Unsupervised Domain Adaptation by Statistics Alignment for Deep Sleep Staging Networks

A Unified Flexible Large Polysomnography Model for Sleep Staging and Mental Disorder Diagnosis

Sleep quality is vital to human health, yet automated sleep staging faces challenges in cross-center generalization due to data scarcity and domain gaps. Traditional scoring is labor-intensive, while deep learning models often fail to generalize across datasets. Here, we present LPSGM, a unified and flexible large polysomnography (PSG) model designed to enhance cross-center generalization in sleep staging and enable fine-tuning for disease diagnosis. Trained on 220,500 hours of PSG data from 16 public datasets, LPSGM integrates domain-adaptive learning and supports variable-channel configurations, achieving performance comparable to models trained directly on target-center data. In a prospective clinical study, LPSGM matches expert-level accuracy with lower variability. When fine-tuned, it attains 88.01% accuracy in narcolepsy detection and 100% in depression detection. These results establish LPSGM as a scalable, plug-and-play solution for automated PSG analysis, bridging the gap between sleep staging and clinical deployment.

Non-Invasive Brain-Computer Interfaces: State of the Art and Trends

Brain-computer interface (BCI) is a rapidly evolving technology that has the potential to widely influence research, clinical and recreational use. Non-invasive BCI approaches are particularly common as they can impact a large number of participants safely and at a relatively low cost. Where traditional non-invasive BCIs were used for simple computer cursor tasks, it is now increasingly common for these systems to control robotic devices for complex tasks that may be useful in daily life. In this review, we provide an overview of the general BCI framework as well as the various methods that can be used to record neural activity, extract signals of interest, and decode brain states. In this context, we summarize the current state-of-the-art of non-invasive BCI research, focusing on trends in both the application of BCIs for controlling external devices and algorithm development to optimize their use. We also discuss various open-source BCI toolboxes and software, and describe their impact on the field at large.

Outlier Handling Strategy of Ensembled-Based Sequential Convolutional Neural Networks for Sleep Stage Classification

The pivotal role of sleep has led to extensive research endeavors aimed at automatic sleep stage classification. However, existing methods perform poorly when classifying small groups or individuals, and these results are often considered outliers in terms of overall performance. These outliers may introduce bias during model training, adversely affecting feature selection and diminishing model performance. To address the above issues, this paper proposes an ensemble-based sequential convolutional neural network (E-SCNN) that incorporates a clustering module and neural networks. E-SCNN effectively ensembles machine learning and deep learning techniques to minimize outliers, thereby enhancing model robustness at the individual level. Specifically, the clustering module categorizes individuals based on similarities in feature distribution and assigns personalized weights accordingly. Subsequently, by combining these tailored weights with the robust feature extraction capabilities of convolutional neural networks, the model generates more accurate sleep stage classifications. The proposed model was verified on two public datasets, and experimental results demonstrate that the proposed method obtains overall accuracies of 84.8% on the Sleep-EDF Expanded dataset and 85.5% on the MASS dataset. E-SCNN can alleviate the outlier problem, which is important for improving sleep quality monitoring for individuals.

Personalization of Automatic Sleep Scoring: How Best to Adapt Models to Personal Domains in Wearable EEG

Wearable EEG enables us to capture large amounts of high-quality sleep data for diagnostic purposes. To make full use of this capacity we need high-performance automatic sleep scoring models. To this end, it has been noted that domain mismatch between recording equipment can be considerable, e.g. PSG to wearable EEG, but a previously observed benefit from personalizing models to individual subjects further indicates a personal domain in sleep EEG. In this work, we have investigated the extent of such a personal domain in wearable EEG, and review supervised and unsupervised approaches to personalization as found in the literature. We investigated the personalization effect of the unsupervised Adversarial Domain Adaptation and implemented an unsupervised method based on statistics alignment. No beneficial personalization effect was observed using these unsupervised methods. We find that supervised personalization leads to a substantial performance improvement on the target subject ranging from 15% Cohen's Kappa for subjects with poor performance ( ) to roughly 2% on subjects with high performance ( ). This improvement was present for models trained on both small and large data sets, indicating that even high-performance models benefit from supervised personalization. We found that this personalization can be beneficially regularized using Kullback-Leibler regularization, leading to lower variance with negligible cost to improvement. Based on the experiments, we recommend model personalization using Kullback-Leibler regularization.

