Skip to main content
. 2017 Sep 15;14:97. doi: 10.1186/s12984-017-0299-x

Table 2.

Overview of mean relative differences (mDIFF), mean standard error of the mean (mSEM), goodness of fit (GOF) and Akaike information criterium (AIC) for each combination of sensitivity functions and models fitted on simulated data

mDIFF (%) mSEM GOF (%) AIC
Basica ACTb Ka Basica ACTb Ka BS T MA
Fitted model without activation dynamics
BS 22.4 0.10 99.5 −1.34
T 22.2 0.11 99.4 −1.36
BS + T 22.2 0.08 99.5 99.4 −0.94
Fitted model with activation dynamics
BS 28.0 47.6 1.76 0.44 99.7 −1.43
T 27.6 47.9 1.66 0.42 99.5 −1.45
BS + T 23.3 59.4 0.46 0.26 99.7 99.5 −1.11
BS + MA 47.4 30.9 1.94e12 0.27 93.5 85.6 1.88
T + MA 47.5 30.9 2.19e12 0.28 90.9 85.6 1.88
BS + T + MA 48.2 31.0 2.01e12 0.22 96.2 95.2 81.6 2.02
Fitted model with activation dynamics and acceleration feedback
BS + MA 15.0 0.8 15.4 0.08 0.09 0.07 99.5 98.7 −0.41
T + MA 15.1 0.8 15.4 0.08 0.09 0.07 99.3 98.7 −0.43
BS + T + MA 13.1 0.7 13.9 0.07 0.08 0.07 99.6 99.5 98.5 −0.25

BS body sway, T ankle torque, MA muscle activation, ACT activation dynamics, K a acceleration feedback

aBasic parameters consist of intrinsic dynamics, reflexive dynamics, time delay and proprioceptive weight

bActivation dynamics parameters consist of the relative damping and eigenfrequency