The anatomical separation of dorsal and ventral pathways seems clear for the visual system. However, the precise functional roles of these putative pathways are still being debated. My starting point for commenting on the present target article was to consider the possibility that a general distinction between action and perception might turn out not to be the correct organizing principle for other sensory systems.

I see three possible organizational principles that together seem to capture a host of findings cited as evidence in support of functional differences between the dorsal and ventral visual pathways. Briefly, these are, for dorsal versus ventral systems respectively: (1) peripersonal versus extrapersonal, referring to the space immediately surrounding the body versus more distant space; (2) egocentric versus allocentric, referring to representations of body coordinates versus environmental coordinates; and (3) online versus memory-based, referring to real-time versus delayed processing. [Some references that provide more precise definitions include Previc (Reference Previc1998), Rizzolatti et al. (Reference Rizzolatti, Fadiga, Fogassi and Gallese1997), and some of those cited in the target article.] Importantly, I see these not as absolute, but, rather, as relative distinctions that might provide guiding principles for defining the general organization of brain systems.

Based on the careful synthesis of findings provided by Dijkerman & de Haan (D&dH), and related work, it seems that for the visual pathways, all three of these principles are supported. Accordingly, these become candidates for consideration for the other sensory systems, as well.

With respect to the somatosensory system, it seems that on the basis of the neuroanatomical findings elaborated by D&dH, a strict separation of dorsal and ventral pathways is somewhat questionable. It might be just as reasonable to suggest that a mosaic of areas are involved in somatosensory processing, with greater reliance on the insula in the case of recognition (i.e., if recognition turns out to be a defining property). In addition, although some of the neuropsychological findings support the authors' proposal, many seem mixed and open to other interpretations.

In terms of the function of putative somatosensory systems, one might consider a different organizational principle than that proposed by D&dH. Consider, for example, the peripersonal versus extrapersonal space distinction. Most actions performed within peripersonal space involve both visual and somatosensory processes on objects manipulated within that space, although other sensory systems such as auditory and gustatory often are also involved. In contrast, when considering processing of objects some distance from the body (e.g., outside of reaching distance), it seems to be the visual system that does most of the perceiving (e.g., perceiving which object to act on next) – although, of course, audition is used to hear distant objects, and other systems might come into play to a lesser extent. Thus, with respect to peripersonal space, the visual and somatosensory systems are both primary; moreover, the involvement of posterior parietal cortex (PPC) for both types of information (visual and somatosensory) implies its crucial role in sensorimotor transformations on multisensory input. In sum, it seems that the visual and somatosensory systems show a rather nice parallel in terms of proposed functions of the dorsal system(s). For the proposed ventral system(s), however, this brief exercise suggests far less similarity for visual and somatosensory systems. One might ask: Does this lack of a parallel suggest that the proposed organizational principle is the wrong one? Or, does it suggest that general organizational principles might more effectively reveal differences rather than similarities across sensory systems? The point is that, perhaps the somatosensory system parallels the visual system in terms of some functions (those related to the dorsal pathways), but for other functions there is no direct analogue.

My second point is basically that a complete discussion of these putative systems might require some reference to the role of attention, particularly given the emphasis on findings in humans. For example, findings from our own neuropsychological studies suggest that the flexible allocation of attention might enable switches between egocentric and allocentric processing systems of the brain (Franz Reference Franz and Johnston-Frey2003). One might extend this idea to the present context by suggesting that the flexible allocation of attention is likely to influence which pathway(s) is (are) utilized at any particular moment in time. Perhaps more importantly, though, attention might to some extent define the functions of different systems; that is, attention itself might be a guiding principle (Franz Reference Franz2004).

In relation to the visual and haptic systems, a recent study from our lab demonstrated that visual and non-visual (haptic) attention processes can show similar kinds of effects on motor output (i.e., attention results in measurable changes in output parameters such as the size and shape of a drawn trajectory), although the effects of visual attention were generally greater than the effects of haptic attention (Franz & Packman Reference Franz and Packman2004). These findings could be accounted for by attentional modulations on dorsal system processing (for example). In a different set of studies that required subjects to estimate object sizes on the basis of internal representations, we found that subjects with congenital or acquired blindness tended to be more accurate in their estimations than were sighted subjects. Perhaps even more interesting, the blind subjects most often reported using strategies that relate to the manual grasping of objects (even though they were never asked to grasp the objects), whereas the sighted subjects reported more visual strategies. In addition, a control condition demonstrated that when visual memory processes were not used, the sighted subjects performed similarly to the blind subjects (Smith et al. Reference Smith, Franz, Joy and Whitehead2005). These findings again suggest that perhaps the dorsal stream mechanisms might actually be the same (or shared) for visual and somatosensory information, with flexible adaptations occurring based on strategy, attentional influences, and the availability of sensory information.

In sum, it seems that preliminary steps toward elucidating whether general principles apply to different sensory systems are well underway. In my view, the present target article will prove very valuable in generating useful discussion and insightful directions toward this aim.