Xinjun Sheng

Invariant Surface EMG Feature Against Varying Contraction Level for Myoelectric Control Based on Muscle Coordination

Variations in muscle contraction effort have a substantial impact on performance of pattern recognition based myoelectric control. Though incorporating changes into training phase could decrease the effect, the training time would be increased and the clinical viability would be limited. The modulation of force relies on the coordination of multiple muscles, which provides a possibility to classify motions with different forces without adding extra training samples. This study explores the property of muscle coordination in the frequency domain and found that the orientation of muscle activation pattern vector of the frequency band is similar for the same motion with different force levels. Two novel features based on discrete Fourier transform and muscle coordination were proposed subsequently, and the classification accuracy was increased by around 11% compared to the traditional time domain feature sets when classifying nine classes of motions with three different force levels. Further analysis found that both features decreased the difference among different forces of the same motion p <; 0.005) and maintained the distance among different motions p > 0.1). This study also provided a potential way for simultaneous classification of hand motions and forces without training at all force levels.

User adaptation in long-term, open-loop myoelectric training: implications for EMG pattern recognition in prosthesis control

OBJECTIVE Recent studies have reported that the classification performance of electromyographic (EMG) signals degrades over time without proper classification retraining. This problem is relevant for the applications of EMG pattern recognition in the control of active prostheses. APPROACH In this study we investigated the changes in EMG classification performance over 11 consecutive days in eight able-bodied subjects and two amputees. MAIN RESULTS It was observed that, when the classifier was trained on data from one day and tested on data from the following day, the classification error decreased exponentially but plateaued after four days for able-bodied subjects and six to nine days for amputees. The between-day performance became gradually closer to the corresponding within-day performance. SIGNIFICANCE These results indicate that the relative changes in EMG signal features over time become progressively smaller when the number of days during which the subjects perform the pre-defined motions are increased. The performance of the motor tasks is thus more consistent over time, resulting in more repeatable EMG patterns, even if the subjects do not have any external feedback on their performance. The learning curves for both able-bodied subjects and subjects with limb deficiencies could be modeled as an exponential function. These results provide important insights into the user adaptation characteristics during practical long-term myoelectric control applications, with implications for the design of an adaptive pattern recognition system.

Combining Motor Imagery With Selective Sensation Toward a Hybrid-Modality BCI

A hybrid modality brain-computer interface (BCI) is proposed in this paper, which combines motor imagery with selective sensation to enhance the discrimination between left and right mental tasks, e.g., the classification between left/ right stimulation sensation and right/ left motor imagery. In this paradigm, wearable vibrotactile rings are used to stimulate both the skin on both wrists. Subjects are required to perform the mental tasks according to the randomly presented cues (i.e., left hand motor imagery, right hand motor imagery, left stimulation sensation or right stimulation sensation). Two-way ANOVA statistical analysis showed a significant group effect (F (2,20) = 7.17, p = 0.0045), and the Benferroni-corrected multiple comparison test (with α = 0.05) showed that the hybrid modality group is 11.13% higher on average than the motor imagery group, and 10.45% higher than the selective sensation group. The hybrid modality experiment exhibits potentially wider spread usage within ten subjects crossed 70% accuracy, followed by four subjects in motor imagery and five subjects in selective sensation. Six subjects showed statistically significant improvement ( Benferroni-corrected) in hybrid modality in comparison with both motor imagery and selective sensation. Furthermore, among subjects having difficulties in both motor imagery and selective sensation, the hybrid modality improves their performance to 90% accuracy. The proposed hybrid modality BCI has demonstrated clear benefits for those poorly performing BCI users. Not only does the requirement of motor and sensory anticipation in this hybrid modality provide basic function of BCI for communication and control, it also has the potential for enhancing the rehabilitation during motor recovery.

