Perceptual Color Space Representations in the Oculomotor System Are Modulated by Surround Suppression and Biased Selection

Abstract

The oculomotor system utilizes color extensively for planning saccades. Therefore, we examined how the oculomotor system actually encodes color and several factors that modulate these representations: attention-based surround suppression and inherent biases in selecting and encoding color categories. We measured saccade trajectories while human participants performed a memory-guided saccade task with color targets and distractors and examined whether oculomotor target selection processing was functionally related to the CIE (x,y) color space distances between color stimuli and whether there were hierarchical differences between color categories in the strength and speed of encoding potential saccade goals. We observed that saccade planning was modulated by the CIE (x,y) distances between stimuli thus demonstrating that color is encoded in perceptual color space by the oculomotor system. Furthermore, these representations were modulated by (1) cueing attention to a particular color thereby eliciting surround suppression in oculomotor color space and (2) inherent selection and encoding biases based on color category independent of cueing and perceptual discriminability. Since surround suppression emerges from recurrent feedback attenuation of sensory projections, observing oculomotor surround suppression suggested that oculomotor encoding of behavioral relevance results from integrating sensory and cognitive signals that are pre-attenuated based on task demands and that the oculomotor system therefore does not functionally contribute to this process. Second, although perceptual discriminability did partially account for oculomotor processing differences between color categories, we also observed preferential processing of the red color category across various behavioral metrics. This is consistent with numerous previous studies and could not be simply explained by perceptual discriminability. Since we utilized a memory-guided saccade task, this indicates that the biased processing of the red color category does not rely on sustained sensory input and must therefore involve cortical areas associated with the highest levels of visual processing involved in visual working memory.

Critical Dynamics Mediate Learning of New Distributed Memory Representations in Neuronal Networks

“…Our results indicate that the brain operates near criticality, possibly slightly sub-critical, and that plasticity plays an active role in reducing the dynamical state away from criticality during learning and consolidation. This is an agreement with previous work that suggests slightly sub-critical dynamics still impart increased tunability, response to external input, and long-range spatial and temporal correlations [ 46 ]. The extension of the Hopfield model we present here suggest that the critical point is indeed shifted ( Figure 3 B), rather than the phase transition region is widened, what would be indicative of emergence of Griffith phase [ 47 ].…”

Higher order, multifeatural object encoding by the oculomotor system

“…Therefore, it is difficult to determine conclusively from this analysis whether saccade curvature observed on trials when the distractors were equally similar to the target returned to baseline levels relative to trials with distractors that were unequally similar to the target. We have observed this type of saccade deviation bias in a similar context where saccades were also executed to stimuli embedded in circular displays and appeared at randomized locations trial to trial (Kehoe et al 2018). The strong bias in baseline curvatures may therefore be related to unresolved distractor-related activity from previous trials, because baseline trials were randomly interleaved into blocks and saccades systematically curve away from previous distractor locations (Belopolsky and Theeuwes 2011;Belopolsky and Van der Stigchel 2013).…”

“…For each corresponding value, the absolute subject means were averaged together. This technique is advantageous because 1) we were interested in examining whether the magnitude of saccade deviation covaries with the OS models and were not concerned with the saccade deviation direction, and 2) averaging across the CCW and CW conditions reduces any potential directional bias (see Kehoe et al 2018). Third, the data were standardized at each unique level of the OS model (x i ) and were trimmed if the absolute standardized score was Ͼ2.58 from the mean (|z| Ͼ 2.58), because this indicates less than a two-tailed, 1% cumulative probability of the observation.…”

Spatial clustering of orientation preference in primary visual cortex of the large rodent agouti

“…Some examples of response profiles of single units to several orientations and directions are depicted in Figures 2 A and 2B (for peri-stimulus time histograms (PSTHs) see Figure S3 ). We computed, based on each profile, an orientation selectivity index and a direction selectivity index (OSI and DSI, respectively,( Wunderle et al., 2013 ) ( Conde-Ocazionez et al., 2018 ), and see Methods ), where an index of 1 indicates a neuron that responds to only one stimulus orientation or direction, and an index of 0 indicates equal responsiveness across all orientations or directions.

Figure 2 Orientation selectivity in primary visual cortex of agouti and cat (A and B) Examples of isolated units for agouti (A) and for cat area 18 (B) stimulated at 0.08 cpd.

…”