Single-channel Electroencephalogram Research Articles

Deep learning-based stress detection for daily life use using single-channel EEG and GSR in a virtual reality interview paradigm.

This research aims to establish a practical stress detection framework by integrating physiological indicators and deep learning techniques. Utilizing a virtual reality (VR) interview paradigm mirroring real-world scenarios, our focus is on classifying stress states through accessible single-channel electroencephalogram (EEG) and galvanic skin response (GSR) data. Thirty participants underwent stress-inducing VR interviews, with biosignals recorded for deep learning models. Five convolutional neural network (CNN) architectures and one Vision Transformer model, including a multiple-column structure combining EEG and GSR features, showed heightened predictive capabilities and an enhanced area under the receiver operating characteristic curve (AUROC) in stress prediction compared to single-column models. Our experimental protocol effectively elicited stress responses, observed through fluctuations in stress visual analogue scale (VAS), EEG, and GSR metrics. In the single-column architecture, ResNet-152 excelled with a GSR AUROC of 0.944 (±0.027), while the Vision Transformer performed well in EEG, achieving peak AUROC values of 0.886 (±0.069) respectively. Notably, the multiple-column structure, based on ResNet-50, achieved the highest AUROC value of 0.954 (±0.018) in stress classification. Through VR-based simulated interviews, our study induced social stress responses, leading to significant modifications in GSR and EEG measurements. Deep learning models precisely classified stress levels, with the multiple-column strategy demonstrating superiority. Additionally, discreetly placing single-channel EEG measurements behind the ear enhances the convenience and accuracy of stress detection in everyday situations.

Neural Envelope Processing at Low Frequencies Predicts Speech Understanding of Children With Hearing Loss in Noise and Reverberation.

Children with hearing loss experience greater difficulty understanding speech in the presence of noise and reverberation relative to their normal hearing peers despite provision of appropriate amplification. The fidelity of fundamental frequency of voice (f0) encoding-a salient temporal cue for understanding speech in noise-could play a significant role in explaining the variance in abilities among children. However, the nature of deficits in f0 encoding and its relationship with speech understanding are poorly understood. To this end, we evaluated the influence of frequency-specific f0 encoding on speech perception abilities of children with and without hearing loss in the presence of noise and/or reverberation. In 14 school-aged children with sensorineural hearing loss fitted with hearing aids and 29 normal hearing peers, envelope following responses (EFRs) were elicited by the vowel /i/, modified to estimate f0 encoding in low (<1.1 kHz) and higher frequencies simultaneously. EFRs to /i/ were elicited in quiet, in the presence of speech-shaped noise at +5 dB signal to noise ratio, with simulated reverberation time of 0.62 sec, as well as both noise and reverberation. EFRs were recorded using single-channel electroencephalogram between the vertex and the nape while children watched a silent movie with captions. Speech discrimination accuracy was measured using the University of Western Ontario Distinctive Features Differences test in each of the four acoustic conditions. Stimuli for EFR recordings and speech discrimination were presented monaurally. Both groups of children demonstrated a frequency-dependent dichotomy in the disruption of f0 encoding, as reflected in EFR amplitude and phase coherence. Greater disruption (i.e., lower EFR amplitudes and phase coherence) was evident in EFRs elicited by low frequencies due to noise and greater disruption was evident in EFRs elicited by higher frequencies due to reverberation. Relative to normal hearing peers, children with hearing loss demonstrated: (a) greater disruption of f0 encoding at low frequencies, particularly in the presence of reverberation, and (b) a positive relationship between f0 encoding at low frequencies and speech discrimination in the hardest listening condition (i.e., when both noise and reverberation were present). Together, these results provide new evidence for the persistence of suprathreshold temporal processing deficits related to f0 encoding in children despite the provision of appropriate amplification to compensate for hearing loss. These objectively measurable deficits may underlie the greater difficulty experienced by children with hearing loss.

