Although most investigators would agree on the existence of functional specialization for processing somatosensory information, we argue here that the nature of such specialization, both within and outside classical somatosensory cortex, is far from clear. In monkey primary somatosensory cortex (SI), focal ablation of Brodmann's areas 1 or 2 has been found to specifically impair perception of texture or form, respectively (Randolph & Semmes Reference Randolph and Semmes1974); whereas, lesions of Brodmann's area 3b (Randolph & Semmes Reference Randolph and Semmes1974) and of secondary somatosensory cortex (SII) (Murray & Mishkin Reference Murray and Mishkin1984) non-selectively impair perception of both texture and form. Partly consistent with these findings are neurophysiological observations in monkeys that neuronal responses in SI and SII encode texture changes (Jiang et al. Reference Jiang, Tremblay and Chapman1997; Pruett et al. Reference Pruett, Sinclair and Burton2000; Sinclair et al. Reference Sinclair, Pruett and Burton1996) and that neurons in Brodmann's area 2 are shape-selective (Koch & Fuster Reference Koch and Fuster1989).

These functional differences described in monkeys do not map neatly onto humans. Somatosensory cortical lesions were formerly thought to produce complex discriminative deficits, as opposed to thalamic lesions resulting in simpler problems with detection of touch or pain, but this clinical dogma has fallen by the wayside in the era of neuroimaging (Sathian Reference Sathian2007). Whereas SI lesions affect both texture and form discrimination (Roland Reference Roland1987), parietal opercular lesions seem to specifically compromise form perception (Roland Reference Roland1987; Caselli Reference Caselli1993; Reed et al. Reference Reed, Caselli and Farah1996). However, another report indicates that SI lesions cause deficits in kinesthesis and discriminative tactile perception, whereas parietal opercular-insular lesions result in thermo-nociceptive deficits and poststroke pain (Kim Reference Kim2007). Thus, the findings from studies of patients with somatosensory cortical lesions have not yet led to a coherent picture.

Functional neuroimaging studies of somatosensory cortex have led to similarly heterogeneous results. Some studies have noted common regions of activation in the postcentral sulcus (PCS) (O'Sullivan et al. Reference O'Sullivan, Roland and Kawashima1994; Servos et al. Reference Servos, Lederman, Wilson and Gati2001) or parietal operculum (Ledberg et al. Reference Ledberg, O'Sullivan, Kinomura and Roland1995) during perception of both texture and form. We reported PCS activity selective for form relative to texture in two independent studies (Peltier et al. Reference Peltier, Stilla, Mariola, LaConte, Hu and Sathian2007; Zhang et al. Reference Zhang, Weisser, Stilla, Prather and Sathian2004). Since the PCS corresponds to Brodmann's area 2 (Grefkes et al. Reference Grefkes, Geyer, Schormann, Roland and Zilles2001), form-selectivity in this region is consistent with the aforementioned neurophysiological and lesion studies in monkeys that implicate area 2 in form perception (Koch & Fuster Reference Koch and Fuster1989; Randolph & Semmes Reference Randolph and Semmes1974). Despite the implication of parietal opercular activity in form perception (Reed et al. Reference Reed, Shoham and Halgren2004), such activity has not been shown to be form-selective. On the contrary, parietal opercular activity is stronger during texture than form perception (Roland et al. Reference Roland, O'Sullivan and Kawashima1998; Stilla et al. Reference Stilla, Mariola and Sathian2006). It should be noted that parietal opercular cortex, which was initially equated with SII (for a review, see Burton Reference Burton, Jones and Peters1986), actually contains multiple somatosensory areas, with evidence now for three areas in monkeys (Fitzgerald et al. Reference Fitzgerald, Lane, Thakur and Hsiao2004) and four in humans (Eickhoff et al. Reference Eickhoff, Amunts, Mohlberg and Zilles2006a; Reference Eickhoff, Schleicher, Zilles and Amunts2006b).

