Slow Feature Analysis on Retinal Waves Leads to V1 Complex Cells

Abstract

The developing visual system of many mammalian species is partially structured and organized even before the onset of vision. Spontaneous neural activity, which spreads in waves across the retina, has been suggested to play a major role in these prenatal structuring processes. Recently, it has been shown that when employing an efficient coding strategy, such as sparse coding, these retinal activity patterns lead to basis functions that resemble optimal stimuli of simple cells in primary visual cortex (V1). Here we present the results of applying a coding strategy that optimizes for temporal slowness, namely Slow Feature Analysis (SFA), to a biologically plausible model of retinal waves. Previously, SFA has been successfully applied to model parts of the visual system, most notably in reproducing a rich set of complex-cell features by training SFA with quasi-natural image sequences. In the present work, we obtain SFA units that share a number of properties with cortical complex-cells by training on simulated retinal waves. The emergence of two distinct properties of the SFA units (phase invariance and orientation tuning) is thoroughly investigated via control experiments and mathematical analysis of the input-output functions found by SFA. The results support the idea that retinal waves share relevant temporal and spatial properties with natural visual input. Hence, retinal waves seem suitable training stimuli to learn invariances and thereby shape the developing early visual system such that it is best prepared for coding input from the natural world.

Author Summary

It is believed that our sensory systems are adapted to statistical properties of behaviorally relevant elements in our natural environments. In the case of vision, one adaptation principle that has been put forward is the so-called slowness principle. However, the visual system is partially structured even before eye opening, when no natural input is available yet. Thus, spontaneous neural activity in the developing visual system of mammals (so-called retinal waves) has been suggested to contribute to shaping connections in early visual areas before the onset of vision. Here we aim to bring these two ideas together. Specifically, we apply an algorithm that implements the slowness principle to simulated retinal waves. The algorithm is set to encode the retinal wave input and thus has to extract relevant features from that input. After encoding, we are able to investigate the emerged representation and we find that the extracted features bear strong similarity to features that are encoded by neurons in the early visual system. These features are the building blocks for an object representation that is independent of the object's position in the visual field.