Data-driven modeling of habituation with its frequency-dependent hallmark based on Fourier Neural Operator

Abstract

Habituation refers to the decrease in response to repetitive stimuli, a key process that helps organisms conserve cognitive and sensory resources by filtering out irrelevant stimuli. While mathematical models capture qualitative aspects of habituation, there is a lack of quantitative models applicable to experimental data. To address this, we propose a data-driven framework for modeling habituation using the Fourier Neural Operator (FNO). The FNO's discretization-invariant property allows it to replicate frequency-dependent behaviors of habituation. Numerical experiments show that the framework accurately predicts and replicates significant hallmarks of habituation, demonstrating its potential for this application.