The relationship between thermal comfort, performance, and electroencephalogram during cognitive tests at normal indoor temperatures in summer

Electroencephalograms have been proposed as potential physiological indicators to gauge the thermal comfort and cognitive performance of individuals in indoor environments, yet further experimental validation is required. In this study, a simulated experiment was conducted in a climatic room during summer, with temperatures set at 24 °C, 26 °C, 28 °C, 30 °C, and 32 °C. Twenty participants, attired in typical summer clothing (short sleeves and long pants totaling 0.39 clo), undertook cognitive tests assessing short-term memory and concentration while seated, alongside providing thermal comfort ratings. Electroencephalogram signals were continuously recorded from 14 electrode channels across the brain during the second cognitive test unit, spanning the 70th to 90th minute of experimental exposure. The relative power of electroencephalogram signals was computed and analyzed in conjunction with thermal comfort voting, revealing a significant impact of indoor temperature variations on α-band activity in the occipital lobe, with peak α-band relative power observed at 26 °C. Moreover, the whole-brain α-band relative power was significantly higher in states of thermal comfort compared to discomfort, particularly concentrated in the occipital lobe. Linear correlation analysis between cognitive test results and electroencephalogram signals indicated significant associations between the relative power of whole-brain β, δ, and θ bands and scores from the concentration-focused Flanker test. Channels O1 and P8 emerged as potential candidates for single-channel electroencephalogram signals, offering insights into thermal comfort status and cognitive performance related to concentration.

Research on Ocular Artifacts Removal from Single-Channel Electroencephalogram Signals in Obstructive Sleep Apnea Patients Based on Support Vector Machine, Improved Variational Mode Decomposition, and Second-Order Blind Identification.

The electroencephalogram (EEG) has recently emerged as a pivotal tool in brain imaging analysis, playing a crucial role in accurately interpreting brain functions and states. To address the problem that the presence of ocular artifacts in the EEG signals of patients with obstructive sleep apnea syndrome (OSAS) severely affects the accuracy of sleep staging recognition, we propose a method that integrates a support vector machine (SVM) with genetic algorithm (GA)-optimized variational mode decomposition (VMD) and second-order blind identification (SOBI) for the removal of ocular artifacts from single-channel EEG signals. The SVM is utilized to identify artifact-contaminated segments within preprocessed single-channel EEG signals. Subsequently, these signals are decomposed into variational modal components across different frequency bands using the GA-optimized VMD algorithm. These components undergo further decomposition via the SOBI algorithm, followed by the computation of their approximate entropy. An approximate entropy threshold is set to identify and remove components laden with ocular artifacts. Finally, the signal is reconstructed using the inverse SOBI and VMD algorithms. To validate the efficacy of our proposed method, we conducted experiments utilizing both simulated data and real OSAS sleep EEG data. The experimental results demonstrate that our algorithm not only effectively mitigates the presence of ocular artifacts but also minimizes EEG signal distortion, thereby enhancing the precision of sleep staging recognition based on the EEG signals of OSAS patients.

Validation of a sleep staging classification model for healthy adults based on two combinations of a single-channel EEG headband and wrist actigraphy.

The aim of this study was to develop a sleep staging classification model capable of accurately performing on different wearable devices. Twenty-three healthy participants underwent a full-night type I polysomnography and used two device combinations: (A) flexible single-channel electroencephalogram (EEG) headband + actigraphy (n = 12) and (B) rigid single-channel EEG headband + actigraphy (n = 11). The signals were segmented into 30-second epochs according to polysomnographic stages (scored by a board-certified sleep technologist; model ground truth) and 18 frequency and time features were extracted. The model consisted of an ensemble of bagged decision trees. Bagging refers to bootstrap aggregation to reduce overfitting and improve generalization. To evaluate the model, a training dataset under 5-fold cross-validation and an 80-20% dataset split was used. The headbands were also evaluated without the actigraphy feature. Participants also completed a usability evaluation (comfort, pain while sleeping, and sleep disturbance). Combination A had an F1-score of 98.4% and the flexible headband alone of 97.7% (error rate for N1: combination A = 9.8%; flexible headband alone = 15.7%). Combination B had an F1-score of 96.9% and the rigid headband alone of 95.3% (error rate for N1: combination B = 17.0%; rigid headband alone = 27.7%); in both, N1 was more confounded with N2. We developed an accurate sleep classification model based on a single-channel EEG device, and actigraphy was not an important feature of the model. Both headbands were found to be useful, with the rigid one being more disruptive to sleep. Future research can improve our results by applying the developed model in a population with sleep disorders. Registry: ClinicalTrials.gov; Name: Actigraphy, Wearable EEG Band and Smartphone for Sleep Staging; URL: https://clinicaltrials.gov/study/NCT04943562; Identifier: NCT04943562. Melo MC, Vallim JRS, Garbuio S, etal. Validation of a sleep staging classification model for healthy adults based on 2 combinations of a single-channel EEG headband and wrist actigraphy. J Clin Sleep Med. 2024;20(6):983-990.

