Time-frequency-space EEG decoding model based on dense graph convolutional network for stroke.

Stroke, a sudden cerebrovascular ailment resulting from brain tissue damage, has prompted the use of motor imagery (MI)-based Brain-Computer Interface (BCI) systems in stroke rehabilitation. However, analyzing EEG signals from stroke patients is challenging because of their low signal-to-noise ratio and high variability. Therefore, we propose a novel approach that combines the modified S-transform (MST) and a dense graph convolutional network (DenseGCN) algorithm to enhance the MI-BCI performance across time, frequency, and space domains. MST is a time-frequency analysis method that efficiently concentrates energy in EEG signals, while DenseGCN is a deep learning model that uses EEG feature maps from each layer as inputs for subsequent layers, facilitating feature reuse and hyper-parameters optimization. Our approach outperforms conventional networks, achieving a peak classification accuracy of 90.22% and an average information transfer rate (ITR) of 68.52 bits per minute. Moreover, we conduct an in-depth analysis of the event-related desynchronization/event-related synchronization (ERD/ERS) phenomenon in the deep-level EEG features of stroke patients. Our experimental results confirm the feasibility and efficacy of the proposed approach for MI-BCI rehabilitation systems.

EMPT: a sparsity Transformer for EEG-based motor imagery recognition.

Introduction: Transformer network is widely emphasized and studied relying on its excellent performance. The self-attention mechanism finds a good solution for feature coding among multiple channels of electroencephalography (EEG) signals. However, using the self-attention mechanism to construct models on EEG data suffers from the problem of the large amount of data required and the complexity of the algorithm. Methods: We propose a Transformer neural network combined with the addition of Mixture of Experts (MoE) layer and ProbSparse Self-attention mechanism for decoding the time-frequency-spatial domain features from motor imagery (MI) EEG of spinal cord injury patients. The model is named as EEG MoE-Prob-Transformer (EMPT). The common spatial pattern and the modified s-transform method are employed for achieving the time-frequency-spatial features, which are used as feature embeddings to input the improved transformer neural network for feature reconstruction, and then rely on the expert model in the MoE layer for sparsity mapping, and finally output the results through the fully connected layer. Results: EMPT achieves an accuracy of 95.24% on the MI EEG dataset for patients with spinal cord injury. EMPT has also achieved excellent results in comparative experiments with other state-of-the-art methods. Discussion: The MoE layer and ProbSparse Self-attention inside the EMPT are subjected to visualisation experiments. The experiments prove that sparsity can be introduced to the Transformer neural network by introducing MoE and kullback-leibler divergence attention pooling mechanism, thereby enhancing its applicability on EEG datasets. A novel deep learning approach is presented for decoding EEG data based on MI.

Combining EEG Features and Convolutional Autoencoder for Neonatal Seizure Detection.

Neonatal epilepsy is a common emergency phenomenon in neonatal intensive care units (NICUs), which requires timely attention, early identification, and treatment. Traditional detection methods mostly use supervised learning with enormous labeled data. Hence, this study offers a semi-supervised hybrid architecture for detecting seizures, which combines the extracted electroencephalogram (EEG) feature dataset and convolutional autoencoder, called Fd-CAE. First, various features in the time domain and entropy domain are extracted to characterize the EEG signal, which helps distinguish epileptic seizures subsequently. Then, the unlabeled EEG features are fed into the convolutional autoencoder (CAE) for training, which effectively represents EEG features by optimizing the loss between the input and output features. This unsupervised feature learning process can better combine and optimize EEG features from unlabeled data. After that, the pre-trained encoder part of the model is used for further feature learning of labeled data to obtain its low-dimensional feature representation and achieve classification. This model is performed on the neonatal EEG dataset collected at the University of Helsinki Hospital, which has a high discriminative ability to detect seizures, with an accuracy of 92.34%, precision of 93.61%, recall rate of 98.74%, and F1-score of 95.77%, respectively. The results show that unsupervised learning by CAE is beneficial to the characterization of EEG signals, and the proposed Fd-CAE method significantly improves classification performance.

Electroencephalography-based biological and functional characteristics of spinal cord injury patients with neuropathic pain and numbness.

