{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/dosed-a-deep-learning-approach-to-detect","title":"DOSED: a deep learning approach to detect multiple sleep micro-events in EEG signal","arxiv_id":"1812.04079","date":"2018-12-07","proceeding":null,"authors":["Stanislas Chambon","Valentin Thorey","Pierrick J. Arnal","Emmanuel Mignot","Alexandre Gramfort"],"abstract":"Background: Electroencephalography (EEG) monitors brain activity during sleep\nand is used to identify sleep disorders. In sleep medicine, clinicians\ninterpret raw EEG signals in so-called sleep stages, which are assigned by\nexperts to every 30s window of signal. For diagnosis, they also rely on shorter\nprototypical micro-architecture events which exhibit variable durations and\nshapes, such as spindles, K-complexes or arousals. Annotating such events is\ntraditionally performed by a trained sleep expert, making the process time\nconsuming, tedious and subject to inter-scorer variability. To automate this\nprocedure, various methods have been developed, yet these are event-specific\nand rely on the extraction of hand-crafted features.\n  New method: We propose a novel deep learning architecure called Dreem One\nShot Event Detector (DOSED). DOSED jointly predicts locations, durations and\ntypes of events in EEG time series. The proposed approach, applied here on\nsleep related micro-architecture events, is inspired by object detectors\ndeveloped for computer vision such as YOLO and SSD. It relies on a\nconvolutional neural network that builds a feature representation from raw EEG\nsignals, as well as two modules performing localization and classification\nrespectively.\n  Results and comparison with other methods: The proposed approach is tested on\n4 datasets and 3 types of events (spindles, K-complexes, arousals) and compared\nto the current state-of-the-art detection algorithms.\n  Conclusions: Results demonstrate the versatility of this new approach and\nimproved performance compared to the current state-of-the-art detection\nmethods.","url_abs":"http://arxiv.org/abs/1812.04079v1","url_pdf":"http://arxiv.org/pdf/1812.04079v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"dosed-a-deep-learning-approach-to-detect","repo_url":"https://github.com/Dreem-Organization/dosed","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"eeg-1","task_name":"EEG"},{"task_slug":"eeg","task_name":"Electroencephalogram (EEG)"},{"task_slug":"k-complex-detection","task_name":"K-complex detection"},{"task_slug":"sleep-arousal-detection","task_name":"Sleep Arousal Detection"},{"task_slug":"sleep-micro-event-detection","task_name":"Sleep Micro-event detection"},{"task_slug":"sleep-quality","task_name":"Sleep Quality"},{"task_slug":"sleep-apnea-detection","task_name":"Sleep apnea detection"},{"task_slug":"spindle-detection","task_name":"Spindle Detection"},{"task_slug":"time-series","task_name":"Time Series Analysis"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"non-maximum-suppression","method_name":"Non Maximum Suppression"},{"method_slug":"ssd","method_name":"SSD"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/k-complex-detection-on-mass-ss2","task":"K-complex detection","dataset":"MASS SS2","model":"DOSED","rank_in_archive_order":3,"of":4,"metrics":{"F1-score (@IoU = 0.3)":"0.6"},"uses_additional_data":false},{"leaderboard":"/sota/sleep-arousal-detection-on-mesa","task":"Sleep Arousal Detection","dataset":"MESA","model":"DOSED (3 EEG + 2 EOG)","rank_in_archive_order":1,"of":3,"metrics":{"F1-score (@IoU = 0.3)":"0.71"},"uses_additional_data":false},{"leaderboard":"/sota/sleep-arousal-detection-on-mesa","task":"Sleep Arousal Detection","dataset":"MESA","model":"DOSED (1 EEG)","rank_in_archive_order":2,"of":3,"metrics":{"F1-score (@IoU = 0.3)":"0.61"},"uses_additional_data":false},{"leaderboard":"/sota/sleep-apnea-detection-on-dreem_nct03657329","task":"Sleep apnea detection","dataset":"Dreem_NCT03657329","model":"DOSED","rank_in_archive_order":1,"of":1,"metrics":{"Accuracy":"81%","F1-score (@IoU = 0.3)":"0.57 ± 0.23","Mean AHI Error":"4.69 ± 4.25"},"uses_additional_data":false},{"leaderboard":"/sota/spindle-detection-on-mass-ss2","task":"Spindle Detection","dataset":"MASS SS2","model":"DOSED","rank_in_archive_order":3,"of":6,"metrics":{"F1-score (@IoU = 0.3)":"0.75"},"uses_additional_data":false},{"leaderboard":"/sota/spindle-detection-on-stanford-sleep-cohort","task":"Spindle Detection","dataset":"Stanford Sleep Cohort (SSC)","model":"DOSED","rank_in_archive_order":1,"of":1,"metrics":{"F1-score (@IoU = 0.3)":"0.48"},"uses_additional_data":false},{"leaderboard":"/sota/spindle-detection-on-wisconsin-sleep-cohort","task":"Spindle Detection","dataset":"Wisconsin Sleep Cohort (WSC)","model":"DOSED","rank_in_archive_order":1,"of":1,"metrics":{"F1-score (@IoU = 0.3)":"0.46"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}