Single-channel EEG Research Articles

An Eye Tracking and Brain–Computer Interface-Based Human–Environment Interactive System for Amyotrophic Lateral Sclerosis Patients

Amyotrophic Lateral Sclerosis (ALS) patients suffer a great inconvenience in their daily lives due to the gradual loss of their motion abilities. In order to help ALS patients regain their self-care abilities at a certain level, an eye-tracking and brain-computer interface (BCI) based human-environment interactive (HEI) system is proposed. Through a single-channel EEG recorder and a pair of eye-tracking glasses, the user’s attention levels and gaze points are detected. Based on the two types of intention information, the target objects are recognized by a YOLOv5 (You Only Look Once v5) model. Then according to voice instructions, users can control the household electric appliances. The system performed a high success rate of 89.3% experimentally, providing a reliable and capable assistive solution for ALS patients.

Dynamical graph neural network with attention mechanism for epilepsy detection using single channel EEG.

Epilepsy is a chronic brain disease, and identifying seizures based on electroencephalogram (EEG) signals would be conducive to implement interventions to help patients reduce impairment and improve quality of life. In this paper, we propose a classification algorithm to apply dynamical graph neural network with attention mechanism to single channel EEG signals. Empirical mode decomposition (EMD) are adopted to construct graphs and the optimal adjacency matrix is obtained by model optimization. A multilayer dynamic graph neural network with attention mechanism is proposed to learn more discriminative graph features. The MLP-pooling structure is proposed to fuse graph features. We performed 12 classification tasks on the epileptic EEG database of the University of Bonn, and experimental results showed that using 25 runs of ten-fold cross-validation produced the best classification results with an average of 99.83[Formula: see text] accuracy, 99.91[Formula: see text] specificity, 99.78[Formula: see text] sensitivity, 99.87[Formula: see text] precision, and 99.47[Formula: see text] [Formula: see text] score for the 12 classification tasks.

L-SeqSleepNet: Whole-cycle Long Sequence Modelling for Automatic Sleep Staging.

Human sleep is cyclical with a period of approximately 90 minutes, implying long temporal dependency in the sleep data. Yet, exploring this long-term dependency when developing sleep staging models has remained untouched. In this work, we show that while encoding the logic of a whole sleep cycle is crucial to improve sleep staging performance, the sequential modelling approach in existing state-of-the-art deep learning models are inefficient for that purpose. We thus introduce a method for efficient long sequence modelling and propose a new deep learning model, L-SeqSleepNet, which takes into account whole-cycle sleep information for sleep staging. Evaluating L-SeqSleepNet on four distinct databases of various sizes, we demonstrate state-of-the-art performance obtained by the model over three different EEG setups, including scalp EEG in conventional Polysomnography (PSG), in-ear EEG, and around-the-ear EEG (cEEGrid), even with a single EEG channel input. Our analyses also show that L-SeqSleepNet is able to alleviate the predominance of N2 sleep (the major class in terms of classification) to bring down errors in other sleep stages. Moreover the network becomes much more robust, meaning that for all subjects where the baseline method had exceptionally poor performance, their performance are improved significantly. Finally, the computation time only grows at a sub-linear rate when the sequence length increases.

Multi-task learning for arousal and sleep stage detection using fully convolutional networks

Objective. Sleep is a critical physiological process that plays a vital role in maintaining physical and mental health. Accurate detection of arousals and sleep stages is essential for the diagnosis of sleep disorders, as frequent and excessive occurrences of arousals disrupt sleep stage patterns and lead to poor sleep quality, negatively impacting physical and mental health. Polysomnography is a traditional method for arousal and sleep stage detection that is time-consuming and prone to high variability among experts. Approach. In this paper, we propose a novel multi-task learning approach for arousal and sleep stage detection using fully convolutional neural networks. Our model, FullSleepNet, accepts a full-night single-channel EEG signal as input and produces segmentation masks for arousal and sleep stage labels. FullSleepNet comprises four modules: a convolutional module to extract local features, a recurrent module to capture long-range dependencies, an attention mechanism to focus on relevant parts of the input, and a segmentation module to output final predictions. Main results. By unifying the two interrelated tasks as segmentation problems and employing a multi-task learning approach, FullSleepNet achieves state-of-the-art performance for arousal detection with an area under the precision-recall curve of 0.70 on Sleep Heart Health Study and Multi-Ethnic Study of Atherosclerosis datasets. For sleep stage classification, FullSleepNet obtains comparable performance on both datasets, achieving an accuracy of 0.88 and an F1-score of 0.80 on the former and an accuracy of 0.83 and an F1-score of 0.76 on the latter. Significance. Our results demonstrate that FullSleepNet offers improved practicality, efficiency, and accuracy for the detection of arousal and classification of sleep stages using raw EEG signals as input.