PSEENet: A Pseudo-Siamese Neural Network Incorporating Electroencephalography and Electrooculography Characteristics for Heterogeneous Sleep Staging

Sleep staging plays a critical role in evaluating the quality of sleep. Currently, most studies are either suffering from dramatic performance drops when coping with varying input modalities or unable to handle heterogeneous signals. To handle heterogeneous signals and guarantee favorable sleep staging performance when a single modality is available, a pseudo-siamese neural network (PSN) to incorporate electroencephalography (EEG), electrooculography (EOG) characteristics is proposed (PSEENet). PSEENet consists of two parts, spatial mapping modules (SMMs) and a weight-shared classifier. SMMs are used to extract high-dimensional features. Meanwhile, joint linkages among multi-modalities are provided by quantifying the similarity of features. Finally, with the cooperation of heterogeneous characteristics, associations within various sleep stages can be established by the classifier. The evaluation of the model is validated on two public datasets, namely, Montreal Archive of Sleep Studies (MASS) and SleepEDFX, and one clinical dataset from Huashan Hospital of Fudan University (HSFU). Experimental results show that the model can handle heterogeneous signals, provide superior results under multimodal signals and show good performance with single modality. PSEENet obtains accuracy of 79.1%, 82.1% with EEG, EEG and EOG on Sleep-EDFX, and significantly improves the accuracy with EOG from 73.7% to 76% by introducing similarity information.

Source-free unsupervised domain adaptation: A survey

Unsupervised domain adaptation (UDA) via deep learning has attracted appealing attention for tackling domain-shift problems caused by distribution discrepancy across different domains. Existing UDA approaches highly depend on the accessibility of source domain data, which is usually limited in practical scenarios due to privacy protection, data storage and transmission cost, and computation burden. To tackle this issue, many source-free unsupervised domain adaptation (SFUDA) methods have been proposed recently, which perform knowledge transfer from a pre-trained source model to the unlabeled target domain with source data inaccessible. A comprehensive review of these works on SFUDA is of great significance. In this paper, we provide a timely and systematic literature review of existing SFUDA approaches from a technical perspective. Specifically, we categorize current SFUDA studies into two groups, i.e., white-box SFUDA and black-box SFUDA, and further divide them into finer subcategories based on different learning strategies they use. We also investigate the challenges of methods in each subcategory, discuss the advantages/disadvantages of white-box and black-box SFUDA methods, conclude the commonly used benchmark datasets, and summarize the popular techniques for improved generalizability of models learned without using source data. We finally discuss several promising future directions in this field.

Dual-Teacher Feature Distillation: A Transfer Learning Method for Insomniac PSG Staging

Insomnia is the most common sleep disorder linked with adverse long-term medical and psychiatric outcomes. Automatic sleep staging plays a crucial role in aiding doctors to diagnose insomnia disorder. Only a few studies have been conducted to develop automatic sleep staging methods for insomniacs, and most of them have utilized transfer learning methods, which involve pre-training models on healthy individuals and then fine-tuning them on insomniacs. Unfortunately, significant differences in feature distribution between the two subject groups impede the transfer performance, highlighting the need to effectively integrate the features of healthy subjects and insomniacs. In this paper, we propose a dual-teacher cross-domain knowledge transfer method based on the feature-based knowledge distillation to improve the performance of sleep staging for insomniacs. Specifically, the insomnia teacher directly learns from insomniacs and feeds the corresponding domain-specific features into the student network, while the health domain teacher guide the student network to learn domain-generic features. During the training process, we adopt the OFD (Overhaul of Feature Distillation) method to build the health domain teacher. We conducted the experiments to validate the proposed method, using the Sleep-EDF database as the source domain and the CAP-Database as the target domain. The results demonstrate that our method surpasses advanced techniques, achieving an average sleep staging accuracy of 80.56% on the CAP-Database. Furthermore, our method exhibits promising performance on the private dataset.

Robust myoelectric pattern recognition methods for reducing users' calibration burden: challenges and future

Myoelectric pattern recognition (MPR) has evolved into a sophisticated technology widely employed in controlling myoelectric interface (MI) devices like prosthetic and orthotic robots. Current MIs not only enable multi-degree-of-freedom control of prosthetic limbs but also demonstrate substantial potential in consumer electronics. However, the non-stationary random characteristics of myoelectric signals poses challenges, leading to performance degradation in practical scenarios such as electrode shifting and switching new users. Conventional MIs often necessitate meticulous calibration, imposing a significant burden on users. To address user frustration during the calibration process, researchers have focused on identifying MPR methods that alleviate this burden. This article categorizes common scenarios that incur calibration burdens as based on data distribution shift and based on dynamic data categories. Then further investigated and summarized the popular robust MPR algorithms used to reduce the user's calibration burden. We categorize these algorithms as based on data manipulate, feature manipulation and, model structure. And describes the scenarios to which each method is applicable and the conditions required for calibration. Finally, this review is concluded with the advantages of robust MPR and the remaining challenges and future opportunities.

