Context-Aware Temporal Modeling for Single-Channel EEG Sleep Staging
Analysis
This paper addresses the critical problem of automatic sleep staging using single-channel EEG, a practical and accessible method. It tackles key challenges like class imbalance (especially in the N1 stage), limited receptive fields, and lack of interpretability in existing models. The proposed framework's focus on improving N1 stage detection and its emphasis on interpretability are significant contributions, potentially leading to more reliable and clinically useful sleep staging systems.
Key Takeaways
- •Proposes a context-aware and interpretable framework for single-channel EEG sleep staging.
- •Addresses class imbalance, especially in the N1 stage, using class-weighted loss and data augmentation.
- •Combines multi-scale feature extraction with temporal modeling to capture local and long-range dependencies.
- •Achieves significant improvements in N1 stage detection compared to previous methods.
“The proposed framework achieves an overall accuracy of 89.72% and a macro-average F1-score of 85.46%. Notably, it attains an F1- score of 61.7% for the challenging N1 stage, demonstrating a substantial improvement over previous methods on the SleepEDF datasets.”