In the target article, Dijkerman & de Haan (D&dH) propose that the visual-tactile modalities interact closely through the polymodal higher-order cortical areas such as the posterior parietal cortex (PPC) and the lateral occipital complex (LOC), whose functions may differ. They suggest that the PPC is involved in multimodal coding of body-related configuration for the guidance of action, whereas the LOC is a multimodal area involved in perceptual representation of object form features. Here we argue that these polymodal areas also show some specificity of the cross-modal interaction for perception.
Mah-Jong is a Chinese game similar to card games, involving two-dimensional plastic tiles with various marks carved on one side; therefore, shape can be identified by both visual and tactile inspection. Visual-tactile cross-modal shape comparison with Mah-Jong tile activated the caudal part of the intraparietal sulcus more prominently than did intramodal comparison (Saito et al. Reference Saito, Okada, Morita, Yonekura and Sadato2003). This is consistent with the notion that the two processing streams in both modalities project to the same higher-order cortical areas that are involved in multimodal sensory integration. Therefore, this finding may suggest that the PPC is related to the comparison of shapes coded by different sensory modalities.
Intramodal tactile shape comparison does not usually recruit the visual cortex in activation (Harada et al. Reference Harada, Saito, Kashikura, Sato, Yonekura, Honda and Sadato2004). In contrast, during three-dimensional object naming, Amedi et al. (Reference Amedi, Malach, Hendler, Peled and Zohary2001) first demonstrated consistent somatosensory activation in the LOC. They suggested that cortical neurons in the LOC in humans may function as a part of the multimodal object-selective network (Amedi et al. Reference Amedi, Malach, Hendler, Peled and Zohary2001). The activation in the LOC reflects stored object-related visual information that can be accessed via cues from somatosensory modalities, and possibly from other modalities as well (Amedi et al. Reference Amedi, Malach, Hendler, Peled and Zohary2001). This argument suggests that direct interactions among modality-specific sensory pathways underlie the multimodal representation of objects (Amedi et al. Reference Amedi, Malach, Hendler, Peled and Zohary2001). According to this view, bimodal activation occurs in the visual cortex rather than in the somatosensory areas because object recognition relies primarily on vision. Hence, the LOC is related to the multimodal shape identification.
The involvement of the multimodal LOC in tactile processing appears to be modified by learning (Saito et al. Reference Saito, Okada, Honda, Yonekura and Sadato2006) or sensory deprivation (Sadato et al. Reference Sadato, Okada, Kubota and Yonekura2004). Some well-trained Mah-Jong players can identify the carved patterns of the tiles by touch. The subjects who were well trained in the tactile comparison of Mah-Jong patterns showed more prominent activation in the visual cortex, including the LOC and V1, when performing this task than did the naïve subjects, who showed activation in the LOC but not in V1 (Saito et al. Reference Saito, Okada, Honda, Yonekura and Sadato2006). Both the LOC and the V1 of the well-trained subjects were activated during Braille tactile comparison tasks to which they are naïve. This is probably due to cross-modal plastic changes produced by the effects of long-term training on one set of cards (Mah-Jong) that generalizes to an untrained set of cards (Braille) (Saito et al. Reference Saito, Okada, Honda, Yonekura and Sadato2006). During tactile shape comparison with Braille characters, lately blind subjects (i.e., those with late onset of blindness) who had never learned Braille showed the activation of the association visual cortex, including the LOC (Sadato et al. Reference Sadato, Okada, Kubota and Yonekura2004). This may be interpreted as that the visual deafferentation caused less demand on bottom-up visual processing, which may in turn introduce an opportunity for the expansion of tactile representations in the visual association cortex.
Burton et al. (Reference Burton, Sinclair and McLaren2004) showed that simple vibrotactile stimulation activates both lower-tier visuotopic (e.g., V1, V2, VP, and V3) and several higher-tier visual areas (e.g., V4v, V8, and BA 37). Early blind participants showed the most extensive distribution of activity. Lately blind participants exhibited activity in mostly similar regions, but the response magnitudes declined with the age of onset of blindness. Three sighted individuals had supra-threshold activity in V1. These results suggest that vibrotactile inputs probably activate the visual cortex through some latent pathway common to both blind and sighted subjects (Burton et al. Reference Burton, McLaren and Sinclair2006). Burton et al. (Reference Burton, McLaren and Sinclair2006) speculated that visual deprivation alone induces reorganization of the visual cortex, particularly in regions with already strong multisensory properties, where a competitive shift to non-visual inputs may readily follow visual deprivation (Sadato et al. Reference Sadato, Okada, Kubota and Yonekura2004). In contrast, cross-modal reorganization of the lower-tier visual areas, which are not cross-modally responsive in sighted people, may be activated particularly through regular attention to selected non-visual inputs. Such learning might be needed to strengthen the more remote connections with multisensory cortical areas (Saito et al. Reference Saito, Okada, Honda, Yonekura and Sadato2006). Cross-modal reorganization of lower-tier visual areas may thus be triggered in sighted subjects by learning such skills as Mah-Jong pattern identification by touch.
In conclusion, visual–tactile interaction for perception is mediated by task-specific pathways which are dynamically modified by sensory deprivation or learning.
