Abstract
Discrimination thresholds for radial deformations of closed circular contours, which may outline simple shapes such as stars, squares and ovals, fall within the hyper-acuity range. This extreme performance of the visual system may be accounted for by a global pooling processing after the local features are extracted in the earlier stages. We employed the pupillary response as a measure of the processing load in a shape discrimination task, where subjects had to detect deviations from circularity on different shapes (RF patterns). Thus we addressed the question of whether different RF patterns or their parts are processed differentially by the visual system. Pupil dilation was sampled at 1000 Hz while 4 subjects discriminated between modulated and non-modulated versions of RF patterns. Subjects sat in a dim room and their chin and forehead were stabilized. Stimuli with radial frequency of 2 (ovals, RF2), 5 (stars, RF5) and one cycle from RF5 (RF1) were presented at 100% contrast in a temporal 2AFC procedure and subjects had to respond in which of the two intervals the modulated stimulus was presented by a key press. Task-evoked pupillary responses (TERP) for high (100% detected stimuli) and low (chance level detection) amplitude modulations were analyzed. Preliminary results show that TERPs were higher (increased pupil dilation) for RF2 than for RF5 at both high and low modulation amplitudes. Further, the TERP was lower (decreased dilation) at low amplitude modulations. TERP functions for RF5 and RF1 exhibited distinct shapes. Pupil dilation decrease at low amplitudes indicates that processing load in the visual system is low when shape differences cannot be detected. Pupil dilation decrease with radial modulation increase and distinct TERP function shapes suggest that the stimuli investigated here are processed differentially by the visual system, possibly by a mechanism that pools global information at higher cortical levels.
Meeting abstract presented at VSS 2017