Modelling suppressed muscle activation by means of an exponential sigmoid function: validation and bounds

The aim of this study was to establish how well a three-parameter sigmoid exponential function, DIFACT, follows experimentally obtained voluntary neural activation-angular velocity profiles and how robust it is to perturbed levels of maximal activation. Six male volunteers (age 26.3±2.73 years) were tested before and after an 8-session, 3-week training protocol. Torque–angular velocity (T–ω) and experimental voluntary neural drive–angular velocity (%VA–ω) datasets, obtained via the interpolated twitch technique, were determined from pre- and post-training testing sessions. Non-linear regression fits of the product of DIFACT and a Hill type tetanic torque function and of the DIFACT function only were performed on the pre- and post-training T–ω and %VA–ω datasets for three different values of the DIFACT upper bound, αmax, 100%, 95% & 90%. The determination coefficients, R2, and the RMS of the fits were compared using a two way mixed ANOVA and results showed that there was no significant difference (p<0.05) due to changing αmax values indicating the DIFACT remains robust to changes in maximal activation. Mean R2 values of 0.95 and 0.96 for pre- and post-training sessions show that the maximal voluntary torque function successfully reproduces the T–ω raw dataset.