Objectives: To identify potential treatment targets for spinal cord injury (SCI)-related neuropathic pain (NP) by analysing the differences in electroencephalogram (EEG) and brain network connections among SCI patients with NP or numbness. Participants and methods: The EEG signals during rest, as well as left- and right-hand and feet motor imagination (MI), were recorded. The power spectral density (PSD) of the θ (4-8 Hz), α (8-12 Hz), and ß (13-30 Hz) bands was calculated by applying Continuous Wavelet Transform (CWT) and Modified S-transform (MST) to the data. We used 21 electrodes as network nodes and performed statistical measurements of the phase synchronisation between two brain regions using a phase-locking value, which captures nonlinear phase synchronisation. Results: The specificity of the MST algorithm was higher than that of the CWT. Widespread non-lateralised event-related synchronization was observed in both groups during the left- and right-hand MI. The PWP (patients with pain) group had lower θ and α bands PSD values in multiple channels of regions including the frontal, premotor, motor, and temporal regions compared with the PWN (patients with numbness) group (all p < 0.05), but higher ß band PSD values in multiple channels of regions including the frontal, premotor, motor, and parietal region compared with the PWN group (all p < 0.05). During left-hand and feet MI, in the lower frequency bands (θ and α bands), the brain network connections of the PWP group were significantly weaker than the PWN group except for the frontal region. Conversely, in the higher frequency bands (ß band), the brain network connections of the PWP group were significantly stronger in all regions than the PWN group. Conclusion: The differences in the power of EEG and network connectivity in the frontal, premotor, motor, and temporal regions are potential biological and functional characteristics that can be used to distinguish NP from numbness. The differences in brain network connections between the two groups suggest that the distinct mechanisms for pain and numbness.

IL-13 facilitates ferroptotic death in asthmatic epithelial cells via SOCS1-mediated ubiquitinated degradation of SLC7A11.

Th2-high asthma is characterized by elevated levels of type 2 cytokines, such as interleukin 13 (IL-13), and its prevalence has been increasing worldwide. Ferroptosis, a recently discovered type of programmed cell death, is involved in the pathological process of Th2-high asthma; however, the underlying mechanisms remain incompletely understood. In this study, we demonstrated that the serum level of malondialdehyde (MDA), an index of lipid peroxidation, positively correlated with IL-13 level and negatively correlated with the predicted forced expiratory volume in 1 s (FEV1%) in asthmatics. Furthermore, we showed that IL-13 facilitates ferroptosis by upregulating of suppressor of cytokine signaling 1 (SOCS1) through analyzing immortalized airway epithelial cells, human airway organoids, and the ovalbumin (OVA)-challenged asthma model. We identified that signal transducer and activator of transcription 6 (STAT6) promotes the transcription of SOCS1 upon IL-13 stimulation. Moreover, SOCS1, an E3 ubiquitin ligase, was found to bind to solute carrier family 7 member 11 (SLC7A11) and catalyze its ubiquitinated degradation, thereby promoting ferroptosis in airway epithelial cells. Last, we found that inhibiting SOCS1 can decrease ferroptosis in airway epithelial cells and alleviate airway hyperresponsiveness (AHR) in OVA-challenged wide-type mice, while SOCS1 overexpression exacerbated the above in OVA-challenged IL-13-knockout mice. Our findings reveal that the IL-13/STAT6/SOCS1/SLC7A11 pathway is a novel molecular mechanism for ferroptosis in Th2-high asthma, confirming that targeting ferroptosis in airway epithelial cells is a potential therapeutic strategy for Th2-high asthma.

Variation in brain connectivity during motor imagery and motor execution in stroke patients based on electroencephalography.

Objective: The objective of this study was to analyze the changes in connectivity between motor imagery (MI) and motor execution (ME) in the premotor area (PMA) and primary motor cortex (MA) of the brain, aiming to explore suitable forms of treatment and potential therapeutic targets. Methods: Twenty-three inpatients with stroke were selected, and 21 right-handed healthy individuals were recruited. EEG signal during hand MI and ME (synergy and isolated movements) was recorded. Correlations between functional brain areas during MI and ME were compared. Results: PMA and MA were significantly and positively correlated during hand MI in all participants. The power spectral density (PSD) values of PMA EEG signals were greater than those of MA during MI and ME in both groups. The functional connectivity correlation was higher in the stroke group than in healthy people during MI, especially during left-handed MI. During ME, functional connectivity correlation in the brain was more enhanced during synergy movements than during isolated movements. The regions with abnormal functional connectivity were in the 18th lead of the left PMA area. Conclusion: Left-handed MI may be crucial in MI therapy, and the 18th lead may serve as a target for non-invasive neuromodulation to promote further recovery of limb function in patients with stroke. This may provide support for the EEG theory of neuromodulation therapy for hemiplegic patients.