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

Sleep staging is crucial for assessing sleep quality and diagnosing sleep disorders. Recent advances in deep learning methods with electroencephalogram (EEG) signals have shown remarkable success in automatic sleep staging. However, the use of deeper neural networks may lead to the issues of gradient disappearance and explosion, while the non-stationary nature and low signal-to-noise ratio of EEG signals can negatively impact feature representation. To overcome these challenges, we proposed a novel lightweight sequence-to-sequence deep learning model, 1D-ResNet-SE-LSTM, to classify sleep stages into five classes using single-channel raw EEG signals. Our proposed model consists of two main components: a one-dimensional residual convolutional neural network with a squeeze-and-excitation module to extract and reweight features from EEG signals, and a long short-term memory network to capture the transition rules among sleep stages. In addition, we applied the weighted cross-entropy loss function to alleviate the class imbalance problem. We evaluated the performance of our model on two publicly available datasets; Sleep-EDF Expanded consists of 153 overnight PSG recordings collected from 78 healthy subjects and ISRUC-Sleep includes 100 PSG recordings collected from 100 subjects diagnosed with various sleep disorders, and obtained an overall accuracy rate of 86.39% and 81.97%, respectively, along with corresponding macro average F1-scores of 81.95% and 79.94%. Our model outperforms existing sleep staging models in terms of overall performance metrics and per-class F1-scores for several sleep stages, particularly for the N1 stage, where it achieves F1-scores of 59.00% and 55.53%. The kappa coefficient is 0.812 and 0.766 for the Sleep-EDF Expanded and ISRUC-Sleep datasets, respectively, indicating strong agreement with certified sleep experts. We also investigated the effect of different weight coefficient combinations and sequence lengths of EEG epochs used as input to the model on its performance. Furthermore, the ablation study was conducted to evaluate the contribution of each component to the model's performance. The results demonstrate the effectiveness and robustness of the proposed model in classifying sleep stages, and highlights its potential to reduce human clinicians' workload, making sleep assessment and diagnosis more effective. However, the proposed model is subject to several limitations. Firstly, the model is a sequence-to-sequence network, which requires input sequences of EEG epochs. Secondly, the weight coefficients in the loss function could be further optimized to balance the classification performance of each sleep stage. Finally, apart from the channel attention mechanism, incorporating more advanced attention mechanisms could enhance the model's effectiveness.

Method for analyzing sequential services using EEG: Micro-meso analysis of emotional changes in real flight service

Capturing customers’ emotional changes in sequential service should be realized using physiological measurements to assess customer delight. Questionnaire-based customer surveys may miss significant and dissipating emotional responses. This study developed a micro‑meso analysis method of capturing emotional changes for sequential service using electroencephalograph (EEG) measurement, dealing with both service encounters (micro-level) and servicescape (meso‑level) over a couple of hours. Customers’ emotion states were defined based on emotional arousal and valence. Emotional responses caused by human interactions were evaluated, and periods of high positive affect throughout the customer journey were visualized. Experiments in actual flight services demonstrated successful emotion estimation across flight phases using a single-channel EEG measurement over two hours. Analysis results on the measurement data revealed emotional peaks outside service encounters that are not captured in customers’ individual self-reports. The results also statistically revealed that two individual services (asking about a refill and conversations started by flight attendants) evoked high positive affect. Temporal dynamic analyses around high positive affect suggested patterns of interplay between joy and surprise, which are key components of customer delight. Compared with questionnaire-based evaluation, the proposed method contributes significantly to empirical studies on sequential services in marketing and design by enabling the extraction of “high positive affect,” which needs to be identified for customer delight. This study supplements existing research on the interactions among physiology (EEG), behavior (emotional changes), and customer service research.

