Evidence from human patient studies suggests that the functional differences of the dorsal and ventral pathways are better explained by vision-for-action and vision-for-perception, respectively (Goodale and Milner, 1992). In fact, V4 receives mixed magnocellular and parvocellular inputs originating from the lateral geniculate nucleus (Ferrera et al., 1994a), as well as input from MT (Maunsell and Van Essen, 1983; Ungerleider and Desimone, 1986). These connections make V4 an area that has rich access to motion information in the visual check details stimulus. Furthermore,
it has been shown that top-down signals to V4 also contain motion information (Ferrera et al., 1994b). As a result, V4 is well activated when monkeys are viewing moving stimuli (Tolias et al., 2001; Vanduffel et al., 2001). Our findings further suggest that the motion information is actively processed
in this area. Note that V4 is much larger than MT, so V4 may contain a comparable number of direction-selective neurons as area MT. This may raise the question, “Why would both dorsal and ventral pathways participate in motion processing?” A reasonable assumption is that the same motion information needs to be processed in different ways in order to serve different purposes. For example, motion perception requires integration of local motions, while distinguishing a moving object c-Met inhibitor from its background requires motion differentiation (Braddick 1993). We found that the motion-processing organization in V4 is different from that in MT. For example, many direction-preferring domains in V4 are scattered singulars, while direction preference maps in MT are more uniform (Malonek et al., 1994; Xu et al., 2004; Kaskan et al., 2010). The mean direction selectivity of neurons recorded in the V4 direction-preferring domains (mean DI = 0.63; this study) is lower than that found in area MT (mean DI = ∼1;
mafosfamide Albright et al., 1984). V4 neurons also tend to be more activated by moving lines than by moving random dots (Baker et al., 1981; Vanduffel et al., 2001). In addition, motion adaptation could induce direction selectivity of V4 neurons (Tolias et al., 2005). These data suggest that the direction-selective neurons in V4 have very different receptive field features than do MT neurons. These differences could give us hints on the functional roles of direction-selective neurons in the ventral pathway. It is also possible that perception of motion might be a distributed process that is not limited to the dorsal areas. This idea is supported by a recent finding that MT does not process global motion (Hedges et al., 2011). Motion information is useful for object identification. Relative motion between an object and its background is an important cue for figure-ground segregation, especially when other types of cues are weak or ambiguous (e.g., camouflaged insects).