The electromotor neural system in weakly electric fish is a network responsible for active electroreception and electrolocation. This system controls the timing of pulse generation in the electrical signals used by these fish for extracting information from the environment and communicating with other specimens.
Ethological studies related to fish mating, exploratory, submissive or aggressive behaviors have described distinct sequences of pulse intervals (SPIs). Accelerations, scallops, rasps, and cessations are four patterns of SPIs reported in pulse mormyrids, each showing characteristic temporal structures and large variability both in timing and duration.
This paper presents a biologically plausible computational model of the electromotor command circuit that reproduces these four SPI patterns as a function of the input to the model while keeping the same internal parameter configuration. The topology of the model is based on a simplified representation of the network as described by morphological and electrophysiological studies. An initial ad hoc tuned configuration (S-T) was build to reproduce all four SPI patterns. Then, starting from S-T, a genetic algorithm (GA) was developed to automatically find the parameters of the model connectivity. Two different configurations obtained from the GA are presented here: one optimized to a set of synthetic examples of SPI patterns based on experimental observations in mormyrids (S-GA), and another configuration adjusted to patterns recorded from freely-behaving Gnathonemus Petersii specimens (R-GA).
A robustness analysis to input variability of these model configurations was performed to discard overfitting and assess validity. Results showed that the four SPI patterns are consistently reproduced, both with synthetic (S-GA) data and with signals recorded from behaving animals (R-GA). This new model can be used as a tool to analyze the electromotor command chain during electrogeneration and assess the role of temporal structure in electroreception.
bioRxiv Subject Collection: Neuroscience