A novel spatiotemporal muscle activity imaging approach based on the Extended Kalman Filter.

Author: Wang J, Zhang Y, Zhu X, Zhou P, Liu C, Rymer WZ.
Conference/Journal: Conf Proc IEEE Eng Med Biol Soc.
Date published: 2012 Aug 20
Other: Volume ID: 2012 , Pages: 6236-8 , Special Notes: doi: 10.1109/EMBC.2012.6347419 , Word Count: 178


A novel spatiotemporal muscle activity imaging (sMAI) approach has been developed using the Extended Kalman Filter (EKF) to reconstruct internal muscle activities from non-invasive multi-channel surface electromyogram (sEMG) recordings. A distributed bioelectric dipole source model is employed to describe the internal muscle activity space, and a linear relationship between the muscle activity space and the sEMG measurement space is then established. The EKF is employed to recursively solve the ill-posed inverse problem in the sMAI approach, in which the weighted minimum norm (WMN) method is utilized to calculate the initial state and a new nonlinear method is developed based on the propagating features of muscle activities to predict the recursive state. A series of computer simulations was conducted to test the performance of the proposed sMAI approach. Results show that the localization error rapidly decreases over 35% and the overlap ratio rapidly increases over 45% compared to the results achieved using the WMN method only. The present promising results demonstrate the feasibility of utilizing the proposed EKF-based sMAI approach to accurately reconstruct internal muscle activities from non-invasive sEMG recordings.
PMID: 23367354