Frequency analysis of electromyogram of the triceps surae muscle taken while the stroke patients walk

Abstract

Background: Many groups have reported results of frequency analysis of electromyogram (EMG) while the muscle with spasticity contracts in stroke patients. How ever, very few reports on characteristics of the muscle with spasticity during exercise have been published, and the majority is static evaluation using Fast Fourier Transform (FFT). Purpose: We carried out and examined dynamic frequency characteristics of contraction characteristics of the triceps surae while the stroke patients walk using Wavelet Transform (WT) which analyzes temporal information and frequency information of conventional FFT at same time. Methods: Subject of the present study was nine healthy men (age 28.0±2.83), healthy group (HG), and nine stroke patients (age 70.17±9.38), stroke group (SG), who were in the recovery stage and the rehabilitation ward of our hospital. These patients have their hemiplegia after stroke, and can walk by themselves or while being supervised. We measured the dominant feet of the HG and the paralyzed legs of the SG. Surface EMG monitored muscle activation while the subjects freely walked 10m on a walkway at the optimum speed. Also, a foot switch sensor was attached on the plantar part to identify a single walking cycle. EMG Master (Ozawa Medical Instruments) was used for surface EMG to monitor medial and lateral heads of gastrocnemius muscle. The data after fourth steps from walking starts were analyzed. Integral EMG analysis and wavelet transform (12.5 Hz to 200 Hz) were carried out to any given data of three walking cycles. Calculated integrated values of both groups were normalized using the integrated values during a single walking cycle. Also, the frequency range was divided into two categories; a low-frequency band (LFB) of 45 Hz and lower and highfrequency band of 75 Hz and higher (HFB) to calculate the power density of HFB to LFB, H/L ratio, of each muscle. As a statistical analysis, theWilcoxon signed-rank test was used to compare the integrated values and H/L ratio between the HG and SG. Results: Each evaluation item of the HG and SG was compared. The HG has a significantly high integrated value (p < 0.05), and the SG has significantly high H/L ratio (p < 0.05). Conclusion(s): Although muscle activity characteristics of the triceps surae of the stroke patients is low as an integrated value, the ratio of HFB is high as dynamic frequency characteristics. This indicates strong activity of fast fibers. By taking the facts of low integrated value and atrophy of the presence of fast-twitch fibers into account, synchronization defect in the motor units possibly caused a frequency rise. Implications: These findings are expected to help treatment to improve quality of muscle activity in rehabilitation for stroke patients to walk.