A recent study showed improved performance of the P300-speller when the flashing row or column was overlaid with translucent pictures of familiar faces (FF spelling paradigm). However, the performance of the P300-speller is not yet satisfactory due to its low classification accuracy and information transfer rate. To investigate whether P300-speller performance is further improved when the chromatic property and the FF spelling paradigm are combined. We proposed a new spelling paradigm in which the flashing row or column is overlaid with translucent green pictures of familiar faces (GFF spelling paradigm). We analyzed the ERP waveforms elicited by the FF and proposed GFF spelling paradigms and compared P300-speller performance between the two paradigms. Significant differences in the amplitudes of four ERP components (N170, VPP, P300, and P600f) were observed between both spelling paradigms. Compared to the FF spelling paradigm, the GFF spelling paradigm elicited ERP waveforms of higher amplitudes and resulted in improved P300-speller performance. Combining the chromatic property (green color) and the FF spelling paradigm led to better classification accuracy and an increased information transfer rate. These findings demonstrate a promising new approach for improving the performance of the P300-speller.

Electromyography (EMG) pattern recognition is one of the widely used methods to control the rehabilitation robots and prostheses. However, the changes in the distribution of EMG data due to electrodes shifting results in classification decline, which hinders its clinical application in repeated uses. Adaptive learning can solve this problem but takes additional time. To address this, an efficient scheme is developed by comparing the performance of 12 combinations of three feature selection methods [no feature selection (NFS), sequential forward search (SFS), and particle swarm optimization (PSO)] and four classification methods [non-adaptive support vector machine (N-SVM), incremental SVM (I-SVM), SVM based on TrAdaBoost (T-SVM), and I-SVM based on TrAdaBoost (TI-SVM)] in the classification of EMG data of 12 subjects for 5 consecutive days. Our results showed that TI-SVM achieved the highest classification accuracy among the classification methods (p < 0.05). The SFS method achieved the same classification accuracy as that of the scheme trained with the feature vectors selected by the NFS method (p = 0.999) while achieving a lower training time than that of TI-SVM combined with the NFS method (p = 0.043). Although the PSO method outperformed the NFS and SFS methods by achieving reduced training and response times (p < 0.05), the PSO method achieved a considerably lower classification accuracy than that of the scheme trained with the feature vectors selected by the NFS (p = 0.001) or SFS (p = 0.001) method. Furthermore, TI-SVM combined with the SFS method outperformed the CNN method with fine-tuning in classification accuracy on a small data set (p = 0.001). The results indicate that TI-SVM combined with the SFS method is suitable for improving the performance of EMG pattern recognition in repeated uses.

Objective: Previous studies have shown that the performance of the famous face P300-speller was better than that of the classical row/column flashing P300-speller. Furthermore, in some studies, the brain was more active when responding to one's own face than to a famous face, and a self-face stimulus elicited larger amplitude event-related potentials (ERPs) than did a famous face. Thus, we aimed to study the role of the self-face paradigm on further improving the performance of the P300-speller system with the famous face P300-speller paradigm as the control paradigm. Methods: We designed two facial P300-speller paradigms based on the self-face and a famous face (Ming Yao, a sports star; the famous face spelling paradigm) with a neutral expression. Results: ERP amplitudes were significantly greater in the self-face than in the famous face spelling paradigm at the parietal area from 340 to 480 ms (P300), from 480 to 600 ms (P600f), and at the fronto-central area from 700 to 800 ms. Offline and online classification results showed that the self-face spelling paradigm accuracies were significantly higher than those of the famous face spelling paradigm at superposing first two times (P < 0.05). Similar results were found for information transfer rates (P < 0.05). Conclusions: The self-face spelling paradigm significantly improved the performance of the P300-speller system. This has significant practical applications for brain-computer interfaces (BCIs) and could avoid infringement issues caused by using images of other people's faces.

