Anatomical organization of ventral and dorsal processing streams

Much of our knowledge of the areas that comprise the ventral and dorsal streams comes from anatomical tract-tracing studies in nonhuman primates. These studies have demonstrated that regions that make up the ventral stream lie directly anterior to V1 in the occipital lobe and in progressively more anterior and ventral portions of the temporal lobe, whereas regions that make up the dorsal stream also include these early occipital-lobe areas but then occupy sites within the posterior STS and include progressively more anterior sites within this sulcus and more dorsal sites within the parietal lobe. Figure 34.1C illustrates the location of these areas within the cortex, and Figure 34.2 diagrams their anatomical connections (see Table 34.1 for a list of abbreviations).

The cortical analysis of objects begins in V1, where information about contour orientation, color composition, brightness, and direction of motion is represented in subsets of neurons responsible for each point in the visual field (Livingstone and Hubel, 1984; Tootell et al., 1988; Ts'o and Gilbert, 1988). Information from V1 is then sent forward to subdivisions or modules—interdigitating thin, thick, and interstripe regions—within V2 (Livingstone and Hubel, 1984, 1987). From the thin and interstripe regions in V2, representing mainly color and form information, respectively (De Yoe and Van Essen, 1985; Hubel and Livingstone, 1985, 1987; Livingstone and Hubel, 1984; Shipp and Zeki, 1985), neural signals proceed forward to area V4 on the lateral and ventromedial surfaces of the hemisphere and to a posterior inferior temporal area just in front of V4, area TEO (Gattass et al., 1997; Nakamura et al., 1993). From both V4 and TEO, signals related to object form, color, and texture proceed forward to area TE (Desimone et al., 1980; Distler et al., 1993), the last exclusively visual area within the ventral stream for object recognition (Desimone and Gross, 1979). Together, areas TEO and TE comprise inferior temporal (IT) cortex.

Progressing forward along the ventral stream, there is a gradual shift in the nature of connectivity. Patterns of cortical projections become successively less topographic or point-to-point. For example, while connections from V1 to V2 maintain topographic order, they are looser between V2 and V4, and inputs to area TE from V4 and TEO retain no obvious retinotopic organization (Desimone et al., 1980; Distler et al., 1993; Gattass et al., 1997; Livingstone and Hubel, 1983; Rockland and Pandya, 1979). This loss of retinotopy within area TE means that single neurons respond to objects anywhere in the visual field. Thus, explicit information about the spatial location of an object is not retained at the highest levels of the ventral stream.

The cortical analysis of spatial perception also begins in V1, deriving largely from the motion-sensitive cells in layer 4B (Maunsell and Van Essen, 1983a). In turn, cells in layer 4B project primarily to the thick stripes within V2 (Roe and Ts'o, 1995; Zeki and Shipp, 1987) and to V3, where direction of motion information is retained. Together, these three early visual areas provide the input to the middle temporal visual area, MT (Ungerleider and Desimone, 1986), also known as the motion-sensitive area of the STS. From MT, information is sent forward within the STS to several additional motion-sensitive areas, including MST (the medial superior temporal area) and FST (the fundus of the superior temporal area; Maunsell and Van Essen, 1983a; Ungerleider and Desimone, 1986), both of which project to area STP (the superior temporal polysensory area) located farther forward on the upper bank of the STS (Boussaoud et al., 1990). Inasmuch as portions of STP also receive ventral stream inputs, these regions may serve as anatomical sites for the integration of visual information about form and motion.

MT is also the source of information for VIP (the ventral intraparietal area; Ungerleider and Desimone, 1986), which in turn projects to several additional parietal areas, including LIP (lateral intraparietal area) and area 7a, located in the inferior parietal lobule (Andersen et al., 1990a; Blatt et al., 1990; Boussaoud et al., 1990). Interestingly, a number of these parietal areas receive direct inputs from the peripheral field representations of V1 and V2, bypassing MT (e.g., Colby et al., 1988), and these inputs may provide rapid activation of regions mediating spatial attention.

Neural processing is not, however, a simple matter of successive elaboration of information from lower-order to higher-order areas. As described earlier, at all stages connections are reciprocal. For example, an area receiving feedforward projections from an area earlier in the ventral stream also provides feedback projections to that area. Whereas feedforward projections provide bottom-up, sensory-driven inputs to subsequent visual areas, and severing these projections disconnects subsequent areas from their visual input (Rocha-Miranda et al., 1975; Schiller and Malpeli, 1977), the precise functions of the reciprocal feedback projections are still unknown. However, they are thought to play a top-down role in vision, such as in selective attention, by modulating activity in earlier areas of the processing stream. In addition to feedforward and feedback projections, there appear to be “intermediate-type” projections that link areas at the same level of the visual hierarchy. These are seen most notably between areas of the ventral and dorsal streams. For example, area V4 is interconnected with areas MT and MST (Maunsell and Van Essen, 1983a; Ungerleider and Desimone, 1986). Thus, there exists an anatomical substrate for interactions between the two processing streams at several levels, which we will discuss when we consider the physiological properties of neurons within the ventral and dorsal streams.

All areas within the ventral and dorsal streams also have heavy interconnections with subcortical structures, including the pulvinar, claustrum, and basal ganglia (Baizer et al., 1993; Benevento and Davis, 1977; Benevento and Rezak, 1976; Boussaoud et al., 1992; Kemp and Powell, 1970; Saint-Cyr et al., 1990; Selemon and Goldman-Rakic, 1985; Ungerleider et al., 1983, 1984; Webster et al., 1993). In addition, each stream receives subcortical modulatory inputs from ascending cholinergic projections from the basal forebrain and ascending noradrenergic projections from the locus coeruleus (Gatter and Powell, 1977; Mesulam and Van Hoesen, 1976; Webster et al., 1993). These projections are thought to play a role in the storage of information in cortex and the influence of arousal on information processing, respectively. Finally, visual information is sent from the last stations of both streams to the most ventral and anterior reaches of the temporal lobe, notably perirhinal cortex and parahippocampal areas TF and TH (Van Hoesen, 1980, 1982). These regions in turn project, via entorhinal cortex, to medial temporal lobe structures, such as the hippocampus, which contribute to forming long-term memories of visual objects and their contexts. Information is also sent from both streams to prefrontal cortex (Barbas and Mesulam, 1985; Cavada and Goldman-Rakic, 1989; Chavis and Pandya, 1976; Ungerleider and Desimone, 1986; Ungerleider et al., 1989; Webster et al., 1994), which plays an important role in working memory (Fuster, 1995; Goldman-Rakic, 1990), that is, holding an object and its location briefly in mind when it is no longer visible. Finally, there are direct projections from the ventral stream to the amygdala (Webster et al., 1993), which is important for attaching emotional valence to a stimulus, and from the dorsal stream to the pons (Boussaoud et al., 1992), which, in conjunction with the cerebellum, likely contributes to the control of eye and head velocity during tracking eye movements.