Item

Abstract

This study introduces a novel neuromechanical model employing a detailed spiking neural network to explore the role of axial proprioceptive sensory feedback in salamander locomotion. Unlike previous studies that often oversimplified the dynamics of the locomotor networks, our model includes detailed simulations of the classes of neurons that are considered responsible for generating movement patterns. The locomotor circuits, modeled as a spiking neural network of adaptive leaky integrate-and-fire neurons, are coupled to a three-dimensional mechanical model of a salamander with realistic physical parameters and simulated muscles. In open-loop simulations (i.e., without sensory feedback) the model accurately replicates locomotor patterns observed in-vitro and in-vivo for swimming and trotting gaits. Additionally, a modular architecture of the descending reticulospinal (RS) drive to the central pattern generation (CPG) network, allows to accurately control the activation, frequency and phase relationship of the different sections of the limb and axial circuits. In closed-loop simulations (i.e. with the inclusion of axial proprioceptive sensory feedback), systematic evaluations reveal that intermediate values of feedback strength significantly enhance the locomotor efficiency and robustness to disturbances during swimming. Specifically, our results show that sensory feedback increases the tail beat frequency and reduces the intersegmental phase lag, contributing to more coordinated and faster movement patterns. Moreover, the presence of feedback expanded the stability region of the closed-loop swimming network, enhancing tolerance to a wider range of external stimulations, internal parameters’ modulation and noise levels. This study provides new insights into the complex interplay between central and peripheral pattern generation mechanisms, offering potential strategies for developing advanced biomimetic robots. Additionally, this study underscores the critical role of detailed, biologically-realistic neural networks to improve our understanding of vertebrate locomotion.