The secondary somatosensory cortex (SII) of nonhuman primates is located on the parietal operculum. In the monkey, electrophysiological and connectivity tracing studies as well as histological investigations provide converging evidence for 3 distinct cortical areas (SII, PV, and VS) within this region, each of which contains a complete somatotopic map. Although the equivalency of the parietal operculum as the location of SII between humans and nonhuman primates is undisputed, the internal organization of the human SII region is still largely unknown. Based on their topography, we have previously argued that the cytoarchitectonic areas OP 1, OP 4, and OP 3 may constitute the human homologues of areas SII, PV, and VS, respectively. To test this hypothesis, we here examined (using functional magnetic resonance imaging) the somatotopic organization of the human parietal operculum by applying tactile stimulation to the skin at 4 different locations on either side of the body (face, hands, trunk, and legs). The locations of the resulting activation foci were then compared with the cytoarchitectonic maps of this region. Data analysis revealed 2 somatotopic body representations on the lateral operculum in areas OP 1 and OP 4. The functional border between these 2 body maps was defined by a mirror reversal in the somatotopic arrangement and coincided with the cytoarchitectonically defined border between these 2 areas. This somatotopic arrangement closely matches that described for SII and PV in nonhuman primates. The data also suggested a third somatotopic map located deeper inside the Sylvian fissure in area OP 3. Based on the observed topographic arrangement and their functional response characteristics, we conclude that cytoarchitectonic areas OP1, OP 4, and OP 3 on the human parietal operculum constitute the human homologues of primate areas SII, PV, and VS, respectively.
The concept of a secondary somatosensory cortex (SII) located ventrally to the primary somatosensory cortex (SI) has been introduced more than 60 years ago based on electrophysiological studies in cats (Adrian 1940). Over the following decades, corresponding areas were described in virtually all examined mammals including nonhuman primates (Burton 1986). In anthropoid primates, SII is located on the upper bank of the Sylvian fissure, that is, the parietal operculum. A correspondingly located human SII region was first described by direct electrical stimulation (Penfield and Jasper 1954). Subsequently, studies using evoked potentials (Woolsey et al. 1979) and positron emission tomography (PET) (Fox et al. 1987; Burton et al. 1993) confirmed these findings. Since then SII activations have been reported consistently for a wide range of experimental conditions, for example, light touch and pain (cf., review and meta-analysis by Eickhoff, Amunts, et al. 2006), but also for more complex tasks such as tactile attention (Burton and Sinclair 2000; Lam et al. 2001) and sensory-motor integration (Inoue et al. 2002; Wasaka et al. 2005). In contrast to SI, the human parietal operculum commonly shows bilateral activation even with unilateral peripheral stimulation (Ruben et al. 2001; Del Gratta et al. 2002; Bingel et al. 2003; Young et al. 2004). This observation is in good accordance with SII response characteristics in nonhuman primates (Robinson and Burton 1980a; Burton et al. 1995; Kaas and Collins 2003).
Nevertheless, the concept of “SII” has changed considerably over the last 15 years. Electrophysiological and tracing studies as well as histological investigations in a variety of species have provided converging evidence that the SII region can be subdivided into several distinct areas (Cusick et al. 1989; Krubitzer and Kaas 1990, 1993; Burton et al. 1995; Krubitzer, Clarey, et al. 1995; Huffman et al. 1999; Slutsky et al. 2000). In this context, one important feature of a cortical area is the presence of a separate, complete somatotopic map (Fink et al. 1997). The idea behind this approach is the assumption that each part of the body is represented only once in each area, a claim that is strongly supported by the role these topographic maps seem to play in the development of sensory cortices (Kaas 1997, 2000). To date, the most commonly used nomenclature (Kaas and Collins 2003) for SII subareas in nonhuman primates is as follows (Fig. 1): The area immediately ventral to SI is the “parietal ventral area” (PV). PV is caudally followed by “area SII,” which must be differentiated from the earlier defined “SII region” as the term SII is used for the whole region as well as for an individual area within it. SII and PV each contain a full somatotopic map. These maps constitute mirror images of each other, bordering at the representations of the most distal parts of the body: face (lateral), hands (intermediate), and feet (medial). More proximal parts of the body, for example, shoulder, trunk, and legs (which are considered a proximal body part because they are distally followed by the feet) are represented further apart from this border in the anterior part of PV and the posterior aspect of SII (Burton et al. 1995; Krubitzer, Clarey, et al. 1995). A third SII subregion, the “ventral somatosensory area” (VS), has been described in the depth of the Sylvian fissure medial to SII and PV. VS contains a crude rostrocaudal somatotopic map in which the head is represented most anteriorly (Cusick et al. 1989; Krubitzer and Calford 1992; Qi et al. 2002).
