Abstract
Motion standstill is a phenomenon in which the pattern, color, texture, and depth of a rapidly moving stimulus can be perceived, but the stimulus appears to be standing still—the motion system fails while the other systems are still functioning. Motion standstill previously has been observed in various types of third-order motion displays, such as color motion (Cavanagh, Tyler, & Favreau, 1984; Lu, Lesmes, & Sperling, 1999) and stereo motion (Julesz & Payne, 1968; Tseng, Kim, Gobell, Lu, & Sperling, 2001). To discover motion standstill of first- and second-order motion, we used moving luminance and moving contrast-modulated-texture sine gratings. They were presented either centrally or peripherally, with various spatial and temporal frequencies (fx, ft). Gratings were slanted either +45 or −45 deg, moving either up or down (90 deg/frame). Centrally viewed gratings were circular discs whose edges were smoothly windowed to remove terminators' effect on motion judgments. Peripheral gratings were circular annuli with similarly windowed edges. In the second-order (contrast-modulated) displays, first-order (luminance) motion components were removed by individual calibration for each observer. Observers made both slant-discrimination and motion-direction judgments on each trial.
For each observer, with both first- and second-order stimuli, we found combinations of fx and ft that lead to phenomenal reports of motion-standstill. Forced-choice slant judgments were close to 100%-correct, whereas motion-direction judgments were at chance. Incidentally, we also found some spatiotemporal frequencies in which slant judgments were nearly perfect whereas motion direction judgments were significantly below chance-level, indicating motion-aliasing. The main standstill results further confirm Lu et al (1999) that pattern systems extract a single, stable representative view from a moving stimulus, a process that probably begins in the complex cells of visual cortex V1.
Supported by Air Force Office of Scientific Research, Human Information Processing Program.