Purpose: This study investigated the effects of different biomechanical constraints on the variability of muscle activation during cycling.
Methods: Fifteen male athletes cycled at a power of 150 and 300 W. Surface EMG was recorded from seven lower limb muscles. Wavelet transformed EMG signals of all muscles were subjected to a principal component analysis to study the variability of the EMG. The full vector space was reduced to the first principal components that explained 90% of the variance. The input data of each cycle revolution were projected onto these principal component vectors. Means and SD of the projections were calculated across all cycles and summed across all time points. The relative variability (RV) was expressed as the ratio between the SD and the mean of the summed projections. The principal angle was calculated between the principal components used for the 150-W condition and those used for the 300-W condition.
Results: The RV could be split into low- and high-variability components. The variability was smaller for the lower ordered eigenvectors compared with the higher ordered ones (P < 0.001) independent of the loading condition. Overall, the 300-W condition showed lower RV compared with the 150-W condition (P < 0.01). The average principal angle between the 150- and 300-W subspaces was 0.4, respectively.
Conclusions: Structured aspects of variability were found in the muscle activation of lower leg muscles during cycling. In the context of the minimum intervention principal, this might be interpreted as a transition into a regime that requires specific necessary muscles where the increased constraints of the task specify the muscle coordination pattern in a more precise way.