An homologous subdivision of the human SII was then proposed (Disbrow et al. 2000) using functional magnetic resonance imaging (fMRI). Disbrow et al. provided data which suggested that the human SII region may contain several somatotopically organized areas by demonstrating multiple activation foci within the SII region for various body parts. Further evidence for multiple areas in the human SII region was provided by fMRI and PET studies showing multiple opercular activation foci (Burton et al. 1993; Ledberg et al. 1995; Ferretti et al. 2003).
A recent cytoarchitectonic study of our own group (Eickhoff, Amunts, et al. 2006; Eickhoff, Schleicher, et al. 2006), based on histological examination in a series of postmortem brains, identified 4 distinct cytoarchitectonic areas (OP 1–4) on the human parietal operculum. Topographically OP 4 corresponds to primate PV as both are located near to the superficially exposed cortical surface within the Sylvian fissure bordering SI. OP 1 is located caudally to OP 4 and thus seems to constitute the human homologue of area SII. Like macaque areas SII and PV, OP 4 and OP 1 share a common border running in a medial to lateral direction. In contrast, OP 3 is located deeper in the Sylvian fissure than both aforementioned areas. It is thus the most likely candidate for a human homologue of area VS. Finally, OP 2 is not part of the SII but rather comparable with the parietal-insular vestibular cortex (PIVC) of nonhuman primates (Eickhoff, Weiss, et al. 2006).
Please note that in order to allow the reader a better orientation and avoid unnecessary confusion due to differences in nomenclature, we will use the terms OP 1 (SII), OP 2 (PIVC), OP 3 (VS), and OP 4 (PV) throughout the text. This denotation is based on the established topological correspondence between these 4 cytoarchitectonic fields on the human parietal operculum (Eickhoff, Amunts, et al. 2006; Eickhoff, Schleicher, et al. 2006) and the respective areas in nonhuman mammals, as well as on the existing evidence for their functional homology (Young et al. 2004; Naito et al. 2005; Eickhoff, Lotze, et al. 2006; Eickhoff, Weiss, et al. 2006).
The aim of the present study was to examine whether the somatotopic organization of OP 1, 3, and 4 corresponds to that of SII, PV, and VS in nonhuman primates. We accordingly analyzed the functional responses elicited on the parietal operculum by tactile stimulation of 4 body parts: face, hand, trunk, and leg. These locations were chosen based on the somatotopic organization of the SII region in nonhuman primates: In macaques, head and hands are represented at the border between there areas. They may therefore be used to define the functional border between the human homologues of these areas. The opercular representation of the trunk and the legs (instead of the feet, which are also represented at the functional border between SII and PV) was analyzed because these more proximal sites are represented at the anterior border of PV and the posterior border of area SII, respectively. They should thus have 2 clearly separable representations in their human homologues. Furthermore, if the human parietal operculum is organized similar to that of nonhuman primates, the head should be represented lateral to the hand, whereas the trunk should be represented lateral to the leg.
Materials and Methods
Subjects and Stimulation
14 healthy subjects (7 males, mean age 25.6 ± 3.4 years) with no history of neurological or psychiatric illness gave informed consent. All subjects were strongly right handed as assessed by the Edinburgh handedness inventory (Oldfield 1971). The study was approved by the ethics committee of the Medical Faculty, RWTH Aachen, Germany.
Somatosensory stimulation was applied to the face, hands, trunk, and legs. The extent of the stimulated skin areas is illustrated in Figure 2. Each location was separately stimulated on either the left or the right side of the body resulting in 8 separate conditions. Stimulation was performed manually by brushing the subject's skin with a sponge in a back-and-forth manner with a frequency of approximately 2 Hz because this method was previously shown to be highly effective in evoking SII activations in both humans (Disbrow et al. 2000) and nonhuman primates (Robinson and Burton 1980b; Krubitzer and Kaas 1990; Krubitzer, Clarey, et al. 1995). Upon debriefing after the experiment, all subjects reported that the stimuli were easy to perceive but had not caused any pain. The fMRI paradigm consisted of 8 sessions of 11 stimulation cycles. Each cycle consisted of approximately 18-s stimulation followed by 18 s rest. In each of the 8 sessions, a different anatomical site was stimulated. The order of the experimental sessions was pseudorandomized across subjects.
