Application of electroencephalogram (EEG) in the study of the influence of different contents of alcohol and Baijiu on brain perception.

Alcoholic beverages flavour is complex and unique with different alcohol content, and the application of flavour perception could improve the objectivity of flavour evaluation. This study utilized electroencephalogram (EEG) to assess brain reactions to alcohol percentages (5 %-53 %) and Baijiu's complex flavours. The findings demonstrate the brain's proficiency in discerning between alcohol concentrations, evidenced by increasing physiological signal strength in tandem with alcohol content. When contrasted with alcohol solutions of equivalent concentrations, Baijiu prompts a more significant activation of brain signals, underscoring EEG's capability to detect subtleties due to flavour complexity. Additionally, the study reveals notable correlations, with δ and α wave intensities escalating in response to alcohol stimulation, coupled with substantial activation in the frontal, parietal, and right temporal regions. These insights verify the efficacy of EEG in charting the brain's engagement with alcoholic flavours, setting the stage for more detailed exploration into the neural encoding of these sensory experiences.

Subject-independent auditory spatial attention detection based on brain topology modeling and feature distribution alignment.

Auditory spatial attention detection (ASAD) seeks to determine which speaker in a surround sound field a listener is focusing on based on the one's brain biosignals. Although existing studies have achieved ASAD from a single-trial electroencephalogram (EEG), the huge inter-subject variability makes them generally perform poorly in cross-subject scenarios. Besides, most ASAD methods do not take full advantage of topological relationships between EEG channels, which are crucial for high-quality ASAD. Recently, some advanced studies have introduced graph-based brain topology modeling into ASAD, but how to calculate edge weights in a graph to better capture actual brain connectivity is worthy of further investigation. To address these issues, we propose a new ASAD method in this paper. First, we model a multi-channel EEG segment as a graph, where differential entropy serves as the node feature, and a static adjacency matrix is generated based on inter-channel mutual information to quantify brain functional connectivity. Then, different subjects' EEG graphs are encoded into a shared embedding space through a total variation graph neural network. Meanwhile, feature distribution alignment based on multi-kernel maximum mean discrepancy is adopted to learn subject-invariant patterns. Note that we align EEG embeddings of different subjects to reference distributions rather than align them to each other for the purpose of privacy preservation. A series of experiments on open datasets demonstrate that the proposed model outperforms state-of-the-art ASAD models in cross-subject scenarios with relatively low computational complexity, and feature distribution alignment improves the generalizability of the proposed model to a new subject.

Can electroencephalography (EEG) identify the different dimensions of pain in fibromyalgia? A pilot study.

BACKGROUND: Electroencephalography (EEG) is a promising tool for identifying the physiological biomarkers of fibromyalgia (FM). Evidence suggests differences in power band and density between individuals with FM and healthy controls. EEG changes appear to be associated with pain intensity; however, their relationship with the quality of pain has not been examined. We aimed to investigate whether abnormal EEG in the frontal and central points of the 10-20 EEG system in individuals with FM is associated with pain's sensory-discriminative and affective-motivational dimensions. The association between EEG and the two dimensions of emotional disorders (depression and anxiety) was also investigated. METHODS: In this cross-sectional pilot study, pain experience (pain rating index [PRI]) and two dimensions of emotional disorders (depression and anxiety) were assessed using the McGill Pain Questionnaire (PRI-sensory and PRI-affective) and Hospital Anxiety and Depression Scale (HADS), respectively. In quantitative EEG analysis, the relative spectral power of each frequency band (delta, theta, alpha, and beta) was identified in the frontal and central points during rest. RESULTS: A negative correlation was found between the relative spectral power for the delta bands in the frontal (r= -0.656; p = 0.028) and central points (r= -0.624; p = 0.040) and the PRI-affective scores. A positive correlation was found between the alpha bands in the frontal (r = 0.642; p = 0.033) and central points (r = 0.642; p = 0.033) and the PRI-affective scores. A negative correlation between the delta bands in the central points and the anxiety subscale of the HADS (r = -0.648; p = 0.031) was detected. CONCLUSION: The affective-motivational dimension of pain and mood disorders may be related to abnormal patterns of electrical activity in patients with FM. TRIAL REGISTRATION: Retrospectively registered on ClinicalTrials.gov (NCT05962658).

