Fig. 5: Dynamic modelling of sensory input and motor output relationships.

This figure displays the mathematical functions derived from system identification techniques, quantifying the dynamic relationship between sensory inputs and motor outputs. Each curve represents a model fit, demonstrating how the system adapts to varying inputs and predicts motor responses, closely mimicking the adaptive responses observed in human neurophysiological processes. The transduction functions in this figure represent the dynamic relationship between afferent tactile signals and efferent motor responses, quantifying how sensory input (S) influences motor output over time and modelling the adaptive responses observed in human neurophysiological processes. The functions were derived using the ‘System Identification’ toolbox in MATLAB®, based on input-output data from our experiments. For active grasping, the function is \(\frac{{{{\rm{a}}}}}{{{{{\rm{S}}}}}^{2}+{{{\rm{bS}}}}+{{{\rm{c}}}}}\), and for reactive grasping, the function is \(\frac{{{{\rm{a}}}}}{{{{{{\rm{bS}}}}}^{3}+{{{\rm{cS}}}}}^{2}+{{{\rm{dS}}}}+{{{\rm{e}}}}}\), where a, b, c, d, and e are parameters optimized to fit the observed data. These functions capture the nonlinear and complex dynamics between sensory inputs and motor outputs. The detailed development process is presented in the ‘Methods’ section, and the MATLAB code used to derive these functions is provided as Data S2 in the supplementary material.