STDA-Net: Spectrogram-Based Domain Adaptation for cross-dataset Sleep Stage Classification

TL;DR

This paper introduces STDA-Net, a spectrogram-based deep learning framework that improves cross-dataset sleep stage classification by combining CNNs, BiLSTM, and domain-adversarial training, outperforming existing methods.

Contribution

The novel integration of spectrogram inputs with temporal modeling and adversarial domain adaptation enhances cross-dataset sleep staging accuracy and stability.

Findings

Achieved 89.03% accuracy and 87.64% macro F1-score across datasets.

Outperformed existing 1D EEG-based methods in accuracy and stability.

Demonstrated robustness and reproducibility in multiple transfer settings.

Abstract

Accurate sleep stage classification across datasets remains challenging due to variability in EEG channel montages, sampling rates, recording environments, and subject populations. Although deep learning has shown considerable promise for automated sleep staging, most existing cross-dataset methods rely on one-dimensional EEG signal representations, whereas the use of two-dimensional spectrogram-based inputs within an unsupervised domain adaptation framework has remained largely unexplored. Here, we propose STDA-Net (Spectrogram-based Temporal Domain Adaptation Network), a framework that combines a convolutional neural network (CNN) for spectrogram-based feature extraction, a bidirectional long short-term memory (BiLSTM) module for temporal modeling of sleep dynamics, and a domain-adversarial neural network (DANN) for source-to-target feature alignment without requiring any labeled target-domain data during training. Experiments are conducted on three publicly available datasets Sleep-EDF, SHHS-1, and SHHS-2 under six cross-dataset transfer settings. Results show that the proposed framework achieves an average accuracy of 89.03% and an average macro F1-score of 87.64%, consistently outperforming existing 1D baseline methods in terms of balanced classification performance, with substantially lower variance across five independent runs, indicating improved stability and reproducibility. Overall, these findings demonstrate that 2D spectrogram-based representations, combined with temporal modeling and adversarial domain adaptation, provide a robust and competitive alternative to conventional 1D EEG inputs for cross-dataset sleep staging.