TY - JOUR
AU - Raichle, M.E.
AB - Abstract We used positron emission tomography to study cortical regions mediating tactile attention. Cues selectively directed subjects to attend to the roughness or duration of contact with embossed gratings that rubbed against a single fingertip with controlled speed and force. The task required discriminating between paired gratings that differed in tactile features of roughness and/or length. For different blocks of trials, cues directed attention to one tactile feature or indicated a divided attention strategy to a change in either feature. All attention conditions unambiguously activated several somatosensory foci in the parietal cortex, including somatotopically appropriate portions of the primary somatosensory cortex in the postcentral gyrus (S1) and the secondary somatosensory region (S2) within parietal opercular regions. There was no evidence for separate processing foci for selective and divided attention strategies, or for selectively attending to roughness versus stimulus duration. We observed a greater magnitude blood flow change in S2 versus S1 during attention tasks, which suggests that S2 might actually influence S1 activity. Despite these differences, modulation of S1 and S2 supports concepts of early selection in tactile attention. There were also examples of non-sensory foci in frontal cortex, anterior cingulate gyrus and bilateral superior parietal regions at the fundus of the postcentral sulcus. Posterior parietal regions observed in this study did not overlap foci seen in studies of visual attention. Thus, the posterior parietal region may be subdivided into modality-specific subregions, each of which processes information needed to attend to a specific modality. These non-sensory areas may constitute a network that provides a source of modulating influences on the earlier stage, sensory areas. Introduction The current experiments investigated the human cortical regions that mediate tactile attention. At issue was identification of a tactile attention network and its resemblance to the organization previously noted for the human visual attention system (Posner and Petersen, 1990; Corbetta et al., 1991, 1993; Corbetta, 1998). The comparison between tactile and visual attention networks was motived by the notion that both systems have limited capacities; both must, therefore, select behaviorally relevant stimuli for processing. One hypothesis, directly derived from studies in the visual system, is that the network is organized hierarchically with the earliest stages in primary somatosensory cortex (S1). Human imaging studies already showed that tactile attention tasks modulate activity in S1 regions along the post-central gyrus (Roland, 1981; Roland, 1982; Meyer et al., 1991; Drevets et al., 1995). Based on results obtained in studies with monkeys (Poranen and Hyvärinen, 1982; Hsiao et al., 1993b; Burton et al., 1997b), tactile attention is also likely to involve presumed analogous higher-order somatosensory foci located along the parietal operculum in humans (e.g. the second somatosensory area, S2) (Burton et al., 1993, 1997a; Coghill et al., 1994; Ledberg et al., 1995). In the present study, we examined whether tactile attention processes engage multiple brain foci, each possibly dedicated to a different aspect of the attention process (Posner et al., 1988). A related issue was potential identification of separate foci within somatosensory areas dedicated to processing different tactile stimulus attributes. In imaging studies of human visual attention, Corbetta and colleagues showed activations of different higher-order visual areas when subjects attended selectively to shape, color or motion speed attributes of a visual display (Corbetta et al., 1991). Although some tactile stimuli also can be parsed into different attributes (e.g. texture, shape, object length), currently only one brain imaging study has reported that separate activation foci might process signals related to different tactile stimulus attributes (Roland et al., 1998). In this study, discrimination tasks involving a two-alternative, forced-choice discrimination of roughness differences between objects activated more anterior parts of the postcentral gyrus than tasks requiring discrimination of object shape or length attributes. Previously, however, we reported no significant differences in the cortical areas activated when embossed grating stimuli preferentially stimulated cutaneous compared to proprioceptive receptors in individual fingertips (Burton et al., 1997b). Further, in this study we examined whether tactile attention engages non-somatosensory cortical areas. Studies of the network associated with visual attention found many cortical regions engaged that occupied nonvisual as well as visual areas (Mesulam, 1990; Corbetta et al., 1991, 1993; Haxby et al., 1994; Posner and Dehaene, 1994; Vandenberghe et al., 1996; Frith and Dolan, 1997; Büchel et al., 1998). Thus, visual attention is a distributed process (Posner and Dehaene, 1994). In question is whether the same or different nonsensory cortical areas are active during a tactile attention task. These may act as cross-modal attentional foci and potential organizers of top-down mechanisms (Posner and Petersen, 1990; Corbetta et al., 1991, 1993; Frith and Dolan, 1997; Nobre et al., 1997; Büchel et al., 1998). One likely common region is the right posterior parietal cortex because lesions here are associated with neglect of visual and somatosensory stimuli located in extra-personal space contralateral to the lesion (Critchley, 1949; Mesulam, 1981; Posner et al., 1984; Moscovitch and Behrmann, 1994). Another issue to be addressed is whether imaging data support the notion that foci activated by passive presentation of a stimulus are more highly activated when attention is directed to the same stimuli. Numerous studies have found evidence of increases in relative cerebral blood flow (rCBF) in appropriate sensory cortical areas during attention tasks engaging specific sensory channels (Corbetta et al., 1991, 1993; Woldorff et al., 1993; Woodruff et al., 1996; Büchel et al., 1998). Blood flow in S1 was 13% higher when subjects attended to vibrotactile stimuli than when they received the same stimuli but were engaged in a distraction task (Meyer et al., 1991). These studies suggest that attention may involve early selective modulation within the appropriate sensory cortical areas. Here we attempted to investigate three aspects of tactile attention: (i) location of foci showing increases in blood flow in sensory and nonsensory cortical areas; (ii) identification of separate activation foci for different attended tactile attributes; and (iii) evidence that some regions showing increased neural activity during control tasks involving passive tactile stimulation were further activated during the attention task (enhancement). The tactile attributes of surface roughness and stimulus duration were manipulated in the present study. A comparison of performance on (valid) trials, when cues predicted which attribute would change, to trials with a neutral cue allowed us to determine the benefit of devoting all resources to one feature and demonstrated that subjects' attention was manipulated by the different attention conditions. The rationale behind this testing was that on trials with valid cues subjects' ability to discriminate between small changes in the selected feature would be improved by selectively directing attention to one feature in a multidimensioned stimulus (Posner, 1986). On trials in which cues did not direct attention to either feature (neutral), resources had to be divided, and thus the probability of correctly discriminating differences between gratings based on each feature was expected to be lower. Importantly, this experimental design ensured balanced difficulty, arousal and vigilance across all trials since identical stimuli are presented across conditions. We report here results from psychophysical experiments aimed at assessing task performance and positron emission tomography (PET) experiments directed toward identifying brain regions that mediate tactile attention. Materials and Methods Subjects We conducted psychophysical experiments on 20 paid volunteers (11 males and 9 females) between the ages of 18 and 30 (mean 22.4). Ten subjects (6 males and 4 females; between the ages 18 and 27; mean 22.5) met performance criteria (as described below) and returned for scan sessions. All reported strong right-handed preference as measured by the Edinburgh Handedness Inventory (Raczkowski et al., 1974) and had no neurological or physical disabilities. Informed