The spread of presaccadic attention depends on the spatial configuration of the visual scene

Abstract

When preparing a saccade, attentional resources are focused at the saccade target and its immediate vicinity. Here we show that this does not hold true when saccades are prepared toward a recently extinguished target. We obtained detailed maps of orientation sensitivity when participants prepared a saccade toward a target that either remained on the screen or disappeared before the eyes moved. We found that attention was mainly focused on the immediate surround of the visible target and spread to more peripheral locations as a function of the distance from the cue and the delay between the target’s disappearance and the saccade. Interestingly, this spread was not accompanied with a spread of the saccade endpoint. These results suggest that presaccadic attention and saccade programming are two distinct processes that can be dissociated as a function of their interaction with the spatial configuration of the visual scene.

A dynamic 1/f noise protocol to assess visual attention without biasing perceptual processing

“…This demonstrates that visual structures pose the risk of compromising the attention measurement in a non-systematic way. In line with previous work showing the impact of placeholder objects on attentional modulations of visual perception (Taylor et al, 2015; Puntiroli et al, 2018; Szinte et al, 2019), these findings underline the necessity of an unbiased, item-free approach to map perceptual dynamics across the visual field in the absence of object-like structures.…”

“…Spatiotemporal maps of visual attention created with conventional test stimuli are affected by these perceptual inhomogeneities, which prevents a direct comparison of attentional effects on foveal and peripheral visual performance. To account for this confound, the discrimination signal strength needs to be adjusted separately for each tested retinal eccentricity (e.g., Koenig-Robert & VanRullen, 2011; Szinte et al, 2019), which is time consuming, comparably more error-prone, and less flexible. Since local 1/f orientation sensitivity is largely constant across visual eccentricities, the noise paradigm is resistant to visual sensitivity changes across space and thus offers the unique opportunity to directly map attentional modulations of visual perception continuously across space, without the need to increase discrimination signal strength with increasing test eccentricity.…”

“…However, these conventional approaches face a common shortcoming: They all rely on localized test stimuli to determine the deployment of visual attention. The potential problem with this approach is that these stimuli could bias visual perception by structuring the visual field (Taylor, Chan, Bennett, & Pratt, 2015; Puntiroli, Kerzel, & Born, 2018; Szinte, Puntiroli, & Deubel, 2019; Shurygina, Pooresmaeili, & Rolfs, 2021) and may thus affect what they are intended to measure – the spatial distribution of attention. In a paradigm using a typical stimulus configuration as displayed in Figure 1A , attention is likely biased towards the presented stimuli (as compared to locations in between or further in- or outside), since those are the only locations containing potentially task-relevant visual information ( Figure 1B ).…”

Where we perceive before we look: The distribution of presaccadic attention assessed with dynamic 1/f noise

“…However, in contrast to the spatially highly specific deployment of attention to a physical saccade target, the presaccadic attention shift is far less focused when no target is presented -and this is not explained by the relatively higher saccade landing variance. In line with previous research (9)(10)(11), this demonstrates that it is not the saccade endpoint that determines the spread of attention, but the presence (or absence) of scenestructuring objects.…”

“…Our present and former evidence that objects mold the distribution of visual attention (8)(9)(10)(11) is in line with an fMRI study investigating how attention modulates perception depending on the size of the 'attention field' (i.e., the attended area), which was manipulated using placeholder objects (12). Similar to our study, attention field size, approximated via the spread of cortical activity measured by fMRI, narrowed when 3 placeholders marked the test location.…”