Online semi-supervised learning for motor imagery EEG classification

Objective. We proposed a novel simultaneous hybrid brain–computer interface (BCI) by incorporating electrical stimulation into a motor imagery (MI) based BCI system. The goal of this study was to enhance the overall performance of an MI-based BCI. In addition, the brain oscillatory pattern in the hybrid task was also investigated. Approach. 64-channel electroencephalographic (EEG) data were recorded during MI, selective attention (SA) and hybrid tasks in fourteen healthy subjects. In the hybrid task, subjects performed MI with electrical stimulation which was applied to bilateral median nerve on wrists simultaneously. Main results. The hybrid task clearly presented additional steady-state somatosensory evoked potential (SSSEP) induced by electrical stimulation with MI-induced event-related desynchronization (ERD). By combining ERD and SSSEP features, the performance in the hybrid task was significantly better than in both MI and SA tasks, achieving a ~14% improvement in total relative to the MI task alone and reaching ~89% in mean classification accuracy. On the contrary, there was no significant enhancement obtained in performance while separate ERD feature was utilized in the hybrid task. In terms of the hybrid task, the performance using combined feature was significantly better than using separate ERD or SSSEP feature. Significance. The results in this work validate the feasibility of our proposed approach to form a novel MI-SSSEP hybrid BCI outperforming a conventional MI-based BCI through combing MI with electrical stimulation.

Motor imagery (MI) has been widely used to operate brain-computer interface (BCI) systems for rehabilitation and some life assistive devices. However, the current performance of an MI-based BCI cannot fully meet the needs of its in-field applications. Most of the BCIs utilizing a generalized feature for all participants have been found to greatly hamper the efficacy of the BCI system. Hence, some attempts have made on the exploration of subject-dependent parameters, but it remains challenging to enhance BCI performance as expected. To this end, in this study, we used the independent component analysis (ICA), which has been proved capable of isolating the pure motor-related component from non-motor-related brain processes and artifacts and extracting the common motor-related component across MI, motor execution (ME), and motor observation (MO) conditions. Then, a sliding window approach was used to detect significant mu-suppression from the baseline using the electroencephalographic (EEG) alpha power time course and, thus, the success rate of the mu-suppression detection could be assessed on a single-trial basis. By comparing the success rates using different parameters, we further quantified the extent of the improvement in each motor condition to evaluate the effectiveness of both generalized and individualized parameters. The results showed that in ME condition, the success rate under individualized latency and that under generalized latency was 90.0% and 77.75%, respectively; in MI condition, the success rate was 74.14% for individual latency and 58.47% for generalized latency, and in MO condition, the success rate was 67.89% and 61.26% for individual and generalized latency, respectively. As can be seen, the success rate in each motor condition was significantly improved by utilizing an individualized latency compared to that using the generalized latency. Moreover, the comparison of the individualized window latencies for the mu-suppression detection across different runs of the same participant as well as across different participants showed that the window latency was significantly more consistent in the intra-subject than in the inter-subject settings. As a result, we proposed that individualizing the latency for detecting the mu-suppression feature for each participant might be a promising attempt to improve the MI-based BCI performance.

Riemannian geometry-based classification (RGBC) gained popularity in the field of brain-computer interfaces (BCIs) lately, due to its ability to deal with non-stationarities arising in electroencephalography (EEG) data. Domain adaptation, however, is most often performed on sample covariance matrices (SCMs) obtained from EEG data, and thus might not fully account for components affecting covariance estimation itself, such as regional trends. Detrended cross-correlation analysis (DCCA) can be utilized to estimate the covariance structure of such signals, yet it is computationally expensive in its original form. A recently proposed online implementation of DCCA, however, allows for its fast computation and thus makes it possible to employ DCCA in real-time applications. In this study we propose to replace the SCM with the DCCA matrix as input to RGBC and assess its effect on offline and online BCI performance. First we evaluated the proposed decoding pipeline offline on previously recorded EEG data from 18 individuals performing left and right hand motor imagery (MI), and benchmarked it against vanilla RGBC and popular MI-detection approaches. Subsequently, we recruited eight participants (with previous BCI experience) who operated an MI-based BCI (MI-BCI) online using the DCCA-enhanced Riemannian decoder. Finally, we tested the proposed method on a public, multi-class MI-BCI dataset. During offline evaluations the DCCA-based decoder consistently and significantly outperformed the other approaches. Online evaluation confirmed that the DCCA matrix could be computed in real-time even for 22-channel EEG, as well as subjects could control the MI-BCI with high command delivery (normalized Cohen's κ: 0.7409 ± 0.1515) and sample-wise MI detection (normalized Cohen's κ: 0.5200 ± 0.1610). Post-hoc analysis indicated characteristic connectivity patterns under both MI conditions, with stronger connectivity in the hemisphere contralateral to the MI task. Additionally, fractal scaling exponent of neural activity was found increased in the contralateral compared to the ipsilateral motor cortices (C4 and C3 for left and right MI, respectively) in both classes. Combining DCCA with Riemannian geometry-based decoding yields a robust and effective decoder, that not only improves upon the SCM-based approach but can also provide relevant information on the neurophysiological processes behind MI.

