An EEG channel selection method for motor imagery based on Fisher score and local optimization.

Objective. Multi-channel electroencephalogram (EEG) technology in brain-computer interface (BCI) research offers the advantage of enhanced spatial resolution and system performance. However, this also implies that more time is needed in the data processing stage, which is not conducive to the rapid response of BCI. Hence, it is a necessary and challenging task to reduce the number of EEG channels while maintaining decoding effectiveness.Approach. In this paper, we propose a local optimization method based on the Fisher score for within-subject EEG channel selection. Initially, we extract the common spatial pattern characteristics of EEG signals in different bands, calculate Fisher scores for each channel based on these characteristics, and rank them accordingly. Subsequently, we employ a local optimization method to finalize the channel selection.Main results. On the BCI Competition IV Dataset IIa, our method selects an average of 11 channels across four bands, achieving an average accuracy of 79.37%. This represents a 6.52% improvement compared to using the full set of 22 channels. On our self-collected dataset, our method similarly achieves a significant improvement of 24.20% with less than half of the channels, resulting in an average accuracy of 76.95%.Significance. This research explores the importance of channel combinations in channel selection tasks and reveals that appropriately combining channels can further enhance the quality of channel selection. The results indicate that the model selected a small number of channels with higher accuracy in two-class motor imagery EEG classification tasks. Additionally, it improves the portability of BCI systems through channel selection and combinations, offering the potential for the development of portable BCI systems.

Face preference detection task based on EEG energy analysis in the source domain.

Facial stimulation can produce specific event-related potential (ERP) component N170 in the fusiform gyrus region. However, the role of the fusiform gyrus region in facial preference tasks is not clear at present, and the current research of facial preference analysis based on EEG signals is mostly carried out in the scalp domain. This paper explores whether the region of the fusiform gyrus is involved in processing face preference emotions in terms of the distribution of energy over the source domain, and finds that the pars orbitalis cortex is most energetically active in the face preference task and that there are significant differences between the left and right hemispheres.Clinical Relevance- The role of pars orbitalis in facial preference may help doctors determine whether the pars orbitalis cortex is lost in clinical practice.

Using Determinant Point Process in Generative Adversarial Networks for SSVEP Signals Synthesis.

Steady-state visual evoked potential (SSVEP) is one of the main paradigms of brain-computer interface (BCI). However, the acquisition method of SSVEP can cause subject fatigue and discomfort, leading to the insufficiency of SSVEP databases. Inspired by generative determinantal point process (GDPP), we utilize the determinantal point process in generative adversarial network (GAN) to generate SSVEP signals. We investigate the ability of the method to synthesize signals from the Benchmark dataset. We further use some evaluation metrics to verify its validity. Results prove that the usage of this method significantly improved the authenticity of generated data and the accuracy (97.636%) of classification using deep learning in SSVEP data augmentation.

Design and Performance Analysis of a Bioelectronic Controlled Hybrid Serial-Parallel Wrist Exoskeleton.

Wrist exoskeletons are increasingly being used in the rehabilitation of stroke and hand dysfunction because of its ability to assist patients in high intensity, repetitive, targeted and interactive rehabilitation training. However, the existing wrist exoskeletons cannot effectively replace the work of therapist and improve hand function, mainly because the existing exoskeletons cannot assist patients to perform natural hand movement covering the entire physiological motor space (PMS). Here, we present a bioelectronic controlled hybrid serial-parallel wrist exoskeleton HrWr-ExoSkeleton (HrWE) which is based on the PMS design guidance, the gear set can carry out forearm pronation/supination (P/S) and the 2-DoF parallel configuration fixed on the gear set can carry out wrist flexion/extension (F/E) and radial/ulnar deviation (R/U). This special configuration not only provides enough range of motion (RoM) for rehabilitation training (85F/85E, 55R/55U, and 90P/90S), but also makes it easier to provide the interface for finger exoskeletons and be adapted to upper limb exoskeletons. In addition, to further improve the rehabilitation effect, we propose a HrWE-assisted active rehabilitation training platform based on surface electromyography signals.

Resting-state SEEG-based brain network analysis for the detection of epileptic area.

BACKGROUND: Most epilepsy research is based on interictal or ictal functional connectivity. However, prolonged electrode implantation may affect patients' health and the accuracy of epileptic zone identification. Brief resting-state SEEG recordings reduce the observation of epileptic discharges by reducing electrode implantation and other seizure-inducing interventions. NEW METHOD: The location coordinates of SEEG in the brain were identified using CT and MRI. Based on undirected brain network connectivity, five functional connectivity measures and data feature vector centrality were calculated. Network connectivity was calculated from multiple perspectives of linear correlation, information theory, phase, and frequency, and the relative influence of nodes on network connectivity was considered. We investigated the potential value of resting-state SEEG for epileptic zone identification by comparing the differences between epileptic and non-epileptic zones, as well as the differences between patients with different surgical outcomes. RESULTS: By comparing the centrality of brain network connectivity between epileptic and non-epileptic zones, we found significant differences in the distribution of brain networks between the two zones. There was a significant difference in brain network between patients with good surgical outcomes and those with poor surgical outcomes (p < 0.01). By combining support vector machines with static node importance, we predicted an AUC of 0.94 ± 0.08 for the epilepsy zone. CONCLUSIONS AND SIGNIFICANCE: The results illustrated that nodes in epileptic zones are distinct from those in non-epileptic zones. Analysis of resting-state SEEG data and the importance of nodes in the brain network may contribute to identifying the epileptic zone and predicting the outcome.

