Changing the Real Viewing Distance Reveals the Temporal Evolution of Size Constancy in Visual Cortex
Document Type
Article
Publication Date
7-8-2019
Journal
Current Biology
Volume
29
Issue
13
First Page
2237
Last Page
2243.e4
URL with Digital Object Identifier
10.1016/j.cub.2019.05.069
Abstract
© 2019 Elsevier Ltd Our visual system provides a distance-invariant percept of object size by integrating retinal image size with viewing distance (size constancy). Single-unit studies with animals have shown that some distance cues, especially oculomotor cues such as vergence and accommodation, can modulate the signals in the thalamus or V1 at the initial processing stage [1–7]. Accordingly, one might predict that size constancy emerges much earlier in time [8–10], even as visual signals are being processed in the thalamus. So far, the studies that have looked directly at size coding have either used fMRI (poor temporal resolution [11–13]) or relied on inadequate stimuli (pictorial illusions presented on a monitor at a fixed distance [11, 12, 14, 15]). Here, we physically moved the monitor to different distances, a more ecologically valid paradigm that emulates what happens in everyday life and is an example of the increasing trend of “bringing the real world into the lab.” Using this paradigm in combination with electroencephalography (EEG), we examined the computation of size constancy in real time with real-world viewing conditions. Our study provides strong evidence that, even though oculomotor distance cues have been shown to modulate the spiking rate of neurons in the thalamus and in V1, the integration of viewing distance cues and retinal image size takes at least 150 ms to unfold, which suggests that the size-constancy-related activation patterns in V1 reported in previous fMRI studies (e.g., [12, 13]) reflect the later processing within V1 and/or top-down input from other high-level visual areas. Chen et al. examine the computation of size constancy using EEG in a real-world setting in which the stimuli are viewed from different distances. They show that even when the distance cues are strong and congruent, none of them has any effect on the early event-related potential (ERP) components. Size constancy takes time to evolve even though that time is short (∼150 ms).
Notes
This article is freely available to read from the publisher