Interocular Grouping in Perceptual Rivalry Localized with fMRI

Abstract

Bistable perception refers to a broad class of dynamically alternating visual illusions that result from ambiguous images. These illusions provide a powerful method to study the mechanisms that determine how visual input is integrated over space and time. Binocular rivalry occurs when subjects view different images in each eye, and a similar experience called stimulus rivalry occurs even when the left and right images are exchanged at a fast rate. Many previous studies have identified with fMRI a network of cortical regions that are recruited during binocular rivalry, relative to non-rivalrous control conditions (termed replay) that use physically changing stimuli to mimic rivalry. However, we show here for the first time that additional cortical areas are activated when subjects experience rivalry with interocular grouping. When interocular grouping occurs, activation levels broadly increase, with a slight shift towards right hemisphere lateralization. Moreover, direct comparison of binocular rivalry with and without grouping highlights strong focused activity in the intraparietal sulcus and lateral occipital areas, such as right-sided retinotopic visual areas LO1 and IP2, as well as activity in left-sided visual areas LO1, and IP0-IP2. The equivalent analyses for comparable stimulus (eye-swap) rivalry showed very similar results; the main difference is greater recruitment of the right superior parietal cortex for binocular rivalry, as previously reported. Thus, we found minimal interaction between the novel networks isolated here for interocular grouping, and those previously attributed to stimulus and binocular rivalry. We conclude that spatial integration (i.e,. image grouping/segmentation) is a key function of lateral occipital/intraparietal cortex that acts similarly on competing binocular stimulus representations, regardless of fast monocular changes.

Keywords: Binocular; Flicker-and-swap; Gestalt; Illusory contours; Lateral occipital; Monocular.

Fig. 1

Stimuli and behavioral results. a Binocular rivalry stimuli were viewed dichoptically and, in the case of interocular grouping, were comprised of two parts in each eye that combined perceptually. b and c Stimulus rivalry conditions were identical to binocular rivalry except for stimulus exchange between eyes at a rate of 6.7 Hz (grouped conditions shown). d and e All four stimulus conditions produced rivalry with comparable alternations rates in pre-fMRI testing (d) and during fMRI scanning (e). The differences between conditions were not significant

Fig. 2

Comparison of binocular rivalry with and without interocular grouping. a Average brain maps for the binocular rivalry task compared to the replay task (N = 6). Bar graphs below the maps indicate the regions of interest (ROI) where the differences between tasks were significant. b Average brain maps for binocular rivalry with grouping task compared to replay task. With false-discovery-rate (FDR) correction for multiple comparisons in Brain Voyager, the exact p-values vary slightly for each hemisphere in each subtraction. The values ranged from p < 0.014 to 0.028. Bar graphs below the maps show the ROIs where the % BOLD differences between tasks were significant with bright color; non-significant regions shown with pale color. Both bar graphs show results for right and left hemisphere separately; error bars plot SED

Fig. 3

Comparison of binocular rivalry with grouping to stimulus rivalry with grouping. a Average brain maps for the binocular rivalry with grouping task compared to the binocular rivalry task (posterior brain view) (N = 6). b Average brain maps for the stimulus rivalry with grouping task compared to the stimulus rivalry task. With false-discovery-rate (FDR) correction for multiple comparisons in Brain Voyager, the exact p-values vary slightly for each hemisphere in each subtraction. The values ranged from p < 0.002–0.006. In both a and b blue outlines indicate the probabilistic location of cortical areas from the Wang atlas, along with area names in white. Bar graphs below the maps show the ROIs where the % BOLD differences between tasks were significant with bright color; non-significant regions shown with pale color. Both bar graphs show results for right and left hemisphere separately; error bars plot SED

Fig. 4

Direct comparison of grouped binocular and stimulus rivalry. Average brain maps for the binocular rivalry with grouping task compared to stimulus rivalry with grouping (N = 6). With false-discovery-rate (FDR) correction for multiple comparisons in Brain Voyager, the exact p-values vary slightly for each hemisphere. The values ranged from p < 0.001–0.005. Bar graphs below the maps show significant ROIs in bright colors, with highest signal change in region SP; error bars plot SED. A very strong right hemisphere lateralization is visible for both the maps and bars

Fig. 5

Topography of functional ROIs and probabilistic areas of Wang et al. (2015). Left panel shows inflated right hemisphere of one individual subject with selected areas from the Wang atlas, all shown filled in with warm colors. Center panel shows the same hemisphere with all of our functionally defined regions of interest outlined in cool colors. Right panel shows the registered overlay of both atlas areas and functionally defined regions. See adjacent text for description