Evaluating EMG Feature and Classifier Selection for Application to Partial-Hand Prosthesis Control

Abstract

Pattern recognition-based myoelectric control of upper-limb prostheses has the potential to restore control of multiple degrees of freedom. Though this control method has been extensively studied in individuals with higher-level amputations, few studies have investigated its effectiveness for individuals with partial-hand amputations. Most partial-hand amputees retain a functional wrist and the ability of pattern recognition-based methods to correctly classify hand motions from different wrist positions is not well studied. In this study, focusing on partial-hand amputees, we evaluate (1) the performance of non-linear and linear pattern recognition algorithms and (2) the performance of optimal EMG feature subsets for classification of four hand motion classes in different wrist positions for 16 non-amputees and 4 amputees. Our results show that linear discriminant analysis and linear and non-linear artificial neural networks perform significantly better than the quadratic discriminant analysis for both non-amputees and partial-hand amputees. For amputees, including information from multiple wrist positions significantly decreased error (p < 0.001) but no further significant decrease in error occurred when more than 4, 2, or 3 positions were included for the extrinsic (p = 0.07), intrinsic (p = 0.06), or combined extrinsic and intrinsic muscle EMG (p = 0.08), respectively. Finally, we found that a feature set determined by selecting optimal features from each channel outperformed the commonly used time domain (p < 0.001) and time domain/autoregressive feature sets (p < 0.01). This method can be used as a screening filter to select the features from each channel that provide the best classification of hand postures across different wrist positions.

Characterization of Forearm Muscle Activation in Duchenne Muscular Dystrophy via High-Density Electromyography: A Case Study on the Implications for Myoelectric Control

“…The 10 steady-state segments for every gesture were concatenated and created a 10-s vector. We trained the classifier by extracting four timedomain features from the raw segmented data including mean absolute value, zero crossing, slope sign change, and waveform length (41). We chose for a feature extraction window of 200 ms (with an overlap of 100 ms), which would be within acceptable range for real-time myoelectric applications (42).…”

Biceps Brachii Muscle Synergy and Target Reaching in a Virtual Environment

“…Similarly, Ameri et al (2014) used an artificial neural network (ANN), where visual training was considered better than force training to simultaneously estimate intended movements of multiple joints. Comparing the classifiers performance, Adewuyi et al (2016) found for non-amputees and partial-hand amputees that LDA and ANN perform better than the quadratic discriminant analysis. Betthauser et al (2018) developed a robust sparsity-based adaptive classification method to get a classification system which is appreciably less sensitive to signal deviations between training and testing.…”

Multi-length Windowed Feature Selection for Surface EMG Based Hand Motion Recognition

“…The classification accuracy under limb variation was improved by [11] with the selected subset out of 10 feature candidates. Adewuyi et al [12] evaluated the optimal sEMG feature subset in varying wrist positions for subjects with partial hand amputation and proved its superiority over the ensemble TD or TDAR features. Let alone the emphasis on the feature selection, the importance of the data segmentation with various window length of sEMG stream is addressed by [13].…”