EEG-based brain-computer interfaces (BCI) are non-invasive approaches for replacing or restoring motor functions in impaired patients, and direct brain-to-device communication in the general population. Motor imagery (MI) is one of the most used BCI paradigms, but its performance varies across individuals and certain users require substantial training to develop control. In this study, we propose to integrate a MI paradigm simultaneously with a recently proposed Overt Spatial Attention (OSA) paradigm, to accomplish BCI control. We evaluated a cohort of 25 human subjects' ability to control a virtual cursor in one- and two-dimensions over 5 BCI sessions. The subjects used 5 different BCI paradigms: MI alone, OSA alone, MI, and OSA simultaneously towards the same target (MI+OSA), and MI for one axis while OSA controls the other (MI/OSA and OSA/MI). Our results show that MI+OSA reached the highest average online performance in 2D tasks at 49% Percent Valid Correct (PVC), and statistically outperforms both MI alone (42%) and OSA alone (45%). MI+OSA had a similar performance to each subject's best individual method between MI alone and OSA alone (50%) and 9 subjects reached their highest average BCI performance using MI+OSA. Integrating MI and OSA leads to improved performance over both individual methods at the group level and is the best BCI paradigm option for some subjects. This work proposes a new BCI control paradigm that integrates two existing paradigms and demonstrates its value by showing that it can improve users' BCI performance.

This paper is focused on the experimental approach to explore the potential of independent component analysis (ICA) in the context of motor imagery (MI)-based brain-computer interface (BCI). We presented a simple and efficient algorithmic framework of ICA-based MI BCI (ICA-MIBCI) for the evaluation of four classical ICA algorithms (Infomax, FastICA, Jade, and Sobi) as well as a simplified Infomax (sInfomax). Two novel performance indexes, self-test accuracy and the number of invalid ICA filters, were employed to assess the performance of MIBCI based on different ICA variants. As a reference method, common spatial pattern (CSP), a commonly-used spatial filtering method, was employed for the comparative study between ICA-MIBCI and CSP-MIBCI. The experimental results showed that sInfomax-based spatial filters exhibited significantly better transferability in session to session and subject to subject transfer as compared to CSP-based spatial filters. The online experiment was also introduced to demonstrate the practicability and feasibility of sInfomax-based MIBCI. However, four classical ICA variants, especially FastICA, Jade, and Sobi, performed much worse as compared to sInfomax and CSP in terms of classification accuracy and stability. We consider that conventional ICA-based spatial filtering methods tend to be overfitting while applied to real-life electroencephalogram data. Nevertheless, the sInfomax-based experimental results indicate that ICA methods have a great space for improvement in the application of MIBCI. We believe that this paper could bring forth new ideas for the practical implementation of ICA-MIBCI.

Brain–computer interface (BCI) based on motor imagery (MI) supplies a communication method between humans and machines independent of external stimulation. MI is based on the imagination quality of the subjects, which resulted in large variability and unreliability of the signals. Improved MI paradigms and data analysis methods would be helpful to reduce the side effects of the individual specificity and temporal variation, and increase the classification performance of MI-based BCI. Visual guidance could improve the MI quality, but there is a lack of a systematic analysis of the forms of visual guidance and their effects on MI. In this article, we conducted a comparative study to evaluate the MI paradigms with different visual guidance manners. Five paradigms were designed, i.e., pure MI, MI with synchronous pursuit (SP), asynchronous pursuit (AP), synchronous saccade (SS), and asynchronous saccade (AS). Furthermore, a convolutional neural network (CNN) architecture with multiple receptive fields and attention module was proposed. Twenty subjects accomplished the experiments. The SP paradigm induced the most significant event-related desynchronization (ERD) phenomenon and the highest classification accuracy for electroencephalograph (EEG) signals. The proposed network achieved an average classification accuracy of 91.89% and standard deviation of 5.55%, which outperformed the compared methods. To test the applicability of the paradigm and the method, six subjects with different performance in the offline experiment then participated in an online experiment and a simulated brain-controlled trajectory tracking flight of the unmanned aerial vehicle (UAV). All the subjects could accomplish the task, and their performance was positively correlated with the classification accuracy and negatively correlated with the complexity of the tracking path. In general, the SP visual guidance effectively helped the subjects modulate their brain activity, leading to increased MI quality. The convolutional network with multiple receptive fields and an attention module showed the improved classification performance and robustness.

