Search papers, labs, and topics across Lattice.
This paper introduces SwinSleepNet, a hierarchical context-aware dual-stream framework designed to enhance automatic sleep staging by addressing the limitations of traditional epoch-level encoders. By employing separate branches for raw EEG signals and their time-frequency transformations, the model captures both fine temporal details and complex spectral dependencies, leading to improved recognition accuracy, particularly for challenging sleep stages like N1. Comprehensive evaluations on multiple datasets demonstrate that SwinSleepNet not only achieves competitive performance but also exhibits superior robustness in identifying ambiguous and transitional sleep stages.
SwinSleepNet outperforms traditional models by effectively capturing both intra-epoch microstructures and inter-epoch dependencies, leading to significant improvements in sleep stage classification accuracy.
Automatic sleep staging is a critical role in sleep disorder diagnosis, sleep quality assessment, and long-term health monitoring; however, existing approaches suffer poor performance on ambiguous and transition-related sleep stages, caused by inadequate modeling of fine-grained intra-epoch structures and complex cross-region spectral dependencies. Traditional epoch-level encoders commonly fail to extract subtle temporal microstructures and intra-epoch cross-region interactions, resulting in unsatisfactory recognition accuracy for hard categories such as the N1 stage. To tackle these drawbacks, we propose SwinSleepNet, a hierarchical context-aware dual-stream framework that separately optimizes intra-epoch representation learning and inter-epoch contextual modeling. Concretely, we characterize each sleep epoch from two complementary perspectives: raw time-domain EEG signal and its time-frequency transformation. The time-domain branch adopts convolutional encoders to capture fine waveform temporal details, and the time-frequency branch uses Swin Transformer to extract local spectro-temporal features, hierarchical multi-scale information and long-range spatial dependencies. The multi-branch extracted features are fused into integrated embeddings, which are optimized by a bidirectional context module to capture cross-epoch temporal dependencies for final sleep stage classification. Comprehensive experiments on Sleep-EDF-20, Sleep-EDF-78 and SHHS datasets verify that our method achieves competitive overall performance, and exhibits stronger robustness and stability on difficult N1 stages and transitional epochs. The results prove that optimized intra-epoch representation learning based on hierarchical architecture greatly benefits automatic sleep staging tasks.