Continuous estimation of finger joint angles under different static wrist motions from surface EMG signals

Abstract In this paper, a solution is proposed to predict the finger joint angle using electromyography (EMG) towards application for partial-hand amputees with functional wrist. In the experimental paradigm, the subject was instructed to continuously move one finger (middle finger for able-bodied subjects and index finger for partial-hand amputees) up to the maximum angle of flexion and extension while the wrist was conducting seven different wrist motions. A switching regime, including one linear discriminant analysis (LDA) classifier and fourteen state-space models, was proposed to continuously decode finger joint angles. LDA classifier was used to recognize which static wrist motion that the subject was conducting and choose the corresponding two state-space models for decoding joint angles of the finger with two degrees of freedom (DOFs). The average classification error rate (CER) was 6.18%, demonstrating that these seven static wrist motions along with the continuous movement of the finger could be classified. To improve the classification performance, a preprocessing method, class-wise stationary subspace analysis (cwSSA), was firstly adopted to extract the stationary components from original EMG data. Consequently, the average CER was reduced by 1.82% ( p R 2 ) of the two joint angles of the finger across seven static wrist motions achieved 0.843. This result shows that the fingers joint angles can be continuously estimated well while the wrist was conducting different static motions simultaneously. The average accuracy of seven static wrist motions with and without cwSSA and the average estimation performance of the two joint angles of the finger prove that the proposed switching regime is effective for continuous estimation of the finger joint angles under different static wrist motions from EMG.

Selective Sensation Based Brain-Computer Interface via Mechanical Vibrotactile Stimulation

In this work, mechanical vibrotactile stimulation was applied to subjects’ left and right wrist skins with equal intensity, and a selective sensation perception task was performed to achieve two types of selections similar to motor imagery Brain-Computer Interface. The proposed system was based on event-related desynchronization/synchronization (ERD/ERS), which had a correlation with processing of afferent inflow in human somatosensory system, and attentional effect which modulated the ERD/ERS. The experiments were carried out on nine subjects (without experience in selective sensation), and six of them showed a discrimination accuracy above 80%, three of them above 95%. Comparative experiments with motor imagery (with and without presence of stimulation) were also carried out, which further showed the feasibility of selective sensation as an alternative BCI task complementary to motor imagery. Specifically there was significant improvement () from near 65% in motor imagery (with and without presence of stimulation) to above 80% in selective sensation on some subjects. The proposed BCI modality might well cooperate with existing BCI modalities in the literature in enlarging the widespread usage of BCI system.

A New Time Synchronization Method for Reducing Quantization Error Accumulation Over Real-Time Networks: Theory and Experiments

In real-time network-based systems with long linear paths, the growth rate of time synchronization error is the major barrier to the scalability of systems even if a transparent clock mechanism of IEEE 1588 is used. This paper is devoted to designing a new time synchronization method for such systems. In the proposed algorithm, a proportional-integral (PI) clock servo is used to achieve the frequency compensation. In order to reduce the growth rate of synchronization error due to the quantization error in timestamping, a Kalman filter is designed based on a state-variable model, which is built for the PI controller-tuned slave clock. In addition, the quantization effect is analyzed and the variance of quantization error is quantitatively estimated for each slave node. Experiments are performed to validate its effectiveness and demonstrate that the peak-to-peak jitter is measured to be only 59.37 ns after four hops, and the growth rate of synchronization error can also be significantly reduced by the presented synchronization method. This indicates that the maximum number of networked nodes can be correspondingly increased.

Improving robustness against electrode shift of high density EMG for myoelectric control through common spatial patterns.

BackgroundMost prosthetic myoelectric control studies have concentrated on low density (less than 16 electrodes, LD) electromyography (EMG) signals, due to its better clinical applicability and low computation complexity compared with high density (more than 16 electrodes, HD) EMG signals. Since HD EMG electrodes have been developed more conveniently to wear with respect to the previous versions recently, HD EMG signals become an alternative for myoelectric prostheses. The electrode shift, which may occur during repositioning or donning/doffing of the prosthetic socket, is one of the main reasons for degradation in classification accuracy (CA).MethodsHD EMG signals acquired from the forearm of the subjects were used for pattern recognition-based myoelectric control in this study. Multiclass common spatial patterns (CSP) with two types of schemes, namely one versus one (CSP-OvO) and one versus rest (CSP-OvR), were used for feature extraction to improve the robustness against electrode shift for myoelectric control. Shift transversal (ST1 and ST2) and longitudinal (SL1 and SL2) to the direction of the muscle fibers were taken into consideration. We tested nine intact-limb subjects for eleven hand and wrist motions. The CSP features (CSP-OvO and CSP-OvR) were compared with three commonly used features, namely time-domain (TD) features, time-domain autoregressive (TDAR) features and variogram (Variog) features.ResultsCompared with the TD features, the CSP features significantly improved the CA over 10 % in all shift configurations (ST1, ST2, SL1 and SL2). Compared with the TDAR features, a. the CSP-OvO feature significantly improved the average CA over 5 % in all shift configurations; b. the CSP-OvR feature significantly improved the average CA in shift configurations ST1, SL1 and SL2. Compared with the Variog features, the CSP features significantly improved the average CA in longitudinal shift configurations (SL1 and SL2).ConclusionThe results demonstrated that the CSP features significantly improved the robustness against electrode shift for myoelectric control with respect to the commonly used features.