Strength of ensemble learning in automatic sleep stages classification using single-channel EEG and ECG signals.

Healthy sleep plays an essential role in human daily life. Classification of sleep stages is a crucial tool for assisting physicians in diagnosing and treating sleep disorders. In this study, a strong ensemble learning model is proposed to enhance the ability of classification models in accurate sleep staging, particularly in multi-class classification. We asserted that high-accuracy sleep classification is achievable using only single-channel electroencephalogram (EEG) and electrocardiogram (ECG) by combining their best-extractable features in the time and frequency domains we recommended. More importantly, the superiority of the recommended method, which is the simultaneous use of stacking and bagging, over conventional machine learning classifiers in sleep staging was demonstrated, using the MIT-BIH Polysomnographic and Sleep-EDF expanded databases. Finally, K-fold cross-validation was used to fairly estimate these models. The best mean test accuracy rates for distinguishing between two classes of "sleep vs. wake," "rapid vs. non-rapid eye movement," and "deep vs. light sleep," were obtained 99.93%, 99.64%, and 99.69%, respectively. Furthermore, our proposed method achieved accuracies of 97.14%, 95.18%, 92.7%, and 85.64% for separating three, four, five, and six sleep classes, respectively. Compared to recent studies, our method outperforms other sleep stage classification schemes, especially in multi-class staging.

Comparison of Polysomnography, Single-Channel Electroencephalogram, Fitbit, and Sleep Logs in Patients With Psychiatric Disorders: Cross-Sectional Study.

Sleep disturbances are core symptoms of psychiatric disorders. Although various sleep measures have been developed to assess sleep patterns and quality of sleep, the concordance of these measures in patients with psychiatric disorders remains relatively elusive. This study aims to examine the degree of agreement among 3 sleep recording methods and the consistency between subjective and objective sleep measures, with a specific focus on recently developed devices in a population of individuals with psychiatric disorders. We analyzed 62 participants for this cross-sectional study, all having data for polysomnography (PSG), Zmachine, Fitbit, and sleep logs. Participants completed questionnaires on their symptoms and estimated sleep duration the morning after the overnight sleep assessment. The interclass correlation coefficients (ICCs) were calculated to evaluate the consistency between sleep parameters obtained from each instrument. Additionally, Bland-Altman plots were used to visually show differences and limits of agreement for sleep parameters measured by PSG, Zmachine, Fitbit, and sleep logs. The findings indicated a moderate agreement between PSG and Zmachine data for total sleep time (ICC=0.46; P<.001), wake after sleep onset (ICC=0.39; P=.002), and sleep efficiency (ICC=0.40; P=.006). In contrast, Fitbit demonstrated notable disagreement with PSG (total sleep time: ICC=0.08; wake after sleep onset: ICC=0.18; sleep efficiency: ICC=0.10) and exhibited particularly large discrepancies from the sleep logs (total sleep time: ICC=-0.01; wake after sleep onset: ICC=0.05; sleep efficiency: ICC=-0.02). Furthermore, subjective and objective concordance among PSG, Zmachine, and sleep logs appeared to be influenced by the severity of the depressive symptoms and obstructive sleep apnea, while these associations were not observed between the Fitbit and other sleep instruments. Our study results suggest that Fitbit accuracy is reduced in the presence of comorbid clinical symptoms. Although user-friendly, Fitbit has limitations that should be considered when assessing sleep in patients with psychiatric disorders.

