Visual crowding involves delayed frontoparietal response and enhanced top‐down modulation

Abstract

Crowding, the disrupted recognition of a peripheral target in the presence of nearby flankers, sets a fundamental limit on peripheral vision perception. Debates persist on whether the limit occurs at early visual cortices or is induced by top‐down modulation, leaving the neural mechanism for visual crowding largely unclear. To resolve the debate, it is crucial to extract the neural signals elicited by the target from that by the target‐flanker clutter, with high temporal resolution. To achieve this purpose, here we employed a temporal response function (TRF) approach to dissociate target‐specific response from the overall electroencephalograph (EEG) recordings when the target was presented with (crowded) or without flankers (uncrowded) while subjects were performing a discrimination task on the peripherally presented target. Our results demonstrated two components in the target‐specific contrast‐tracking TRF response—an early component (100–170 ms) in occipital channels and a late component (210–450 ms) in frontoparietal channels. The late frontoparietal component, which was delayed in time under the crowded condition, was correlated with target discrimination performance, suggesting its involvement in visual crowding. Granger causality analysis further revealed stronger top‐down modulation on the target stimulus under the crowded condition. Taken together, our findings support that crowding is associated with a top‐down process which modulates the low‐level sensory processing and delays the behavioral‐relevant response in the high‐level region.

Interference in visual crowding stimuli with small separations between target and flanking stimuli

“…The term "visual crowding" refers to phenomenon that adjacent flanking stimuli reduce the visibility of a target stimulus (Levi, 2008). Crowding has generally been interpreted in terms of mechanisms or elements in the visual system (Bacigalupo & Luck, 2019;Battaglini et al, 2019;Bi et al, 2009;Chaney et al, 2014;Clarke et al, 2014;Coates et al, 2018;Crutch & Warrington, 2007Doron et al, 2015;Fang & He, 2008;Freeman & Simoncelli, 2011;Han & Luo, 2019;He et al, 2019: Leaf et al, 1999Lev & Polat, 2015;Levi, 2008Levi, , 2011Liu et al, 2019;Manassi et al, 2012Manassi et al, , 2016Manassi & Whitney, 2018;Motter, 2018;Ogata et al, 2019;Parkes et al, 2001;Patro, & Huckauf, 2019;Pelli et al, 2004;Reddy & VanRullen, 2007;Sacchi et al, 2018;Siderov et al, 2020;Soo et al, 2018;Whitney & Levi, 2011;Yashar et al, 2019). However, analyses have demonstrated that effects similar to crowding may be generated by interference in the stimuli (Skottun, 2018a, b).…”

Different patterns of foreground and background processing contribute to texture segregation in humans: an electrophysiological study

“…EEG data were preprocessed using the EEGLAB toolbox ( Delorme & Makeig, 2004 ). EEG data were offline re-referenced to the averaged mastoids (TP9 and TP10) and were band-pass filtered between 1 and 40 Hz using a Butterworth IIR filter with the order of two ( Han & Luo, 2019 ). Independent components analysis was performed to identify and filter out eye-movements and artifact components.…”

“…The TRF method estimates the neural impulse response function evoked by certain stimuli to reflect the brain’s response to a unit change in the stimulus, such as a luminance change from ongoing visual stimuli ( Lalor et al, 2006 ; Jia, Fang & Luo, 2019 ; Jia et al, 2017 ). It has been used to dissociate the neural correlates of individual auditory or visual streams from the mixture of concurrent streams ( Crosse et al, 2016 ; Ding & Simon, 2012 ; Goncalves et al, 2014 ; Han & Luo, 2019 ; Jia, Fang & Luo, 2019 ; Jia et al, 2017 ; Liu et al, 2017 ).…”

Attentional rhythms are generated in binocular cells

“…The temporal response function (TRF) is an effective method used to measure neural activity underlying attentional rhythms (Jia, Fang, & Luo, 2019; Jia et al, 2017; Vanrullen & MacDonald, 2012). This method has been successfully used to study the temporal dynamics of working memory (Huang, Jia, Han, & Luo, 2018) and visual crowding (Han & Luo, 2019). This method assumes that the visual system is linear and that the neuron response at each moment is the convolution of the neural activity induced by all previous stimuli.…”