LWSleepNet: A lightweight attention-based deep learning model for sleep staging with singlechannel EEG

Abstract

Introduction Sleep is vital to human health, and sleep staging is an essential process in sleep assessment. However, manual classification is an inefficient task. Along with the increased demand for portable sleep quality detection devices, lightweight automatic sleep staging needs to be developed. Methods This study proposes a novel attention-based lightweight deep learning model called LWSleepNet. A depthwise separable multi-resolution convolutional neural network is introduced to analyze the input feature map and captures features at multiple frequencies using two different sized convolutional kernels. The temporal feature extraction module divides the input into patches and feeds them into a multi-head attention block to extract time-dependent information from sleep recordings. The model's convolution operations are replaced with depthwise separable convolutions to minimize its number of parameters and computational cost. The model's performance on two public datasets (Sleep-EDF-20 and Sleep-EDF-78) was evaluated and compared with those of previous studies. Then, an ablation study and sensitivity analysis were performed to evaluate further each module. Results LWSleepNet achieves an accuracy of 86.6% and Macro-F1 score of 79.2% for the Sleep-EDF-20 dataset and an accuracy of 81.5% and Macro-F1 score of 74.3% for the Sleep-EDF-78 dataset with only 55.3 million floating-point operations per second and 180 K parameters. Conclusion On two public datasets, LWSleepNet maintains excellent prediction performance while substantially reducing the number of parameters, demonstrating that our proposed Light multiresolution convolutional neural network and temporal feature extraction modules can provide excellent portability and accuracy and can be easily integrated into portable sleep monitoring devices.

SeriesSleepNet: an EEG time series model with partial data augmentation for automatic sleep stage scoring

“…Specifically, only overall F1-score was lower than XSleepNet2 ( Phan et al, 2021 ) at 0.1%. It is also 1.1% lower in overall accuracy than LWSleepNet ( Yang et al, 2023 ). Finally, the overall Kappa values of the proposed model were 0.02 lower than XSleepNet2 ( Phan et al, 2021 ) and LWSleepNet ( Yang et al, 2023 ).…”

“…It is also 1.1% lower in overall accuracy than LWSleepNet ( Yang et al, 2023 ). Finally, the overall Kappa values of the proposed model were 0.02 lower than XSleepNet2 ( Phan et al, 2021 ) and LWSleepNet ( Yang et al, 2023 ). On the other hand, in most performance metrics, the performance is similar to other models, and in particular, our model has the highest performance in the F1-score of the REM stage.…”

Deep learning-based sleep stage classification with cardiorespiratory and body movement activities in individuals with suspected sleep disorders

“…The low performance at N1 indicates the difficulty of discriminating N1 by the model. Previous studies have also pointed out that the model tended to misclassify the N1 period as N2 and WK 28 – 31 . When N1 and N2 were combined to form a light NREM sleep stage, accuracy and Cohen’s κ were almost the same as those at N2 in the five-class classification (accuracy = 0.68 ± 0.07, Cohen’s κ = 0.36 ± 0.14).…”

Automated Sleep Stage Classification in Home Environments: An Evaluation of Seven Deep Neural Network Architectures

“…Inspired by these advances, Yao et al [9] implemented a convolutional network along with transformers, which were used to capture changing temporal features in the data. Attention mechanisms, combined with a CNN network, were also employed in the studies of Zhou et al [10] and Yang et al [11], demonstrating the feasibility of their integration into portable sleepmonitoring devices.…”