Angeline Yang¹ (), Dongcheng He², Daniel R. Coates³, Haluk Ogmen⁴, Susana T.L. Chung¹; ¹University of California, Berkeley, ²Hefei Comprehensive National Science Center, ³University of Houston, ⁴University of Denver

Contour integration is a process whereby the visual system integrates local contour elements into a percept of a global shape. Previous electroencephalography (EEG) research has implicated the middle temporal N300 event-related potential (ERP) as a neural correlate of the contour integration process, along with the frontal P200, parietal contralateral and occipital N200, and parietal and occipital P400, which reflect attentional awareness, stimulus visibility, and task-related efforts, respectively. We aimed to investigate how these ERP signals associated with contour integration might be affected under binocular vs. monocular viewing, and in the presence of stimulus blur. Nine participants (6 females, 20.11±1.27 years) completed a contour integration task binocularly (no blur) and monocularly (no blur, 1, and 2 diopter of optical blur). The stimulus was a 27x15 array of 2x0.5˚ line segments that varied in orientation independently and randomly at 15 Hz, presented over 1,670 ms. At 735 ms, 12 line segments would form a 3x3 perfectly- or partially-aligned (by introducing orientation-jitter noise to the 12 segments) contour of a square for 152 ms at 6˚ to the right or left of a central fixation target. Participants indicated the contour location (right or left), under eye-position monitoring, while EEG activity was recorded. The amount of orientation-jitter noise was determined psychometrically for each viewing x blur condition before EEG was performed. Across viewing conditions, participants’ behavioral performance averaged 99.12% and 77.87% for the perfectly-aligned and partially-aligned conditions, respectively. There were no significant differences in the effects of viewing condition on behavioral performance or the signal strengths of the six associated ERPs. Additional Bayesian analysis favored the null effect over the effect of viewing condition on the ERP magnitudes. These results suggest that the neural correlates of contour integration are largely independent of the source (binocular vs monocular) and spatial-frequency content of local contour signals.

Acknowledgements: NIH Grant T32EY007043, UC Berkeley Undergraduate Research Apprenticeship Program