FMRI Procedure and Image Preprocessing
Functional magnetic resonance (MR) images were acquired on a Siemens Sonata 1.5-T whole-body scanner (Erlangen, Germany) using blood oxygen level–dependent (BOLD) contrast (gradient-echo echo planar imaging [EPI] pulse sequence, time repetition = 3 s, resolution = 3.1 × 3.1 × 3.1 mm, 30 axial slices for whole-brain coverage). Each session consisted of 132 EPI images. The fMRI scanning was preceded by the acquisition of 4 dummy images allowing the MR scanner to reach a steady state which were discarded prior to further analysis. Additional high-resolution anatomical images (voxel size 1 × 1 × 1 mm) were acquired using a standard T1-weighted 3-dimensional magnetization-prepared rapid gradient-echo sequence. Images were analyzed on a Pentium 4 Windows XP system using SPM5 (http://www.fil.ion.ucl.ac.uk/spm). The EPI images were corrected for head movement between scans by an affine registration (Ashburner and Friston 2003b). One subject was removed from further analysis due to excessive head motion (more than 1.5 mm movement between scans). The T1 scan was coregistered to the mean of the realigned EPIs and subsequently normalized to the Montreal Neurological Institute (MNI) single-subject template (Evans et al. 1992; Collins et al. 1994; Holmes et al. 1998) using linear proportions and a nonlinear sampling algorithm (Ashburner and Friston 2003a, 2003c). The resulting normalization parameters were then also applied to the EPI volumes. These were hereby transformed into standard stereotaxic space and resampled at 1.5 × 1.5 × 1.5 mm voxel size. The normalized images were spatially smoothed using a 6-mm full width half maximum Gaussian kernel to meet the statistical requirements of the general linear model and to compensate for residual anatomical variations across subjects.
The data were analyzed in the context of the general linear model employed by SPM5 (Kiebel and Holmes 2003). Each experimental condition was modeled using a boxcar reference vector convolved with a canonical hemodynamic response function. Low-frequency signal drifts were filtered using a set of discrete cosine functions with a cutoff period of 72 s. No global scaling was applied. The main effects for the 8 stimulation conditions were computed by applying appropriate baseline contrasts.
The corresponding contrasts from different subjects were then analyzed in a second-level Bayesian mixed-effects model to allow inference to the general population (Penny and Holmes 2003) using the probabilistic empirical Bayes algorithm implemented in SPM5. This algorithm calculates the conditional distribution for the parameter estimates (across subjects) at each voxel using the variance across voxels as Bayesian prior (Friston 2002; Friston, Glaser, et al. 2002; Friston, Penny, et al. 2002; Friston and Penny 2003a, 2003b). The resulting posterior probability maps were thresholded at a probability of 99.99% for an effect size greater than 2 prior standard deviations (γ threshold). That is, only those voxels were considered as significantly activated that had parameter estimates larger than γ with at least 99.99% confidence. The rationale for using the prior standard deviation as the effect size threshold γ is that it equates to a “background noise level,” that is, to a level of activation that is generic to the brain as a whole. The chosen threshold thus allows directing Bayesian inference to only show those voxels that are almost certainly more active than this generic response (Friston, Glaser, et al. 2002; Friston and Penny 2003a, 2003b). For the considerably weaker activations resulting from stimulation of trunk and legs, the threshold was lowered to 99% confidence for activation greater than background noise.
In order to delineate those parts of the opercular cortex that were tuned most specifically to input from the examined body parts, we subsequently identified those clusters of voxels where stimulation of a particular body part (irrespectively of the stimulated side) was more likely to evoke activation than stimulation of any of the 3 other examined body parts. First, the joint probability for activation following either left- or right-sided stimulation of a particular body parts was computed for each voxel to quantify its responsiveness to stimulation of each body parts. To delineate the core of the SII representation of the head, hands, trunk, and legs, these joint probabilities were then compared across stimulation sites, resulting in a maximum likelihood representation.