Detection of dynamic changes of electrodermal activity to predict the classroom performance of college students.

It is emphasized in the Self-regulated learning (SRL) framework that self-monitoring of learning state is vital for students to keep effective in studying. However, it's still challenging to get an accurate and timely understanding of their learning states during classes. In this study, we propose to use electrodermal activity (EDA) signals which are deemed to be associated with physiological arousal state to predict the college student's classroom performance. Twenty college students were recruited to attend eight lectures in the classroom, during which their EDA signals were recorded simultaneously. For each lecture, the students should complete pre- and after-class tests, and a self-reported scale (SRS) on their learning experience. EDA indices were extracted from both time and frequency domains, and they were furtherly mapped to the student's learning efficiency. As a result, the indices relevant to the dynamic changes of EDA had significant positive correlations with the learning efficiency. Furthermore, compared with only using SRS, a combination with EDA indices had significantly higher accuracy in predicting the learning efficiency. In conclusion, our findings demonstrate that the EDA dynamics are sensitive to the changes in learning efficiency, suggesting a promising approach to predicting the classroom performance of college students.

Multilevel Laser-Induced Pain Measurement with Wasserstein Generative Adversarial Network - Gradient Penalty Model.

Pain is an experience of unpleasant sensations and emotions associated with actual or potential tissue damage. In the global context, billions of people are affected by pain disorders. There are particular challenges in the measurement and assessment of pain, and the commonly used pain measuring tools include traditional subjective scoring methods and biomarker-based measures. The main tools for biomarker-based analysis are electroencephalography (EEG), electrocardiography and functional magnetic resonance. The EEG-based quantitative pain measurements are of immense value in clinical pain management and can provide objective assessments of pain intensity. The assessment of pain is now primarily limited to the identification of the presence or absence of pain, with less research on multilevel pain. High power laser stimulation pain experimental paradigm and five pain level classification methods based on EEG data augmentation are presented. First, the EEG features are extracted using modified S-transform, and the time-frequency information of the features is retained. Based on the pain recognition effect, the 20-40[Formula: see text]Hz frequency band features are optimized. Afterwards the Wasserstein generative adversarial network with gradient penalty is used for feature data augmentation. It can be inferred from the good classification performance of features in the parietal region of the brain that the sensory function of the parietal lobe region is effectively activated during the occurrence of pain. By comparing the latest data augmentation methods and classification algorithms, the proposed method has significant advantages for the five-level pain dataset. This research provides new ways of thinking and research methods related to pain recognition, which is essential for the study of neural mechanisms and regulatory mechanisms of pain.

Self-Supervised EEG Representation Learning with Contrastive Predictive Coding for Post-Stroke Patients.

Stroke patients are prone to fatigue during the EEG acquisition procedure, and experiments have high requirements on cognition and physical limitations of subjects. Therefore, how to learn effective feature representation is very important. Deep learning networks have been widely used in motor imagery (MI) based brain-computer interface (BCI). This paper proposes a contrast predictive coding (CPC) framework based on the modified s-transform (MST) to generate MST-CPC feature representations. MST is used to acquire the temporal-frequency feature to improve the decoding performance for MI task recognition. EEG2Image is used to convert multi-channel one-dimensional EEG into two-dimensional EEG topography. High-level feature representations are generated by CPC which consists of an encoder and autoregressive model. Finally, the effectiveness of generated features is verified by the k-means clustering algorithm. It can be found that our model generates features with high efficiency and a good clustering effect. After classification performance evaluation, the average classification accuracy of MI tasks is 89% based on 40 subjects. The proposed method can obtain effective feature representations and improve the performance of MI-BCI systems. By comparing several self-supervised methods on the public dataset, it can be concluded that the MST-CPC model has the highest average accuracy. This is a breakthrough in the combination of self-supervised learning and image processing of EEG signals. It is helpful to provide effective rehabilitation training for stroke patients to promote motor function recovery.

ARRDC5 deficiency impairs spermatogenesis by affecting SUN5 and NDC1.

