Estimating smooth and sparse neural receptive fields with a flexible spline basis

TL;DR

This paper introduces a computationally efficient spline basis method for estimating smooth and sparse neural receptive fields, outperforming traditional approaches in high-dimensional sensory neuroscience data.

Contribution

It proposes using natural cubic splines as basis functions for receptive field estimation, offering a flexible and efficient alternative to existing empirical Bayes methods.

Findings

Spline-based methods outperform non-spline methods on simulated data.

Spline approach is computationally efficient in high-dimensional settings.

Method is applicable to both simulated and experimental neural data.

Abstract

Spatio-temporal receptive field (STRF) models are frequently used to approximate the computation implemented by a sensory neuron. Typically, such STRFs are assumed to be smooth and sparse. Current state-of-the-art approaches for estimating STRFs based on empirical Bayes are often not computationally efficient in high-dimensional settings, as encountered in sensory neuroscience. Here we pursued an alternative approach and encode prior knowledge for estimation of STRFs by choosing a set of basis functions with the desired properties: natural cubic splines. Our method is computationally efficient and can be easily applied to a wide range of existing models. We compared the performance of spline-based methods to non-spline ones on simulated and experimental data, showing that spline-based methods consistently outperform the non-spline versions.