Reduced Daily Recalibration of Myoelectric Prosthesis Classifiers Based on Domain Adaptation

Control scheme design based on surface electromyography (sEMG) pattern recognition has been the focus of much research on a myoelectric prosthesis (MP) technology. Due to inherent nonstationarity in sEMG signals, prosthesis systems may need to be recalibrated day after day in daily use applications; thereby, hindering MP usability. In order to reduce the recalibration time in the subsequent days following the initial training, we propose a domain adaptation (DA) framework, which automatically reuses the models trained in earlier days as input for two baseline classifiers: a polynomial classifier (PC) and a linear discriminant analysis (LDA). Two novel algorithms of DA are introduced, one for PC and the other one for LDA. Five intact-limbed subjects and two transradial-amputee subjects participated in an experiment lasting ten days, to simulate the application of a MP over multiple days. The experiment results of four methods were compared: PC-DA (PC with DA), PC-BL (baseline PC), LDA-DA (LDA with DA), and LDA-BL (baseline LDA). In a new day, the DA methods reuse nine pretrained models, which were calibrated by 40 s training data per class in nine previous days. We show that the proposed DA methods significantly outperform nonadaptive baseline methods. The improvement in classification accuracy ranges from 5.49% to 28.48%, when the recording time per class is 2 s. For example, the average classification rates of PC-BL and PC-DA are 83.70% and 92.99%, respectively, for intact-limbed subjects with a nine-motions classification task. These results indicate that DA has the potential to improve the usability of MPs based on pattern recognition, by reducing the calibration time.

Improving Myoelectric Control for Amputees through Transcranial Direct Current Stimulation

Most prosthetic myoelectric control studies have shown good performance for unimpaired subjects. However, performance is generally unacceptable for amputees. The primary problem is the poor quality of electromyography (EMG) signals of amputees compared with healthy individuals. To improve clinical performance of myoelectric control, this study explored transcranial direct current stimulation (tDCS) to modulate brain activity and enhance EMG quality. We tested six unilateral transradial amputees by applying active and sham anodal tDCS separately on two different days. Surface EMG signals were acquired from the affected and intact sides for 11 hand and wrist motions in the pre-tDCS and post-tDCS sessions. Autoregression coefficients and linear discriminant analysis classifiers were used to process the EMG data for pattern recognition of the 11 motions. For the affected side, active anodal tDCS significantly reduced the average classification error rate (CER) by 10.1%, while sham tDCS had no such effect. For the intact side, the average CER did not change on the day of sham tDCS but increased on the day of active tDCS. These results demonstrated that tDCS could modulate brain function and improve EMG-based classification performance for amputees. It has great potential in dramatically reducing the length of learning process of amputees for effectively using myoelectrically controlled multifunctional prostheses.

Simultaneously Optimizing Spatial Spectral Features Based on Mutual Information for EEG Classification

High performance of the brain-computer interface (BCI) needs efficient algorithms to extract discriminative features from raw electroencephalography (EEG) signals. In this paper, we present a novel scheme to extract spatial spectral features for the motor imagery-based BCI. The learning task is formulated by maximizing the mutual information between spatial spectral features (MMISS) and class labels, by which a unique objective function directly related to Bayes classification error is optimized. The spatial spectral features are assumed to follow a parametric Gaussian distribution, which has been validated by the normal distribution Mardias test, and under this assumption the estimation of mutual information is derived. We propose a gradient based alternative and iterative learning algorithm to optimize the cost function and derive the spatial and spectral filters simultaneously. The experimental results on dataset IVa of BCI competition III and dataset IIa of BCI competition IV show that the proposed MMISS is able to efficiently extract discriminative features from motor imagery-based EEG signals to enhance the classification accuracy compared to other existing algorithms.