Multi-Action Knee Contact Force Prediction by Domain Adaptation

Most recent musculoskeletal dynamics estimation methods are designed for predefined actions, such as gait, and don't generalize to various tasks. In this work, we address the problem of estimating internal biomechanical forces during more than one actions by introducing unsupervised domain adaptation into a deep learning model. More specifically, we developed a Bidirectional Long Short-Term Memory network for knee contact force prediction, enhanced with correlation alignment layers, in order to minimize the domain shift between kinematic data from different actions. Furthermore, we used the novel Neural State Machine (NSM) as a simulation platform to test and visualize our model predictions in a wide range of trajectories adapted to different 3D scene geometries in real-time. We conducted multiple experiments, including comparison with previous models, model alignment across action classes and real-to-synthetic data alignment. The results showed that the proposed deep learning architecture with domain adaptation performs better than the benchmark in terms of NRMSE and t-test. Overall, our method is capable of predicting knee contact forces for more than one action classes using a single architecture and thereby opens the path for estimating internal forces for intermediate actions, while the knowledge of the hidden state of motion may be used to support personalized rehabilitation. Moreover, our model can be easily integrated into any human motion simulation environment, which shows its potential in enabling biomechanical analysis in an automated and computationally efficient way.

SFDA: Domain Adaptation With Source Subject Fusion Based on Multi-Source and Single-Target Fall Risk Assessment

In cross-subject fall risk classification based on plantar pressure, a challenge is that data from different subjects have significant individual information. Thus, the models with insufficient generalization ability can't perform well on new subjects, which limits their application in daily life. To solve this problem, domain adaptation methods are applied to reduce the gap between source and target domain. However, these methods focus on the distribution of the source and the target domain, but ignore the potential correlation among multiple source subjects, which deteriorates domain adaptation performance. In this paper, we proposed a novel method named domain adaptation with subject fusion (SFDA) for fall risk assessment, greatly improving the cross-subject assessment ability. Specifically, SFDA synchronously carries out source target adaptation and multiple source subject fusion by domain adversarial module to reduce source-target gap and distribution distance within source subjects of same class. Consequently, target samples can learn more task-specific features from source subjects to improve the generalization ability. Experiment results show that SFDA achieved mean accuracy of 79.17 % and 73.66 % based on two backbones in a cross-subject classification manner, outperforming the state-of-the-art methods on continuous plantar pressure dataset. This study proves the effectiveness of SFDA and provides a novel tool for implementing cross-subject and few-gait fall risk assessment.

Decentralized Data-Privacy Preserving Deep-Learning Approaches for Enhancing Inter-Database Generalization in Automatic Sleep Staging

Automatic sleep staging has been an active field of development. Despite multiple efforts, the area remains a focus of research interest. Indeed, while promising results have reported in past literature, uptake of automatic sleep scoring in the clinical setting remains low. One of the current issues regards the difficulty to generalization performance results beyond the local testing scenario, i.e. across data from different clinics. Issues derived from data-privacy restrictions, that generally apply in the medical domain, pose additional difficulties in the successful development of these methods. We propose the use of several decentralized deep-learning approaches, namely ensemble models and federated learning, for robust inter-database performance generalization and data-privacy preservation in automatic sleep staging scenario. Specifically, we explore four ensemble combination strategies (max-voting, output averaging, size-proportional weighting, and Nelder-Mead) and present a new federated learning algorithm, so-called sub-sampled federated stochastic gradient descent (ssFedSGD). To evaluate generalization capabilities of such approaches, experimental procedures are carried out using a leaving-one-database-out direct-transfer scenario on six independent and heterogeneous public sleep staging databases. The resulting performance is compared with respect to two baseline approaches involving single-database and centralized multiple-database derived models. Our results show that proposed decentralized learning methods outperform baseline local approaches, and provide similar generalization results to centralized database-combined approaches. We conclude that these methods are more preferable choices, as they come with additional advantages concerning improved scalability, flexible design, and data-privacy preservation.