consent, obtained before participation in each part of the study, followed guidelines of the Human Studies Committee and the Radioactive Drug Research Committee of Washington University. Stimulus Attributes Tactile stimulation consisted of rubbing the right index finger with embossed gratings of alternating ridges and grooves, which were made by photo-etching polyamide plastic sheets (Process Color Plate, Chicago, IL). Ridge widths (0.25 mm) and groove depths (~0.5 mm) were constant, but groove width varied (see below). All gratings were 2 cm wide. Gratings were presented in discontinuous pairs. A smooth area of 4.5 cm separated one pair of surfaces from the next; a shorter smooth area of 1 cm separated the first pair member from the second (Fig. 1). The surfaces in each pair differed in length (perceived as duration of stimulation) and/or groove width (perceived as surface roughness). These attributes were chosen because objects differing in both could be presented passively and identically without altering the stimulus set, and small changes in the dimensions of one feature did not alter the discrimination of the other feature. Surface lengths were between 2.64 and 4.64 cm. Groove widths were between 1132 and 1925 μm. These values were selected on the basis of psychometric testing in 10 subjects (separate from the current subject group) that established the magnitude differences needed to obtain at least 80% correct discrimination of differences between the first and second grating of a pair for each dimension (duration or roughness) when subjects selectively attended to just one feature. Apparatus and Stimulus Application For each of the behavioral tasks (see below), 32 different grating pairs were randomly distributed into four continuous strips (tracks), each consisting of eight pairs. These were wrapped around the perimeter of a cylinder (~10 cm height and 32 cm diameter). Three cylinders were constructed for a total of 96 grating pairs. Eight sequential pairs comprising a track were presented in one rotation of the cylinder, and through manual shifts each of the four tracks of grating pairs could be aligned under the fingertip. The three cylinders were exchanged during a testing session depending on the task. On the ‘passive stimulation control cylinder' (explained below), the two gratings within a pair were identical in both features. The range of roughnesses and lengths for the different pairs matched those used during the attention tasks (described below). On the ‘selective cylinder', the values of both attributes changed in 75% of the pairs. On 12.5% of the pairs only roughness changed, and on the remaining 12.5% of the pairs, only length changed. On the ‘divided cylinder', the values of both attributes changed on 50% of the pairs. On 25% of the pairs only length changed, and on 25% of the pairs, only roughness changed. The grating velocity was 30 mm/s. A constant force load of ∼100 g was maintained by balancing the cylinder against a torsion counterbalance with an adjustable fulcrum. The finger was positioned so that the ridges aligned perpendicular to the long axis of the finger and moved down the finger from proximal to distal end. A restraint, made of thermoplastic and foam, supported the subject's right forearm and hand. The right index finger extended through an opening in the restraint and was additionally supported in an individually fashioned mold made from Rolyan 50/50 Elastomer Putty (Smith and Nephew, WI). This finger-cast passively held the finger in a slightly flexed position, which aligned the distal fingertip against the grating surfaces. We instructed subjects to relax their arm, hand and fingers throughout the trials. Behavioral Tasks Subjects performed both control and attention tasks before and during PET scans. Feature Attention Conditions Using a two-alternative, forced-choice paradigm, each trial consisted of stimulation with a pair of gratings. Before each set of trials, verbal instructions cued the subject in one of three ways (Fig. 1). During the ‘selective attention' task, instructions cued the subjects to detect and respond to changes in a single feature (either duration of contact or surface roughness) and to ignore changes in the other feature. The subject was asked to determine which surface felt rougher or was in contact with the finger for a longer period of time. Following each trial, the subject responded verbally with ‘one' if the first (‘reference') surface in the pair was longer/rougher or ‘two' if the second (‘test') surface in the pair was longer/rougher (Fig. 1). Responses were tape-recorded using an audio cassette. In a selective attention trial, both features would change (a validly cued trial), the selective feature alone would change (a validly cued trial), or the unspecified feature alone would change (an invalidly cued trial). Subjects were not informed of the invalid trials, and therefore responses from these trials were not considered further. In the ‘divided attention' task, subjects were instructed to detect and respond to changes in either feature (neutrally cued trial). Thus, subjects had to divide their attention between the two features in order to determine which of the two surfaces in a pair was in contact longer or was rougher. The same values of duration and groove width were used in the selective and divided attention tasks. Control Conditions There were three control conditions during PET scans, the last of which was practiced during the psychophysical testing sessions. The fingertip was restrained in a finger mold over the surfaces during all controls. In the least demanding control condition, named ‘eyes-closed-rest', subjects were verbally instructed to close their eyes and not pay any attention to their fingertips. The grating pairs did not touch the subject's finger although the wheel rotated during the scan. In the next control condition, named ‘passive stimulation', subjects also kept their eyes closed and were told to ignore their fingers while a series of paired surfaces whose members were identical in both features were presented. Subjects learned to perform the third control condition, named ‘passive stimulation/counting', through several steps. First, subjects were told to ignore stimulation from rubbing paired gratings against the fingertip. Next, and without tactile stimulation, we instructed subjects to say ‘one' or ‘two' in random order at a rate of once per 4 s. Initially a timer beeped every 4 s to assist learning the desired interval; then the timer was turned on and off several times until the subjects' timing approximated the desired interval. Finally, the subject's finger was stimulated again by the control surfaces while the subject made random verbal responses. Subjects were informed that the order in which they verbalized the numbers was unimportant, that they should not base their choices on the surfaces (since the members of each pair were identical), and that they needed to respond at 4 s intervals, even though their responses would not coordinate with the onset or offset of surface pairs. Subjects practiced this task through multiple presentations until they made randomly chosen responses in a manner unaffected by the concurrent stimulation. Psychophysical Testing Testing extended over 2 days. On day 1 subjects first practiced the passive stimulation/counting control task and then did the attention tasks. On day 2 this order was reversed. Subjects received specific instructions to attend selectively to one of the features or to divide their attention between features before each of the attention tasks. We then presented eight trials (one rotation of the appropriate cylinder) during which the experimenter provided correct answers. The cylinder was rotated to a different starting position and subjects next practiced with their own answers for another set of eight trials. Feedback on performance was provided at the end of these trials. Each day, the first and third attention treatment blocks contained selective attention trials in a random sequence between duration and roughness. The middle block contained only divided attention trials (neutrally cued trials). In counterbalanced order, the grating pairs for each attention task were rotated in both the forward and reverse drum directions across the two test days. This ensured that subjects encountered all surfaces in the first or second positions, thereby reducing the possibility of memorizing the pair relations and being able to answer based solely on the first member of a pair. On each day subjects performed 80 trials (10 complete cylinder rotations) for each of the selective attention tasks in one