Each voxel was hereby assigned to that body part whose stimulation resulted in the highest probability for activation at this particular voxel, given that the joint posterior probability for activation following left- or right-sided stimulation was at least 99%. The resulting clusters of maximum likelihood representation thus define those parts of the SII region, which were most responsive to stimulation of the respective body part. It has to be noted that the respective clusters are mutually exclusive because in a particular voxel, only a single body part can be represented most likely, although there will be some apparent overlap in the surface projection used to illustrate the location of these regions.
Comparison with Cytoarchitectonic Data
The localization of significant activations with respect to cytoarchitectonic areas was analyzed on the basis of a summary map (maximum probability map) that identifies the most likely anatomical area at each voxel of the MNI single-subject template (Eickhoff et al. 2005; Eickhoff, Heim, et al. 2006). The definition of this map is based on probabilistic cytoarchitectonic maps derived from the analysis of cortical areas in a sample of 10 human postmortem brains (Schleicher et al. 2005), which were subsequently normalized to the MNI reference space (Amunts et al. 2004; Eickhoff, Schleicher, et al. 2006). To quantify the correspondence between anatomical and the functional data, the maximum likelihood clusters defined above were then compared with the probabilistic maps of the parietal operculum (Eickhoff et al. 2005; Eickhoff, Amunts, et al. 2006) using the SPM Anatomy toolbox (www.fz-juelich.de/ime/spm_anatomy_toolbox).
Probability of Activation within OP 1–4
This analysis represented a complementary approach to the just-described identification of the significant activation foci and was carried out in order to characterize the responsiveness of OP 1–4 to tactile stimulation. First, the mean posterior probabilities for activation greater than background noise following the 8 experimental conditions (4 body parts × 2 sides) were calculated within each area. The probabilities for activation of this area following stimulation of a specific body part regardless of the stimulated side were subsequently derived by the union of the probabilities for left- and right-sided stimulations. These probabilities were then statistically compared by means of a 2-way analysis of variance (ANOVA). The 2 factors of this ANOVA were “area” and “body part.” The level of significance was P < 0.05. If the effect of a factor was significant, we used a subsequent pairwise multiple comparison procedure (Tukey test) to isolate the conditions in which the levels of this factor differed significantly (P < 0.05, corrected for multiple comparisons).
We will first describe the significant activations on the parietal operculum separately for each body part. To assess the effectiveness and specificity of applied stimulation, we also examined the evoked activation in the SI, that is, areas 3a, 3b, 1, and 2 (Geyer et al. 1999, 2000; Grefkes et al. 2001). We will then combine these individual results into a somatotopic map of the human parietal operculum and analyze the mean probabilities for activations following somatosensory stimulation in areas OP 1–4.
As expected (Boling et al. 2002; Iannetti et al. 2003), the face representation within SI was located bilaterally on the inferior lateral aspect of the postcentral gyrus, close to but clearly separated from the Sylvian fissure (Fig. 3A,C).
The contralateral SII activation to left- and right-sided face stimulations consisted of 2 clusters on the lateral parietal operculum (Fig. 3B,D). One of them was found at the lateral aspect of the border between OP 1 (SII) and OP 4 (PV), the other was located more posteriorily and medially, that is, within OP 1 (SII). In the left face condition, the anterior lateral cluster showed a second maximum located in OP 3 (VS) (Fig. 3D). Stimulation of the left face also resulted in a significant cluster of activation on the left lateral operculum at the border between OP 1 (SII) and OP 4 (PV) (Fig. 3B). Stimulation of the right face, on the other hand, caused more scattered activation on the ipsilateral operculum (Fig. 3B). Four of these clusters were located in the superficial parts of OP 1 (SII) and OP 4 (PV) while the fifth cluster was observed more medially within OP 4 (PV) extending into OP 3 (VS) (24% of its volume were allocated to OP 3 [VS]).
The location of the anterior parietal activations (Fig. 4A,C) was in good accordance with the well-established location of the “SI hand area” between the middle and superior third of the postcentral gyrus (Bingel et al. 2003; Blankenburg et al. 2003; Young et al. 2004).
Stimulation of both hands resulted in a single yet extended cluster of activation in contralateral SII. These clusters were located medially to the activation observed for face stimulation. On both hemispheres, the opercular hand activations were located at the border between OP 1 (SII) and OP 4 (PV) (Fig. 4B,D). Moreover, both activations also extended into OP 3 (VS) (25% of cluster volume in the left hand condition, 20% in the right hand condition). Ipsilateral activation, on the other hand, was much smaller and only significant at a more lenient threshold (99% confidence) where it appeared at the same location as the contralateral activations.
