Nonlinear Switching Dynamics in Surface Electromyography of the Spine

E.A. Jonckheere and P. Lohsoonthorn
Dept. of Electrical Engineering-Systems
University of Southern California
Los Angeles, CA 90089-2563

This paper develops a switching dynamics model of the surface Electromyographic (sEMG) signal generated during a condition which the mechanical attachment of the spinal dura to the cervical vertebra creates an unstable nonlinear feedback coupling between the biomechanics of the spine and the central nervous system (CNS). The sEMG signal recorded on the paraspinal muscles during this condition reveals "bursts" of accrued sEMG activity interrupting an otherwise quiet "background" signal. Statistical analysis of the autocorrelation and partial correlation functions of the burst and background parts of the signal reveals that the overall signal indeed switches between two models. Both the burst and the background models are dynamically modeled by ARIMA and ACE, and a switching logic, driven by the autocorrelation and the partial correlation, is designed, resulting in a switching model that matches the experimental sEMG signal fairly well.
more...

(In order to view the file above, it is necessary that you have Adobe Acrobat Reader. If you do not have this please Click Here to download your free copy now.)