Modeling the weber fraction of vibrotactile amplitudes using gain control through global feedforward inhibition

Abstract

Weber’s law describes the linear drop of discriminative performance with increased base intensity of a stimulus. So far, this phenomenon has been modeled using multistable attractor decision networks based on the principle of biased competition between two mutually inhibiting recurrent neural populations. Due to the sensitive balance in a multistable fluctuation-driven regime, these decision models can only account for Weber’s law in a narrow stimulus range. Psychophysical data shows though that the human exhibits this characteristic for a broad stimulus range. Recent neurophysiological evidence suggests that global feedforward inhibition expands the dynamic range of cortical neuron populations and acts as a gain control. In this paper, we introduce a computational model that exploits this type of inhibition and shows through a fit between simulation results and psychophysical data that it is a potential explanation for the principle mechanism behind Weber’s law.