A Hidden Markov, Multivariate Autoregressive (HMM-mAR) Network Framework for Analysis of Surface EMG (sEMG) Data

Abstract

As the primary noninvasive means to assess muscle activation, the surface electromyogram (sEMG) is of central importance for the study of motor behavior in both clinical and biomedical applications. However, multivariate sEMG analysis is complicated by the fact that data recorded during dynamic contractions are inherently nonstationary. To model this nonstationarity and to determine the dynamic muscle activity patterns during reaching movements, we propose combining hidden Markov models (HMMs) and multivariate autoregressive (mAR) models into a joint HMM-mAR framework. We further propose constructing muscle networks statistically by performing a second level, group analysis on the subject-specific models. Network structural features are subsequently investigated as input features for the purpose of classification. The proposed approach was applied to real sEMG recordings collected from healthy and stroke subjects during reaching movements. When examining group muscle networks, we note that specific muscle connection patterns were selectively recruited during reaching movements and were differentially recruited after stroke compared to healthy subjects. As the analysis was performed on the raw data, the amplitude and the underlying "carrier data" of sEMG signals, we notice that the HMM-mAR model fits the amplitude data well, but not the raw or carrier data. The proposed sEMG analysis framework represents a fundamental departure from existing methods where only the amplitude is typically analyzed or the mAR coefficients are directly used for classification. As the method may provide additional insights into motor control, it appears a promising approach warranting further study.Index Terms-Classification tree, expectation maximization (EM) algorithm, hidden Markov model (HMM), multivariate autoregressive (mAR) model, stroke, surface electromyography (sEMG).

Time series modeling of surface EMG based hand manipulation identification via expectation maximization algorithm

“…In order to gain better control performance, the primary thing to do is to extract the main feature of the test sEMG signal. At the same time, a sEMG signal is assumed to be the output of a stochastic dynamic process driven by the physiological properties of the muscle and the form of the contraction [7,19]. Also, a hidden Markov model (HMM) is a representation of a type of random process, it can describe time-varying processes in signal processing.…”

Self-adaptive Method for sEMG Movement Classification Based on Continuous Optimal Electrode Assortment

“…Typically, a sEMG signal movement classification consists on a pattern recognition / classification algorithm, which includes several popular methods such as LDA [2,3], Artificial Neural Networks (ANN) [4,5], Fuzzy Logic [6,7], Neuro Fuzzy [8], Genetic Algorithms, Support Vector Machines [9], Bayesian Networks [10][11][12] and Logistic Regression [13]. There are also some approaches using Independent Component Analysis (ICA) [14] and Principal Component Analysis (PCA) [15,16] focusing on dimensionality reduction and efficient computation, techniques focused on provide more efficiency to classification stage.…”

“…Recently, there have been some attempts to apply Hidden Markov Models (HMM) [11] and the Gaussian Mixture Model (GMM) [12] to upperlimb movement classification using myoelectric signals. Moreover, Bayesian approaches have as characteristics being good at assimilate prior data and construct an adaptive process without concrete information [10].…”

“…Even though a several methods have been attempted so far, achieving a powerful algorithm is still quite challenging because sEMG signals are commonly affected by the environmental changes, movement performed and subjects [10]. Due to the sensitive nature, external noise sources and artifacts can influence sEMG signals, which frequently mislead the classification algorithm, resulting in poor classification accuracy.…”

Identification of motion from multi-channel EMG signals for control of prosthetic hand

“…Different scientists worked for feature extraction using various features like variance (VAR) [15], Willison amplitude (WA) [15], mean absolute value slope (MAVS) [4,16], mean absolute value (MAV) [7,16], number of zero crossings (ZC) [7,16], wave form length (WL) [7,16], number of slope sign changes (SSC) [7,16], root means square (RMS) [26,27], auto regression (AR) coefficients [5,15,31], cepstrum coefficients [18], fast Fourier transform (FFT) coefficients [36], short time Fourier transform coefficients (STFT) [12], wavelet transform coefficients (WT) [19], and wavelet packet transform (WPT) coefficients [6]. Naik et al [27,28] extracted RMS feature from the EMG signal preprocessed with independent component analysis (ICA) to improve the reliability of identification of small movements and gestures of hand.…”