Filtering spinal cord stimulation artifacts while preserving muscle activity during exoskeleton assisted walking after spinal cord injury

Abstract

Analyzing surface electromyography in the presence of transcutaneous spinal cord stimulation requires the filtering of electrical stimulation artifacts. Limitations of previous filtering approaches include the use of a uniform filter and excessive removal of muscle activity of interest. Here, we adapted a prior filtering strategy to create a stimulation artifact-specific filter and assessed the filter’s performance for removing stimulation artifact while preserving muscle activity. Six individuals with chronic spinal cord injury (SCI) stood and walked with exoskeleton assistance while surface electromyography signals were recorded from the muscles of each participant’s weaker leg. In the same session, muscle activity was recorded with and without transcutaneous spinal cord stimulation during standing and walking. In quiet standing with exoskeleton assistance, the differences in muscle activity between the stimulation conditions were calculated to create a stimulation artifact-specific filter that was customized and tuned for each participant and muscle. Our results show that this stimulation artifact-specific filter could remove stimulation artifacts while largely preserving even very low amplitude muscle activity in participants with SCI during exoskeleton-assisted standing and walking. By updating a previously developed stimulation filter, we created a stimulation artifact-specific filter that effectively removes stimulation artifacts from weak surface electromyography signals. Trial registration : ClinicalTrials.gov NCT03509558, registration date: 2018-04-26.