Wavelet Analysis of Surface EMG in Cerebral Palsy

Several methods exist to evaluate motor function in the child with cerebral palsy and are used to assess the outcome of a clinical intervention. However, these scales are not directed towards measuring the changes in muscle activity patterns that can result from the intervention. For example, there are classification scales aimed at measuring motor function and functional abilities, and indices of gait function. These scores, while providing a way to quantify function and mechanics, do not directly measure muscle activation characteristics. Therefore, these tests may be insensitive to how the intervention has directly affected muscle function, which is usually the focus of the intervention (i.e. botulinum toxin, functional electrical stimulation, dorsal rhizotomy). Muscle biopsies and motor evoked potentials can provide information about the muscle activation characteristics, however, they are invasive and there are concerns about using these techniques on the pediatric population and/or the practicality of clinical implementation, especially since they do not provide insight into how the muscle behaves during a functional task. One method that can be used to provide insight into muscle activity in a non-invasive and clinically meaningful manner is the use of surface electromyography (sEMG). Surface EMG is typically a routine part of clinical assessment and the evaluation of motor impairment in CP. However, the analysis of the data has been limited in most cases to examination of signal amplitude or differences in muscle onset and offset timing.

The long-term goal of this research is to develop an analysis method for sEMG that can be used during functional tasks for treatment planning, diagnostic, assessment purposes in CP. This is to be accomplished through the use of the continuous wavelet transform (CWT). By developing an assessment method based on muscle activity, it is believed that a clinically viable measurement tool can be devised that will provide a level of insight into the effects of an intervention on muscle pathophysiology that is not currently available. The first step in progressing towards this long-term goal is to determine the variability and range of expected time-frequency patterns that can be expressed in a given population (i.e., cerebral palsy) during the execution of a meaningful task (gait), and relate the time-frequency information back to more standard assessments