Multi-domain feature joint optimization based on multi-view learning for improving the EEG decoding.

Background: Brain-computer interface (BCI) systems based on motor imagery (MI) have been widely used in neurorehabilitation. Feature extraction applied by the common spatial pattern (CSP) is very popular in MI classification. The effectiveness of CSP is highly affected by the frequency band and time window of electroencephalogram (EEG) segments and channels selected. Objective: In this study, the multi-domain feature joint optimization (MDFJO) based on the multi-view learning method is proposed, which aims to select the discriminative features enhancing the classification performance. Method: The channel patterns are divided using the Fisher discriminant criterion (FDC). Furthermore, the raw EEG is intercepted for multiple sub-bands and time interval signals. The high-dimensional features are constructed by extracting features from CSP on each EEG segment. Specifically, the multi-view learning method is used to select the optimal features, and the proposed feature sparsification strategy on the time level is proposed to further refine the optimal features. Results: Two public EEG datasets are employed to validate the proposed MDFJO method. The average classification accuracy of the MDFJO in Data 1 and Data 2 is 88.29 and 87.21%, respectively. The classification result of MDFJO was significantly better than MSO (p < 0.05), FBCSP32 (p < 0.01), and other competing methods (p < 0.001). Conclusion: Compared with the CSP, sparse filter band common spatial pattern (SFBCSP), and filter bank common spatial pattern (FBCSP) methods with channel numbers 16, 32 and all channels as well as MSO, the MDFJO significantly improves the test accuracy. The feature sparsification strategy proposed in this article can effectively enhance classification accuracy. The proposed method could improve the practicability and effectiveness of the BCI system.

Brain oscillations in reflecting motor status and recovery induced by action observation-driven robotic hand intervention in chronic stroke.

Hand rehabilitation in chronic stroke remains challenging, and finding markers that could reflect motor function would help to understand and evaluate the therapy and recovery. The present study explored whether brain oscillations in different electroencephalogram (EEG) bands could indicate the motor status and recovery induced by action observation-driven brain-computer interface (AO-BCI) robotic therapy in chronic stroke. The neurophysiological data of 16 chronic stroke patients who received 20-session BCI hand training is the basis of the study presented here. Resting-state EEG was recorded during the observation of non-biological movements, while task-stage EEG was recorded during the observation of biological movements in training. The motor performance was evaluated using the Action Research Arm Test (ARAT) and upper extremity Fugl-Meyer Assessment (FMA), and significant improvements (p < 0.05) on both scales were found in patients after the intervention. Averaged EEG band power in the affected hemisphere presented negative correlations with scales pre-training; however, no significant correlations (p > 0.01) were found both in the pre-training and post-training stages. After comparing the variation of oscillations over training, we found patients with good and poor recovery presented different trends in delta, low-beta, and high-beta variations, and only patients with good recovery presented significant changes in EEG band power after training (delta band, p < 0.01). Importantly, motor improvements in ARAT correlate significantly with task EEG power changes (low-beta, c.c = 0.71, p = 0.005; high-beta, c.c = 0.71, p = 0.004) and task/rest EEG power ratio changes (delta, c.c = -0.738, p = 0.003; low-beta, c.c = 0.67, p = 0.009; high-beta, c.c = 0.839, p = 0.000). These results suggest that, in chronic stroke, EEG band power may not be a good indicator of motor status. However, ipsilesional oscillation changes in the delta and beta bands provide potential biomarkers related to the therapeutic-induced improvement of motor function in effective BCI intervention, which may be useful in understanding the brain plasticity changes and contribute to evaluating therapy and recovery in chronic-stage motor rehabilitation.

A novel multiple time-frequency sequential coding strategy for hybrid brain-computer interface.

Background: For brain-computer interface (BCI) communication, electroencephalography provides a preferable choice due to its high temporal resolution and portability over other neural recording techniques. However, current BCIs are unable to sufficiently use the information from time and frequency domains simultaneously. Thus, we proposed a novel hybrid time-frequency paradigm to investigate better ways of using the time and frequency information. Method: We adopt multiple omitted stimulus potential (OSP) and steady-state motion visual evoked potential (SSMVEP) to design the hybrid paradigm. A series of pre-experiments were undertaken to study factors that would influence the feasibility of the hybrid paradigm and the interaction between multiple features. After that, a novel Multiple Time-Frequencies Sequential Coding (MTFSC) strategy was introduced and explored in experiments. Results: Omissions with multiple short and long durations could effectively elicit time and frequency features, including the multi-OSP, ERP, and SSVEP in this hybrid paradigm. The MTFSC was feasible and efficient. The preliminary online analysis showed that the accuracy and the ITR of the nine-target stimulator over thirteen subjects were 89.04% and 36.37 bits/min. Significance: This study first combined the SSMVEP and multi-OSP in a hybrid paradigm to produce robust and abundant time features for coding BCI. Meanwhile, the MTFSC proved feasible and showed great potential in improving performance, such as expanding the number of BCI targets by better using time information in specific stimulated frequencies. This study holds promise for designing better BCI systems with a novel coding method.

