USSD: Unsupervised Sleep Spindle Detector

作     者：Ramirez, Edgardo Estevez, Pablo A. Adams, Martin D. Perez, Claudio A. Garrido Gonzalez, Marcelo Peirano, Patricio 

作者机构：Univ Chile Dept Elect Engn Santiago 8330015 Chile Millennium Inst Intelligent Healthcare Engn Santiago 1025000 Chile Ctr Intervent Med Precis & Adv Cellular Therapy IMPACT Santiago 1025000 Chile Univ Chile Inst Nutr & Food Technol Sleep Lab Santiago 8330015 Chile Univ Diego Portales UDP Fac Educ Santiago 8370109 Chile 

出 版 物：《IEEE ACCESS》 (IEEE Access)

年 卷 期：2025年第13卷

页      面：18644-18659页

核心收录：

基　　金：National Agency for Research and Development (ANID)/Scholarship Program/DOCTORADO/NACIONAL [2018-21181277] ANID-Chile through Millennium Science Initiative Program [ICN2021-004] Basal Funding for Scientific and Technological Center of Excellence [IMPACT FB210024] FONDECYT Agencia Nacional de Investigacion y Desarrollo (ANID)-National Research Agency, Chile/Programa de Investigacion Asociativa (PIA) [AFB230001] ANID Fondo Nacional de Desarrollo Cientifico y Tecnologico (FONDECYT) 

主　　题：Sleep Electroencephalography Signal processing algorithms Brain modeling Detectors Dictionaries Matching pursuit algorithms Machine learning Recording Filtering algorithms EEG sleep spindle unsupervised learning dictionary learning 

摘      要：Sleep spindles (SSs) appear in electroencephalogram (EEG) recordings during sleep stage N2, and they are usually detected through visual inspection by an expert. Labeling SSs in large datasets is time-consuming and depends on the expert criteria. In this work, we propose an unsupervised SS detector based on dictionary learning called the Unsupervised Sleep Spindle Detector (USSD). The proposed detector learns prototype SSs of different lengths (called atoms). An unsupervised adaptive threshold method based on the distribution of the automatically detected SS lengths is developed, which allows the adaptation of the USSD algorithm to different datasets in an unsupervised way. For each detection, the USSD estimates the probability of being an SS. The USSD performances on the labeled MASS-SS2 and INTA-UCH datasets yield F1-scores of $0.72 \pm 0.02$ and $0.72 \pm 0.04$ , respectively. The USSD outperforms the A7 and LUNA detectors, which are traditional unsupervised models. Next, we fine-tune the resulting USSD model with 20% of the labeled MASS-SS2 and INTA-UCH datasets, achieving F1 scores of $0.78 \pm 0.06$ and $0.75 \pm 0.05$ , respectively. In addition, the SSs detected by USSD on the unlabeled CAP dataset are used to pre-train a supervised deep learning method, which after fine-tuning with 20% of the MODA dataset, reaches an F1-score of $0.81 \pm 0.02$ .