Abstract
Crowding refers to the impairment of object identification in peripheral vision due to clutter. Theories on crowding have implicated the involvement of a range of cortical regions. While it is quite probable that crowding occurs at multiple stages of visual processing, identifying the earliest site of crowding can provide the strongest constraint on the basic mechanism of crowding. Previous attempts with fMRI were limited by the need to use indirect and less sensitive methods such as adaptation or functional connectivity to avoid confounds, and large peripheral targets to obtain good isolation of target-evoked activity, but at the cost of a weak crowding effect. In contrast, the present study measures the average signal evoked jointly by the target and flankers within a region of interest (ROI). Using stimuli that yield a robust crowding effect, we made simple and direct measurements of BOLD signal amplitude. We found that when a subject’s attention was directed away from the stimulus, adding a center letter between two flankers in a non-crowded configuration naturally led to a signal increase in the inclusive ROI. However, in a crowded configuration, adding a center letter either led to no increase in V1 for tangentially arranged flankers (a configuration that induces weaker crowding) or a net decrease for radially arranged flankers (stronger crowding). This net decrease found in the ROI that included the flanking and center letters rules out local response saturation as an explanation. Surround suppression also cannot be responsible since there was no significant difference in contrast threshold for detecting the center letter in tangential and radial flanker arrangements. These results are consistent with our earlier finding that crowding reduced BOLD amplitude in V1 while subjects identified the center letter. Crowding clearly modulates activity in V1, implying a basic mechanism that involves elementary features and low-level visual processing.
Meeting abstract presented at VSS 2012