Abstract
In this study, we measured the processing time of luminance-, texture-, and motion-defined contours (LDC, TDC, MDC) with a SAT(Speed-Accuracy Trade-off) task (cf. Sutter & Graham, 1995). The stimulus consisted of 200 line segments(length: 27.4 min; LDC, TDC) or 300 dots (MDC) scattered randomly within a stimulus field (4.4 × 4.4 deg). The stimulus field was divided into two parts (horizontal/vertical) by modulating luminance, orientation, or motion direction for elements within a target area. A stimulus duration was fixed to 100 ms. After measuring the orientation discrimination threshold for LDC, TDC and MDC while varying S/N ratio for each subject, we measured the SAT latency for the three types of contours at equalized visibilities. Subjects were asked to discriminate contour orientation with a 2AFC method within 300 ms after the fixation point enlarged and signaled them to response. An interval between the stimulus offset and the response cue (cue lag) was varied in 10 steps from 50 to 1500 ms. As a result, the average correct rates for MDC and TDC increased more moderately, and exceeded 75 % at cue lags around 150 ms. These results indicate that a speed of contour processing varies among contours defined by different properties and suggest multiple contour mechanisms operate in our visual system.