Sperm with normal morphology and motility are essential for successful fertilization, and the strong attachment of the sperm head-tail coupling apparatus to the nuclear envelope during spermatogenesis is required to ensure the integrity of sperm for capacitation and fertilization. Here, we report that Arrdc5 is associated with spermatogenesis. The Arrdc5 knockout mouse model showed male infertility characterized by a high bent-head rate and reduced motility in sperm, which led to capacitation defects and subsequent fertilization failure. Through mass spectrometry, we found that ARRDC5 affects spermatogenesis by affecting NDC1 and SUN5. We further found that ARRDC5 might affect the vesicle-trafficking protein SEC22A-mediated transport and localization of NDC1, SUN5 and other head-tail coupling apparatus-related proteins that are responsible for initiating the attachment of the sperm head and tail. We finally performed intracytoplasmic sperm injection as a way to explore therapeutic strategies. Our findings demonstrate the essential role and the underlying molecular mechanism of ARRDC5 in anchoring the sperm head to the tail during spermatogenesis.

A Light Vehicle License-Plate-Recognition System Based on Hybrid Edge-Cloud Computing.

With the world moving towards low-carbon and environmentally friendly development, the rapid growth of new-energy vehicles is evident. The utilization of deep-learning-based license-plate-recognition (LPR) algorithms has become widespread. However, existing LPR systems have difficulty achieving timely, effective, and energy-saving recognition due to their inherent limitations such as high latency and energy consumption. An innovative Edge-LPR system that leverages edge computing and lightweight network models is proposed in this paper. With the help of this technology, the excessive reliance on the computational capacity and the uneven implementation of resources of cloud computing can be successfully mitigated. The system is specifically a simple LPR. Channel pruning was used to reconstruct the backbone layer, reduce the network model parameters, and effectively reduce the GPU resource consumption. By utilizing the computing resources of the Intel second-generation computing stick, the network models were deployed on edge gateways to detect license plates directly. The reliability and effectiveness of the Edge-LPR system were validated through the experimental analysis of the CCPD standard dataset and real-time monitoring dataset from charging stations. The experimental results from the CCPD common dataset demonstrated that the network's total number of parameters was only 0.606 MB, with an impressive accuracy rate of 97%.

Rhythmic temporal prediction enhances neural representations of movement intention for brain-computer interface.

Objective.Detecting movement intention is a typical use of brain-computer interfaces (BCI). However, as an endogenous electroencephalography (EEG) feature, the neural representation of movement is insufficient for improving motor-based BCI. This study aimed to develop a new movement augmentation BCI encoding paradigm by incorporating the cognitive function of rhythmic temporal prediction, and test the feasibility of this new paradigm in optimizing detections of movement intention.Methods.A visual-motion synchronization task was designed with two movement intentions (left vs. right) and three rhythmic temporal prediction conditions (1000 ms vs. 1500 ms vs. no temporal prediction). Behavioural and EEG data of 24 healthy participants were recorded. Event-related potentials (ERPs), event-related spectral perturbation induced by left- and right-finger movements, the common spatial pattern (CSP) and support vector machine, Riemann tangent space algorithm and logistic regression were used and compared across the three temporal prediction conditions, aiming to test the impact of temporal prediction on movement detection.Results.Behavioural results showed significantly smaller deviation time for 1000 ms and 1500 ms conditions. ERP analyses revealed 1000 ms and 1500 ms conditions led to rhythmic oscillations with a time lag in contralateral and ipsilateral areas of movement. Compared with no temporal prediction, 1000 ms condition exhibited greater beta event-related desynchronization (ERD) lateralization in motor area (P< 0.001) and larger beta ERD in frontal area (P< 0.001). 1000 ms condition achieved an averaged left-right decoding accuracy of 89.71% using CSP and 97.30% using Riemann tangent space, both significantly higher than no temporal prediction. Moreover, movement and temporal information can be decoded simultaneously, achieving 88.51% four-classification accuracy.Significance.The results not only confirm the effectiveness of rhythmic temporal prediction in enhancing detection ability of motor-based BCI, but also highlight the dual encodings of movement and temporal information within a single BCI paradigm, which is promising to expand the range of intentions that can be decoded by the BCI.

Food packaging films based on ionically crosslinked konjac glucomannan incorporating zein-pectin nanoparticle-stabilized corn germ oil-oregano oil Pickering emulsion.