DeltaScan for the Assessment of Acute Encephalopathy and Delirium in ICU and non-ICU Patients, a Prospective Cross-Sectional Multicenter Validation Study

To measure the diagnostic accuracy of DeltaScan: a portable real-time brain state monitor for identifying delirium, a manifestation of acute encephalopathy (AE) detectable by polymorphic delta activity (PDA) in single-channel electroencephalograms (EEGs). Prospective cross-sectional study. Six Intensive Care Units (ICU's) and 17 non-ICU departments, including a psychiatric department across 10 Dutch hospitals. 494 patients, median age 75 (IQR:64-87), 53% male, 46% in ICUs, 29% delirious. DeltaScan recorded 4-minute EEGs, using an algorithm to select the first 96 seconds of artifact-free data for PDA detection. This algorithm was trained and calibrated on two independent datasets. Initial validation of the algorithm for AE involved comparing its output with an expert EEG panel's visual inspection. The primary objective was to assess DeltaScan's accuracy in identifying delirium against a delirium expert panel's consensus. DeltaScan had a 99% success rate, rejecting 6 of the 494 EEG's due to artifacts. Performance showed and an Area Under the Receiver Operating Characteristic Curve (AUC) of 0.86 (95% CI: 0.83-0.90) for AE (sensitivity: 0.75, 95%CI=0.68-0.81, specificity: 0.87 95%CI=0.83-0.91. The AUC was 0.71 for delirium (95%CI=0.66-0.75, sensitivity: 0.61 95%CI=0.52-0.69, specificity: 72, 95%CI=0.67-0.77). Our validation aim was an NPV for delirium above 0.80 which proved to be 0.82 (95%CI: 0.77-0.86). Among 84 non-delirious psychiatric patients, DeltaScan differentiated delirium from other disorders with a 94% (95%CI: 87-98%) specificity. DeltaScan can diagnose AE at bedside and shows a clear relationship with clinical delirium. Further research is required to explore its role in predicting delirium-related outcomes.

SleePyCo: Automatic sleep scoring with feature pyramid and contrastive learning

Automatic sleep scoring is essential for the diagnosis and treatment of sleep disorders and enables longitudinal sleep tracking in home environments. Conventionally, learning-based automatic sleep scoring on single-channel electroencephalogram (EEG) is actively studied because obtaining multi-channel signals during sleep is difficult. However, learning representation from raw EEG signals is challenging owing to the following issues: (1) sleep-related EEG patterns occur on different temporal and frequency scales and (2) sleep stages share similar EEG patterns. To address these issues, we propose an automatic Sleep scoring framework that incorporates (1) a feature Pyramid and (2) supervised Contrastive learning, named SleePyCo. For the feature pyramid, we propose a backbone network named SleePyCo-backbone to consider multiple feature sequences on different temporal and frequency scales. Supervised contrastive learning allows the network to extract class discriminative features by minimizing the distance between intra-class features and simultaneously maximizing that between inter-class features. Comparative analyses on four public datasets demonstrate that SleePyCo consistently outperforms existing frameworks based on single-channel EEG. Extensive ablation experiments show that SleePyCo exhibited an enhanced overall performance, with significant improvements in discrimination between sleep stages, especially for N1 and rapid eye movement (REM). Source code is available at https://github.com/gist-ailab/SleePyCo.

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.

MultiChannelSleepNet: A Transformer-Based Model for Automatic Sleep Stage Classification With PSG.

Automatic sleep stage classification plays an essential role in sleep quality measurement and sleep disorder diagnosis. Although many approaches have been developed, most use only single-channel electroencephalogram signals for classification. Polysomnography (PSG) provides multiple channels of signal recording, enabling the use of the appropriate method to extract and integrate the information from different channels to achieve higher sleep staging performance. We present a transformer encoder-based model, MultiChannelSleepNet, for automatic sleep stage classification with multichannel PSG data, whose architecture is implemented based on the transformer encoder for single-channel feature extraction and multichannel feature fusion. In a single-channel feature extraction block, transformer encoders extract features from time-frequency images of each channel independently. Based on our integration strategy, the feature maps extracted from each channel are fused in the multichannel feature fusion block. Another set of transformer encoders further capture joint features, and a residual connection preserves the original information from each channel in this block. Experimental results on three publicly available datasets demonstrate that our method achieves higher classification performance than state-of-the-art techniques. MultiChannelSleepNet is an efficient method to extract and integrate the information from multichannel PSG data, which facilitates precision sleep staging in clinical 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.

