Abstract
There is still no firm conclusion about whether the brain mechanism for second-order motion is separate from that for first-order motion. While initial fMRI work suggested that second-order motion is coded later in the visual pathway (Smith et al., J Neurosci 1998), more recent studies reported that both types of motion can be represented as early as V1 (Nishida et al., J Neurophysiol 2003; Seiffert et al., Cereb Cortex 2003). To overcome the relatively coarse spatial resolution of fMRI, we used a fast event-related fMRI adaptation paradigm that allows dissociation of different neural populations within a particular area. We examined the time course of the BOLD signal to a stimulus sequence (S1 of 2s, blank of 2s, and S2 of 1s) obtained using a 3T MR scanner (Siemens Trio). S1 and S2 consisted of a radial luminance-modulated (first-order) or contrast-modulated (second-order) grating that rotated clockwise or counterclockwise. In the MT/V5 complex, direction-selective adaptation was found when the stimulus order was the same for S1 and S2, that is, signals were smaller when the motion directions of S1 and S2 were the same than when they were opposite. No such difference was found when the stimulus order changed between S1 and S2. In other words, cross-adaptation between the two types of motion did not occur while order-specific adaptation did. Our results indicate that separate neural populations are responsible for first-order and second-order motions, even if these populations are located in the same visual areas.
Supported by JSPS grant-in-aid for scientific research for H.A.