Graph-based Analysis to Predict Repetitive Transcranial Magnetic Stimulation Treatment Response in Patients With Major Depressive Disorder Using EEG Signals.

Introduction: Repetitive transcranial magnetic stimulation (rTMS) is a non-pharmacological treatment for drug-resistant major depressive disorder (MDD) patients. Since the success rate of rTMS treatment is about 50%-55%, it is essential to predict the treatment outcome before starting based on electroencephalogram (EEG) signals, leading to identifying effective biomarkers and reducing the burden of health care centers. Methods: To this end, pretreatment EEG data with 19 channels in the resting state from 34 drug-resistant MDD patients were recorded. Then, all patients received 20 sessions of rTMS treatment, and a reduction of at least 50% in the total beck depression inventory (BDI-II) score before and after the rTMS treatment was defined as a reference. In the current study, effective brain connectivity features were determined by the direct directed transfer function (dDTF) method from patients' pretreatment EEG data in all frequency bands separately. Then, the brain functional connectivity patterns were modeled as graphs by the dDTF method and examined with the local graph theory indices, including degree, out-degree, in-degree, strength, out-strength, in-strength, and betweenness centrality. Results: The results indicated that the betweenness centrality index in the Fp2 node and the δ frequency band are the best biomarkers, with the highest area under the receiver operating characteristic curve value of 0.85 for predicting the rTMS treatment outcome in drug-resistant MDD patients. Conclusion: The proposed method investigated the significant biomarkers that can be used to predict the rTMS treatment outcome in drug-resistant MDD patients and help clinical decisions.

Prefrontal, Frontal, and Temporal Theta EEG Asymmetries and Self-Reports of Emotional Regulation.

Previous studies have shown that right-sided frontal alpha asymmetry (fAA) is an electroencephalography (EEG) marker for negatively valenced emotions and a marker for negative self-perceptions of a person's psychosocial interactions. Alpha activity is affected by the changes in visual stimulation associated with eye-opening and eye-closing; theta activity is not so affected. Therefore, this analysis investigates the relationship between an individual's theta asymmetry and self-perceptions of their psychosocial interactions. We used quantitative electroencephalographic (qEEG) data from eight right-handed male medical students aged between 19 and 38 years, recorded under eyes-open (EO) and eyes-closed (EC) conditions. Significant correlations were found between self-reported measures of psychosocial interactions via the Interactive Self-Report Inventory (ISI). The main finding was that greater left-sided frontal temporal asymmetry (fTA) under both EO and EC conditions was associated with lower "regulated" ISI scores and lower "dependent" ISI scores. Greater left-sided temporal theta asymmetry (tTA), under EC conditions, was associated with higher "anxious" ISI scores. Greater left-sided prefrontal theta symmetry (pfTA), under EO conditions, was associated with lower "relaxed" ISI scores. These findings suggest that theta asymmetries in the frontal, prefrontal, and temporal cortices may be indicative of negative emotional states. The results of this study underscore the potential of pfTA, fTA, and tTA to be used as biomarkers for cognitive-emotional balance. The implications for mental health interventions, particularly personalized therapeutic approaches, are significant.

History and Evolution of the Electroencephalogram.

This paper summarizes the history and evolution of the electroencephalogram (EEG). The EEG, used to record the electrical activity of the brain, is a pivotal tool in neuroscience and medicine. Its history and evolution reflect significant advancements in our understanding of brain function and our ability to diagnose and treat neurological conditions. This tool has revolutionized our understanding of the brain's electrical activity and is the cornerstone for the diagnosis and treatment of epilepsy and related disorders. The evolution of the EEG from early experimental observations to sophisticated modern applications highlights the profound progress in our ability to monitor and interpret brain activity. The EEG remains an invaluable tool in clinical and research settings, continually evolving with technological advancements to expand our understanding of the human brain. This review traces the journey of this iconic tool.

Improving classification performance of motor imagery BCI through EEG data augmentation with conditional generative adversarial networks.

