Touch mechanoreceptors: Modeling and simulating the skin and receptors to predict the timing of action potentials

Abstract

The sense of touch, and the neural code that underlies it, makes it possible to execute everyday tasks such as picking up a glass or buttoning a shirt. Recreating touch in upper limb prostheses requires an in-depth understanding of how indentation of the skin is transformed to the time of occurrence of neural action potentials by tactile mechanoreceptors, and this understanding may come from modeling and simulation. Typical models transform either sustained indentation to firing rates, or vibration to the timing of action potentials. In contrast to those approaches, the skin-receptor model over-viewed here simulates the transfer function of the slowly adapting type I (SA-I) receptor with input of indentation and output of the timing of action potentials. To achieve this transformation, the model couples a finite element model of skin, a sigmoidal transduction function, and a leaky integrate-and-fire neuron model. Sub-units of the computational model are then enhanced and combined with a physical sensor to produce action potential discharge that is differentially responsive to the ramp and hold phases of indentation. These are compared to recordings from a mouse SA-I receptor. The overall approach may aid sensor and systems designers building touch prosthetics in an effort to restore perception at the peripheral afferent.