rotation direction. Thus, over the two days of testing, we collected responses to 160 trials for selectively attending to roughness and 160 trials for selectively attending to duration presented in each rotation direction. We presented 128 trials (16 rotations) for the divided attention task per test day. Following testing, subjects were questioned about their attention strategies and insights into the stimulus set. Scan Procedures Task-specific instructions were given before each of 10 scans. Task performance was started several minutes prior to the scans so that during the attention and ‘passive stimulation/counting' control tasks 16–24 trials were completed before the scan, leaving another 8–12 trials during the scan. Task-scan order was as follows: (i) scan 1 was always an ‘eyes-closed-rest' control; (ii) scans 3, 5, 8 and 10 were appropriately counterbalanced to include selective attention to roughness or stimulus duration; (iii) scans 4 and 9 were the divided attention task; (iv) scans 2 and 7 were the ‘passive stimulation /counting' task; and (v) scan 6 was the ‘passive stimulation' control. PET Scanning Techniques We used a Siemens 953B scanner in the three-dimensional (septa retracted) mode (31 transaxial 3.375 mm slices, reconstructed pixel size of 2.086 mm2). We used previously described PET scanning activation methodology developed at Washington University (Fox et al., 1984, 1988; Mintun et al., 1989; Petersen et al., 1989). A thermoplastic mask stabilized the subject's head. Black curtains surrounded the head and blocked all light. 15O-Labeled water (10 ml saline with ~15 mCi of radioactivity) was administered via an i.v. bolus injection to the left forearm. Forty second scans were acquired. Usually, we conducted 10 scans, each separated by 10 min to allow for almost complete decay of 15O between scans (five half-lives). We first performed an ‘eyes-closed-rest' task, ‘sham scan' during which only physiological saline was injected to reduce anxiety associated with injection of the radioactive tracer or other scan procedures (Coghill et al., 1994). PET difference images were computed by subtraction across scan pairs and averaged over subjects (Fox et al., 1988). This was accomplished with a single resampling (to 2 mm cubic voxels) using previously computed transformations and intensity scaling that (i) corrected for interscan head movement (Snyder, 1996) and global activity variations, and (ii) transformed the images to a standard atlas space (Talairach and Tournoux, 1988). Our atlas transformation strategy is similar to that described by Collins and colleagues (Collins et al., 1994). Our atlas representative target PET image consisted of averaged data from 12 normal, young adult subjects (not included in the present study), corrected, as above, for interscan head movement, and mutually co-registered (12 parameter affine transforms) to a selected individual. We similarly aligned structural MP-RAGE (Mugler and Brookeman, 1990) images obtained in the same 12 subjects. The averaged MP-RAGE image was then made to conform to the Talairach atlas using the SN spatial normalization procedure (Lancaster et al., 1995). The averaged PET image then was aligned to the MP-RAGE image to produce the atlas representative, target PET image. This PET atlas representative image then was scaled to a whole-brain mean intensity of 1000 using boundaries defined on the co-registered MP-RAGE. Although the above sequence was described as if successive transformations were applied to cumulatively resampled data, it should be understood that the final product was produced by algebraic composition of transforms followed by one resampling (to 2 mm cubic voxels) and averaging of the originally reconstructed PET data. To identify changes in blood flow, images from each subject were grouped into activation minus control pairs. The resulting difference images show foci of task-related changes in rCBF (Fox et al., 1984; Mintun et al., 1989). Because the subject performed different behavioral tasks in the two scans that made up each pair, the subtracted image represents the blood flow change induced by the different processing demands of the two tasks. PET Statistical Methods We used a replication approach for data analyses to cross-validate identified regions of change in rCBF and to investigate their reliability (Squire et al., 1992; Burton et al., 1993; Hunton et al., 1996; Shulman et al., 1997). This approach, which is based on testing the reproducibility of rCBF changes across data sets, ensures that regions identified as significant are unlikely to result from irrelevant factors (like movement or vessel artifacts) or image noise (Corbetta et al., 1991; Buckner et al., 1995a; Buckner et al., 1995b; Buckner et al., 1996). This strict analysis method was implemented to reduce type I errors in which false positives are accepted as significant. Difference images (scans from each attention task minus scans from ‘passive stimulation/counting' control) were summed together in one composite image. The subtraction pairs comprising this composite were then assigned to one of two data sets, the first of which (hypothesis generating) was used to define regions of interest. The second data set (hypothesis testing) was used for statistical testing of candidate focal responses identified in the first set. Only those subtraction pairs with repeated measures from the same subject were used in this analysis so that subjects contributed equally to both data sets. The subtraction pairs in each data set were proportionally weighted to equate contributions from each attention condition. The hypothesis-generating, averaged-difference image was passed through a three-dimensional, fifth-order, low-pass Butterworth filter with a cut-off frequency of 0.7 cycles/cm. This processing produced an image with a resolution of ~11 mm full width at half maximum in all dimensions. We used a computer algorithm to identify all rCBF maxima in the averaged difference images from the hypothesis-generating data sets (Fox et al., 1988; Mintun et al., 1989). This search identified the location of foci with peak PET counts in a single 8 mm3 voxel at least 10% higher than the counts in neighboring pixels and at least 11 mm away from adjacent peaks. The resulting list identified the top maxima and showed stereo-taxic locations of potentially significant increases in rCBF. We selectively omitted from further analysis loci that were anatomically known to be within blood vessels, white matter, the ventricles or outside the brain. Next, to investigate whether the maxima identified in the hypothesis-generating data set were replicated in the hypothesis-testing data set, total PET counts from the individual (not averaged) difference images from the hypothesis-testing data set were determined. PET counts for each of the maxima identified in the hypothesis-generating data sets were then assessed with a one-sample Student t-test (upper for maxima) against the null hypothesis of 0 PET count differences in the hypothesis-testing data set. Mean PET counts within the identified regions whose t-values exceeded null hypothesis expectations with a P < 0.05 were considered leniently significant replicas of the focal peaks identified from searching the hypothesis-generating half of the data. Those regions whose t-values exceeded the null hypothesis with a Bonferroni correction for the number of t-tests performed were considered strictly significant replicas of the focal peaks. Those foci that reached significance could be attributed to at least one of the three attention tasks. We report both leniently and strictly significant foci because of the conservative nature of the replication analysis, which might tend toward type II errors. However, the reliability of reported foci with strictly significant peaks is extremely high given the low probability of a type I error in these tests. The spatial resolution of the analysis was also close to that of our scanner, and therefore showed potential examples of multiple, yet proximate regions. In comparisons to voxel-by-voxel approaches (Friston et al., 1995) to assess the probability of observed PET difference counts at selected foci, cross-validation, two-stage procedures are extremely robust and have comparable statistical power (Holmes, 1994). Spatial smoothing of the data is also unnecessary in cross-validation analyses as there are no assumptions about the underlying shape of the PET count distributions (i.e. Gaussian random fields). Hence, results can be viewed at resolutions approximating that of the scanner. The methods used to identify interesting regions in the hypothesis-generating