EEG-based sleep staging via self-attention based capsule network with Bi-LSTM model

Sleep staging via electroencephalogram is essential for determining the quality of sleep. Manual sleep staging is expensive and time-consuming. Recently, many deep learning-based sleep staging methods are demonstrated to outperform traditional methods. However, most methods do not fully exploit the temporal correlation between features of electroencephalogram signals. In this paper, we propose a self-attention routing-based capsule network with bi-directional long short-term memory model to extract more discriminative features from electroencephalogram signals and improve the accuracy of sleep staging. First, a convolutional neural network is used to extract salient features from the electroencephalogram signal. Second, to learn the transition rules between different sleep epochs, a bi-directional long short-term memory is used to capture the temporal dependence between the encoded electroencephalogram signals. Finally, to fully explore the temporal correlation between the features from the electroencephalogram signals, a self-attention routing-based capsule network is utilized to recode the importance based on the intrinsic temporal similarity of electroencephalogram signals. We evaluated our model by two different single-channel electroencephalogram signals (i.e., Fpz-Cz and Pz-Oz electroencephalogram channels) from two public sleep datasets, named Sleep-EDF-39 and Sleep-EDF-153. Our overall accuracies on the Sleep-EDF-39 and Sleep-EDF-153 datasets are 85.8% and 83.4%, with a kappa of 0.8 and 0.77, respectively. The results show that our proposed method achieves the state-of-the-art level of sleep staging using a single-channel electroencephalogram and offers the possibility of widespread application of capsule networks for sleep staging.

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.

Automatic sleep staging model based on single channel electroencephalogram signal

Sleep staging is the basis for solving sleep problems. There's an upper limit for the classification accuracy of sleep staging models based on single-channel electroencephalogram (EEG) data and features. To address this problem, this paper proposed an automatic sleep staging model that mixes deep convolutional neural network (DCNN) and bi-directional long short-term memory network (BiLSTM). The model used DCNN to automatically learn the time-frequency domain features of EEG signals, and used BiLSTM to extract the temporal features between the data, fully exploiting the feature information contained in the data to improve the accuracy of automatic sleep staging. At the same time, noise reduction techniques and adaptive synthetic sampling were used to reduce the impact of signal noise and unbalanced data sets on model performance. In this paper, experiments were conducted using the Sleep-European Data Format Database Expanded and the Shanghai Mental Health Center Sleep Database, and achieved an overall accuracy rate of 86.9% and 88.9% respectively. When compared with the basic network model, all the experimental results outperformed the basic network, further demonstrating the validity of this paper's model, which can provide a reference for the construction of a home sleep monitoring system based on single-channel EEG signals.

Drivers’ drowsiness detection based on an optimized random forest classification and single-channel electroencephalogram

&lt;p&gt;The state of functioning (posture) of a driver at the wheel of a car involves a complex set of psychological, physiological, and physical parameters. This combination induces fatigue, which manifests itself in repeated yawning, stinging eyes, a frozen gaze, a stiff and painful neck, back pain, and other signs. The driver may fight fatigue for a few moments, but it inevitably leads to drowsiness, periods of micro-sleep, and then falling asleep. At the first signs of drowsiness, the risk of an accident becomes immense. In Morocco, drowsiness at the wheel is the cause of 1/3 of fatal accidents on the freeways. Thus, in this paper, a new hybrid data analysis and an efficient machine learning algorithm are designed to detect the drowsiness of drivers who spend most of their time behind the wheel over long distances (older than 35 years). This analysis is based on a single channel of electroencephalogram (EEG) recordings using time, frequency fast Fourier transform (FFT), and power spectral density (PSD) analysis. To distinguish between the two states of alertness and drowsiness, several features were extracted from each domain (time, FFT, and PSD), and subjected to different classifier architectures to conduct a general comparison and achieve the highest detection accuracy (98.5%) and best time consumption (13 milliseconds).&lt;/p&gt;

Anesthesia and the Merry-go-round of Information in the Brain.

