, 2004), and in the thick/pale stripes of macaque V2 (Lu et al., 2010). The direction preference map in V4 showed some common properties as

those found in MT and V2, including similarity in domain size and the orthogonal relationship between preferred BMN 673 datasheet direction and orientation angles (Lu et al., 2010; Kaskan et al., 2010). However, the direction preference maps in V4 also showed features that are distinct from those found in all these other areas. First, V4 direction-preferring domains only appear in restricted regions, rather than in the entire V4 area. Second, many V4 direction-preferring domains appear to be isolated singulars, without any neighboring domains for other directions. Third, these domains overlap not only with orientation-preferring domains but also with color-preferring domains. It is understandable that a direction click here map in the ventral pathway may have a different clustering architecture than its counterparts in the dorsal pathway. A precedent for this principle was found, for example, in motion maps of V2 (Lu et al., 2010), where the direction maps differ in architecture from those found in MT (Malonek et al., 1994; Xu et al., 2004; Kaskan et al., 2010) or cat area 18 (Shmuel and Grinvald, 1996). Thus, this functional architecture may suggest a distinct functional computation in the visual system. Direction-preferring

domains found in previous studies either have been shown in an area not considered to signal color (e.g., MT; Malonek et al., 1994; Xu et al., 2004; Kaskan et al., 2010) or avoid color-preferring regions (e.g., in thick/pale stripes of V2; Lu et al., 2010). others Our data show that, in V4, about one fourth of direction-preferring pixels overlap with color-preferring pixels, suggests that these direction-selective neurons may be involved in detection of color motion. Another possibility is that motion cues in V4 are used for surface definition and thus are processed by surface-processing neurons, which were revealed by color versus luminance imaging (e.g., Figure 1D). These results,

however, differ from the findings in a recent fMRI-guided recording study in which color cells recorded from globs rarely showed direction selectivity (Conway et al., 2007). We noted that the V4 color glob neurons they recorded from were mostly from anterior wall of the lunate sulcus, while our imaging and recordings were all from dorsal part of the lunate gyrus. Therefore, it is possible that different parts of V4 may have different color-direction interactions. The separation of motion and color/form information in the primate visual system has been considered to be strong support for the concept of parallel processing of visual information. In particular, areas MT and V4 are often referred to as motion and color/form centers, respectively (Zeki et al., 1991).