Following right trunk stimulation, 2 distinct activation clusters were identified on the contralateral operculum (Fig. 5B). The contralateral SII activation following left-sided stimulation was considerably weaker but in a similar location (Fig. 5D). These 2 clusters on the respective contralateral sides were located in the superficial part of the parietal operculum similar to those observed for face stimulation. However, instead of being located at the border between OP 1 (SII) and OP 4 (PV), the trunk activations were located at the anterior border of OP 4 (PV) and the posterior border of OP 1 (SII), that is, to either side of the head representation. Ipsilateral to the stimulated side, the same pattern was observed for the stimulation of the right trunk (Fig. 5B). Stimulation of the left trunk, on the other hand, failed to evoke activation in left OP 4 (PV), that is, the anterior focus was missing (Fig. 5D).
In SI, the legs were represented even more medially than the trunk (Fig. 6A,C). Their location close to the midline, extending into the interhemispheric fissure, was similar to the one described in previous imaging studies (Del Gratta et al. 2000; Ruben et al. 2001).
SII activations following leg stimulation were weaker than those resulting from stimulation of any other body part. Nevertheless, 3 distinct activation clusters were observed for both left- and right-sided stimulations on the contralateral operculum (Fig. 6B,D). Similar to the activations following trunk stimulation, 2 of these foci were located at the anterior border of OP 4 (PV) and the posterior border of OP 1 (SII). However, activation evoked by leg stimulation was located more medially than the trunk representations. Stimulation of either leg also resulted in a third, more medial activation cluster in the contralateral operculum extending into OP 3 (VS). Ipsilateral activation was mainly found on the posterior parietal operculum (OP 1 [SII]). The anterior focus, on the other hand, was missing following right leg stimulation (Fig. 6B).
Somatotopic Maps on the Human Parietal Operculum
The opercular representation of a specific body part was delineated by computing its clusters of maximum likelihood representation, that is, by identifying those voxels where the probability for activation following stimulation of that body part was higher than those following stimulation of any of the 3 other parts of the body (Fig. 7 and Table 1). A synopsis of these clusters clearly outlines the somatotopic organization of OP 1 (SII) and OP 4 (PV). The face and hand areas were mainly located centrally on the parietal operculum at the border between OP 1 (SII) and OP 4 (PV) (Fig. 7A,B). In contrast, the trunk and the legs were represented twice on the parietal operculum. That is, they were represented anterior (in the anterior part of OP 4 [PV]) as well as posterior (in the posterior part of area OP 1 [SII]) to the representations of the distal body parts, which were found at the OP 1 (SII)/OP 4 (PV) border (Fig. 7C,D). Moreover, the representations of the different body parts were also arranged in a rostrocaudal sequence from the superficial parietal operculum to its deeper aspect. That is, the head was represented lateral to the hand and the trunk was represented lateral to the leg.
|Center||Voxel||% P (OP 1) ( = 27%)||% P (OP 2) ( = 26%)||% P (OP 3) ( = 24%)||% P (OP 4) ( = 25%)||∑|
|Center||Voxel||% P (OP 1) ( = 27%)||% P (OP 2) ( = 26%)||% P (OP 3) ( = 24%)||% P (OP 4) ( = 25%)||∑|
Note: The representation for each body part comprised those voxels, which were tuned most specifically to input from the examined body parts. The center of gravity (in anatomical MNI space—Evans et al. 1992; Eickhoff et al. 2005) and the extent and the mean probability for anatomically defined areas OP 1–4 are given for each of these representations. For comparison, the mean probabilities for each area averaged over its entire probability map are given in the top row. The numbers in the left hand column correspond to those used to label the respective clusters in Figure 7.
The somatotopic organization of the deeper parietal operculum was less well defined. A distinct representation of the leg was observed on the medial operculum at the border between OP 2 (PIVC) and OP 3 (VS) (Fig. 7D), clearly separated from activation in OP 1 (SII) and OP 4 (PV). No such distinct representation in the deep parietal operculum was observed for any of the other examined body parts. However, the representations of both hands and both sides of the face extended into OP 3 (VS) (Fig. 7A,B). All these activations in OP 3 (VS) were located anterior to its leg representation. However, a clear differentiation between them, corresponding to a full somatotopic arrangement within OP 3 (VS), was impossible. Whereas the head was represented anterior to the hand on the right hemisphere, these activations largely overlapped on the left. In the latter hemisphere, the trunk activation also extended into OP 3 (VS) at roughly the same anterior–posterior location as that following leg stimulation.