data set also make little difference (Holmes, 1994) although our searches for peaks in 8 mm3 spheres, with minimum separations between peaks of 11 mm, were consistent with the volumes studied in simulations. The major disadvantage of two-stage testing is reduced sensitivity from halving the degrees of freedom due to dividing the data. However, we compensated by using a replication design where each subject was studied twice for each attention and passive/counting task condition. Across all of the attention tasks, data were available from 60 scans (2 repetitions × 3 attention tasks × 10 subjects), which, when divided, provided samples of 30 for hypothesis generation and testing. This was ~6 times the sample shown in simulation studies to yield nearly comparable statistical power to voxel-based methods (Holmes, 1994). Thus, the increased sample compensated for any possible loss in sensitivity while maintaining the better spatial resolution of an assumption-free approach to the distribution of PET counts To determine the most accurate stereotaxic location for each of the maxima, the locations identified in the hypothesis-generating data set were visually compared to peak loci on a summed difference image composed of both hypothesis-generating and hypothesis-testing subtraction pairs (again weighed to balance conditions). This was done because the combined data set provides a better estimate of an activation's true location than either of the smaller data subsets used in replication analysis. These coordinates, which were all within millimeters of the coordinates used in the cross-validation analysis, are presented in the tables. Within the text, coordinates are referred to in the x,y,z format (positive x = right; positive y = anterior; positive z = superior). Results Experiment 1: Psychophysics We first determined (by analyzing data from the 20 subjects psychophysically tested with a repeated-measures analysis of variance) whether our methods ensured that subjects were both sufficiently attentive and effective in ignoring uncued features on selected attention trials or dividing their attention between multiple features on divided attention trials. This analysis involved two measures. First, we compared performance on selective and divided attention trials in which only one feature changed. Subjects who performed correctly on more trials with valid than neutral cues probably divided their attention between the two features on divided trials, and thus lowered their performance. Second, we compared performance on validly cued trials in which a possible response to the uncued attribute changed in a negatively correlated, uncorrelated or positively correlated fashion relative to the cued attribute. Subjects who showed little difference in the percentage of correct trials in relation to changes in the uncued feature probably more effectively ignored the uncued feature. Ten subjects, who met these performance criteria in the initial psychophysical tests, returned for scan sessions. The performance measure for each attention task was the percentage of correct responses. Data from those subjects chosen to return for scan sessions are shown in Figure 2. The main finding was that percentage correct was significantly higher (Bonferroni, P < 0.05) on validly cued trials (75%) than neutrally cued trials (68%). There were no significant effects of dimension or the specific values of roughness or duration on percentage correct. These results support the conclusion that subjects performed the task as instructed. Deficits in the neutral cue condition suggest a divided attention strategy. Experiment 2: Positron Emission Tomography Psychophysics Behavioral data obtained during the scan session was collected in the same way as during psychophysical testing. A statistical analysis of the results was not possible, however, because the 40 s scan interval did not allow for sufficient data collection. Qualitatively, the results were similar to those obtained during initial testing. The following results discuss foci of change in rCBF that were tested for cross-validation between randomly separated halves of the data. Figure 3 shows, on matching columns of sagittal sections, the distribution of PET difference counts for the passive stimulation/counting versus eyes-closed-rest (leftmost column) and the summed image across all attention tasks versus passive stimulation/counting (middle column). The foci are correspondingly labeled in Tables 1 and 2. [In addition to descriptions of gross anatomical features, like major gyri, the tables and text note Brodmann area (BA) designations as labeled in the Talairach atlas (Talairach and Tournoux, 1988). These BA listings provide a convenient summary labeling convention that, in addition to the Talairach atlas coordinates, usefully points to results from other studies.] (1) ‘Passive Stimulation/Counting' Control Task Minus ‘Eyes-closed-rest'. Baseline activity present when the subjects remained still with eyes closed was subtracted from activity evoked during conjoint passive tactile stimulation of the fingertip of the right index finger and randomly chosen verbal responses (either ‘one' or ‘two'). This subtraction image reveals any areas involved in planning and executing the motor (verbal) responses and areas activated by simple passive stimulation. Significant increases in blood flow were identified in five foci (Table 1), all of which satisfied the strictest, Bonferroni-corrected criteria. Four of the foci were found in the left hemisphere, contralateral to the stimulated finger. There were three foci of rCBF change in the vicinity of the precentral gyrus. One occupied the crown of the precentral gyrus ipsilateral to the stimulated fingertip (labeled D in Table 1). The other foci were also centered in precentral gyral cortex but contralateral to stimulation (A and E in Table 1 and Fig. 3). The foci in opposite hemispheres involved partially analogous regions although the activated sites on the contralateral side were more extensive and included portions of the postcentral gyrus, and were therefore labeled as peri-Rolandic foci. Two additional foci were found in frontal cortex. One was located on the medial wall in the paracentral lobule (C in Table 1 and Fig. 3) contralateral to the stimulated finger. Significant rCBF changes were also found in the putamen that was contralateral to stimulation (Putamen, B in Table 1). Surprisingly, no isolated foci of significant increases in rCBF were found in either hemisphere within somatosensory cortex previously identified in the postcentral gyrus and parietal operculum (Burton et al., 1997a). (2) Attention Tasks Minus ‘Passive Stimulation/Counting' Control: Composite Analysis. Activity present during the ‘passive stimulation/counting' control task was subtracted from activity present when the subject performed the attention tasks (selectively cued to roughness or to duration, or no cued attribute). In these subtraction images, any common activations due to verbal responses, passive tactile stimulation and extraneous effects of being scanned were removed. The remaining positive foci reflected activity from several possible processes. First, the foci may be part of an activated attention network that generates signals which modulate sensory processing in multiple modalities. Second, the foci may be components of the somatosensory system that are modulated by signals generated elsewhere, possibly by a nonsensory attention network. Third, the foci may be activated as a result of the increased arousal and greater memory demands of the attention tasks over the control task. An ANOVA, performed for each significant focus from all subtraction pairs (hypothesis-generating and hypothesis-testing data sets), assessed whether the change in rCBF could be attributed to one attention condition or a subset of attention conditions rather than equally to all three attention conditions. For example, the selective tasks might have caused an increase in rCBF that replicated across data sets despite finding no change in rCBF for that region in the divided attention task. The dependent variable was change in rCBF and the main independent variables were subject, condition (selected roughness, selected duration, divided) and subject-by-condition. These ANOVA analyses yielded no significant effects, demonstrating that for each focus under investigation, no attention task accounted for significantly more of the rCBF changes. Thus, since the three attention conditions activated a common set of foci, further discussion mainly considers data collapsed across attention conditions; this has the added advantage of