Even though it has long been felt that psychological state influences the performance of brain-computer interfaces (BCI), formal analysis to support this hypothesis has been scant. This study investigates the inter-relationship between motor imagery (MI) and mental fatigue using EEG: a. whether prolonged sequences of MI produce mental fatigue and b. whether mental fatigue affects MI EEG class separability. Eleven participants participated in the MI experiment, 5 of which quit in the middle because of experiencing high fatigue. The growth of fatigue was monitored using the Kernel Partial Least Square (KPLS) algorithm on the remaining 6 participants which shows that MI induces substantial mental fatigue. Statistical analysis of the effect of fatigue on motor imagery performance shows that high fatigue level significantly decreases MI EEG separability. Collectively, these results portray an MI-fatigue inter-connection, emphasizing the necessity of developing adaptive MI BCI by tracking mental fatigue.

A novel Morse code-inspired method for multiclass motor imagery brain–computer interface (BCI) design

The use of robotic devices to provide active motor support and sensory feedback of ongoing motor intention, by means of a Brain Computer Interface (BCI), has received growing support by recent literature, with particular focus on neurorehabilitation therapies. At the same time, performance in the use of the BCI has become a more critical factor, since it directly influences congruency and consistency of the provided sensory feedback. As motor imagery is the mental simulation of a given movement without depending on residual function, training of patients in the use of motor imagery BCI can be extended beyond each rehabilitation session, and practiced by using simpler devices than rehabilitation robots available in the hospital. In this work, we investigated the use of haptic stimulation provided by vibrating electromagnetic motors to enhance BCI system training. The BCI is based on motor imagery of hand grasping and designed to operate a hand exoskeleton. We investigated whether haptic stimulation at fingerpads proves to be more effective than stimulation at wrist, already experimented in literature, due to the higher density of mechano-receptors. Our results did not show significant differences between the two body locations in BCI performance, yet a wider and more stable event-relateddesynchronization appeared for the finger-located stimulation. Future investigations will put in relation training with haptic feedback at fingerpads with BCI performance using the handexoskeleton, in grasping tasks that naturally involve haptic feedback at fingerpads.

Motor imagery (MI) is a cognitive process wherein an individual mentally rehearses a specific movement without physically executing it. Recently, MI-based brain-computer interface (BCI) has attracted widespread attention. However, accurate decoding of MI and understanding of neural mechanisms still face huge challenges. These seriously hinder the clinical application and development of BCI systems based on MI. Thus, it is very necessary to develop new methods to decode MI tasks. In this work, we propose a multi-branch convolutional neural network (MBCNN) with a temporal convolutional network (TCN), an end-to-end deep learning framework to decode multi-class MI tasks. We first used MBCNN to capture the MI electroencephalography signals information on temporal and spectral domains through different convolutional kernels. Then, we introduce TCN to extract more discriminative features. The within-subject cross-session strategy is used to validate the classification performance on the dataset of BCI Competition IV-2a. The results showed that we achieved 75.08% average accuracy for 4-class MI task classification, outperforming several state-of-the-art approaches. The proposed MBCNN-TCN-Net framework successfully captures discriminative features and decodes MI tasks effectively, improving the performance of MI-BCIs. Our findings could provide significant potential for improving the clinical application and development of MI-based BCI systems.