Parameterized aperiodic and periodic components of single-channel EEG enables reliable seizure detection.

Although it is clinically important, a reliable and economical solution to automatic seizure detection for patients at home is yet to be developed. Traditional algorithms rely on multi-channel EEG signals and features of canonical EEG power description. This study is aimed to propose an effective single-channel EEG seizure detection method centered on novel EEG power parameterization and channel selection algorithms.We employed the publicly available multi-channel CHB-MIT Scalp EEG database to gauge the effectiveness of our approach. We first adapted a power spectra parameterization algorithm to characterize the aperiodic and periodic components of the ictal and inter-ictal EEGs. We selected four features based on their statistical significance and interpretability, and developed a ranking approach to channel selection for each patient. We then tested the effectiveness of our approaches to channel and feature selection for automatic seizure detection using support vector machine (SVM) as the classifier. The performance of our algorithm was evaluated using five-fold cross-validation and compared to those methods of comparable complexity (using one or two channels of EEG), in terms of accuracy, specificity, sensitivity, precision and F1 score.Some channels of EEG signals show strikingly different distributions of PSD features between the ictal and inter-ictal states. Four features including the offset and exponent parameters for the aperiodic component and the first and second highest total power (TPW1 and TPW2) form the basis of channel selection and the input of SVM classifier. The selected channel is found to be patient-specific. Our approach has achieved a mean sensitivity of 95.6%, specificity of 99.2%, accuracy of 98.6%, precision of 95.5%, and F1 score of 95.5%. Compared with algorithms in previous studies that used one or two channels of EEG signals, ours outperforms in specificity and accuracy with comparable sensitivity.EEG power spectra parameterization to feature extraction and feature ranking-based channel selection are found to enable efficient and effective automatic seizure detection based on single-channel EEG signal.

Automated insomnia detection using wavelet scattering network technique with single-channel EEG signals

Sleep is crucial for both the physical and mental well-being of human life. As the sleep pattern varies in every individual, it is essential to develop a methodology that enables us to detect sleep-related disorders precisely. Insomnia is one such disorder that affects humans mentally and physically. Signals from Polysomnograms (PSGs) are typically used to identify different sleep disorders. The PSG signals need a lot of handling time and are not patient-friendly. This work proposes to develop a method to automatically identify insomnia using single-channel Electroencephalogram (EEG) signals. We have employed a Wavelet Scattering Network (WSN), a variant of deep convolutional networks, to extract features effectively WSN is an optimized network capable of learning features that assist in discriminating patterns concealed within signals. In addition, WSNs are insensitive to local disturbances, enhancing the network’s dependability and efficiency. We have used single-channel EEG derivation, C4-A1, to create six separate subsets variants based on the sleep-stage annotations, developing the model individually and collectively for the wake, N1, N2, N3, and Rapid Eye Movement (REM) sleep stages To build the proposed model, 12,576 epochs from the Cyclic Alternating Pattern (CAP) sleep database and 20,523 epochs from the Sleep Disorders Research Center (SDRC) dataset were considered. Our proposed model has attained the highest classification Accuracy (Acc) of 97% and area under the curve (AUC) of 0.98 using a Trilayered Neural Network (TNN) classifier considering the wake-sleep stage of the CAP database. Our devised method is straightforward and computationally efficient. Hence, it could be used for clinical applications.

REM Sleep Stage Identification with Raw Single-Channel EEG.

This paper focused on creating an interpretable model for automatic rapid eye movement (REM) and non-REM sleep stage scoring for a single-channel electroencephalogram (EEG). Many methods attempt to extract meaningful information to provide to a learning algorithm. This method attempts to let the model extract the meaningful interpretable information by providing a smaller number of time-invariant signal filters for five frequency ranges using five CNN algorithms. A bi-directional GRU algorithm was applied to the output to incorporate time transition information. Training and tests were run on the well-known sleep-EDF-expanded database. The best results produced 97% accuracy, 93% precision, and 89% recall.