reducing background noise and increasing statistical power. Increases in blood flow were identified in 11 foci; with the exception of a focus in the thalamus (labeled 1 in Table 2 and Figs 3 and 4), these predominantly were in parietal cortex (Table 2). Three foci (labeled 1, 2 and 11 in Table 2 and Figs 3 and 4) satisfied the strictest, Bonferroni-corrected criteria and five (labeled 3, 4, 5, 8 and 9) passed the lenient criteria of 0.05. Three foci did not cross-validate (labeled 6, 7 and 10), but were prominent on the images, and were therefore included in the following description. Eight of the foci were found in the left hemisphere, contralateral to the stimulated finger. Many of the foci in parietal cortex occupied regions previously identified as parts of the somatosensory system. Blood Flow Increases in Parietal Cortex A pronounced focus was seen along the contralateral parietal operculum, close to the posterior third of the insula (labeled 2 on Figs 3 and 4). A smaller, less conspicuous locus of increased rCBF occupied a comparable position in the right hemisphere (labeled 7 in Fig. 4). The ipsilateral focus did not cross-validate (Table 2). The same contralateral site was identified in all three attention tasks but was not found during the control tasks (Fig. 3). Two activation foci lay along the postcentral gyrus contralateral to the stimulated fingertip. One occupied the crown of the postcentral gyrus, extending anteriorly toward the central sulcus and posteriorly to the postcentral sulcus (labeled 6 in Table 2 and Figs 3 and 4). Although the same focus was observed across all attention conditions, it did not cross-validate between the two halves of the data (Table 2). However, a significant single t-test result was obtained when all of the subject data (~50 subtraction pairs from 10 subjects) was included and with the blood flow measurements centered on the coordinates of focus 6 (t = 5.35, df = 9, P = 0.0005). The same region was not significantly activated when contrasting blood flow during the passive stimulation/counting control task to an eyes-closed-rest condition (Fig. 3, x = –49 to –53). Another postcentral gyrus region was also located on the crown of the postcentral gyrus (labeled 3 in Table 2 and Figs 3 and 4). It was lateral and posterior to focus 6. This region cross-validated using lenient criteria for all three attention tasks (Table 2). It was not seen during the control task (Fig. 3). Focus 3 was contiguous to site 2 along the parietal operculum although, as shown in Figure 4 (y = –21 and –25), the two peaks were separable. Figure 5 shows blood flow measurements obtained from comparable 11 mm diameter spheres, centered on foci 6 and 2, which respectively included S1 and S2. Predictably, across all tested conditions the highest blood flows appeared during the attention tasks, which principally accounted for significant F ratios for the task variable for each region [S1: F(3) = 4.32, P = 0.017; S2: F(3) = 10.09, P = 0.0003]. However, post hoc comparisons within regions found differences across the tasks. Thus, there was significantly greater blood flow through S2 in the attention tasks than all other tasks; blood flow measures were equal during the other three tasks. In S1 significantly greater blood flow similarly occurred during the attention tasks. In addition, significantly lower blood flow occurred during the passive stimulation/counting task; blood flow measures were equal during the eyes-closed-rest and passive-stimulation tasks. Blood flows during the attention tasks were significantly greater over S2 than S1. Two foci were identified within the depths of the postcentral sulcus. On the contralateral side the focus labeled 4 (Figs 3 and 4) had a peak that occupied the medial aspect of the sulcus (Fig. 4, y = –39). Figure 4 (y = –35) shows that, in a more anterior location, blood flow changes associated with focus 4 are distinct from the posterior extension of focus 6. A matching postcentral sulcal region was noted on the ipsilateral side (labeled 5 in Fig. 4). Both regions cross-validated using lenient criteria (Table 2) and were activated by all three attention tasks. The same regions were not activated during the control tasks (Fig. 3). Blood Flow Increases in Frontal Cortex Three foci were activated in the frontal cortex during the attention tasks. One focus occupied contralateral cortex within the precentral sulcus (labeled 9 in Table 2 and Fig. 3). This focus, which cross-validated using lenient criteria (Table 2), was present in all three attention conditions. The region covered by this activation was >1 cm medial to focus A found on the precentral gyrus in the control task (Fig. 3). Thus, these foci do not overlap. Another focus appeared on the anterior third of the insula on the contralateral side (labeled 10 in Table 2 and Fig. 3). Possibly due to intersubject variability, it did not satisfy even the lenient criteria although it was consistently observed, especially during the roughness discrimination task. There was no evidence of activation on the insula in the control task. The attention tasks activated a third frontal cortex focus that occupied a portion of the middle frontal gyrus, ipsilateral to the stimulated finger (labeled 8 in Table 2). This rCBF satisfied lenient cross-validation criteria and was present in all attention tasks. Blood Flow Increases in the Cingulate Cortex One focus was found near the superior margin of the cingulate gyrus in all three attention conditions, contralateral to stimulation (labeled 11 in Table 2 and Figs 3 and 4). There was no comparable site activated during the control scans. (3) Attention Tasks Minus ‘Passive Stimulation/Counting' Control: Single Condition Analysis. In separate cross-validation analyses for each attention condition, we identified no regions that were only significantly activated by one of the conditions. (4) Attention Task Comparisons. Potential differences in blood flow changes for the different attention tasks were analyzed with a cross-validation procedure that examined PET difference images created by contrasting scans obtained during all selective attention trials with scans when subjects divided their attention between both tactile attributes. Only bilateral cortical foci in the orbito-frontal region showed increased blood flow during the divided attention task. Only the left focus passed our stricter criterion for significance. Both foci occupied coordinates that placed them within the region labeled Brodmann area 10 in the Talairach atlas (left BA 10: –29, 61, –4; t = 3.3, df = 7; P = 0.013; right BA 10: 17, 53, 0; t = 2.26, df = 7; P = 0.058). Discussion Psychophysical Observations and a Tactile Attention Network The psychophysical results showed the advantages of a selective attention strategy (Posner, 1986; Allport, 1989; Posner and Driver, 1992). As in prior studies with visual stimuli in humans (Corbetta et al., 1991), the percentage correct was significantly higher when attention was selectively focused on a validly cued feature of a multidimensional tactile stimulus than when attention was divided between two features. Stimulus values and subjects' verbal responses were identical in the selected and divided conditions, ensuring that perceptual discrimination difficulty remained constant across trial types. However, there was a greater memory demand during the divided attention trials as both stimulus attributes required attention. The decreases in percentage correct on neutrally cued trials demonstrate the benefit of focusing limited resources on a single cued feature. It is unlikely that the selected attention advantage resulted from a reflexive process that drew attention to one of the stimulus features because performance was not significantly different on valid trials cued to texture or duration. A cortical network of sensory and nonsensory regions has been proposed for visual attention processes (Posner and Petersen, 1990; Posner and Driver, 1992; Posner and Dehaene, 1994; Frith and Dolan, 1997; Büchel et al., 1998; Coull, 1998). In a comparable arrangement the imaging data were expected to reveal components of the tactile attention network as the same tactile attention task was performed by the same subjects during scans. First, we discuss possible sites of attentional modulation within somatosensory foci. Next, we consider blood flow increases in several nonsensory cortical foci. Several of these were near or overlapped foci previously identified as nonsensory components of the visual attention network. Blood Flow Increases Within the Somatosensory System The current study did not find significant differences within the