In brain-computer interface (BCI), building accurate electroencephalogram (EEG) classifiers for specific mental tasks is critical for BCI performance. The classifiers are developed by machine learning (ML) and deep learning (DL) techniques, requiring a large dataset for training to build reliable and accurate models. However, collecting large enough EEG datasets is difficult due to intra-/inter-subject variabilities and experimental costs. This leads to the data scarcity problem, which causes overfitting issues to training samples, resulting in reducing generalization performance. To solve the EEG data scarcity problem and improve the performance of the EEG classifiers, we propose a novel EEG data augmentation (DA) framework using conditional generative adversarial networks (cGANs). An experimental study is implemented with two public EEG datasets, including motor imagery (MI) tasks (BCI competition IV IIa and III IVa), to validate the effectiveness of the proposed EEG DA method for the EEG classifiers. To evaluate the proposed cGAN-based DA method, we tested eight EEG classifiers for the experiment, including traditional MLs and state-of-the-art DLs with three existing EEG DA methods. Experimental results showed that most DA methods with proper DA proportion in the training dataset had higher classification performances than without DA. Moreover, applying the proposed DA method showed superior classification performance improvement than the other DA methods. This shows that the proposed method is a promising EEG DA method for enhancing the performances of the EEG classifiers in MI-based BCIs.

SleepGCN: A transition rule learning model based on Graph Convolutional Network for sleep staging.

BACKGROUND AND OBJECTIVE: Automatic sleep staging is essential for assessing and diagnosing sleep disorders, serving millions of people who suffer from them. Numerous sleep staging models have been proposed recently, but most of them have not fully explored the sleep transition rules that are essential for sleep experts to identify sleep stages. Therefore, one objective of this paper is to develop an automatic sleep staging model to capture the transition rules between sleep stages. METHODS: In this paper, we propose a novel sleep staging model named SleepGCN. It utilizes the deep features of electroencephalogram (EEG) and electrooculogram (EOG) signals extracted by the sleep representation learning (SRL) module, in conjunction with the transition rules learned by the sleep transition rule learning (STRL) module to identify sleep stages. Specifically, the SRL module utilizes the residual network (ResNet) and Long Short Term Memory (LSTM) structure to capture the deep time-invariant features and temporal information of each sleep stage from the two-channel EEG-EOG, and then applies a feature enhancement block to obtain the refined features. The STRL module employs a Graph Convolutional Network (GCN) and a transition rule matrix to capture transition rules between sleep stages based on the sequence labels of the input signals. RESULTS: We evaluate SleepGCN on five public datasets: SleepEDF-20, SleepEDF-78, SHHS, DOD-H and DOD-O. Overall, SleepGCN achieves an accuracy of 89.70%, 87.70%, 86.16%, 82.07%, and 81.20%, alongside a macro-average F1-score of 85.20%, 82.70%, 77.69%, 72.44%, and 72.93% across these datasets, respectively. CONCLUSIONS: The results achieved by our proposed model are much better than those of all other compared models. The ablation study validates the contributions of the SRL and STRL modules proposed in SleepGCN to the sleep staging tasks. Additionally, it shows that the sleep staging model using two-channel EEG-EOG outperforms the model using single-channel EEG or EOG. Overall, SleepGCN is an effective solution for sleep staging using two-channel EEG-EOG.

The impact of isoflurane anesthesia on brain metabolism in mice: An MRI and electroencephalography study.

Isoflurane is one of the most widely used anesthetic agents in rodent imaging studies. However, the impact of isoflurane on brain metabolism has not been fully characterized to date, primarily due to a scarcity of noninvasive technologies to quantitatively measure the brain's metabolic rate in vivo. In this study, using noncontrast MRI techniques, we dynamically measured cerebral metabolic rate of oxygen (CMRO2) under varying doses of isoflurane anesthesia in mice. Concurrently, systemic parameters of heart and respiration rates were recorded alongside CMRO2. Additionally, electroencephalogram (EEG) recording was used to identify changes in neuronal activities under the same anesthetic regimen employed in the MRI experiments. We found suppression of the CMRO2 by isoflurane in a dose-dependent manner, concomitant with a diminished high-frequency EEG activity. The degree of metabolic suppression by isoflurane was strongly correlated with the respiration rate, which offers a potential approach to calibrate CMRO2 measurements. Furthermore, the metabolic level associated with neural responses of the somatosensory and motor cortices in mice was estimated as 308.2 µmol/100 g/min. These findings may facilitate the integration of metabolic parameters into future studies involving animal disease models and anesthesia usage.

