11 March 2015 in IEEE Transactions on Neural Systems and Rehabilitation Engineering
This study presents a progressive FastICA peel-off (PFP) framework for high-density surface electromyogram (EMG) decomposition. The novel framework is based on a shift-invariant model for describing surface EMG. The decomposition process can be viewed as progressively expanding the set of motor unit spike trains, which is primarily based on FastICA. To overcome the local convergence of FastICA, a peel-off strategy, i.e., removal of the estimated motor unit action potential trains from the previous step, is used to mitigate the effects of the already identified motor units, so more motor units can be extracted. A constrained FastICA is applied to assess the extracted spike trains and correct possible erroneous or missed spikes. These procedures work together to improve decomposition performance. The proposed framework was validated using simulated surface EMG signals with different motor unit numbers (30, 70, 91) and SNRs (20, 10, and 0 dB). The results demonstrated relatively large numbers of extracted motor units and high accuracies (high F1-scores). The framework was tested with 111 trials of 64-channel electrode array experimental surface EMG signals during the first dorsal interosseous muscle contraction at different intensities. On average 14.1 ±5.0 motor units were identified from each trial of experimental surface EMG signals.
7 December 2020 in IEEE Transactions on Neural Systems and Rehabilitation Engineering
Estimation of muscle contraction force based on the macroscopic feature of surface electromyography (SEMG) has been widely reported, but the use of microscopic neural drive information has not been thoroughly investigated. In this study, a novel method is proposed to process individual motor unit (MU) activities (firing sequences and action potential waveforms) derived from the decomposition of high density SEMG (HD-SEMG), and it is applied to muscle force estimation. In the proposed method, a supervised machine learning approach was conducted to determine the twitch force of each MU according to its action potential waveforms, which enables separate calculation of every MU’s contribution to force. Thus, the muscle force was predicted through a physiologically meaningful muscle force model. In the experiment, HD-SEMG data were recorded from the abductor pollicis brevis muscles of eight healthy subjects during their performance of thumb abduction with the force increasing gradually from zero to four force levels (10%, 20%, 30%, 40% of the maximal voluntary contraction), while the true muscle force was measured simultaneously. When the proposed method was used, the root mean square difference (RMSD) of the error of the estimated force with respect to the measured force was reported to be 8.3% ± 2.8%. The proposed method also significantly outperformed the other four common methods for force estimation (RMSD: from 11.7% to 20%, p < 0.001), demonstrating its effectiveness. This study offers a useful tool for exploiting the neural drive information towards muscle force estimation with improved precision. The proposed method has wide applications in precise motor control, sport and rehabilitation medicine.
9 November 2023 in IEEE Transactions on Biomedical Engineering
This study presents a method for adaptive online decomposition of high-density surface electromyogram (SEMG) signals to overcome the performance degradation during long-term recordings. The proposed method utilized the progressive FastICA peel-off (PFP) method and integrated a practical double-thread-parallel algorithm into the conventional two-stage calculation approach. During the offline initialization stage, a set of separation vectors was computed. In the subsequent online decomposition stage, a backend thread was implemented to periodically update the separation vectors using the constrained FastICA algorithm and the automatic PFP method. Concurrently, the frontend thread employed the newly updated separation vectors to accurately extract motor unit (MU) spike trains in real time. To assess the effectiveness of the proposed method, simulated and experimental SEMG signals from abductor pollicis brevis muscles of ten subjects were used for evaluation. The results demonstrated that the proposed method outperformed the conventional method, which relies on fixed separation vectors. Specifically, the proposed method showed an improved matching rate by 3.63% in simulated data and 1.98% in experimental data, along with an increased motor unit number by 2.39 in simulated data and 1.30 in experimental data. These findings illustrated the feasibility of the proposed method to enhance the performance of online SEMG decomposition. As a result, this work holds promise for various applications that require accurate MU firing activities in decoding neural commands and building neural-machine interfaces.
11 July 2023 in IEEE Transactions on Biomedical Engineering
Surface electromyogram (SEMG) decomposition provides a promising tool for decoding and understanding neural drive information non-invasively. In contrast to previous SEMG decomposition methods mainly developed in offline conditions, there are few studies on online SEMG decomposition. A novel method for online decomposition of SEMG data is presented using the progressive FastICA peel-off (PFP) method. The proposed online method utilized a two-stage approach, consisting of an offline prework stage for initializing high-quality separation vectors through the offline PFP algorithm, and an online decomposition stage for estimating source signals of different motor units by applying these vectors to the input SEMG data stream. Specifically, a new successive multi-threshold Otsu algorithm was developed in the online stage for determining each motor unit spike train (MUST) precisely with fast and simple computations (to replace a time-consuming iterative threshold setting in the original PFP method). The performance of the proposed online SEMG decomposition method was evaluated by both simulation and experimental approaches. When processing simulated SEMG data, the online PFP method achieved a decomposition accuracy of 97.37%, superior to that (95.1%) of an online method with a traditional k-means clustering algorithm for MUST extraction. Our method was also found to achieve superior performance at higher noise levels. For decomposing experimental SEMG data, the online PFP method was able to extract an average of 12.00 ± 3.46 MUs per trial, with a matching rate of 90.38%, with respect to the expert-guided offline decomposition results. Our study provides a valuable way of online decomposition of SEMG data with advanced applications in movement control and health.
13 June 2016 in BioMedical Engineering OnLine
Multi-channel recording of surface electromyographyic (EMG) signals is very likely to be contaminated by electrocardiographic (ECG) interference, specifically when the surface electrode is placed on muscles close to the heart. A novel fast independent component analysis (FastICA) based peel-off method is presented to remove ECG interference contaminating multi-channel surface EMG signals. Although demonstrating spatial variability in waveform shape, the ECG interference in different channels shares the same firing instants. Utilizing the firing information estimated from FastICA, ECG interference can be separated from surface EMG by a “peel off” processing. The performance of the method was quantified with synthetic signals by combining a series of experimentally recorded “clean” surface EMG and “pure” ECG interference. It was demonstrated that the new method can remove ECG interference efficiently with little distortion to surface EMG amplitude and frequency. The proposed method was also validated using experimental surface EMG signals contaminated by ECG interference. The proposed FastICA peel-off method can be used as a new and practical solution to eliminating ECG interference from multichannel EMG recordings.
20 November 2018 in IEEE Transactions on Neural Systems and Rehabilitation Engineering
The progressive FastICA peel-off (PFP) is a recently developed blind source separation approach for high-density surface EMG decomposition. This study explores a novel application of PFP for automatic decomposition of multi-channel intramuscular electromyogram signals. The automatic PFP (APFP) was used to decompose an open access multichannel intramuscular EMG dataset, simultaneously collected from the brachioradialis muscle using 6 to 8 fine wire or needle electrodes. Given usually limited number of intramuscular electrodes compared with high-density surface EMG recording, a modification was made to the original APFP framework to dramatically increase the decomposition yield. A total of 131 motor units were automatically decomposed by the APFP framework from 10 multichannel intramuscular EMG signals, among which 128 motor units were also manually identified from the expert interactive EMGLAB decomposition. The average matching rate of discharge instants for all the common motor units was (98.71±1.73) %. The outcomes of this study indicate that the APFP framework can also be used to automatically decompose multichannel intramuscular EMG with high accuracies, even though the number of recording channels is relatively small compared with high-density surface EMG.