somatosensory network between selected and divided attention. Similarly, Corbetta and colleagues (Corbetta et al., 1991) found comparable activation of visual cortical foci for their selected and divided attention tasks. Thus, both types of attention tasks activate overlapping sensory cortex within their respective visual or tactile domains. In contrast, selected and divided attention tasks activated different foci outside the sensory regions (see below). Modulation of rCBF in the Postcentral Gyrus The activated focus across the contralateral postcentral gyral regions resembled the pattern reported in previous PET (Burton et al., 1997a; Roland et al., 1997) and functional magnetic resonance imaging (Lin et al., 1996) studies that involved tactile stimuli evoking a roughness perception. The most anterior part of this area, which extends into the central sulcus, possibly corresponds to the finger representation of area 3b (Woolsey et al., 1979; Lin et al., 1996; Burton et al., 1997a; Roland et al., 1997; White et al., 1997). As noted previously (Burton et al., 1997a), the more posterior aspects of this focus, which covers the postcentral gyrus and the anterior bank of the postcentral sulcus, suggest that in humans, like nonhuman primates (Pons et al., 1985), the finger representations of Brodmann areas 1 and 2 were also engaged. However, separable foci were not observed and Brodmann area distinctions based on these PET imaging data were not possible. Therefore, the following discusses the collective activation across the postcentral gyrus as S1. We observed significantly smaller changes in CBF in this region of the postcentral gyrus during all of the control tasks (eyes-closed-rest, passive tactile stimulation and passive stimulation/counting). In addition, the lowest blood flows over S1 occurred during the passive stimulation/counting task. There are several possible interpretations for these contrasting findings in S1. An unlikely technical explanation for observing lower increases in rCBF over S1 in the passive stimulation/counting control trials is that a dominant region of increased rCBF in the precentral gyrus, which probably reflected mouth and face movements during verbalizations, overwhelmed the resolution of our analysis routines. This implies that blood flow increases within the postcentral gyrus were subsumed under the motor-activated changes represented in the precentral gyrus. However, we observed two distinct foci, one on the precentral and one postcentral, on difference images created from contrasting the attention tasks with eyes-closed-rest control condition. These images showed no quenching of postcentral activation sites by blood flow changes on the precentral gyrus. In addition, Figure 3 shows that the precentral focus in the control task was entirely confined and did not spread across the central sulcus to the region of the S1 representation labeled 6 in the figures. The significantly smaller local blood flow increases relative to the postcentral gyrus during the passive stimulation/counting control task possibly reflects engagement in the verbalization task. Thus, because S1 might have been inactivated when subjects engaged in a non-tactile task, we could not observe rCBF changes in S1 despite the presence of tactile stimulation during the passive stimulation/counting control task. Similarly, PET studies in which attention was directed to visual (Haxby et al., 1994), auditory (Fiez et al., 1995) or other tactile (Drevets et al., 1995) tasks also reported decreases in blood flow in S1. Anecdotally, most subjects reported being unaware of the tactile stimulation during control trials. When subjects receive passive tactile stimulation, but are not engaged in another task, limited resources are not being taxed, so S1, and possibly other components of the somatosensory system, are not depressed. Thus, passive stimulation with gratings, without a superimposed counting task (Lin et al., 1996; Burton et al., 1997a), or other studies that used a variety of somatosensory stimuli and response measurement or brain-imaging techniques, identified foci in this region (Woolsey et al., 1979; Roland, 1981; Hari et al., 1984; Fox et al., 1987; Seitz and Roland, 1992; Burton et al., 1993; Forss et al., 1994). During the attention tasks, increased rCBF was observed in the S1 cortex. However, these blood flow changes failed to reach significance in the cross-validation analysis, possibly because only the tip of a single digit received stimulation, and this may have activated too small a focus to survive the more conservative (and less sensitive) tests used in this study. Associated with this is the difficulty of cross-realigning data from several subjects in a region of highly varied anatomy (Rademacher et al., 1993; Damasio, 1995; Zilles et al., 1997). Thus, possible partial volume effects from having to realign around the central sulcus, coupled with an already small active focus, could have reduced reproducibility across subjects. As the cross-validation methods used maybe too insensitive to detect small foci in anatomically variable regions without having larger data sets, we reanalyzed the significance of blood flow changes in the volume centered on focus labeled 6. This t-test, which used all of the data, and therefore had twice the degrees of freedom, was significant, indicating that rCBF over focus 6 was substantially greater than background levels. There was a smaller blood flow increase in S1, in comparison with S2 (see below), and this might reflect more limited involvement of S1 in attention. This suggestion is consistent with single-neuron data which show that cells from S1 more frequently display comparatively more minor modulation of responses and these effects appear in a smaller proportion of the cell sample (Hsiao et al., 1993b; Burton et al., 1997c). Increased rCBF in S1 during the attention tasks does not directly imply operation of an attention mechanism within an early hierarchical component of the somatosensory cortex because the subtraction from task conditions was a control state (passive stimulation/counting) with depressed blood flows in comparison to baseline conditions during eyes-closed-rest. The observed rCBF increases in S1 possibly resulted from passive stimulation activity that was not selectively depressed because tactile information was now behaviorally relevant. Attention-related modulation of S1 was previously described in a PET study in which blood flow to S1 was 13% higher when subjects attended to vibrotactile stimuli than when they received the same stimuli but also performed a non-tactile, distracting counting task (Meyer et al., 1991). The present data, therefore, suggest that S1 activity is boosted when tactile information is behaviorally relevant and depressed when information about other modalities or tasks must be processed. The occurrence of similar changes in blood flow in the posterior thalamus suggests that top-down mechanisms might affect the input of information to somatosensory cortex from the thalamus. Thus, significant increases appeared in the thalamus only when contrasting attention tasks with the passive stimulation/counting control condition but not when the latter was contrasted with results from eyes-closed-rest scans. Further study of the role of subcortical structures during attention tasks are clearly needed. Modulation of rCBF in the Parietal Opercular Cortex Increases in rCBF bilaterally in parietal opercular cortex (e.g. S2) during the attention tasks confirm numerous reports of S2's role in somatosensory processing of innocuous tactile stimulation (Garcha and Ettlinger, 1978; LaMotte and Mountcastle, 1979; Robinson and Burton, 1980; Garcha et al., 1982; Roland, 1987; Carlson and Burton, 1988; Burton and Sinclair, 1990; Burton et al., 1993, 1997a, Burton et al., b; Hsiao et al., 1993a,b; Sinclair and Burton, 1993; Ledberg et al., 1995; Huttunen et al., 1996; Roland et al., 1998). Nearly bilaterally matched foci along the parietal operculum showed increases in rCBF during the attention tasks (labeled 2 and 7 in Table 2), but only the changes in the contralateral cortex were significant. The increase during the attention task reflected a change over a zero baseline because no blood flow increases were detected during the passive stimulation/ counting task and because the blood flow measurements for S2 during all control tasks were equal. This enhanced rCBF suggests that when attention was directed to tactile information, S2 activity was even greater than when the subject received but ignored passive stimulation. This result was similar to the changes observed in S1, and further