Probability of Activation within OP 1–4
The mean posterior probabilities for activation greater than background noise within OP 1–4 following tactile stimulation were compared by means of a 2-way ANOVA. This analysis revealed a significant main effect of the factor body part (F = 5.8, degrees of freedom [df] = 3, P < 0.05). In the subsequent pairwise comparison, the probabilities for activation following leg stimulation were significantly lower than those following stimulation of the face, hand, or trunk (P < 0.05, corrected for multiple comparisons). No significant difference in the probabilities for SII activation was found between any other body parts. The main effect for factor area was also significant (F = 11.1, df = 3, P < 0.05). The pairwise comparison between OP 1–4 revealed that the probability for activation following tactile stimulation was significantly higher in OP 1 (SII) than in any other area (P < 0.05 corrected, Fig. 8). Similarly, the mean probability for activation was significantly lower in OP 2 (PIVC) than in OP 1 (SII), 3, or 4 (P < 0.05 corrected). There was, however, no significant difference in the posterior probabilities between OP 3 (VS) and OP 4 (PV) (Fig. 8).
In the present study, we examined the location of significant BOLD signal changes on the parietal operculum (i.e., SII) following tactile stimulation of the face, hands, trunk, and legs. Two mirror image somatotopic representations were observed within cytoarchitectonically defined areas OP 1 (SII) and OP 4 (PV). The functional border between these 2 body maps, defined by the mirror reversal in their somatotopic arrangement, coincided with the architectonic border between these 2 areas. There was also some evidence for a third somatotopic map deeper in the Sylvian fissure, corresponding most likely to cytoarchitectonically defined area OP 3 (VS). This interpretation was corroborated by the observation that the mean probability for somatosensory activations in OP 3 (VS) was not significantly smaller than that for OP 4 (PV).
Previous Maps of the Parietal Operculum in Monkeys and Humans
The observation of multiple representations for each body part in the lateral sulcus has challenged the traditional view of the SII as an uniform region since the late 80s of the last century (Krubitzer et al. 1986; Cusick et al. 1989). Subsequent data obtained in nonhuman primates by electrophysiology (Krubitzer et al. 1986; Cusick et al. 1989; Krubitzer and Calford 1992; Krubitzer, Clarey, et al. 1995; Beck et al. 1996; Huffman et al. 1999; Qi et al. 2002), tracer injection (Krubitzer et al. 1986, 1993; Cusick et al. 1989; Krubitzer and Kaas 1990; Burton et al. 1995; Beck et al. 1996; Qi et al. 2002), and histological studies (Krubitzer et al. 1986, 1993; Krubitzer and Calford 1992; Burton et al. 1995; Krubitzer, Manger, et al. 1995; Huffman et al. 1999; Kaas and Collins 2001; Qi et al. 2002) revealed converging evidence for the existence of at least 3 distinct cortical areas (SII, PV, and VS) within the SII region. Therefore, these areas are now regarded as part of the general organization of somatosensory cortices in primates and possibly also other mammals (Krubitzer 1995; Kaas and Collins 2001, 2003). An overview on their topography and somatotopic organization is given in Figure 9 next to a summary of the somatotopic layout within the cytoarchitectonic areas of the human parietal operculum as observed in the present study.