Automated seizure detection based on scalp electroencephalography (EEG) can significantly accelerate the epilepsy diagnosis process. However, most existing deep learning-based epilepsy detection methods are deficient in mining the local features and global time series dependence of EEG signals, limiting the performance enhancement of the models in seizure detection. Our study proposes an epilepsy detection model, CMFViT, based on a Multi-Stream Feature Fusion (MSFF) strategy that fuses a Convolutional Neural Network (CNN) with a Vision Transformer (ViT). The model converts EEG signals into time-frequency domain images using the Tunable Q-factor Wavelet Transform (TQWT), and then utilizes the CNN module and the ViT module to capture local features and global time-series correlations, respectively. It fuses different feature representations through the MSFF strategy to enhance its discriminative ability, and finally completes the classification task through the average pooling layer and the fully connected layer. The effectiveness of the model was validated by experimental evaluations on the publicly available CHB-MIT dataset and the Kaggle 121 people epilepsy dataset. The model achieved 98.85% classification accuracy and other excellent metrics in single-subject experiments on the CHB-MIT dataset, and also demonstrated strong performance in cross-subject experiments on the Kaggle dataset. Ablation experiments demonstrate the complementary roles of the CNN and ViT modules, and their integration significantly improves detection accuracy and generalization. Comparisons with other methods highlight the advantages of the CMFViT model. The CMFViT model provides an efficient, accurate, and innovative solution for complex EEG signal analysis and seizure detection tasks for single and cross-subjects while laying the foundation for developing real-time, accurate seizure detection systems.

Objective: P300-speller is the most commonly used brain-computer interface (BCI) for providing a means of communication to patients with amyotrophic lateral sclerosis. However, the performance of the P300-speller BCI is still inadequate. We investigated whether the performance of P300-speller can be further improved by increasing the mental effort required of the user. Methods: We designed two active mental tasks for a P300-speller based on a differently colored smiling cartoon-face paradigm. The tasks were based on color distinction, and their difficulty was modulated. One of the active mental tasks (DC task) required participants to focus on and distinguish the color of a target, while the other task (CN + DC task) required participants to simultaneously count the number of times a target flashed and distinguish its color. Results: The amplitudes of the event-related potentials (ERPs) in both DC and CN + DC tasks were higher than that in the CN task. The significant difference in the amplitudes between the DC and CN tasks was observed around the parietal-central area from 440 to 800 ms (late positive component, LPC), and that between the CN + DC and CN tasks was observed around the left-frontal and right-frontal areas from 320 to 480 ms (P3a) and the parietal-central area from 480 to 800 ms (P3b and LPC). The latency of the P300 potential in the CN + DC task was significantly longer than that in the CN task at F3, Fz, F4, C4, Pz, and P4 (P < 0.05). Offline (P < 0.05 at superposing once, twice, and thrice) and online (P < 0.001) classification results showed that the average accuracies in the CN + DC task were significantly greater than that in the CN task. Similar results were found for online information transfer rates (ITRs; P < 0.001). In addition, we found that the average online accuracies in the DC task were greater than those in the CN task, although the difference was not statistically significant (P = 0.051). Significance: The active mental task based on task difficulty modulation can significantly improve the performance of the P300-speller, and that based on color distinction shows a trend of improved performance.

The variation in the distributions of recorded data between individuals leads to low classification accuracy. To address this issue, we introduce a multimodal fusion convolutional neural network (MFCNN). This network extracts common information from surface electromyography (sEMG) and accelerometer signals of different subjects using a two-stream CNN. To enhance the classification accuracy of a particular subject, a fine-tuning approach was implemented. The performance of the proposed method was assessed in four different scenarios, which include inter-subject classification, inter-subject classification when training data from multiple subjects, fine-tuned inter-subject classification, and fine-tuned inter-subject classification when training data from multiple subjects. The results demonstrate that in the inter-subject scenario, when multiple subjects are available for training, the MFCNN achieves higher classification accuracy (p < 0.05) than other neural networks and support vector machines (SVMs) that use sEMG signals (NN and SVM), accelerometer signals (accNN and accSVM), sEMG and accelerometer signals (MFNN and MFSVM) as inputs, as well as a CNN that uses sEMG signals as input after fine-tuning. Furthermore, compared with an MFCNN model trained with data from a single subject and an accCNN model trained with data from a single subject or multiple subjects, an MFCNN trained with multiple subjects demonstrated better performance on new subjects after fine-tuning (p < 0.05). This method can learn common features among different subjects and improve the performance of classification among subjects. Our proposed method demonstrates the innovation of using a multimodal fusion approach and two-stream CNN to improve inter-subject classification accuracy in upper limb movements.