Abstract
Purpose: Humans can perceive illusory contours in the absence of any real physical boundaries (e.g., between abutting gratings). Single-unit recordings in cats and monkeys have found neurons selective for illusory contour orientation. We used an adaptation protocol to identify orientation-selective neural responses to illusory contours in human visual cortex.
Methods: Event-related fMRI was used to measure post-adaptation responses to illusory grating stimuli. Each trial consisted of a 4s adapter (vertical or horizontal illusory contours defined by abutting line gratings), a 1s ISI, a 1s probe (illusory contours parallel or orthogonal to the adapter, or a blank) and 1.2s behavioral response period. Every 500 ms, inducer line orientation (±45°), inducer phase, and illusory contour phase were varied to minimize adaptation to the inducers. Stimuli were presented in an annulus (1.5–315 deg) around the fixation point. A highly demanding task at fixation was used to equate attentional load across stimulus conditions. Data were analyzed independently in nine visual areas (V1, V2, V3, V3A/B, V4, V7, VO1, LO1 and LO2), each defined by standard retinotopic mapping procedures.
Results: We found orientation-selective adaptation to illusory contours in most visual areas, evidenced by reduced fMRI responses for probe stimuli parallel to the adapter compared to those orthogonal to the adapter. This difference disappeared if inducer line positions were jittered so as to eliminate the illusory contour percept. This suggests that both early and higher visual areas are involved in the processing of illusory contours in humans.
Support: NEI (R01-EY11794, R01-EY16165) and the Seaver Foundation