Challenges of Applying Automated Polysomnography Scoring at Scale

Automatic polysomnography analysis can be leveraged to shorten scoring times, reduce associated costs, and ultimately improve the overall diagnosis of sleep disorders. Multiple and diverse strategies have been attempted for implementation of this technology at scale in the routine workflow of sleep centers. The field, however, is complex and presents unsolved challenges in a number of areas. Recent developments in computer science and artificial intelligence are nevertheless closing the gap. Technological advances are also opening new pathways for expanding our current understanding of the domain and its analysis.

Automatic Upper-Limb Brunnstrom Recovery Stage Evaluation via Daily Activity Monitoring

Motor function assessment is crucial for post-stroke rehabilitation. Conventional evaluation methods are subjective, heavily depending on the experience of therapists. In light of the strong correlation between the stroke severity level and the performance of activities of daily living (ADLs), we explored the possibility of automatically evaluating the upper-limb Brunnstrom Recovery Stage (BRS) via three typical ADLs (tooth brushing, face washing and drinking). Multimodal data (acceleration, angular velocity, surface electromyography) were synchronously collected from 5 upper-limb-worn sensor modules. The performance of BRS evaluation system is known to be variable with different system parameters (e.g., number of sensor modules, feature types and classifiers). We systematically searched for the optimal parameters from different data segmentation strategies (five window lengths and four overlaps), 42 types of features, 12 feature optimization techniques and 9 classifiers with the leave-one-subject-out cross-validation. To achieve reliable and low-cost monitoring, we further explored whether it was possible to obtain a satisfactory result using a relatively small number of sensor modules. As a result, the proposed approach can correctly recognize the stages of all 27 participants using only three sensor modules with the optimized data segmentation parameters (window length: 7s, overlap: 50%), extracted features (simple square integral, slope sign change, modified mean absolute value 1 and modified mean absolute value 2), the feature optimization method (principal component analysis) and the logistic regression classifier. According to the literature, this is the first study to comprehensively optimize sensor configuration and parameters in each stage of the BRS classification framework. The proposed approach can serve as a factor-screening tool towards the automatic BRS classification and is promising to be further used at home.

From unsupervised to semi-supervised adversarial domain adaptation in EEG-based sleep staging

OBJECTIVE

The recent breakthrough of wearable sleep monitoring devices results in large amounts of sleep data. However, as limited labels are available, interpreting these data requires automated sleep stage classification methods with a small need for labeled training data. Transfer learning and domain adaptation offer possible solutions by enabling models to learn on a source dataset and adapt to a target dataset.

APPROACH

In this paper, we investigate adversarial domain adaptation applied to real use cases with wearable sleep datasets acquired from diseased patient populations. Different practical aspects of the adversarial domain adaptation framework \hl{are examined}, including the added value of (pseudo-)labels from the target dataset and the influence of domain mismatch between the source and target data. The method is also implemented for personalization to specific patients.

MAIN RESULTS

The results show that adversarial domain adaptation is effective in the application of sleep staging on wearable data. When compared to a model applied on a target dataset without any adaptation, the domain adaptation method in its simplest form achieves relative gains of 7%-27% in accuracy. The performance on the target domain is further boosted by adding pseudo-labels and real target domain labels when available, and by choosing an appropriate source dataset. Furthermore, unsupervised adversarial domain adaptation can also personalize a model, improving the performance by 1%-2% compared to a non-personal model.

SIGNIFICANCE

In conclusion, adversarial domain adaptation provides a flexible framework for semi-supervised and unsupervised transfer learning. This is particularly useful in sleep staging and other wearable EEG applications.

Automatic sleep staging of EEG signals: recent development, challenges, and future directions

Modern deep learning holds a great potential to transform clinical practice on human sleep. Teaching a machine to carry out routine tasks would be a tremendous reduction in workload for clinicians. Sleep staging, a fundamental step in sleep practice, is a suitable task for this and will be the focus in this article. Recently, automatic sleep staging systems have been trained to mimic manual scoring, leading to similar performance to human sleep experts, at least on scoring of healthy subjects. Despite tremendous progress, we have not seen automatic sleep scoring adopted widely in clinical environments. This review aims to give a shared view of the authors on the most recent state-of-the-art development in automatic sleep staging, the challenges that still need to be addressed, and the future directions for automatic sleep scoring to achieve clinical value.