Abstract
Watanabe et al (2001, VSS) showed that in MT+ and V3A of humans fMRI signals were higher when the subjects were viewing a global motion flow in a so-called Sekuler display (in which dots moved spatiotemporally locally within a certain range of directions, Williams & Sekuler, 1984) than random motion. In contrast, in V1 no significant difference was found between the two types of displays. The results suggested that that global flow motion might be generated in V3A and MT+ and local motion in V1. However, it was not clear whether the activation in MT+ and V3A was due to the global motion signal or due to a smaller amount of suppression between opponent motion directions in the Sekuler display than in the random motion display. To control the motion opponency (Heeger et al., 1999), here we used a transparent Sekuler display (TSD) in which half dots moved randomly within a 45 degree range and the other half within the opposite 45 degree range (e.g., 0 deg-45deg and 180deg-225deg), so that two global motion flows in opposite directions were clearly perceived to move transparently. TSD gives greater opponent-motion suppression than the random motion display. Siemens 3T scanner was used along with flattened occipital format. The subjects were instructed to maintain fixation at a central spot while the TSD and the random motion display were presented alternately every 16 seconds. Despite the greater opponent suppression, the same results as in the previous study were obtained. That is, TSD induced significantly higher activity in MT+ and V3A than the random motion display, whereas no significant difference was found in V1 activity between the two conditions. We conducted another experiment in which a speed increment detection task was performed during the exposure to the two displays to control attention (Huk et al., 2001). The same tendency was obtained. We conclude that global flow motion is indeed processed in V3A and MT+ and local motion in lower stages including V1.