Do changes in cerebral blood flow modulate the amplitudes of P300 during cognitive task?

A single session of aerobic or resistance training transiently enhances cognitive function, making it a valuable strategy for dementia prevention in the older people. Despite its acknowledged benefits, the precise mechanism behind exercise-induced cognitive improvement remains controversial. In the present study, we investigated the impact of altered cerebral blood flow (CBF) on brain neural activity originating from motor executive and inhibitory processing using electroencephalographic event-related potentials (EEG-ERPs). Sixteen healthy subjects participated in four sessions, with EEG-ERPs measured during somatosensory Go/No-go tasks. The sessions were conducted under four distinct respiratory conditions presented in random order: normal breathing (NB) and rapid breathing (RB) with room air, normal breathing with hypercapnic gas (5% CO2, 21% O2, and balanced N2) (NB+Gas), and rapid breathing with the same gas (RB+Gas). Changes in CBF were evaluated based on the middle cerebral artery mean blood velocity (MCA Vmean) using transcranial Doppler. PETCO2 was decreased under the RB condition but increased under the NB+Gas condition, thereby decreasing and increasing MCA Vmean, respectively. Under the NB+Gas condition, MCA Vmean significantly increased, but it had no effect on either the executive or inhibitory function. In contrast, the reduction in MCA Vmean induced by RB decreased the peak amplitudes of Go-P300 and No-go-P300. However, even under the RB+Gas condition while MCA Vmean increased, the peak amplitudes of both also decreased. These findings suggest that neither increases nor decreases in CBF affected cognitive function.

Sensorimotor brain-computer interface performance depends on signal-to-noise ratio but not connectivity of the mu rhythm in a multiverse analysis of longitudinal data.

OBJECTIVE: Serving as a channel for communication with locked-in patients or control of prostheses, sensorimotor brain-computer interfaces (BCIs) decode imaginary movements from the recorded activity of the user's brain. However, many individuals remain unable to control the BCI, and the underlying mechanisms are unclear. The user's BCI performance was previously shown to correlate with the resting-state signal-to-noise ratio (SNR) of the mu rhythm and the phase synchronization (PS) of the mu rhythm between sensorimotor areas. Yet, these predictors of performance were primarily evaluated in a single BCI session, while the longitudinal aspect remains rather uninvestigated. In addition, different analysis pipelines were used to estimate PS in source space, potentially hindering the reproducibility of the results. APPROACH: To systematically address these issues, we performed an extensive validation of the relationship between pre-stimulus SNR, PS, and session-wise BCI performance using a publicly available dataset of 62 human participants performing up to 11 sessions of BCI training. We performed the analysis in sensor space using the surface Laplacian and in source space by combining 24 processing pipelines in a multiverse analysis. This way, we could investigate how robust the observed effects were to the selection of the pipeline. MAIN RESULTS: Our results show that SNR had both between- and within-subject effects on BCI performance for the majority of the pipelines. In contrast, the effect of PS on BCI performance was less robust to the selection of the pipeline and became non-significant after controlling for SNR. SIGNIFICANCE: Taken together, our results demonstrate that changes in neuronal connectivity within the sensorimotor system are not critical for learning to control a BCI, and interventions that increase the SNR of the mu rhythm might lead to improvements in the user's BCI performance.

Brain connectivity and time-frequency fusion-based auditory spatial attention detection.