Noninvasive neuroimaging and spatial filter transform enable ultra low delay motor imagery EEG decoding.

Objective.The brain-computer interface (BCI) system based on sensorimotor rhythm can convert the human spirit into instructions for machine control, and it is a new human-computer interaction system with broad applications. However, the spatial resolution of scalp electroencephalogram (EEG) is limited due to the presence of volume conduction effects. Therefore, it is very meaningful to explore intracranial activities in a noninvasive way and improve the spatial resolution of EEG. Meanwhile, low-delay decoding is an essential factor for the development of a real-time BCI system.Approach.In this paper, EEG conduction is modeled by using public head anatomical templates, and cortical EEG is obtained using dynamic parameter statistical mapping. To solve the problem of a large amount of computation caused by the increase in the number of channels, the filter bank common spatial pattern method is used to obtain a spatial filter kernel, which reduces the computational cost of feature extraction to a linear level. And the feature classification and selection of important features are completed using a neural network containing band-spatial-time domain self-attention mechanisms.Main results.The results show that the method proposed in this paper achieves high accuracy for the four types of motor imagery EEG classification tasks, with fairly low latency and high physiological interpretability.Significance.The proposed decoding framework facilitates the realization of low-latency human-computer interaction systems.

N170 component analysis of single-trial EEG based on electrophysiological source imaging.

Event-related potentials (ERP) are brain-evoked potentials that reflect the neural activity of the brain. However, it is difficult to isolate the ERP components of our interest because single-trial EEG is disturbed by other signals, and the average ERP analysis in turn loses single-trial information. In this paper, we used electrophysiological source imaging (ESI) to analyze the N170 component of single-trial EEG triggered by face stimulation. The results show that ESI is feasible for the analysis of N170 and that there are left-right differences in the area of the fusiform gyrus associated with face stimulation in the brain. Clinical Relevance- Analysis of the N170 of single-trial EEG by ESI may help in the diagnosis of patients with prosopagnosia and may also help physicians clinically in determining whether the fusiform gyrus region is damaged.

Comparison of MI-EEG Decoding in Source to Sensor Domain.

Brain-computer interface (BCI) system based on sensorimotor rhythm (SMR) is a more natural brain-computer interaction system. In this paper, we propose a new multi-task motor imagery EEG (MI-EEG) classification framework. Unlike traditional EEG decoding algorithms, we perform the decoding task in the source domain rather than the sensor domain. In the proposed algorithm, we first build a conduction model of the signal using the public ICBM152 head model and the boundary element method (BEM). The sensor domain EEG was then mapped to the selected cortex region using standardized low-resolution electromagnetic tomography (sLORETA) technology, which benefit to address volume conduction effects problem. Finally, the source domain features are extracted and classified by combining FBCSP and simple LDA. The results show that the classification-decoding algorithm performed in the source domain can well solve the classification task of MI-EEG. In addition, we found that the source imaging method can significantly increase the number of available EEG channels, which can be expanded at least double. The preliminary results of this study encourage the implementation of EEG decoding algorithms in the source domain. Clinical Relevance- This confirms that better results can be obtained by performing MI-EEG decoding in the source domain than in the sensor domain.

EEG-Based Brain Network Analysis of Chronic Stroke Patients After BCI Rehabilitation Training.

Traditional rehabilitation strategies become difficult in the chronic phase stage of stroke prognosis. Brain-computer interface (BCI) combined with external devices may improve motor function in chronic stroke patients, but it lacks comprehensive assessments of neurological changes regarding functional rehabilitation. This study aimed to comprehensively and quantitatively investigate the changes in brain activity induced by BCI-FES training in patients with chronic stroke. We analyzed the EEG of two groups of patients with chronic stroke, one group received functional electrical stimulation (FES) rehabilitation training (FES group) and the other group received BCI combined with FES training (BCI-FES group). We constructed functional networks in both groups of patients based on direct directed transfer function (dDTF) and assessed the changes in brain activity using graph theory analysis. The results of this study can be summarized as follows: (i) after rehabilitation training, the Fugl-Meyer assessment scale (FMA) score was significantly improved in the BCI-FES group (p < 0.05), and there was no significant difference in the FES group. (ii) Both the global and local graph theory measures of the brain network of patients with chronic stroke in the BCI-FES group were improved after rehabilitation training. (iii) The node strength in the contralesional hemisphere and central region of patients in the BCI-FES group was significantly higher than that in the FES group after the intervention (p < 0.05), and a significant increase in the node strength of C4 in the contralesional sensorimotor cortex region could be observed in the BCI-FES group (p < 0.05). These results suggest that BCI-FES rehabilitation training can induce clinically significant improvements in motor function of patients with chronic stroke. It can improve the functional integration and functional separation of brain networks and boost compensatory activity in the contralesional hemisphere to a certain extent. The findings of our study may provide new insights into understanding the plastic changes of brain activity in patients with chronic stroke induced by BCI-FES rehabilitation training.