Surface electromyography (sEMG) records muscle activities from the surface of muscles, which offers a wealth of information concerning muscle activation patterns in both research and clinical settings. A key principle underlying sEMG analyses is the decomposition of the signal into a number of motor unit action potentials (MUAPs) that capture most of the relevant features embedded in a low-dimensional space. Toward this, the principal component analysis (PCA) has extensively been sought after, whereby the original sEMG data are translated into low-dimensional MUAP components with a reduced level of redundancy. The objective of this paper is to disseminate the role of PCA in conjunction with the quantitative sEMG analyses. Following the preliminaries on the sEMG methodology and a statement of PCA algorithm, an exhaustive collection of PCA applications related to sEMG data is in order. Alongside the technical challenges associated with the PCA-based sEMG processing, the envisaged research trend is also discussed.
- artificial neural network (ANN)
- motor unit action potential (MUAP)
- Myoelectric signal
- principal component analysis (PCA)
- self-organizing feature map (SOFM)
- support vector regression (SVR)
- Surface electromyography (sEMG)