A federated semi-supervised automatic sleep staging method based on relationship knowledge sharing

Sleep staging is essential in assessing sleep quality and diagnosing sleep-related disorders, but the lack of labeled data impedes the development of automatic sleep staging models. Generally, institutions rely on semi-supervised approaches to enhance the utilization of their own unlabeled data. However, the task knowledge obtained from a limited amount of labeled data is often insufficient to guide the learning based on large amounts of unlabeled data, which may even lead to catastrophic forgetting and further degrade the performance of most existing methods. In this paper, we propose a novel strategy of building secure collaboration among multiple institutions, to achieve the implicit augmentation of labeled data and expansion of task knowledge for each participating institution by acquiring external knowledge from others. We adopt the Federated Learning (FL) to facilitate secure collaboration and propose a federated semi-supervised sleep staging method based on knowledge sharing, which enables the automatic scoring of sleep stages using only single-channel EEG data. The task knowledge in our method is contained in relationships, which exist naturally among sleep stages and can be extracted from both local labeled and unlabeled data. Furthermore, the knowledge sharing among participating institutions can be achieved by aligning the local relationships to the aggregated global relationships. Additionally, we employ prototype-contrastive learning to enhance the clarity of relationships extracted from labeled data, and propose pseudo-labeling optimization to generate reliable pseudo-labels for subsequent relationship extraction from unlabeled data. Our method is shown to be effective and outperforms compared methods in extensive experiments conducted on two publicly available datasets.

A CNN-BiLSTM Deep Learning Model for Automatic Scoring of EEG Signals

Abstract Recently, several automatic sleep stage classification methods have been proposed based on deep learning using convolutional (CNN) and recurrent (RNN) neural networks. However, the state of the art CNN methods are still complex which usually require significant time and considerable computational resources in order to set up and sufficiently train a deep CNN from scratch. This study eliminates the need to establish and train a deep CNN from scratch by leveraging a pre-trained deep architecture that has been previously trained from sufficient labeled data in a different context. A convolutional neural network (CNN) and a Bidrectional long short term memory network (BiLSTM) are integrated for automatic feature extraction and sleep stage scoring using only a singlechannel EEG signal. To demonstrate the generalizability of our results, the proposed model was evaluated using PSG records of 81 patients that were collected in different environments, through different recording hardware, and annotated with different sleep experts. The use of a single EEG source and a one-to-one classification scheme in the proposed model can allow further development towards wearable systems and online in home monitoring applications.

An explainable deep-learning model to stage sleep states in children and propose novel EEG-related patterns in sleep apnea

Automatic deep-learning models used for sleep scoring in children with obstructive sleep apnea (OSA) are perceived as black boxes, limiting their implementation in clinical settings. Accordingly, we aimed to develop an accurate and interpretable deep-learning model for sleep staging in children using single-channel electroencephalogram (EEG) recordings. We used EEG signals from the Childhood Adenotonsillectomy Trial (CHAT) dataset (n = 1637) and a clinical sleep database (n = 980). Three distinct deep-learning architectures were explored to automatically classify sleep stages from a single-channel EEG data. Gradient-weighted Class Activation Mapping (Grad-CAM), an explainable artificial intelligence (XAI) algorithm, was then applied to provide an interpretation of the singular EEG patterns contributing to each predicted sleep stage. Among the tested architectures, a standard convolutional neural network (CNN) demonstrated the highest performance for automated sleep stage detection in the CHAT test set (accuracy = 86.9% and five-class kappa = 0.827). Furthermore, the CNN-based estimation of total sleep time exhibited strong agreement in the clinical dataset (intra-class correlation coefficient = 0.772). Our XAI approach using Grad-CAM effectively highlighted the EEG features associated with each sleep stage, emphasizing their influence on the CNN's decision-making process in both datasets. Grad-CAM heatmaps also allowed to identify and analyze epochs within a recording with a highly likelihood to be misclassified, revealing mixed features from different sleep stages within these epochs. Finally, Grad-CAM heatmaps unveiled novel features contributing to sleep scoring using a single EEG channel. Consequently, integrating an explainable CNN-based deep-learning model in the clinical environment could enable automatic sleep staging in pediatric sleep apnea tests.