First evidence for human homologues of these areas was provided by an fMRI study examining the representation of 5 body parts in SII (Disbrow et al. 2000). This data indicated a double representation of the shoulder and the hips within SII and thereby supported the notion of a homology of the SII region between monkeys and humans. The present study confirms the findings by Disbrow et al. but extends them in several important aspects: First, we used a genuine random-effects model for the analysis of our data, which compares the observed effect sizes with their variation across subjects and thus allows inference on the general population from which the subjects were drawn (Friston, Penny, et al. 2002; Penny and Holmes 2003). Second, in contrast to Disbrow et al., we examined the entire brain, not just the perisylvian cortex, allowing the additional examination of activations in SI. The importance of this supplementary information lies in the fact that by comparison of the evoked activations in SI with its well-known somatotopic organization, we could monitor the efficacy and specificity of the stimulation. Given the good correspondence between the SI activation observed in our study and their previously described location, we demonstrated that the applied stimulation was appropriate for somatotopic mapping. Finally and most important, due to the recent progress in cytoarchitectonic mapping of the human cerebral cortex (Zilles et al. 2002, 2003; Schleicher et al. 2005; Eickhoff, Schleicher, et al. 2006), we were now able to relate the observed activations and their somatotopic arrangement to distinct cytoarchitectonically defined cortical areas. This comparison allows for the first time to examine a crucial aspect of homology: the correspondence of both structure and function. Our results show that SII and PV do not only contain separate somatotopic organizations but also can be differentiated on the basis of their cytoarchitectonic features as previously described for nonhuman primates (Cusick et al. 1989; Krubitzer and Kaas 1990, 1993; Burton et al. 1995; Krubitzer, Clarey, et al. 1995; Qi et al. 2002; Wu and Kaas 2003).
Ruben et al. (2001), on the other hand, examined the somatotopic organization of the human SII region by stimulating the subjects' hands and feet. The authors demonstrated a lateral to medial arrangement of these 2 body parts, which is in line with primate data, the results by Disbrow et al. (2000) and our results. As acknowledged by the authors, the design of their study did not, however, allow to demonstrate the double somatotopy within the parietal operculum because hands and feet are both represented at the SII/PV border.
The Human Homologues of SII and PV
The hypothesis that OP 1 and OP 4 are the human homologues of monkey SII and PV was initially based on their location and topographic relationship to neighboring cortical areas (Eickhoff, Schleicher, et al. 2006). It received further support from a meta-analysis of functional imaging studies on hand activations in SII. This analysis showed that the most likely location for SII hand activations was the border between OP 1 and OP 4, similar to its position at the border between SII and PV in monkeys (Eickhoff, Amunts, et al. 2006). The present study confirms these results by showing significant activation following tactile hand stimulation at the same location, thereby again closely matching the location of the SII/PV hand region in monkeys (Fig. 9B). The head was represented more laterally at the border between OP 1 (SII) and OP 4 (PV). This also follows closely the predictions from nonhuman primates where the functional border between SII and PV is defined by a mirror reversal of the somatotopic map along the representations of the head (lateral), hands (intermediate), and feet (medial) (Cusick et al. 1989; Krubitzer and Kaas 1990, 1993; Burton et al. 1995; Krubitzer, Clarey, et al. 1995; Qi et al. 2002; Wu and Kaas 2003).
This mirror reversal along distal body parts implies that proximal body parts are represented to either side of this functional border, that is, separately in SII and PV (Fig. 9A). In the present experiment, we stimulated 2 more proximal locations, the trunk and the leg (which is a proximal body part because it is followed more distally by the foot). Both were represented twice on the parietal operculum: One of these 2 representations for either body part was located in the anterior part of OP 4 (PV) and the other in the posterior part of area OP 1 (SII). Representations of proximal body parts were thus found to either side of the functional boundary between SII and PV as expected from animal data (Fig. 9B). Moreover, the topographic relationship between trunk and leg activations is also identical to that observed in monkeys where the leg is represented medial to the trunk (Cusick et al. 1989; Krubitzer and Kaas 1990, 1993; Burton et al. 1995; Krubitzer, Clarey, et al. 1995; Qi et al. 2002; Wu and Kaas 2003). However, differences between the human data and the known somatotopic maps from nonhuman primates were also observed. These differences were mainly related to the representation of proximal body parts. In humans, the trunk representation appears to be more lateral, which is, toward the free surface of the parietal cortex, than in monkeys. In particular, the trunk representation was located lateral to the SII/PV hand representation which is not the case in other species. Also, in humans, the representation of the legs was further separated from the functional and anatomical border between SII (OP 1) and PV (OP 4) compared with nonhuman primates. Such variations of a common topological scheme, however, are well recognized in a wide variety of species (Krubitzer, Clarey, et al. 1995; Kaas and Collins 2003) and can thus also be expected when comparing human and nonhuman primate data. The observed differences can therefore be interpreted as an evolutionary modulation of the “general” primate SII map in the human species.
Thus, in summay, OP 1 (SII) and OP 4 (PV) both appear to contain a complete somatotopic map. These 2 maps are mirror images of each other with the reversal occurring at the location of the anatomical border between these areas. We therefore have found additional evidence that OP 1 represents the human homologue of macaque area SII, whereas OP 4 represents the human homologue of macaque PV.