This study addresses the limitations of konjac glucomannan (KGM) films in mechanical properties, hydrophobicity and antibacterial activities. For the first time, a zein-pectin nanoparticle-stabilized corn germ oil-oregano essential oil Pickering emulsion (ZPCEO) was incorporated into KGM, with the resulting film being further ionically crosslinked with Ca2+, Cu2+ or Fe3+. FTIR, SEM and EDS results showed that the metal ions were crosslinked with the hydroxyl and carbonyl groups of polysaccharides and uniformly distributed throughout the films (degree of crosslinking: Fe3+ > Cu2+ > Ca2+). Compared with pure KGM films, the ionic crosslinked ZPCEO/KGM (IL-ZPCEO/KGM) films have superior water resistance mechanical properties, and exhibit unique UV-blocking properties, antioxidant and antibacterial activities. The ZPCEO/KGM-Fe3+ film offered the best all-round properties, including the highest tensile strength, water resistance, UV-blocking capacity, and antimicrobial activity. Thus, ionic crosslinking of ZPCEO/KGM films can be applied to the preparation of food packaging for use in high humidity environments.

Epileptic EEG Classification via Graph Transformer Network.

Deep learning-based epileptic seizure recognition via electroencephalogram signals has shown considerable potential for clinical practice. Although deep learning algorithms can enhance epilepsy identification accuracy compared with classical machine learning techniques, classifying epileptic activities based on the association between multichannel signals in electroencephalogram recordings is still challenging in automated seizure classification from electroencephalogram signals. Furthermore, the performance of generalization is hardly maintained by the fact that existing deep learning models were constructed using just one architecture. This study focuses on addressing this challenge using a hybrid framework. Alternatively put, a hybrid deep learning model, which is based on the ground-breaking graph neural network and transformer architectures, was proposed. The proposed deep architecture consists of a graph model to discover the inner relationship between multichannel signals and a transformer to reveal the heterogeneous associations between the channels. To evaluate the performance of the proposed approach, the comparison experiments were conducted on a publicly available dataset between the state-of-the-art algorithms and ours. Experimental results demonstrate that the proposed method is a potentially valuable instrument for epoch-based epileptic EEG classification.

A novel visual brain-computer interfaces paradigm based on evoked related potentials evoked by weak and small number of stimuli.

Introduction: Traditional visual Brain-Computer Interfaces (v-BCIs) usually use large-size stimuli to attract more attention from users and then elicit more distinct and robust EEG responses, which would cause visual fatigue and limit the length of use of the system. On the contrary, small-size stimuli always need multiple and repeated stimulus to code more instructions and increase separability among each code. These common v-BCIs paradigms can cause problems such as redundant coding, long calibration time, and visual fatigue. Methods: To address these problems, this study presented a novel v-BCI paradigm using weak and small number of stimuli, and realized a nine-instruction v-BCI system that controlled by only three tiny stimuli. Each of these stimuli were located between instructions, occupied area with eccentricities subtended 0.4°, and flashed in the row-column paradigm. The weak stimuli around each instruction would evoke specific evoked related potentials (ERPs), and a template-matching method based on discriminative spatial pattern (DSP) was employed to recognize these ERPs containing the intention of users. Nine subjects participated in the offline and online experiments using this novel paradigm. Results: The average accuracy of the offline experiment was 93.46% and the online average information transfer rate (ITR) was 120.95 bits/min. Notably, the highest online ITR achieved 177.5 bits/min. Discussion: These results demonstrate the feasibility of using a weak and small number of stimuli to implement a friendly v-BCI. Furthermore, the proposed novel paradigm achieved higher ITR than traditional ones using ERPs as the controlled signal, which showed its superior performance and may have great potential of being widely used in various fields.

Design and Test of Spatial Cognitive Training and Evaluation System Based on Virtual Reality Head-Mounted Display With EEG Recording.

The field of spatial cognitive training and evaluation has rapidly evolved. However, the low learning motivation and engagement of the subjects hinder the widespread use of spatial cognitive training. This study designed a home-based spatial cognitive training and evaluation system (SCTES), which aimed to train subjects on spatial cognitive tasks for 20 days, and compared the brain activities before and after the training. This study also evaluated the feasibility of using a portable all-in-one prototype for cognitive training that combined a virtual reality (VR) head-mounted display with high-quality electroencephalogram (EEG) recording. During the course of training, the length of the navigation path and the distance between the starting position and the platform position revealed significant behavioral differences. In the testing sessions, the subjects showed significant behavioral differences in the time it took to complete the test task before and after training. After only four days of training, the subjects demonstrated significant differences in the Granger causality analysis (GCA) characteristics of brain regions in the δ , θ , α1 , ß2 , and γ frequency bands of the EEG, as well as significant differences in the GCA of the EEG in the ß1 , ß2 , and γ frequency bands between the two test sessions. The proposed SCTES used a compact and all-in-one form factor to train and evaluate spatial cognition and collect EEG signals and behavioral data simultaneously. The recorded EEG data can be used to quantitatively assess the efficacy of spatial training in patients with spatial cognitive impairments.