A Combination of Caffeine, TeaCrine, and Dynamine Improves the Neurophysiological and Performance Measures of Electronic (E)-Gamers.

Video games require precise motor skills, quick reaction times, and cognitive engagement. The tremendous growth of the electronic (e)-gaming industry has increased the demands for cognitive supplements (e.g., nootropics) to help e-athletes gain a competitive edge. The primary aim of this study was to evaluate the effects of combined caffeine + TeaCrine + Dynamine measures of neurophysiological and first-person shooter game performance in e-gamers. Using a randomized double-blinded, crossover design, we assessed the effects of an acute, single-dose treatment of caffeine (200 mg) vs. caffeine (200 mg) + TeaCrine (10 mg) + Dynamine (50 mg) (CTD) vs. Ppacebo (maltodextrin). Each participant was tested under all three conditions one week apart. Baseline and post-dose measures were tested one hour apart. Participants [n = 49 male (24.4 ±, 4.5 yr)] were amateur e-gamers who play a first-person video game for at least 10 hours/week. Gaming performance was assessed through a series of first-person shooter training games through AIMLAB (State Space Labs, Inc., New York,USA). These included Reflex Shot (RS) standard, speed, and precision. The neurophysiological activity was captured while participants played three games through a single-channel EEG. In the standard game, the caffeine and the CTD conditions shot significantly more targets relative to the placebo, and both caffeine and the CTD condition had significantly greater targets post-dose compared to pre-dose. However, both the placebo and caffeine conditions had significantly slower reaction times post-dose compared to pre-dose. In the speed game, both the caffeine and placebo conditions shot a significantly greater number of targets, while the placebo and caffeine conditions had significantly more shots post-dose compared to pre-dose. Only the CTD condition had a significant increase in total kills post-dose compared to pre-dose. In the precision game, only the CTD condition significantly improved the number of kills per second post-dose, while only the caffeine condition had more shots post-dose. EEG data collected concomitantly with game playing showed that the CTD condition resulted in significantly lower alpha power compared to the placebo condition. The CTD group also showed increased theta activity post-dose during game playing compared to both the placebo caffeine conditions. CTD appears to improve overall shooting gaming performance and neurophysiological measures of cognitive activity compared to caffeine and placebo. Collectively, these findings suggest that CTD assists with speed-accuracy tradeoffs where caffeine-only can lead to erratic play; thus, CTD may be particularly beneficial for shooting precision. The EEG data support this notion since the CTD exhibited lower alpha power suggesting increased cognitive flexibility and arousal and higher theta power suggesting greater cognitive control and decision-making under pressure.

Predictors of Low Risk for Delirium during Anesthesia Emergence.

Processed EEG is used to monitor the level of anesthesia and it has shown the potential to predict the occurrence of delirium. While emergence trajectories of relative eeg band power identified post-hoc show promising results in predicting a risk for a delirium, they are not easily transferable into an online predictive application. In this article we describe a low-resource and easily applicable method to differentiate between patients at high risk and low risk for delirium, with patients at low risk expected to show decreasing eeg power during emergence. We included data from 169 patients (median age: 61 [49, 73]), who underwent surgery with general anesthesia, maintained either with propofol, sevoflurane, or desflurane. The data were derived from a previously published study. We chose a single frontal channel and calculated the total and spectral band power from the EEG and calculated a linear regression model to observe the parameters' change during anesthesia emergence, described as slope. The slope of total power and single band power was correlated with the occurrence of delirium. Of 169 patients, 32 (19%) showed delirium. We observed that patients whose total EEG power diminished the most during emergence were less likely to screen positive for delirium in the PACU. A positive slope in total power and band power evaluated by using a regression model was associated with a higher risk-ratio (total: RR 2.83 (95%CI: 1.46 to 5.51) alpha/beta band: RR 7.79 (95%CI: 2.24 to 27.09)) for delirium. Furthermore, a negative slope in multiple bands during emergence was specific for patients without delirium and allowed to define a test for patients at low risk. In this study we developed an easily applicable exploratory method to analyze a single frontal EEG channel and to identify patterns specific for patients at low risk for delirium.

Mapping sleep's oscillatory events as a biomarker of Alzheimer's disease.