Auditory spatial attention detection (ASAD) aims to decipher the spatial locus of a listener's selective auditory attention from electroencephalogram (EEG) signals. However, current models may exhibit deficiencies in EEG feature extraction, leading to overfitting on small datasets or a decline in EEG discriminability. Furthermore, they often neglect topological relationships between EEG channels and, consequently, brain connectivities. Although graph-based EEG modeling has been employed in ASAD, effectively incorporating both local and global connectivities remains a great challenge. To address these limitations, we propose a new ASAD model. First, time-frequency feature fusion provides a more precise and discriminative EEG representation. Second, EEG segments are treated as graphs, and the graph convolution and global attention mechanism are leveraged to capture local and global brain connections, respectively. A series of experiments are conducted in a leave-trials-out cross-validation manner. On the MAD-EEG and KUL datasets, the accuracies of the proposed model are more than 9% and 3% higher than those of the corresponding state-of-the-art models, respectively, while the accuracy of the proposed model on the SNHL dataset is roughly comparable to that of the state-of-the-art model. EEG time-frequency feature fusion proves to be indispensable in the proposed model. EEG electrodes over the frontal cortex are most important for ASAD tasks, followed by those over the temporal lobe. Additionally, the proposed model performs well even on small datasets. This study contributes to a deeper understanding of the neural encoding related to human hearing and attention, with potential applications in neuro-steered hearing devices.

Group-member selection for RSVP-based collaborative brain-computer interfaces.

Objective: The brain-computer interface (BCI) systems based on rapid serial visual presentation (RSVP) have been widely utilized for the detection of target and non-target images. Collaborative brain-computer interface (cBCI) effectively fuses electroencephalogram (EEG) data from multiple users to overcome the limitations of low single-user performance in single-trial event-related potential (ERP) detection in RSVP-based BCI systems. In a multi-user cBCI system, a superior group mode may lead to better collaborative performance and lower system cost. However, the key factors that enhance the collaboration capabilities of multiple users and how to further use these factors to optimize group mode remain unclear. Approach: This study proposed a group-member selection strategy to optimize the group mode and improve the system performance for RSVP-based cBCI. In contrast to the conventional grouping of collaborators at random, the group-member selection strategy enabled pairing each user with a better collaborator and allowed tasks to be done with fewer collaborators. Initially, we introduced the maximum individual capability and maximum collaborative capability (MIMC) to select optimal pairs, improving the system classification performance. The sequential forward floating selection (SFFS) combined with MIMC then selected a sub-group, aiming to reduce the hardware and labor expenses in the cBCI system. Moreover, the hierarchical discriminant component analysis (HDCA) was used as a classifier for within-session conditions, and the Euclidean space data alignment (EA) was used to overcome the problem of inter-trial variability for cross-session analysis. Main results: In this paper, we verified the effectiveness of the proposed group-member selection strategy on a public RSVP-based cBCI dataset. For the two-user matching task, the proposed MIMC had a significantly higher AUC and TPR and lower FPR than the common random grouping mode and the potential group-member selection method. Moreover, the SFFS with MIMC enabled a trade-off between maintaining performance and reducing the number of system users. Significance: The results showed that our proposed MIMC effectively optimized the group mode, enhanced the classification performance in the two-user matching task, and could reduce the redundant information by selecting the sub-group in the RSVP-based multi-user cBCI systems.

Decrease of the peak heights of EEG bicoherence indicated insufficiency of analgesia during surgery under general anesthesia.

BACKGROUND: Studies show that the two peak heights of electroencephalographic bicoherence (pBIC-high, pBIC-low) decrease after incision and are restored by fentanyl administration. We investigated whether pBICs are good indicators for adequacy of analgesia during surgery. METHODS: After local ethical committee approval, we enrolled 50 patients (27-65 years, ASA-PS I or II) who were scheduled elective surgery. Besides standard anesthesia monitors, to assess pBICs, we used a BIS monitor and freeware Bispectrum Analyzer for A2000. Fentanyl 5 µg/kg was completely administered before incision, and anesthesia was maintained with sevoflurane. After skin incision, when the peak of pBIC-high or pBIC-low decreased by 10% in absolute value (named LT10-high and LT10-low groups in order) or when either peak decreased to below 20% (BL20-high and BL20-low groups), an additional 1 g/kg of fentanyl was administered to examine its effect on the peak that showed a decrease. RESULTS: The mean values and standard deviation for pBIC-high 5 min before fentanyl administration, at the time of fentanyl administration, and 5 min after fentanyl administration for LT10-high group were 39.8% (10.9%), 26.9% (10.5%), and 35.7% (12.5%). And those for pBIC-low for LT10-low group were 39.5% (6.0%), 26.8% (6.4%) and 35.0% (7.0%). Those for pBIC-high for BL20-high group were 26.3% (5.6%), 16.5% (2.6%), and 25.7% (7.0%). And those for pBIC-low for BL20-low group were 26.7% (4.8%), 17.4% (1.8%) and 26.9% (5.7%), respectively. Meanwhile, at these trigger points, hemodynamic parameters didn't show significant changes. CONCLUSION: Superior to standard anesthesia monitoring, pBICs are better indicators of analgesia during surgery. TRIAL REGISTRY: Clinical trial Number and registry URL: UMIN ID: UMIN000042843 https://center6.umin.ac.jp/cgi-open-bin/ctr/ctr_view.cgi?recptno = R000048907.