Hybrid Attention Network for Epileptic EEG Classification.

Automatic seizure detection from electroencephalography (EEG) based on deep learning has been significantly improved. However, existing works have not adequately excavate the spatial-temporal information between EEG channels. Besides, most works mainly focus on patient-specific scenarios while cross-patient seizure detection is more challenging and meaningful. Regarding the above problems, we propose a hybrid attention network (HAN) for automatic seizure detection. Specifically, the graph attention network (GAT) extracts spatial features at the front end, and Transformer gets time features as the back end. HAN leverages the attention mechanism and fully extracts the spatial-temporal correlation of EEG signals. The focal loss function is introduced to HAN to deal with the imbalance of the dataset accompanied by seizure detection based on EEG. Both patient-specific and patient-independent experiments are carried out on the public CHB-MIT database. Experimental results demonstrate the efficacy of HAN in both experimental settings.

One-Dimensional Local Binary Pattern and Common Spatial Pattern Feature Fusion Brain Network for Central Neuropathic Pain.

Central neuropathic pain (CNP) after spinal cord injury (SCI) is related to the plasticity of cerebral cortex. The plasticity of cortex recorded by electroencephalogram (EEG) signal can be used as a biomarker of CNP. To analyze changes in the brain network mechanism under the combined effect of injury and pain or under the effect of pain, this paper mainly studies the changes of brain network functional connectivity in patients with neuropathic pain and without neuropathic pain after SCI. This paper has recorded the EEG with the CNP group after SCI, without the CNP group after SCI, and a healthy control group. Phase-locking value has been used to construct brain network topological connectivity maps. By comparing the brain networks of the two groups of SCI with the healthy group, it has been found that in the [Formula: see text] and [Formula: see text] frequency bands, the injury increases the functional connectivity between the frontal lobe and occipital lobes, temporal, and parietal of the patients. Furthermore, the comparison of brain networks between the group with CNP and the group without CNP after SCI has found that pain has a greater effect on the increased connectivity within the patients' frontal lobes. Motor imagery (MI) data of CNP patients have been used to extract one-dimensional local binary pattern (1D-LBP) and common spatial pattern (CSP) features, the left and right hand movements of the patients' MI have been classified. The proposed LBP-CSP feature method has achieved the highest accuracy of 98.6% and the average accuracy of 91.5%. The results of this study have great clinical significance for the neural rehabilitation and brain-computer interface of CNP patients.

Transformed common spatial pattern for motor imagery-based brain-computer interfaces.

Objective: The motor imagery (MI)-based brain-computer interface (BCI) is one of the most popular BCI paradigms. Common spatial pattern (CSP) is an effective algorithm for decoding MI-related electroencephalogram (EEG) patterns. However, it highly depends on the selection of EEG frequency bands. To address this problem, previous researchers often used a filter bank to decompose EEG signals into multiple frequency bands before applying the traditional CSP. Approach: This study proposed a novel method, i.e., transformed common spatial pattern (tCSP), to extract the discriminant EEG features from multiple frequency bands after but not before CSP. To verify its effectiveness, we tested tCSP on a dataset collected by our team and a public dataset from BCI competition III. We also performed an online evaluation of the proposed method. Main results: As a result, for the dataset collected by our team, the classification accuracy of tCSP was significantly higher than CSP by about 8% and filter bank CSP (FBCSP) by about 4.5%. The combination of tCSP and CSP further improved the system performance with an average accuracy of 84.77% and a peak accuracy of 100%. For dataset IVa in BCI competition III, the combination method got an average accuracy of 94.55%, which performed best among all the presented CSP-based methods. In the online evaluation, tCSP and the combination method achieved an average accuracy of 80.00 and 84.00%, respectively. Significance: The results demonstrate that the frequency band selection after CSP is better than before for MI-based BCIs. This study provides a promising approach for decoding MI EEG patterns, which is significant for the development of BCIs.