Memory-associated neural circuits produce oscillatory events including theta bursts (TBs), sleep spindles (SPs), and slow waves (SWs) in sleep electroencephalography (EEG). Changes in the"coupling" of these events may indicate early Alzheimer's disease (AD) pathogenesis. We analyzed 205 aging adults using single-channel sleep EEG, cerebrospinal fluid (CSF) AD biomarkers, and Clinical Dementia Rating® (CDR®) scale. We mapped SW-TB and SW-SP neural circuit coupling precision to amyloid positivity, cognitive impairment, and CSF AD biomarkers. Cognitive impairment correlated with lower TB spectral power in SW-TB coupling. Cognitively unimpaired, amyloid positive individuals demonstrated lower precision in SW-TB and SW-SP coupling compared to amyloid negative individuals. Significant biomarker correlations were found in oscillatory event coupling with CSF Aβ42 /Aβ40 , phosphorylated- tau181 , and total-tau. Sleep-dependent memory processing integrityin neural circuits can be measured for both SW-TB and SW-SP coupling. This breakdownassociateswith amyloid positivity, increased AD pathology, and cognitive impairment. At-home sleep EEG is a potential biomarker of neural circuits linked to memory. Circuit precision is associated with amyloid positivity in asymptomatic aging adults. Levels of CSF amyloid and tau also correlate with circuit precision in sleep EEG. Theta burst EEG power is decreased in very early mild cognitive impairment. This technique may enable inexpensive wearable EEGs for monitoring brain health.

Combining temporal and spatial attention for seizure prediction.

Approximately 1% of the world population is currently suffering from epilepsy. Successful seizure prediction is necessary for those patients. Influenced by neurons in their own and surrounding locations, the electroencephalogram (EEG) signals collected by scalp electrodes carry information of spatiotemporal interactions. Therefore, it is a great challenge to exploit the spatiotemporal information of EEG signals fully. In this paper, a new seizure prediction model called Gatformer is proposed by fusing the graph attention network (GAT) and the Transformer. The temporal and spatial attention are combined to extract EEG information from the perspective of spatiotemporal interactions. The model aims to explore the temporal dependence of single-channel EEG signals and the spatial correlations among multi-channel EEG signals. It can automatically identify the most noteworthy interaction in brain regions and achieve accurate seizure prediction. Compared with the baseline models, the performance of our model is significantly improved. The false prediction rate (FPR) on the private dataset is 0.0064/h. The average accuracy, specificity and sensitivity are 98.25%, 99.36% and 97.65%. The proposed model is comparable to the state of the arts. Experiments on different datasets show that it has good robustness and generalization performance. The high sensitivity and low FPR prove that this model has great potential to realize clinical assistance for diagnosis and treatment.

Early EEG-burst sharpness and 2-year disability in extremely preterm infants

BackgroundAutomated computational measures of EEG have the potential for large-scale application. We hypothesised that a predefined measure of early EEG-burst shape (increased burst sharpness) could predict neurodevelopmental impairment (NDI) and mental developmental index (MDI) at 2 years of age over-and-above that of brain ultrasound.MethodsWe carried out a secondary analysis of data from extremely preterm infants collected for an RCT (SafeBoosC-II). Two hours of single-channel cross-brain EEG was used to analyse burst sharpness with an automated algorithm. The co-primary outcomes were moderate-or-severe NDI and MDI. Complete data were available from 58 infants. A predefined statistical analysis was adjusted for GA, sex and no, mild–moderate, and severe brain injury as detected by cranial ultrasound.ResultsNine infants had moderate-or-severe NDI and the mean MDI was 87 ± 17.3 SD. The typical burst sharpness was low (negative values) and varied relatively little (mean –0.81 ± 0.11 SD), but the odds ratio for NDI was increased by 3.8 (p = 0.008) and the MDI was reduced by –3.2 points (p = 0.14) per 0.1 burst sharpness units increase (+1 SD) in the adjusted analysis.ConclusionThis study confirms the association between EEG-burst measures in preterm infants and neurodevelopment in childhood. Importantly, this was by a priori defined analysis.ImpactA fully automated, computational measure of EEG in the first week of life was predictive of neurodevelopmental impairment at 2 years of age.This confirms many previous studies using expert reading of EEG.Only single-channel EEG data were used, adding to the applicability.EEG was recorded by several different devices thus this measure appears to be robust to differences in electrodes, amplifiers and filters.The likelihood ratio of a positive EEG test, however, was only about 2, suggesting little immediate clinical value.

