Presaccadic attention enhances and reshapes the Contrast Sensitivity Function differentially around the visual field

Abstract

The visual world is processed with a great degree of non-uniformity across eccentricity and around polar angle locations. Contrast sensitivity, a fundamental visual dimension constraining our vision, drops drastically from foveal to peripheral locations. It also decreases from the horizontal to the vertical meridian, and from the lower vertical to the upper vertical meridian (i.e., polar angle asymmetries). Moreover, contrast sensitivity is a function of spatial frequency, and the general relation between the two dimensions is referred to as the Contrast Sensitivity Function (CSF). To overcome constraints of the visual system throughout the visual field, we constantly make rapid saccadic eye movements to foveate on relevant objects in the visual scene. Already before saccade onset, attention shifts to the saccade target location (upcoming fovea) and enhances visual processing at that location. Thispresaccadic shift of attentiontypically benefits contrast sensitivity, but it is unknown whether and how it modulates the CSF and if this effect varies around polar angle locations. Contrast sensitivity enhancement may result from a horizontal or vertical shift of the CSF, increase in bandwidth, or any of their combinations, and the magnitude may differ around the visual field. In the current study, we investigated these possibilities by extracting key attributes of the CSF using Hierarchical Bayesian Modeling. First, we characterized how the CSF varies around polar angle. Then we compared the CSF attributes during fixation and saccade preparation (presaccadic attention). The results reveal that presaccadic attention (1) enhances contrast sensitivity across a wide range of spatial frequencies, (2) increases both the most preferred and the highest discernable spatial frequency, and (3) narrows the bandwidth. Importantly, the presaccadic enhancement in contrast sensitivity was more pronounced along the horizontal than the vertical meridian, exacerbating the polar angle asymmetries present during fixation. In conclusion, presaccadic attention bridges the gap between presaccadic peripheral and post-saccadic foveal input by rendering more of the world visible at the saccade target, and more so at the horizontal than the vertical meridian.

How the window of visibility varies around polar angle

“…The group level sensitivity indicated by the key attributes was best for the horizontal meridian, followed by the lower vertical, and then the upper vertical meridian, consistent with previous studies (Figure 2; Jigo et al, 2023;Kwak et al, 2024). Therefore, although the shape of the CSF is preserved around polar angle, CSFs need to be shifted and scaled to account for these differences.…”

“…Given the prominence of the CSF as a building block of spatial vision, obtaining accurate CSF estimates is essential (Glassman et al, 2024;Hou et al, 2010;Lesmes, Lu, Baek, & Albright, 2010). We extracted key CSF attributes at the upper vertical, lower vertical, and horizontal meridian (Figure 1a,c), using Hierarchical Bayesian Modeling (HBM; Methods and Figure S1), which enables accurate and precise CSF estimates with relatively few observations (Zhao, Lesmes, Hou, & Lu, 2021): Peak-CS is the maximum contrast sensitivity, peak-SF is the SF corresponding to the peak-CS and thus the most preferred SF, and bandwidth is the width of the CSF (cutoff-SF -the highest perceivable SF-and AULCSF -area under the log CSF, the total sensitivityare in Supporting Information) We found that across polar angle locations, the shape of the CSF is constant, but that sensitivity (e.g., peak-CS and peak-SF) is higher at the horizontal than the vertical (HVA), and at the lower than the upper vertical meridian (VMA), consistent with the two studies from our group that have measured CSF attributes around polar angle (Jigo et al, 2023;Kwak et al, 2024).…”

“…However, given the implicit or explicit assumption that CSF attributes generalize across iso-eccentric locations, scarce attention has been paid to variations of the CSF around the polar angle (Jigo et al, 2023;Kwak et al, 2024). Here, we conducted a detailed investigation of the key CSF attributes not only at a group level, line in our previous studies (Jigo et al, 2023;Kwak et al, 2024), but also at an individual level using covariance analysis. This analysis revealed critical differences within and across CSF attributes around the polar angle.…”

“…Thus, its variation with stimulus parameters has been widely assessed (DeValois & DeValois, 1990;Kelly, 1977;Graham, 1989;Watson, 2018), for example across eccentricity (Hilz & Cavonius, 1974;Jigo & Carrasco, 2020;Rovamo et al, 1978;. However, given the implicit or explicit assumption that CSF attributes generalize across iso-eccentric locations, scarce attention has been paid to variations of the CSF around the polar angle (Jigo et al, 2023;Kwak et al, 2024). Here, we conducted a detailed investigation of the key CSF attributes not only at a group level, line in our previous studies (Jigo et al, 2023;Kwak et al, 2024), but also at an individual level using covariance analysis.…”

“…However, scarce attention has been paid to characterizing the CSF at iso-eccentric locations around the visual field. There are only two studies that systematically assessed the entire CSF at polar angle meridians and showed that the these asymmetries are present for contrast sensitivity across a wide range of SFs (Jigo, Tavdy, Himmelberg, & Carrasco, 2023;Kwak, Zhao, Lu, Hanning, & Carrasco, 2024). Interestingly, M scaling -scaling the stimulus to equate its V1 cortical representation at different locations ) -eliminates eccentricity, but not polar angle differences in CSF (Jigo et al, 2023).…”