaches, such as deep neural networks, can reduce scoring time and costs. However, most methods require prior filtering and preprocessing of the raw signal. Our work presents a novel method for diagnosing obstructive sleep apnea using a transformer neural network with learnable positional encoding, which outperforms existing state-of-the-art solutions. This approach has the potential to improve the diagnostic performance of oximetry for obstructive sleep apnea and reduce the time and costs associated with traditional polysomnography. Contrary to existing approaches, our approach performs annotations at one-second granularity. Allowing physicians to interpret the model’s outcome. In addition, we tested different positional encoding designs as the first layer of the model, and the best results were achieved using a learnable positional encoding based on an autoencoder with structural novelty. In addition, we tried different temporal resolutions with various granularity levels from 1 to 360 s. All experiments were carried out on an independent test set from the public OSASUD dataset and showed that our approach outperforms current state-of-the-art solutions with a satisfactory AUC of 0.89, accuracy of 0.80, and F1-score of 0.79.
format Online
Article
Text
id pubmed-10536445
institution National Center for Biotechnology Information
language English
publishDate 2023
publisher MDPI
record_format MEDLINE/PubMed
spelling pubmed-105364452023-09-29 Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea Almarshad, Malak Abdullah Al-Ahmadi, Saad Islam, Md Saiful BaHammam, Ahmed S. Soudani, Adel Sensors (Basel) Article Scoring polysomnography for obstructive sleep apnea diagnosis is a laborious, long, and costly process. Machine learning approaches, such as deep neural networks, can reduce scoring time and costs. However, most methods require prior filtering and preprocessing of the raw signal. Our work presents a novel method for diagnosing obstructive sleep apnea using a transformer neural network with learnable positional encoding, which outperforms existing state-of-the-art solutions. This approach has the potential to improve the diagnostic performance of oximetry for obstructive sleep apnea and reduce the time and costs associated with traditional polysomnography. Contrary to existing approaches, our approach performs annotations at one-second granularity. Allowing physicians to interpret the model’s outcome. In addition, we tested different positional encoding designs as the first layer of the model, and the best results were achieved using a learnable positional encoding based on an autoencoder with structural novelty. In addition, we tried different temporal resolutions with various granularity levels from 1 to 360 s. All experiments were carried out on an independent test set from the public OSASUD dataset and showed that our approach outperforms current state-of-the-art solutions with a satisfactory AUC of 0.89, accuracy of 0.80, and F1-score of 0.79. MDPI 2023-09-15 /pmc/articles/PMC10536445/ /pubmed/37765980 http://dx.doi.org/10.3390/s23187924 Text en © 2023 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
spellingShingle Article
Almarshad, Malak Abdullah
Al-Ahmadi, Saad
Islam, Md Saiful
BaHammam, Ahmed S.
Soudani, Adel
Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea
title Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea
title_full Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea
title_fullStr Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea
title_full_unstemmed Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea
title_short Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea
title_sort adoption of transformer neural network to improve the diagnostic performance of oximetry for obstructive sleep apnea
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10536445/
https://www.ncbi.nlm.nih.gov/pubmed/37765980
http://dx.doi.org/10.3390/s23187924
work_keys_str_mv AT almarshadmalakabdullah adoptionoftransformerneuralnetworktoimprovethediagnosticperformanceofoximetryforobstructivesleepapnea
AT alahmadisaad adoptionoftransformerneuralnetworktoimprovethediagnosticperformanceofoximetryforobstructivesleepapnea
AT islammdsaiful adoptionoftransformerneuralnetworktoimprovethediagnosticperformanceofoximetryforobstructivesleepapnea
AT bahammamahmeds adoptionoftransformerneuralnetworktoimprovethediagnosticperformanceofoximetryforobstructivesleepapnea
AT soudaniadel adoptionoftransformerneuralnetworktoimprovethediagnosticperformanceofoximetryforobstructivesleepapnea