suggested that blood flow increases appear in somatosensory cortex for behaviorally relevant stimuli. Unlike the changes in S1, however, those over the contralateral S2 were of sufficient magnitude and consistency that they cross-validated. The blood flow increases over S2 were also significantly greater than those over S1. Thus, it is possible that S2 activity during these demanding attention tasks was enhanced above the level that would result from a less difficult vigilance task for passive tactile stimulation. This finding is consistent with results from neurophysiological studies in monkeys that described significant modulations of S2 activity during attention-demanding tasks (Poranen and Hyvärinen, 1982; Hsiao et al., 1993b; Burton et al., 1997b). Additionally, a recent PET study found activation near our parietal operculum focus during a roughness discrimination task (Roland et al., 1998). However, there has been little investigation of the role of S2 in duration discrimination. The current results demonstrate that S2 is activated by attention tasks related to stimulus duration as well as stimulus texture. Collectively these results suggest that S2 provides somatosensory processing for tactile attention to a variety of stimulus features. The greater increases in blood flow in S2 than S1 might reflect that S2 is an essential component of a descending control path within the tactile attention system. A top-down path, acting through an enhancement mechanism that primes the somatosensory regions, would enable nonsensory regions to regulate sensory regions or higher-level sensory regions like S2 to regulate lower-level sensory responses in S1. This network organization has been proposed in the visual attention system (Corbetta et al., 1991; Frith and Dolan, 1997). A similar hypothesis regarding the predominance of S2 was obtained previously in a study of how limb movements affected sensory responses in S1 and S2. When sensory input was delivered to a moving limb, S1 activity decreased (Huttunen and Homberg, 1991; Chapman, 1994), but S2 activity increased (Huttunen et al., 1996). The greater increases in rCBF observed in the current study, together with the results with a moving limb, suggested that signals which modulate S2 activity do not only ascend from S1 to S2, but possibly also descend to S2 from nonsensory areas. Thus, modulation from nonsensory areas might influence S2, which in the attention tasks, enabled S1 activity. If further studies confirm this speculation, it complicates interpretation of prior hypotheses about the serial and hierarchical flow of processing from S1 to S2 (Pons et al., 1987). Blood Flow Increases in the Posterior Parietal Cortex A region in the posterior parietal cortex was activated bilaterally by the attention tasks. Clinical literature describing neglect syndromes (Mesulam, 1981; Moscovitch and Behrmann, 1994) and neuroimaging studies of visuospatial attention (Haxby et al., 1991; Corbetta et al., 1993; Vandenberghe et al., 1996; Nobre et al., 1997; Corbetta, 1998), vibrotactile attention (Johannsen et al., 1997) and vigilance (Pardo et al., 1991) have emphasized a role for posterior parietal region(s) in attention processes. Thus, a posterior parietal region may be capable of mediating attention to signal streams from multiple modalities (Andersen, 1997). However, the exact location of the identified region differs between studies and modalities, suggesting that the different tasks possibly recruit slightly different anatomical regions (Vandenberghe et al., 1996). For example, the focus found using the current paradigm was located in the depths of the postcentral sulcus at a distance from the more anteriorly located postcentral gyral site. According to the Tailarach atlas, the posterior parietal focus was likely in Brodmann area 7; it was ~1.5 cm anterior to the most anterior site described in visuospatial experiments, which occupied the posterior wall of the postcentral sulcus and extended posteriorly to the intraparietal sulcus (Nobre et al., 1997; Corbetta, 1998). Neurophysiological studies in area 7 of monkeys have shown activity reflecting body or visual spatial coordinate systems (Andersen, 1989; Colby et al., 1993). These observations, together with clinical (Mesulam, 1990) and recent imaging data of activity in frontal cortex serving gaze control (Corbetta, 1998), suggest that mechanisms for selecting stimuli in reference to body or visual spatial coordinates (extrapersonal space) may be crucial to attention processes. However, such spatial representations for the somatosensory system may occupy more anterior portions of the posterior parietal cortex than are involved with visually guided arm movements, which activate sites within the medial bank of the intraparietal sulcus (Kertzman et al., 1997). Consistent with reports of more frequent attention disorders in patients with lesions to the right parietal cortex (Mesulam, 1981, 1990; Moscovitch and Behrmann, 1994), visuospatial attention studies found that the right hemisphere was dominant (Corbetta et al., 1993; Nobre et al., 1997). Thus, the bilateral activations observed in the present study with subjects attending to right hemispace, relative to their bodies, are consistent with the role of the right hemisphere in directing attention to extrapersonal space irrespective of the body hemispace of selected stimuli (Mesulam, 1990). Other Blood Flow Increases Cingulate Gyrus Activation of an anterior cingulate site during visual attention tasks prompted the hypothesis that this region acts as part of an executive attention network which distributes resources between multiple dimensions of a stimulus when all are pertinent to a task the subject is motivated to perform (Posner and Petersen, 1990). Supporting evidence for the anterior cingulate's role in an attention system comes from clinical assessment of attention disorders in patients with strokes in this region (Laplane et al., 1981; Mesulam, 1981) or after anterior cingulotomies (Watson et al., 1973; Janer and Pardo, 1991). Common to all brain-imaging studies showing activation in the anterior cingulate was performance of a cognitively demanding stimulus–response task that required selection between competing stimuli. Examples involving competing stimuli include the Stroop task conflict condition (Pardo et al., 1990; Derbyshire et al., 1998) and, most similar to the current experiment, during divided, but not selected, attention conditions in visual discriminations (Corbetta et al., 1991, 1993; Nobre et al., 1997). In the present study, subjects encountered greater processing demands during the divided attention task in which information about both surface texture and stimulus duration had to be processed and retained in memory to make the discrimination. However, all of our attention tasks activated region 11, which involved parts of the anterior cingulate that included Brodmann area 32. In this study all of the tasks were demanding as shown by performance levels that seldom exceeded 75% correct responses. Therefore, despite significantly more errors during the divided attention trials, there were high processing demands throughout, and this might have resulted in consistent activation of the anterior cingulate region even during selective attention trials. The anterior cingulate regions identified in most attention studies include Brodmann areas 24 and 32. This site was distinct from a posterior and superior site seen during the control scans (labeled C); the latter occupied a midline portion of Brodmann area 6 and probably involved the supplementary motor area. The coordinates of the anterior cingulate focus in the current study (–1, 7, 48) overlap the mean coordinates found in visual spatial attention studies: –2.3, 22.4, 37.3; SD: 9.1, 9.9, 4.5 (Corbetta et al., 1993) and 8, 16, 44 (Nobre et al., 1997). Combined, these activations are consistent with the theory that attention requires some nonsensory foci which process information from multiple modalities. However, the cingulate gyrus also contains a heterogeneous collection of cortical areas (Vogt et al., 1992, 1996; Picard and Strick, 1996; Whalen et al., 1998) that, in the mid-cingulate region, includes sites also activated when subjects receive painful stimuli (Vogt et al., 1996; Derbyshire et al., 1998). Frontal/Motor Areas Activation of several frontal regions during the control scans possibly reflects subjects' verbal responses, and hence the corresponding motor representation of the mouth (labeled A and D in Table 1) and a paracentral lobule region, corresponding to the supplementary motor area (SMA) (labeled C in Table 1). Subtracting these