Analyse der Veränderungen von Wavelet-transformierten elektromyographischen Signalen, wie sie beim Tragen einer Kniebandage entstehen / Analysis of Wavelet Transformed Electromyographic Signals that were Altered by Wearing a Knee Brace

Abstract

The comparison of electromyograms represents a challenge for data analysis. The aim of the project was to present a method that uses a minimal computational effort to resolve small but significant changes in the muscular activity that occur while walking with and without a knee brace. The wavelet transformed electromyograms were represented as intensity patterns that resolve the power of the signal in time and frequency. The intensity pattern of each electromyogram defines single points in a pattern space. The distance between these points in pattern space were used to detect and show the separation between the groups of electromyograms that were recorded while walking with and without a knee brace. The method proposes a distance versus angle representation to visually discriminate the intensity patterns. Once it has been shown that the differences are statistically significant, one can visualize the result in a difference intensity pattern that indicates at what time and at what frequency the electromyograms vary between the two conditions tested. It is to be expected that interventions that are more intrusive than a knee brace will reveal even more distinct differences.

Wavelet analyses of electromyographic signals derived from lower extremity muscles while walking or running: A systematic review

“…The application of wavelet analysis together with SVMs was used to find the boundaries to discriminate patients with total knee arthroplasty from healthy people while walking [ 15 ] or to correctly assign the EMG patterns of different running tasks to each subject [ 17 , 18 ]. Eight studies computed the recognition rate [ 9 , 14 – 19 ].…”

“…The recognition rate discriminated between different running, walking or footwear conditions measured with different muscles of the leg [ 9 , 15 , 17 – 19 ]. The maximum recognition rate was 78 to 94.4% and varied between different muscles and methods.…”

“…Non-linearly scaled wavelets, as originally described by von Tscharner (2000) [ 1 ] have been shown to discriminate not only between walking and running, but can also detect difference in muscle activity during ambulation in patients with different orthopedic conditions [ 9 , 14 – 19 ]. Therefore, these techniques might be clinically useful to detect locomotor dysfunction, as well as help design therapeutic interventions (such as orthotic devices, insoles, etc.)…”