Two Potential Mechanisms of Spatial Attention in Early Visual Areas

TL;DR

This paper presents a theoretical investigation of how spatial attention influences contrast-response and orientation-tuning in early visual areas, proposing two mechanisms—feedback modulation and neuron input-output changes—that can produce observed attentional effects.

Contribution

It introduces a model extending a hypercolumn framework to two visual areas, exploring two distinct mechanisms of attentional modulation and their effects on visual response properties.

Findings

Both mechanisms can produce contrast-gain and response-gain effects.

The mechanisms preserve orientation-tuning width and invariance.

A simple test is proposed to distinguish between the mechanisms.

Abstract

We investigate theoretically the effect of spatial attention on the contrast-response function (CRF) and orientation-tuning curves in early visual areas.We look at a model of a hypercolumn developed recently (Persi et al., 2008), that accounts for both the contrast response and tuning properties in the primary visual cortex, and extend it to two visual areas. The effect of spatial attention is studied in a model of two inter-connected visual areas, under two hypothesis that do not necessarily contradict. The first hypothesis is that attention alters inter-areal feedback synaptic strength, as has been proposed by many previous studies. A second new hypothesis is that attention effectively alters single neuron input-output properties. We show that with both mechanisms it is possible to achieve attentional effects similarly to those observed in experiments, namely contrast-gain and response-gain effects, while keeping the orientation-tuning curves width approximately contrast-invariant and attention-invariant. Nevertheless, some differences occur and are discussed. We propose a simple test on existing data based on the second hypothesis.