An EEG Channel Selection Framework for Driver Drowsiness Detection via Interpretability Guidance.

Drowsy driving has a crucial influence on driving safety, creating an urgent demand for driver drowsiness detection. Electroencephalogram (EEG) signal can accurately reflect the mental fatigue state and thus has been widely studied in drowsiness monitoring. However, the raw EEG data is inherently noisy and redundant, which is neglected by existing works that just use single-channel EEG data or full-head channel EEG data for model training, resulting in limited performance of driver drowsiness detection. In this paper, we are the first to propose an Interpretability-guided Channel Selection (ICS) framework for the driver drowsiness detection task. Specifically, we design a two-stage training strategy to progressively select the key contributing channels with the guidance of interpretability. We first train a teacher network in the first stage using full-head channel EEG data. Then we apply the class activation mapping (CAM) to the trained teacher model to highlight the high-contributing EEG channels and further propose a channel voting scheme to select the top N contributing EEG channels. Finally, we train a student network with the selected channels of EEG data in the second stage for driver drowsiness detection. Experiments are designed on a public dataset, and the results demonstrate that our method is highly applicable and can significantly improve the performance of cross-subject driver drowsiness detection.

Mixed-Input Deep Learning Approach to Sleep/Wake State Classification by Using EEG Signals.

Sleep stage classification plays a pivotal role in predicting and diagnosing numerous health issues from human sleep data. Manual sleep staging requires human expertise, which is occasionally prone to error and variation. In recent times, availability of polysomnography data has aided progress in automatic sleep-stage classification. In this paper, a hybrid deep learning model is proposed for classifying sleep and wake states based on a single-channel electroencephalogram (EEG) signal. The model combines an artificial neural network (ANN) and a convolutional neural network (CNN) trained using mixed-input features. The ANN makes use of statistical features calculated from EEG epochs, and the CNN operates on Hilbert spectrum images generated during each epoch. The proposed method is assessed using single-channel Pz-Oz EEG signals from the Sleep-EDF database Expanded. The classification performance on four randomly selected individuals shows that the proposed model can achieve accuracy of around 96% in classifying between sleep and wake states from EEG recordings.

BTCRSleep: a boundary temporal context refinement-based fully convolutional network for sleep staging with single-channel EEG

Objective. Sleep staging studies on single-channel EEG mainly exploit deep learning methods that combine convolutional neural networks (CNNs) and recurrent neural networks. However, when typical brain waves (such as K-complexes or sleep spindles) that identify sleep stages span two epochs, the abstract process of a CNN extracting features from each sleep stage may cause the loss of boundary context information. This study attempts to capture the boundary context, which contains the characteristics of brain waves during sleep stage transition, to improve the performance of sleep staging. Approach. In this paper we propose a fully convolutional network with boundary temporal context refinement, called BTCRSleep (Boundary Temporal Context Refinement Sleep). The boundary temporal context refinement module refines the boundary information on sleep stages on the basis of extracting multi-scale temporal dependences between epochs and enhances the abstract capability of the boundary temporal context. In addition, we design a class-aware data augmentation method to effectively learn the boundary temporal context between the minority class and other sleep stages. Main results. We evaluate the performance of our proposed network using four public datasets: the 2013 version of Sleep-EDF Expanded (SEDF), the 2018 version of Sleep-EDF Expanded (SEDFX), the Sleep Heart Health Study (SHHS) and CAP Sleep Database (CAP). The evaluation results on the four datasets showed that our model obtains the best total accuracy and kappa score compared with state-of-the-art methods. On average, accuracies of 84.9% in SEDF, 82.9% in SEDFX, 85.2% in SHHS and 76.9% in CAP are obtained under subject-independent cross-validation. We demonstrate that the boundary temporal context contributes to the improvement in capturing the temporal dependences across different epochs.