BiT-MamSleep: Bidirectional Temporal Mamba for EEG Sleep Staging

Xinliang Zhou,Yuzhe Han,Chenyu Liu,Yi Ding,Ziyu Jia,Yang Liu

2024-11-03

Abstract:In this paper, we address the challenges in automatic sleep stage classification, particularly the high computational cost, inadequate modeling of bidirectional temporal dependencies, and class imbalance issues faced by Transformer-based models. To address these limitations, we propose BiT-MamSleep, a novel architecture that integrates the Triple-Resolution CNN (TRCNN) for efficient multi-scale feature extraction with the Bidirectional Mamba (BiMamba) mechanism, which models both short- and long-term temporal dependencies through bidirectional processing of EEG data. Additionally, BiT-MamSleep incorporates an Adaptive Feature Recalibration (AFR) module and a temporal enhancement block to dynamically refine feature importance, optimizing classification accuracy without increasing computational complexity. To further improve robustness, we apply optimization techniques such as Focal Loss and SMOTE to mitigate class imbalance. Extensive experiments on four public datasets demonstrate that BiT-MamSleep significantly outperforms state-of-the-art methods, particularly in handling long EEG sequences and addressing class imbalance, leading to more accurate and scalable sleep stage classification.

Signal Processing,Machine Learning

What problem does this paper attempt to address?

This paper attempts to solve several key problems in automatic sleep - stage classification: 1. **High computational cost**: When dealing with long - time high - resolution EEG sequences, the Transformer model has a quadratic increase in computational complexity with the sequence length, resulting in excessive computational cost and difficulty in efficiently processing complex datasets. 2. **Insufficient bidirectional time - dependence modeling**: Existing Transformer models mainly perform one - way processing within a fixed time window and cannot fully capture the bidirectional (past and future) time - dependence required in the sleep - stage transition process. This is especially important for processing long - time EEG sequences because these sequences contain subtle and gradual changes. 3. **Class imbalance problem**: The occurrence frequency of sleep stages such as REM is much lower than that of other stages (such as NREM), which leads to class imbalance in the data. This imbalance makes the model tend to predict the majority class, thus affecting the overall classification accuracy, especially performing poorly in detecting minority - class sleep stages. To solve these problems, the authors propose the BiT - MamSleep architecture. Specific improvements include: - **Triple - Resolution CNN (TRCNN)**: It is used to efficiently extract multi - scale features to capture the characteristics of EEG signals in different frequency bands. - **Bidirectional Mamba (BiMamba) mechanism**: By bidirectionally processing EEG data, it simultaneously captures short - term and long - term time - dependence, thus representing sleep - stage transitions more accurately. - **Adaptive Feature Recalibration (AFR) module**: Dynamically adjusts the importance of features to optimize classification accuracy without increasing computational complexity. - **Optimization techniques**: Such as Focal Loss and SMOTE, are used to alleviate the class imbalance problem, increase the model's attention to difficult - to - classify samples, and generate synthetic minority - class data. Through these improvements, BiT - MamSleep significantly improves its performance in processing long EEG sequences and dealing with class imbalance. Experimental results on multiple public datasets show that it outperforms existing advanced methods.

BiT-MamSleep: Bidirectional Temporal Mamba for EEG Sleep Staging

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

MSSC-BiMamba: Multimodal Sleep Stage Classification and Early Diagnosis of Sleep Disorders with Bidirectional Mamba

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

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

CareSleepNet: A Hybrid Deep Learning Network for Automatic Sleep Staging

A CNN-Transformer-ConvLSTM-CRF Hybrid Network for Sleep Stage Classification

EEGMamba: Bidirectional State Space Model with Mixture of Experts for EEG Multi-task Classification

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

Intelligent automatic sleep staging model based on CNN and LSTM

A Biomathematical Model for Classifying Sleep Stages Using Deep Learning Techniques

Deep Learning in Automatic Sleep Staging With a Single Channel Electroencephalography

A convolutional transformer for EEG sleep stage classification

Automatic Classification of Sleep Stages from EEG Signals Using Riemannian Metrics and Transformer Networks

A Feature Fusion Model Based on Temporal Convolutional Network for Automatic Sleep Staging Using Single-Channel EEG

A Single-Channel EEG Based Automatic Sleep Stage Classification Method Leveraging Deep One-Dimensional Convolutional Neural Network and Hidden Markov Model.

Research of Automatic Staging of Sleep Based on Period Segmentation

Transformer-Based Contrastive Learning Method for Automated Sleep Stages Classification

3DSleepNet: A Multi-Channel Bio-Signal Based Sleep Stages Classification Method Using Deep Learning

RecSleepNet: an Automatic Sleep Staging Model Based on Feature Reconstruction

An End-to-End Multi-Channel Convolutional Bi-LSTM Network for Automatic Sleep Stage Detection