Anesthesia and the Merry-go-round of Information in the Brain.

Design of a Sleep Modulation System with FPGA-Accelerated Deep Learning for Closed-loop Stage-Specific In-Phase Auditory Stimulation.

Closed-loop sleep modulation is an emerging research paradigm to treat sleep disorders and enhance sleep benefits. However, two major barriers hinder the widespread application of this research paradigm. First, subjects often need to be wire-connected to rack-mount instrumentation for data acquisition, which negatively affects sleep quality. Second, conventional real-time sleep stage classification algorithms give limited performance. In this work, we conquer these two limitations by developing a sleep modulation system that supports closed-loop operations on the device. Sleep stage classification is performed using a lightweight deep learning (DL) model accelerated by a low-power field-programmable gate array (FPGA) device. The DL model uses a single channel electroencephalogram (EEG) as input. Two convolutional neural networks (CNNs) are used to capture general and detailed features, and a bidirectional long-short-term memory (LSTM) network is used to capture time-variant sequence features. An 8-bit quantization is used to reduce the computational cost without compromising performance. The DL model has been validated using a public sleep database containing 81 subjects, achieving a state-of-the-art classification accuracy of 85.8% and a F1-score of 79%. The developed model has also shown the potential to be generalized to different channels and input data lengths. Closed-loop in-phase auditory stimulation has been demonstrated on the test bench.

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

Sleep stage detection from polysomnography (PSG) recordings is a widely used method of monitoring sleep quality. Despite significant progress in the development of machine-learning (ML)-based and deep-learning (DL)-based automatic sleep stage detection schemes focusing on single-channel PSG data, such as single-channel electroencephalogram (EEG), electrooculogram (EOG), and electromyogram (EMG), developing a standard model is still an active subject of research. Often, the use of a single source of information suffers from data inefficiency and data-skewed problems. Instead, a multi-channel input-based classifier can mitigate the aforementioned challenges and achieve better performance. However, it requires extensive computational resources to train the model, and, hence, a tradeoff between performance and computational resources cannot be ignored. In this article, we aim to introduce a multi-channel, more specifically a four-channel, convolutional bidirectional long short-term memory (Bi-LSTM) network that can effectively exploit spatiotemporal features of data collected from multiple channels of the PSG recording (e.g., EEG Fpz-Cz, EEG Pz-Oz, EOG, and EMG) for automatic sleep stage detection. First, a dual-channel convolutional Bi-LSTM network module has been designed and pre-trained utilizing data from every two distinct channels of the PSG recording. Subsequently, we have leveraged the concept of transfer learning circuitously and have fused two dual-channel convolutional Bi-LSTM network modules to detect sleep stages. In the dual-channel convolutional Bi-LSTM module, a two-layer convolutional neural network has been utilized to extract spatial features from two channels of the PSG recordings. These extracted spatial features are subsequently coupled and given as input at every level of the Bi-LSTM network to extract and learn rich temporal correlated features. Both Sleep EDF-20 and Sleep EDF-78 (expanded version of Sleep EDF-20) datasets are used in this study to evaluate the result. The model that includes an EEG Fpz-Cz + EOG module and an EEG Fpz-Cz + EMG module can classify sleep stage with the highest value of accuracy (ACC), Kappa (Kp), and F1 score (e.g., 91.44%, 0.89, and 88.69%, respectively) on the Sleep EDF-20 dataset. On the other hand, the model consisting of an EEG Fpz-Cz + EMG module and an EEG Pz-Oz + EOG module shows the best performance (e.g., the value of ACC, Kp, and F1 score are 90.21%, 0.86, and 87.02%, respectively) compared to other combinations for the Sleep EDF-78 dataset. In addition, a comparative study with respect to other existing literature has been provided and discussed in order to exhibit the efficacy of our proposed model.