Acquisition and processing of Motor Imagery and Motor Execution Dataset (MIMED) for six movement activities.

The MIMED dataset is a dataset that provides raw electroencephalogram signal data for activities: raising the right-hand, lowering the right-hand, raising the left-hand, lowering the left-hand, standing, and sitting. In addition to raw data, this dataset provides feature data that undergoes a baseline reduction process. The baseline reduction process is a process to increase the value of EEG signal features. The feature values ​​of the enhanced EEG signal can be easily recognized in the classification process. The device used is Emotiv Epoc X, which consists of 14 channels. Participants involved in this experiment were 30 students from the Bali region in Indonesia. Four recording scenarios were carried out on the first day and four further scenarios on the second day. Two datasets were obtained based on the recording scenario: the motor movement and image datasets. The duration of motor execution is 40 minutes, while motor imagery is 8 minutes for each scenario.

Infant sleep spindle measures from EEG improve prediction of cerebral palsy.

OBJECTIVE: Early identification of infants at risk of cerebral palsy (CP) enables interventions to optimize outcomes. Central sleep spindles reflect thalamocortical sensorimotor circuit function. We hypothesized that abnormal infant central spindle activity would predict later contralateral CP. METHODS: We trained and validated an automated detector to measure spindle rate, duration, and percentage from central electroencephalogram (EEG) channels in high-risk infants (n = 35) and age-matched controls (n = 42). Neonatal magnetic resonance imaging (MRI) findings, infant motor exam, and CP outcomes were obtained from chart review. Using univariable and multivariable logistic regression models, we examined whether spindle activity, MRI abnormalities, and/or motor exam predicted future contralateral CP. RESULTS: The detector had excellent performance (F1 = 0.50). Spindle rate (p = 0.005, p = 0.0004), duration (p < 0.001, p < 0.001), and percentage (p < 0.001, p < 0.001) were decreased in hemispheres corresponding to future CP compared to those without. In this cohort, PLIC abnormality (p = 0.004) and any MRI abnormality (p = 0.004) also predicted subsequent CP. After controlling for MRI findings, spindle features remained significant predictors and improved model fit (p < 0.001, all tests). Using both spindle duration and MRI findings had highest accuracy to classify hemispheres corresponding to future CP (F1 = 0.98, AUC 0.999). CONCLUSION: Decreased central spindle activity improves the prediction of future CP in high-risk infants beyond early MRI or clinical exam alone. SIGNIFICANCE: Decreased central spindle activity provides an early biomarker for CP.

The Utility of 24-h Video-EEG Monitoring in the Diagnosis of Epilepsy in Children.

Objectives: Evaluate the diagnostic yield of 24-h video-EEG monitoring in a group of children admitted in our epilepsy monitoring unit (EMU). Methods: 232 children who underwent 24-h video-EEG monitoring was analysed. We divided each patient's monitoring duration into the first 1, 2, 4, 8, 16 h, relative to the whole 24 h monitoring period. The detection of the first interictal epileptiform discharges (IEDs), epileptic seizures (ES), and psychogenic non-epileptic seizures (PNES) were analysed relative to the different monitoring time subdivision. Results: Our findings revealed that: (1) there was no significant difference in the prevalence of detecting initial IEDs between the first 4-h and 24-h monitoring periods (73.7% vs 81%); (2) clinical events detection rate was statistically similar between the first 8-h and 24-h monitoring periods (15.5% vs 19.3%); (4) an 8-h monitoring was sufficient to capture IEDs, ES and PNES in focal epilepsy children; (5) a 1-h monitoring was sufficient to capture IEDs, ES and PNES in generalized epilepsy children; and (6) IEDs were detected within the first 1-h of monitoring in 96.7% self-limited focal epilepsies (SeLFEs) patient. Conclusion: Our study suggests that a 4-h monitoring has more value in increasing the detection rate of IEDs compared to the traditional shorter routine EEG. And in the case of SeLFEs, a 1-h of monitoring might be sufficient in detecting IEDs. A 24-h VEEG monitoring can detect clinical events in 19.3% of patients. Overall, the yield of IEDs and clinical events detection is adequate in children in children undergoing 24-h video-EEG monitoring.