Adaptive binary multi-objective harmony search algorithm for channel selection and cross-subject generalization in motor imagery-based BCI.

Objective.Multi-channel electroencephalogram data containing redundant information and noise may result in low classification accuracy and high computational complexity, which limits the practicality of motor imagery (MI)-based brain-computer interface (BCI) systems. Therefore, channel selection can improve BCI performance and contribute to user convenience. Additionally, cross-subject generalization is a key topic in the channel selection of MI-based BCI.Approach.In this study, an adaptive binary multi-objective harmony search (ABMOHS) algorithm is proposed to select the optimal set of channels. Furthermore, a new adaptive cross-subject generalization model (ACGM) is proposed. Three public MI datasets were used to validate the effectiveness of the proposed method.Main results.The Wilcoxon signed-rank test was performed on the test accuracies, and the results indicated that the ABMOHS method significantly outperformed all channels (p< 0.001), the C3-Cz-C4 channels (p< 0.001), and 20 channels (p< 0.001) in the sensorimotor cortex. The ABMOHS algorithm based on Fisher's linear discriminant analysis (FLDA) and support vector machine (SVM) classifiers greatly reduces the number of selected channels, especially for larger channel sizes (Dataset 2), and obtains a comparative classification performance. Although there was no significant difference in test classification performance between ABMOHS and non-dominated sorting genetic algorithm II (NSGA-II) when FLDA and SVM were used, ABMOHS required less computational time than NSGA-II. Furthermore, the number of channels obtained by ABMOHS algorithm were significantly smaller than those obtained by common spatial pattern-Rank and correlation-based channel selection algorithm. Additionally, the generalization of ACGM to untrained subjects shows that the mean test classification accuracy of ACGM created by a small sample of trained subjects is significantly better than that of Special-16 and Special-32.Significance.The proposed method can reduce the calibration time in the training phase and improve the practicability of MI-BCI.

Sensorimotor Rhythm-Brain Computer Interface With Audio-Cue, Motor Observation and Multisensory Feedback for Upper-Limb Stroke Rehabilitation: A Controlled Study.

Several studies have shown the positive clinical effect of brain computer interface (BCI) training for stroke rehabilitation. This study investigated the efficacy of the sensorimotor rhythm (SMR)-based BCI with audio-cue, motor observation and multisensory feedback for post-stroke rehabilitation. Furthermore, we discussed the interaction between training intensity and training duration in BCI training. Twenty-four stroke patients with severe upper limb (UL) motor deficits were randomly assigned to two groups: 2-week SMR-BCI training combined with conventional treatment (BCI Group, BG, n = 12) and 2-week conventional treatment without SMR-BCI intervention (Control Group, CG, n = 12). Motor function was measured using clinical measurement scales, including Fugl-Meyer Assessment-Upper Extremities (FMA-UE; primary outcome measure), Wolf Motor Functional Test (WMFT), and Modified Barthel Index (MBI), at baseline (Week 0), post-intervention (Week 2), and follow-up week (Week 4). EEG data from patients allocated to the BG was recorded at Week 0 and Week 2 and quantified by mu suppression means event-related desynchronization (ERD) in mu rhythm (8-12 Hz). All functional assessment scores (FMA-UE, WMFT, and MBI) significantly improved at Week 2 for both groups (p < 0.05). The BG had significantly higher FMA-UE and WMFT improvement at Week 4 compared to the CG. The mu suppression of bilateral hemisphere both had a positive trend with the motor function scores at Week 2. This study proposes a new effective SMR-BCI system and demonstrates that the SMR-BCI training with audio-cue, motor observation and multisensory feedback, together with conventional therapy may promote long-lasting UL motor improvement. Clinical Trial Registration: [http://www.chictr.org.cn], identifier [ChiCTR2000041119].

Feature optimization based on improved novel global harmony search algorithm for motor imagery electroencephalogram classification.

Background: Effectively decoding electroencephalogram (EEG) pattern for specific mental tasks is a crucial topic in the development of brain-computer interface (BCI). Extracting common spatial pattern (CSP) features from motor imagery EEG signals is often highly dependent on the selection of frequency band and time interval. Therefore, optimizing frequency band and time interval would contribute to effective feature extraction and accurate EEG decoding. Objective: This study proposes an approach based on an improved novel global harmony search (INGHS) to optimize frequency-time parameters for effective CSP feature extraction. Methods: The INGHS algorithm is applied to find the optimal frequency band and temporal interval. The linear discriminant analysis and support vector machine are used for EEG pattern decoding. Extensive experimental studies are conducted on three EEG datasets to assess the effectiveness of our proposed method. Results: The average test accuracy obtained by the time-frequency parameters selected by the proposed INGHS method is slightly better than artificial bee colony (ABC) and particle swarm optimization (PSO) algorithms. Furthermore, the INGHS algorithm is superior to PSO and ABC in running time. Conclusion: These superior experimental results demonstrate that the optimal frequency band and time interval selected by the INGHS algorithm could significantly improve the decoding accuracy compared with the traditional CSP method. This method has a potential to improve the performance of MI-based BCI systems.