Functional Differentiation between OP 1 (SII) and OP 4 (PV)
The interpretation of OP 1 and OP 4 as human SII and PV, respectively, also provides an explanation for the observation that area OP 1 (SII) was generally more responsive to tactile stimuli than OP 4 (PV). Studies in nonhuman primates have shown that the functional properties and connectivity patterns of SII (OP 1) and PV (OP 4) are similar albeit not identical. Area SII is a purely somatosensory area and strongly interconnected with the parietal somatosensory network. PV, on the other hand, is more involved in sensory-motor integration and has dense connections with the frontal motor and premotor cortices (Qi et al. 2002; Disbrow et al. 2003; Kaas and Collins 2003). The purely somatosensory task used in our study can therefore be expected to activate the human homologue of area SII (OP 1) more strongly than the PV homologue OP 4. As shown by the individual activation maps (Figs 3–6) and most pronounced by the overall probability of activation (Fig. 8), our results do once more follow closely the predictions from animal physiology.
Functional Evidence for a Human VS
To date, no functional description of a human VS homologue has been reported. However, OP 3 has been regarded as its most likely structural homologue based on its location in the depth of the Sylvian fissure medially to OP 1 (SII) and OP 4 (PV) (Eickhoff, Amunts, et al. 2006; Eickhoff, Schleicher, et al. 2006). The results of the current study support this hypothesis for the following reasons: First, our fMRI experiment provided evidence for a potential third somatotopic map in the human parietal operculum. The somatotopic arrangement of this area, however, could not be defined as conclusively as those in the more lateral opercular areas OP 1 (SII) and OP 4 (PV). Rather it was mainly based on an additional activation following leg stimulation that was clearly distinct from the leg-associated activations in OP 1 (SII) and OP 4 (PV). For the other body parts, however, a clear somatotopic order could not be elucidated in OP3. The second evidence for the hypothesis that OP 3 corresponds to human VS was provided by the mean probabilities for activation within OP 1–4. OP 1 (SII), OP 4 (PV), and OP 3 (VS) all showed significantly higher probabilities for activation following tactile stimulation than OP 2. These results are well in line with previous data showing that OP 2 constitutes the human PIVC and thus is not part of the SII region (Eickhoff, Weiss, et al. 2006). The likelihood for somatosensory activation within OP 3 (VS), on the other hand, was not significantly smaller than within OP 4 (i.e., PV). This suggests that OP 3 is most likely also part of the SII. Considering this observation, its topography and the—limited—evidence for a somatotopic organization of this area, we suggest that OP 3 might represent the human homologue of VS in nonhuman primates. Although our data thus provide the first evidence for the existence of a human VS, it has to be noted that, in spite of a considerable amount of studies on nunhuman primates (reviewed by Kaas and Collins 2003), the functional relevance of this region is still largely elusive and more work is needed to fully understand its role in the somatosensory network.
The data of the present study imply that cytoarchitectonic areas OP1, OP 4, and OP 3 in the human parietal operculum are equivalent to macaque areas SII, PV, and VS, respectively. This claim is based on the activations elicited by stimulation of different body parts, their somatotopic organization, and importantly their relation to cytoarchitectonically defined areas OP1 (SII), OP4 (PV), and OP3 (VS). These multiple somatotopic maps strongly suggest that several functionally distinct areas exist in human parietal operculum as has been described in the macaque. The functional relevance of the different subregions of the human SII region with respect to, for example, pain perception, tactile discrimination, sensory-motor integration, and tactile attention needs to be investigated in more detail in future studies. By matching probabilistic cytoarchitectonic maps with functional imaging data, such analyses is likely to help establishing a correlation between different stimulation paradigms and the individual cortical areas in the human parietal operculum and will thereby advance our understanding of the integrated functional and cytoarchitectonic organization of the human SII region.
This Human Brain Project/Neuroinformatics research was funded jointly by the National Institute of Mental Health, of Neurological Disorders and Stroke, and of Drug Abuse and the National Cancer Centre. We are grateful to our colleagues from the MR, Architectonic Brain Mapping, and Cognitive Neurology groups for valuable support and advice. GRF is supported by the Deutsche Forschungsgemeinschaft (DFG—KFO 112, TP 1). Conflict of Interest: None declared.