Purpose: Wavelet transform is a time-frequency analysis method that quantifies temporal changes of the frequency content of nonstationary signals without losing resolution in time or frequency. It may be used to analyze surface EMG (sEMG) signals under dynamic conditions. However, this method is difficult to apply clinically because it generates large quantity of data in a very short time, and the change in muscle fiber length and diameter during dynamic contraction induces a large variability of the data at various wavelet domains and joint positions. This study aimed to determine a wavelet domain and a joint range that has the lowest variance and highest sensitivity to document the shift of the frequency intensity that relates to the decline of dynamic muscle power.
Methods: Eleven active young males were tested for maximal isokinetic knee extension and flexion exercise for 50 repetitions at 180°·s−1 and at a range of 100°-0°. sEMG measurement of vastus medialis (VM), vastus lateralis (VL), and rectus femoris (RF) were recorded during the exercise. Wavelet transform was used to filter the sEMG data into 11 wavelet domains. The movement was divided into five 20°-angle range groups.
Results: On the basis of the statistical analysis, the most significant and consistent trend of decrease in sEMG intensity power under maximal dynamic exercise for VM, VL, and RF was identified at the angle range of 40°-20° in wavelet domain 4.
Conclusion: We concluded that the shift in intensity within this angle range and wavelet domain might be used to document fatigue of the quadriceps during dynamic maximal knee extension exercise.
1Department of Rehabilitation Sciences, The Hong Kong Polytechnic University, Hung Hom, HONG KONG; 2Hong Kong Sports Institute, Sha Tin, HONG KONG; and 3Department of Electronic and Information Engineering, The Hong Kong Polytechnic University, Hung Hom, HONG KONG
Address for correspondence: Joseph K.-F. Ng, Ph.D., Department of Rehabilitation Sciences, The Hong Kong Polytechnic University, Hung Hom, Hong Kong; E-mail: Joseph.Ng@polyu.edu.hk.
Submitted for publication February 2008.
Accepted for publication September 2008.