Abstract:Objective: Automatic sleep scoring is crucial for diagnosing sleep disorders. Existing frameworks based on Polysomnography often rely on long sequences of input signals to predict sleep stages, which can introduce complexity. Moreover, there is limited exploration of simplifying representation learning in sleep scoring methods. Methods: In this study, we propose NeuroSleepNet, an automatic sleep scoring method designed to classify the current sleep stage using only the microevents in the current input signal, without the need for past inputs. Our model employs supervised spatial and multi-scale temporal context learning and incorporates a transformer encoder to enhance representation learning. Additionally, NeuroSleepNet is optimized for balanced performance across five sleep stages by introducing a logarithmic scale-based weighting technique as a loss function. Results: NeuroSleepNet achieved similar and comparable performance with current state-of-the-art results. The best accuracy, macro-F1 score, and Cohen's kappa were 86.1 percent, 80.8 percent, and 0.805 for Sleep-EDF expanded; 82.0 percent, 76.3 percent, and 0.753 for MESA; 80.5 percent, 76.8 percent, and 0.738 for Physio2018; and 86.7 percent, 80.9 percent, and 0.804 for the SHHS database. Conclusion: NeuroSleepNet demonstrates that even with a focus on computational efficiency and a purely supervised learning approach, it is possible to achieve performance that is comparable to state-of-the-art methods. Significance: Our study simplifies automatic sleep scoring by focusing solely on microevents in the current input signal while maintaining remarkable performance. This offers a streamlined alternative for sleep diagnosis applications.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to solve several key problems in automatic sleep staging: 1. **Complexity**: Existing automatic sleep staging methods based on polysomnography (PSG) usually rely on long - time - series input signals to predict sleep stages, which increases the complexity and computational burden of the model. 2. **Simplified representation learning**: Current methods have limited exploration in simplified representation learning, especially when classifying only using microevents in the current input signal. 3. **Class imbalance problem**: The number of samples in different sleep stages in the sleep dataset is unbalanced, resulting in poor performance of the model in some stages. 4. **Dependence on historical information**: Many existing methods rely on past input signals for prediction, which limits the real - time and independence of the model. To solve these problems, the paper proposes **NeuroSleepNet**, a new automatic sleep staging method. The main features of this method include: - **Using only the current input signal**: By classifying only using microevents in the current input signal without the need for past input signals. - **Multi - scale spatio - temporal context learning**: Combining supervised spatial and multi - scale temporal context learning to enhance feature representation ability. - **Transformer encoder**: Introducing a Transformer encoder to further improve the learning of feature representation. - **Log - weighted loss function**: Proposing a weighting technique based on the log scale as a loss function to deal with the class imbalance problem and optimize the performance of the model in each sleep stage. Through these improvements, NeuroSleepNet can significantly simplify the process of automatic sleep staging while maintaining high performance and reducing dependence on historical information. ### Key contributions of the paper 1. **Multi - scale temporal representation learning**: Through multi - scale temporal representation learning of virtual channel outputs, subsequent convolution operations are made simpler and more computationally efficient. 2. **Using only the current epoch**: Most existing methods rely on temporal context across epochs, while NeuroSleepNet uses only the current epoch (one - to - one) and provides comparable performance. 3. **Feature representation adjustment**: Proposing a method to adjust the feature representation into a form more suitable for the multi - head encoder network, thereby obtaining a more refined representation. 4. **Log - weighted loss function**: Introducing a log - weighted loss function to solve the class imbalance problem in the dataset and achieve more balanced results. 5. **Extensive verification**: The performance of NeuroSleepNet has been verified on four public datasets, and the results show that its performance is comparable to or better than that of the existing state - of - the - art methods. Through these innovations, NeuroSleepNet provides an efficient automatic sleep staging solution with important clinical application potential.

NeuroSleepNet: A Multi-Head Self-Attention Based Automatic Sleep Scoring Scheme with Spatial and Multi-Scale Temporal Representation Learning

Automatic Classification of Sleep Stages Based on Raw Single-Channel EEG

CareSleepNet: A Hybrid Deep Learning Network for Automatic Sleep Staging

DeepSleepNet: A Model for Automatic Sleep Stage Scoring Based on Raw Single-Channel EEG

SleepEEGNet: Automated sleep stage scoring with sequence to sequence deep learning approach

Sleep Staging Framework with Physiologically Harmonized Sub-Networks

Automated multi-model deep neural network for sleep stage scoring with unfiltered clinical data

NeuroNet: A Novel Hybrid Self-Supervised Learning Framework for Sleep Stage Classification Using Single-Channel EEG

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

SLEEPNET: Automated Sleep Staging System via Deep Learning

Sleep Staging Based on Multi Scale Dual Attention Network

Convolution- and Attention-Based Neural Network for Automated Sleep Stage Classification

0447 ResTNet: A Robust End-to-End Deep Learning Approach to Sleep Staging of Self Applied Somnography Studies

DeepSleepNet-Lite: A Simplified Automatic Sleep Stage Scoring Model with Uncertainty Estimates

Multi-Modal Sleep Stage Classification With Two-Stream Encoder-Decoder

Extracting Multi-Scale and Salient Features by MSE Based U-Structure and CBAM for Sleep Staging

Automatic sleep staging by a hybrid model based on deep 1D-ResNet-SE and LSTM with single-channel raw EEG signals

ZleepAnlystNet: a novel deep learning model for automatic sleep stage scoring based on single-channel raw EEG data using separating training

A Multi Constrained Transformer-BiLSTM Guided Network for Automated Sleep Stage Classification from Single-Channel EEG

Computation-Efficient Multi-Model Deep Neural Network for Sleep Stage Classification.

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