A motor imagery (MI) brain-computer interface (BCI) plays an important role in the neurological rehabilitation training for stroke patients. Electroencephalogram (EEG)-based MI BCI has high temporal resolution, which is convenient for real-time BCI control. Therefore, we focus on EEG-based MI BCI in this paper. The identification of MI EEG signals is always quite challenging. Due to high inter-session/subject variability, each subject should spend long and tedious calibration time in collecting amounts of labeled samples for a subject-specific model. To cope with this problem, we present a supervised selective cross-subject transfer learning (sSCSTL) approach which simultaneously makes use of the labeled samples from target and source subjects based on Riemannian tangent space. Since the covariance matrices representing the multi-channel EEG signals belong to the smooth Riemannian manifold, we perform the Riemannian alignment to make the covariance matrices from different subjects close to each other. Then, all aligned covariance matrices are converted into the Riemannian tangent space features to train a classifier in the Euclidean space. To investigate the role of unlabeled samples, we further propose semi-supervised and unsupervised versions which utilize the total samples and unlabeled samples from target subject, respectively. Sequential forward floating search (SFFS) method is executed for source selection. All our proposed algorithms transfer the labeled samples from most suitable source subjects into the feature space of target subject. Experimental results on two publicly available MI datasets demonstrated that our algorithms outperformed several state-of-the-art algorithms using small number of the labeled samples from target subject, especially for good target subjects.

There is the need for enhanced processing techniques that aid the development of Brain-Computer Interfaces (BCIs), considering their wide use for communication and control. Several paradigms exist for developing BCIs. One of such is motor imagery (MI). MI-based BCIs have been implemented in a variety of ways. Key factors such as the device type, task paradigm, preprocessing, feature extraction and selection and classification techniques, must be properly considered in building BCIs. Also, factors such as the task at hand, target population, processing rate and usability of the online system must be considered. Considering this need, this review presents a summary of the existing techniques for motor imagery classification, stating common trends and challenges facing MI studies, with potential improvements that might be seen. Specifically, the review focuses on electroencephalography (EEG)-based MI BCIs, with works sampled over a wide range of time.

For individuals with high degrees of motor disability or locked-in syndrome, it is impractical or impossible to use mechanical switches to interact with electronic devices. Brain computer interfaces (BCIs) can use motor imagery to detect interaction intention from users but lack the accuracy of mechanical switches. Hence, there exists a strong need to improve the accuracy of EEG-based motor imagery BCIs attempting to implement an on/off switch. Here, we investigate how monitoring the pupil diameter of a person as a psycho-physiological parameter in addition to traditional EEG channels can improve the classification accuracy of a switch-like BCI. We have recently noticed in our lab (work not yet published) how motor imagery is associated with increases in pupil diameter when compared to a control rest condition. The pupil diameter parameter is easily accessible through video oculography since most gaze tracking systems report pupil diameter invariant to head position. We performed a user study with 30 participants using a typical EEG based motor imagery BCI. We used common spatial patterns to separate motor imagery, signaling movement intention, from a rest control condition. By monitoring the pupil diameter of the user and using this parameter as an additional feature, we show that the performance of the classifier trying to discriminate motor imagery from a control condition improves over the traditional approach using just EEG derived features. Given the limitations of EEG to construct highly robust and reliable BCIs, we postulate that multi-modal approaches, such as the one presented here that monitor several psycho-physiological parameters, can be a successful strategy in making BCIs more accurate and less vulnerable to constraints such as requirements for long training sessions or high signal to noise ratio of electrode channels.

In recent years, deep-learning-based motor imagery (MI) electroencephalography (EEG) decoding methods have shown great potential in the field of the brain–computer interface (BCI). The existing literature is relatively mature in decoding methods for two classes of MI tasks. However, with the increase in MI task classes, decoding studies for four classes of MI tasks need to be further explored. In addition, it is difficult to obtain large-scale EEG datasets. When the training data are limited, deep-learning-based decoding models are prone to problems such as overfitting and poor robustness. In this study, we design a data augmentation method for MI-EEG. The original EEG is slid along the time axis and reconstructed to expand the size of the dataset. Second, we combine the gated recurrent unit (GRU) and convolutional neural network (CNN) to construct a parallel-structured feature fusion network to decode four classes of MI tasks. The parallel structure can avoid temporal, frequency and spatial features interfering with each other. Experimenting on the well-known four-class MI dataset BCI Competition IV 2a shows a global average classification accuracy of 80.7% and a kappa value of 0.74. The proposed method improves the robustness of deep learning to decode small-scale EEG datasets and alleviates the overfitting phenomenon caused by insufficient data. The method can be applied to BCI systems with a small amount of daily recorded data.

Cortical activation and BCI performance during brief tactile imagery: A comparative study with motor imagery