There is, surprisingly, greater consensus on findings outside classical somatosensory cortex, in regions now identified as multisensory. Conjoint tactile and visual form-selectivity characterize the visual cortical region known as the lateral occipital complex (LOC) (Amedi et al. Reference Amedi, Malach, Hendler, Peled and Zohary2001; Peltier et al. Reference Peltier, Stilla, Mariola, LaConte, Hu and Sathian2007; Zhang et al. Reference Zhang, Weisser, Stilla, Prather and Sathian2004), the anterior intraparietal sulcus (aIPS) (Grefkes et al. Reference Grefkes, Weiss, Zilles and Fink2002; Peltier et al. Reference Peltier, Stilla, Mariola, LaConte, Hu and Sathian2007; Zhang et al. Reference Zhang, Weisser, Stilla, Prather and Sathian2004), and foci located more caudally in the IPS (Peltier et al. Reference Peltier, Stilla, Mariola, LaConte, Hu and Sathian2007; Saito et al. Reference Saito, Okada, Morita, Yonekura and Sadato2003; Zhang et al. Reference Zhang, Weisser, Stilla, Prather and Sathian2004). Tactile texture-selectivity outside somatosensory cortex appears to be confined to medial occipital cortex at the border between visual areas V1 and V2, where it overlaps with visual texture-selectivity (Stilla et al. Reference Stilla, Mariola and Sathian2006). Task-specific, modality-independent engagement of particular cerebral cortical regions during perception seems to be a general rule; that is, regions known to be specialized for processing particular aspects of vision are also recruited during performance of corresponding tactile tasks (for a review, see Sathian Reference Sathian2005).

So, what about the proposed dichotomy of processing? One study (Reed et al. Reference Reed, Klatzky and Halgren2005) suggested distinct processing streams for form and extrapersonal spatial processing, although selectivity was found only outside somatosensory cortex, involving fronto-cingulate cortex and superior parietal cortex, respectively. Another study (van Boven et al. Reference van Boven, Ingeholm, Beauchamp, Bikle and Ungerleider2005) described hemispheric specialization within posterior parietal cortex, rather than a dorso–ventral dichotomy, the left aIPS being preferentially active during discrimination of grating orientation, and the right temporo-parietal junction, during discrimination of small differences in grating location on the fingerpad. We have reported somewhat similar specializations, involving left aIPS and left parieto-occipital cortex for tactile discrimination of grating orientation, and the right angular gyrus for tactile discrimination of grating groove width (Sathian et al. Reference Sathian, Zangaladze, Hoffman and Grafton1997; Zangaladze et al. Reference Zangaladze, Epstein, Grafton and Sathian1999; Zhang et al. Reference Zhang, Mariola, Stilla, Stoesz, Mao, Hu and Sathian2005). Our findings of selectivity for form in PPC foci and for texture in parietal opercular cortex (Stilla et al. Reference Stilla, Mariola and Sathian2006) could be interpreted as consistent with the proposed action/perception dichotomy of somatosensory processing: The dorsally directed pathway into PPC can be considered as specialized for grasping, which fits with findings for the aIPS region in monkeys (Murata et al. Reference Murata, Gallese, Luppino, Kaseda and Sakata2000) and humans (Frey et al. Reference Frey, Vinton, Norlund and Grafton2005) – grasping is an action closely linked to haptic form perception. The ventral pathway via the parietal operculum into the insula, in contrast, might be particularly concerned with material object properties, such as texture, which are especially the province of touch (Klatzky et al. Reference Klatzky, Lederman and Reed1987). Even the activity of the left aIPS during mental rotation of tactile (Prather et al. Reference Prather, Votaw and Sathian2004) and visual (Alivisatos & Petrides Reference Alivisatos and Petrides1997) forms, which we suggested represents activation of the dorsal visual pathway on account of spatial processing (Prather et al. Reference Prather, Votaw and Sathian2004), could be viewed in terms of manipulation of images, a kind of internal action. However, given the multifaceted nature of somatosensory specialization, we think the proposed action/perception dichotomy should be considered a working hypothesis of somatosensory processing to motivate further research, rather than being regarded as established.