Integrating spatial and temporal features for enhanced artifact removal in multi-channel EEG recordings.

Objective.Various artifacts in electroencephalography (EEG) are a big hurdle to prevent brain-computer interfaces from real-life usage. Recently, deep learning-based EEG denoising methods have shown excellent performance. However, existing deep network designs inadequately leverage inter-channel relationships in processing multi-channel EEG signals. Typically, most methods process multi-channel signals in a channel-by-channel way. Considering the correlations among EEG channels during the same brain activity, this paper proposes utilizing channel relationships to enhance denoising performance.Approach.We explicitly model the inter-channel relationships using the self-attention mechanism, hypothesizing that these correlations can support and improve the denoising process. Specifically, we introduce a novel denoising network, named spatial-temporal fusion network (STFNet), which integrates stacked multi-dimension feature extractor to explicitly capture both temporal dependencies and spatial relationships.Main results.The proposed network exhibits superior denoising performance, with a 24.27% reduction in relative root mean squared error compared to other methods on a public benchmark. STFNet proves effective in cross-dataset denoising and downstream classification tasks, improving accuracy by 1.40%, while also offering fast processing on CPU.Significance.The experimental results demonstrate the importance of integrating spatial and temporal characteristics. The computational efficiency of STFNet makes it suitable for real-time applications and a potential tool for deployment in realistic environments.

Noninvasive brain stimulation during EEG improves machine learning classification in chronic stroke.

Background: In individuals with chronic stroke and hemiparesis, noninvasive brain stimulation (NIBS) may be used as an adjunct to therapy for improving motor recovery. Specific states of movement during motor recovery are more responsive to brain stimulation than others, thus a system that could auto-detect movement state would be useful in correctly identifying the most effective stimulation periods. The aim of this study was to compare the performance of different machine learning models in classifying movement periods during EEG recordings of hemiparetic individuals receiving noninvasive brain stimulation. We hypothesized that transcranial direct current stimulation, a form of NIBS, would modulate brain recordings correlating with movement state and improve classification accuracies above those receiving sham stimulation. Methods: Electroencephalogram data were obtained from 10 participants with chronic stroke and 11 healthy individuals performing a motor task while undergoing transcranial direct current stimulation. Eight traditional machine learning algorithms and five ensemble methods were used to classify two movement states (a hold posture and an arm reaching movement) before, during and after stimulation. To minimize compute times, preprocessing and feature extraction were limited to z-score normalization and power binning into five frequency bands (delta through gamma). Results: Classification of disease state produced significantly higher accuracies in the stimulation (versus sham) group at 78.9% (versus 55.6%, p < 0.000002). We observed significantly higher accuracies when classifying stimulation state in the chronic stroke group (77.6%) relative to healthy controls (64.1%, p < 0.0095). In the chronic stroke cohort, classification of hold versus reach was highest during the stimulation period (75.2%) as opposed to the pre- and post-stimulation periods. Linear discriminant analysis, logistic regression, and decision tree algorithms classified movement state most accurately in participants with chronic stroke during the stimulation period (76.1%). For the ensemble methods, the highest classification accuracy for hold versus reach was achieved using low gamma frequency (30-50 Hz) as a feature (74.5%), although this result did not achieve statistical significance. Conclusions: Machine learning algorithms demonstrated sufficiently high movement state classification accuracy in participants with chronic stroke performing functional tasks during noninvasive brain stimulation. tDCS improved disease state and movement state classification in participants with chronic stroke.