Time-frequency coherence of categorized sEMG data during dynamic contractions of biceps, triceps, and brachioradialis as an approach for spasticity detection

Abstract

The assessment of muscular interactions between biceps, triceps, and brachioradialis can be used as an approach for the detection of spasticity in the upper limbs. A crucial prerequisite for the aforementioned validation of muscular interactions is the calculation of time frequencies due to the non-stationary characteristics of electromyographic (EMG) signals and thus the estimation of coherences. Adding biomechanical parameters increases the validity of the assessment process and simplifies the comparison of EMG data as a result of categorization. In this numerical-experimental study, a method will be introduced by using the smoothed pseudo Wigner-Ville distribution and a categorization algorithm to estimate and categorize coherences between biceps, triceps, and brachioradialis during dynamic contractions. The categorization will be performed according to the type of contraction, external load, joint angle, and angular velocity and will be used to assess 10 healthy subjects and 6 patients with spasticity. Generally, the introduced method shows the velocity dependence of coherence during spasticity in extension movements as well as much stronger muscular co-activation between triceps, biceps, and brachioradialis in spastic patients in comparison to healthy subjects. Furthermore, the influence of variables e.g. as joint angle, angular velocities, and type of contraction on the coherence is quantified.

Keywords: Categorization; Coherence; Spasticity; Time-frequency analysis; sEMG.