Abstract
Recent studies have shown that an edge defined by motion accretion (the origin or trailing edge of a moving pattern) produces stronger fMRI activity than the same edge defined by motion deletion (the trailing edge or disappearance of the moving pattern; Koyama et al., 2003; Liu et al.,2004; Whitney, Goltz, et al., 2003). For example, a stimulus moving toward the fovea produces stronger activity in the periphery, whereas a pattern moving away from the fovea produces more activity near the fovea. Here we tested whether this difference in activity is due to a mechanism that selectively processes first-order (luminance-defined) motion. Using a 4T fMRI scanner, we measured visual cortex activity in seven subjects while they viewed randomly interleaved first-order and second-order (texture-based) moving patterns that moved toward or away from fovea. Subjects performed a letter discrimination task at the fixation point to control for attention.
GLM contrasts of first-order motion toward versus away from the fovea revealed a significant increase in activation at the origin of motion in all subjects, consistent with previous reports. The same GLM contrast for second order stimuli, however, produced significantly less activity at the origin of motion (the edge defined by motion accretion).
These results demonstrate a difference in processing of first and second order motion in primary visual cortex. While we found a significant overall response to second-order motion in early visual areas (consistent with previous studies; Seiffert et al., 2002; Smith et al., 1998), only first-order motion produced a selective response at the origin of motion. The results suggest that the mechanism that operates on the trailing edge or origin of a moving pattern is due to a luminance-based motion detection process, and not to a generalized alerting or salience-based mechanism.