Explainable vision transformer for automatic visual sleep staging on multimodal PSG signals
Abstract Polysomnography (PSG) is crucial for diagnosing sleep disorders, but manual scoring of PSG is time-consuming and subjective, leading to high variability. While machine-learning models have improved PSG scoring, their clinical use is hindered by the ‘black-box’ nature. In this study, we pres...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | npj Digital Medicine |
| Online Access: | https://doi.org/10.1038/s41746-024-01378-0 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832585357845069824 |
|---|---|
| author | Hyojin Lee You Rim Choi Hyun Kyung Lee Jaemin Jeong Joopyo Hong Hyun-Woo Shin Hyung-Sin Kim |
| author_facet | Hyojin Lee You Rim Choi Hyun Kyung Lee Jaemin Jeong Joopyo Hong Hyun-Woo Shin Hyung-Sin Kim |
| author_sort | Hyojin Lee |
| collection | DOAJ |
| description | Abstract Polysomnography (PSG) is crucial for diagnosing sleep disorders, but manual scoring of PSG is time-consuming and subjective, leading to high variability. While machine-learning models have improved PSG scoring, their clinical use is hindered by the ‘black-box’ nature. In this study, we present SleepXViT, an automatic sleep staging system using Vision Transformer (ViT) that provides intuitive, consistent explanations by mimicking human ‘visual scoring’. Tested on KISS–a PSG image dataset from 7745 patients across four hospitals–SleepXViT achieved a Macro F1 score of 81.94%, outperforming baseline models and showing robust performances on public datasets SHHS1 and SHHS2. Furthermore, SleepXViT offers well-calibrated confidence scores, enabling expert review for low-confidence predictions, alongside high-resolution heatmaps highlighting essential features and relevance scores for adjacent epochs’ influence on sleep stage predictions. Together, these explanations reinforce the scoring consistency of SleepXViT, making it both reliable and interpretable, thereby facilitating the synergy between the AI model and human scorers in clinical settings. |
| format | Article |
| id | doaj-art-6a05e602155b401193aefabe86d3df3a |
| institution | Kabale University |
| issn | 2398-6352 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | npj Digital Medicine |
| spelling | doaj-art-6a05e602155b401193aefabe86d3df3a2025-01-26T12:53:48ZengNature Portfolionpj Digital Medicine2398-63522025-01-018111410.1038/s41746-024-01378-0Explainable vision transformer for automatic visual sleep staging on multimodal PSG signalsHyojin Lee0You Rim Choi1Hyun Kyung Lee2Jaemin Jeong3Joopyo Hong4Hyun-Woo Shin5Hyung-Sin Kim6Graduate School of Data Science, Seoul National UniversityGraduate School of Data Science, Seoul National UniversityObstructive Upper Airway Research (OUaR) Laboratory, Department of Pharmacology, Seoul National University College of MedicineDepartment of Computer Engineering, School of Software, Hallym UniversityGraduate School of Data Science, Seoul National UniversityObstructive Upper Airway Research (OUaR) Laboratory, Department of Pharmacology, Seoul National University College of MedicineGraduate School of Data Science, Seoul National UniversityAbstract Polysomnography (PSG) is crucial for diagnosing sleep disorders, but manual scoring of PSG is time-consuming and subjective, leading to high variability. While machine-learning models have improved PSG scoring, their clinical use is hindered by the ‘black-box’ nature. In this study, we present SleepXViT, an automatic sleep staging system using Vision Transformer (ViT) that provides intuitive, consistent explanations by mimicking human ‘visual scoring’. Tested on KISS–a PSG image dataset from 7745 patients across four hospitals–SleepXViT achieved a Macro F1 score of 81.94%, outperforming baseline models and showing robust performances on public datasets SHHS1 and SHHS2. Furthermore, SleepXViT offers well-calibrated confidence scores, enabling expert review for low-confidence predictions, alongside high-resolution heatmaps highlighting essential features and relevance scores for adjacent epochs’ influence on sleep stage predictions. Together, these explanations reinforce the scoring consistency of SleepXViT, making it both reliable and interpretable, thereby facilitating the synergy between the AI model and human scorers in clinical settings.https://doi.org/10.1038/s41746-024-01378-0 |
| spellingShingle | Hyojin Lee You Rim Choi Hyun Kyung Lee Jaemin Jeong Joopyo Hong Hyun-Woo Shin Hyung-Sin Kim Explainable vision transformer for automatic visual sleep staging on multimodal PSG signals npj Digital Medicine |
| title | Explainable vision transformer for automatic visual sleep staging on multimodal PSG signals |
| title_full | Explainable vision transformer for automatic visual sleep staging on multimodal PSG signals |
| title_fullStr | Explainable vision transformer for automatic visual sleep staging on multimodal PSG signals |
| title_full_unstemmed | Explainable vision transformer for automatic visual sleep staging on multimodal PSG signals |
| title_short | Explainable vision transformer for automatic visual sleep staging on multimodal PSG signals |
| title_sort | explainable vision transformer for automatic visual sleep staging on multimodal psg signals |
| url | https://doi.org/10.1038/s41746-024-01378-0 |
| work_keys_str_mv | AT hyojinlee explainablevisiontransformerforautomaticvisualsleepstagingonmultimodalpsgsignals AT yourimchoi explainablevisiontransformerforautomaticvisualsleepstagingonmultimodalpsgsignals AT hyunkyunglee explainablevisiontransformerforautomaticvisualsleepstagingonmultimodalpsgsignals AT jaeminjeong explainablevisiontransformerforautomaticvisualsleepstagingonmultimodalpsgsignals AT joopyohong explainablevisiontransformerforautomaticvisualsleepstagingonmultimodalpsgsignals AT hyunwooshin explainablevisiontransformerforautomaticvisualsleepstagingonmultimodalpsgsignals AT hyungsinkim explainablevisiontransformerforautomaticvisualsleepstagingonmultimodalpsgsignals |