Wearable Ankle Robots in Post-stroke Rehabilitation of Gait: A Systematic Review.

Background: Stroke causes weak functional mobility in survivors and affects the ability to perform activities of daily living. Wearable ankle robots are a potential intervention for gait rehabilitation post-stroke. Objective: The aim of this study is to provide a systematic review of wearable ankle robots, focusing on the overview, classification and comparison of actuators, gait event detection, control strategies, and performance evaluation. Method: Only English-language studies published from December 1995 to July 2018 were searched in the following databases: PubMed, EMBASE, Web of Science, Scopus, IEEE Xplore, Science Direct, SAGE journals. Result: A total of 48 articles were selected and 97 stroke survivors participated in these trials. Findings showed that few comparative trials were conducted among different actuators or control strategies. Moreover, mixed sensing technology which combines kinematic with kinetic information was effective in detecting motion intention of stroke survivors. Furthermore, all the selected clinical studies showed an improvement in the peak dorsiflexion degree of the swing phase, propulsion on the paretic side during push-off, and further enhanced walking speed after a period of robot-assisted ankle rehabilitation training. Conclusions: Preliminary findings suggest that wearable ankle robots have certain clinical benefits for the treatment of hemiplegic gait post-stroke. In the near future, a multicenter randomized controlled clinical trial is extremely necessary to enhance the clinical effectiveness of wearable ankle robots.

Brain Functional Networks Study of Subacute Stroke Patients With Upper Limb Dysfunction After Comprehensive Rehabilitation Including BCI Training.

Brain computer interface (BCI)-based training is promising for the treatment of stroke patients with upper limb (UL) paralysis. However, most stroke patients receive comprehensive treatment that not only includes BCI, but also routine training. The purpose of this study was to investigate the topological alterations in brain functional networks following comprehensive treatment, including BCI training, in the subacute stage of stroke. Twenty-five hospitalized subacute stroke patients with moderate to severe UL paralysis were assigned to one of two groups: 4-week comprehensive treatment, including routine and BCI training (BCI group, BG, n = 14) and 4-week routine training without BCI support (control group, CG, n = 11). Functional UL assessments were performed before and after training, including, Fugl-Meyer Assessment-UL (FMA-UL), Action Research Arm Test (ARAT), and Wolf Motor Function Test (WMFT). Neuroimaging assessment of functional connectivity (FC) in the BG was performed by resting state functional magnetic resonance imaging. After training, as compared with baseline, all clinical assessments (FMA-UL, ARAT, and WMFT) improved significantly (p < 0.05) in both groups. Meanwhile, better functional improvements were observed in FMA-UL (p < 0.05), ARAT (p < 0.05), and WMFT (p < 0.05) in the BG. Meanwhile, FC of the BG increased across the whole brain, including the temporal, parietal, and occipital lobes and subcortical regions. More importantly, increased inter-hemispheric FC between the somatosensory association cortex and putamen was strongly positively associated with UL motor function after training. Our findings demonstrate that comprehensive rehabilitation, including BCI training, can enhance UL motor function better than routine training for subacute stroke patients. The reorganization of brain functional networks topology in subacute stroke patients allows for increased coordination between the multi-sensory and motor-related cortex and the extrapyramidal system. Future long-term, longitudinal, controlled neuroimaging studies are needed to assess the effectiveness of BCI training as an approach to promote brain plasticity during the subacute stage of stroke.

Hand Rehabilitation Robotics on Poststroke Motor Recovery.

The recovery of hand function is one of the most challenging topics in stroke rehabilitation. Although the robot-assisted therapy has got some good results in the latest decades, the development of hand rehabilitation robotics is left behind. Existing reviews of hand rehabilitation robotics focus either on the mechanical design on designers' view or on the training paradigms on the clinicians' view, while these two parts are interconnected and both important for designers and clinicians. In this review, we explore the current literature surrounding hand rehabilitation robots, to help designers make better choices among varied components and thus promoting the application of hand rehabilitation robots. An overview of hand rehabilitation robotics is provided in this paper firstly, to give a general view of the relationship between subjects, rehabilitation theories, hand rehabilitation robots, and its evaluation. Secondly, the state of the art hand rehabilitation robotics is introduced in detail according to the classification of the hardware system and the training paradigm. As a result, the discussion gives available arguments behind the classification and comprehensive overview of hand rehabilitation robotics.

Robotics in Lower-Limb Rehabilitation after Stroke.

With the increase in the elderly, stroke has become a common disease, often leading to motor dysfunction and even permanent disability. Lower-limb rehabilitation robots can help patients to carry out reasonable and effective training to improve the motor function of paralyzed extremity. In this paper, the developments of lower-limb rehabilitation robots in the past decades are reviewed. Specifically, we provide a classification, a comparison, and a design overview of the driving modes, training paradigm, and control strategy of the lower-limb rehabilitation robots in the reviewed literature. A brief review on the gait detection technology of lower-limb rehabilitation robots is also presented. Finally, we discuss the future directions of the lower-limb rehabilitation robots.