activations from those present during the attention tasks, however, showed little remaining activation of motor cortex. Thus, the motor patterns for speaking were the same in both conditions and probably involved bilateral control. After subtracting the precentral activations present in the control trials from the attention trials, significant blood flow increases were still seen in a medial sector of the contralateral precentral gyrus (labeled 9 in Table 2), which is included as part of Brodmann's area 6 in the Tailarach atlas (Talairach and Tournoux, 1988). This increase during the attention tasks might reflect that more meaningful verbal responses were required because during these trials they were related to discrimination decisions rather than practiced repetitions of the same words. Three additional frontal regions (anterior insula, Brodmann area 9 in the middle frontal gyrus and Brodmann area 10 on the orbital frontal cortex) showed increased blood flows only during the attention tasks. The affected sites overlapped with those previously activated in episodic retrieval tasks (Buckner and Tulving, 1995; MacLeod et al., 1998). In all of the attention trials subjects had to remember some property of the first stimulus (its roughness or length) in order to make a paired comparison to the second stimulus. During the divided attention trials the memory load was probably greater because properties of both attributes had to be retained when comparing the second stimulus to the first. This might have contributed to bilateral activation of area 10, which was uniquely associated with the divided attention task when compared to trials with selective cuing. Notes We are indebted to Dr M. Corbetta for his comments and suggestions on an initial draft of this paper. Special thanks are extended to Dr Heather Drury for carrying out spatial normalization of the MP-RAGE atlas representative image, Len Lich and the staff of the Cyclotron Unit at Washington University School of Medicine for technical assistance during the PET experiments, Matt Nissing for his efforts in pilot psychophysical studies that determined the surface parameters of the gratings and for designing the computer graphics routines used to create the surfaces, and John Kreitler and his staff for manufacturing the stimulation apparatus and hand restraint devices used in this study. This work was supported by the National Institutes of Health grants, NS31005 to H.B. and HL13851 to M.E.R. Additional support was obtained from the McDonnell Center for the Study of Higher Brain Function. Address correspondence to Dr H. Burton, Department of Anatomy and Neurobiology, Campus Box Number 8108, Washington University School of Medicine, 4566 Scott Avenue, St Louis, MO 63110, USA. Email: harold@touch.wustl.edu. Table 1 Cross-validation of PET counts in regions activated by control task (‘passive stimulation/counting' – ‘rest') Anatomical regiona  Coordinatesb  Magnitudec  td  Pe      x  y  z        aLetters for named anatomical regions match labels in figures.  bStereotaxic coordinates from atlas (Talairach and Tournoux, 1988) for peak PET difference count that was >10% greater than the count in any neighboring pixel within a radius of 11 mm. The coordinates are in millimeters from a 0, 0, 0 point situated at the level of the anterior–posterior commissures (y = 0), at the midline of the brain (x = 0), and anteroposteriorly halfway between the commissures (z = 0). Coordinates presented are from peaks identified in the combined hypothesis-generating and hypothesis-testing data set that are located in closest proximity to peaks identified in the hypothesis-generating data set that were used in cross-validation analysis.  cMagnitude of activation in normalized PET counts that are linearly correlated with blood flow (see Materials and Methods).  dStudentized t-value for testing whether magnitude of PET difference count in ~5 pixel diameter voxels for the hypothesis testing data set equaled zero. The volume was centered on peak coordinates identified in the hypothesis-generating half of the data.  eProbability values of t-values. Strictly significant foci (Bonferonni corrected by number of regions tested; P < 0.01) are indicated by *.  A  peri-Rolandic (medial)  –47  –9  44  169  6.44  0.0004*  B  putamen  –25  –7  2  172  5.25  0.0012*  C  paracentral lobule (BA 6)  –1  –1  56  243  4.98  0.0016*  D  precentral gyrus (BA 6)  53  –11  38  153  4.14  0.0043*  E  peri-Rolandic (lateral)  –51  –13  28  160  3.25  0.014*  Anatomical regiona  Coordinatesb  Magnitudec  td  Pe      x  y  z        aLetters for named anatomical regions match labels in figures.  bStereotaxic coordinates from atlas (Talairach and Tournoux, 1988) for peak PET difference count that was >10% greater than the count in any neighboring pixel within a radius of 11 mm. The coordinates are in millimeters from a 0, 0, 0 point situated at the level of the anterior–posterior commissures (y = 0), at the midline of the brain (x = 0), and anteroposteriorly halfway between the commissures (z = 0). Coordinates presented are from peaks identified in the combined hypothesis-generating and hypothesis-testing data set that are located in closest proximity to peaks identified in the hypothesis-generating data set that were used in cross-validation analysis.  cMagnitude of activation in normalized PET counts that are linearly correlated with blood flow (see Materials and Methods).  dStudentized t-value for testing whether magnitude of PET difference count in ~5 pixel diameter voxels for the hypothesis testing data set equaled zero. The volume was centered on peak coordinates identified in the hypothesis-generating half of the data.  eProbability values of t-values. Strictly significant foci (Bonferonni corrected by number of regions tested; P < 0.01) are indicated by *.  A  peri-Rolandic (medial)  –47  –9  44  169  6.44  0.0004*  B  putamen  –25  –7  2  172  5.25  0.0012*  C  paracentral lobule (BA 6)  –1  –1  56  243  4.98  0.0016*  D  precentral gyrus (BA 6)  53  –11  38  153  4.14  0.0043*  E  peri-Rolandic (lateral)  –51  –13  28  160  3.25  0.014*  View Large Table 2 Cross-validation of PET counts in regions activated by all attention conditions (attention tasks – ‘passive stimulation/counting') Anatomical regiona  Coordinatesb  Magnitudec  td  Pe      x  y  z        aNumbers for named anatomical regions match labels in figures.  b,c,d See Table 1 for descriptions.  eProbability of t-values. Leniently significant foci (P < 0.05) are reported unmarked. Strictly significant foci (Bonferonni corrected by the number of regions tested; P < 0.0045) are indicated by *.  Subcortical  1  thalamus  –5  –21  2  87  7.92  <0.0001*  Parietal lobe  2  parietal operculum  –53  –21  16  106  6.34  0.0004*  3  postcentral gyrus (posterior)  –59  –23  22  97  3.41  0.0112  4  fundus of the postcentral sulcus (contra)  –35  –39  36  91  3.31  0.0132  5  fundus of the postcentral sulcus (ipsi)  41  –39  38  63  3.15  0.0161  6  postcentral gyrus  –49  –31  40  82  1.71  0.1305  7  parietal operculum  51  –23  24  63  1.07  0.319  Frontal lobe  8  middle frontal gyrus (BA 9)  45  5  38  79  2.61  0.0351  9  precentral gyrus (BA 6)  –35  –9  46  83  2.53  0.0395  10  insula  –31  11  2  97  1.54  0.1675  Cingulate  11  cingulate gyrus (BA 32)  –1  7  48  119  6.18  0.0005*  Anatomical regiona  Coordinatesb  Magnitudec  td  Pe      x  y  z        aNumbers for named anatomical regions match labels in figures.  b,c,d See Table 1 for descriptions.  eProbability of t-values. Leniently significant foci (P < 0.05) are reported unmarked. Strictly significant foci (Bonferonni corrected by the number of regions tested; P < 0.0045) are indicated by *.  Subcortical  1  thalamus  –5  –21  2  87  7.92  <0.0001*  Parietal lobe  2  parietal operculum  –53  –21  16  106  6.34  0.0004*  3  postcentral gyrus (posterior)  –59  –23  22  97  3.41  0.0112  4  fundus of the postcentral sulcus (contra)  –35  –39  36  91  3.31  0.0132  5  fundus of the postcentral sulcus (ipsi)  41  –39  38  63  3.15  0.0161  6  postcentral gyrus  –49  –31  40  82  1.71  0.1305  7  parietal operculum  51  –23  24  63  1.07  0.319  Frontal lobe  8  middle frontal gyrus (BA 9)  45  5  38  79  2.61  0.0351  9  precentral gyrus (BA 6)  –35  –9  46  83  2.53  0.0395  10  insula  –31  11  2  97  1.54  0.1675  Cingulate  11  cingulate gyrus (BA 32)  –1  7  48  119  6.18  0.0005*  View Large Figure 1. View largeDownload slide Schematic representation of surfaces and task conditions for three examples of paired gratings (see Materials and Methods). The subject received an attention-directing cue, was stimulated by a pair of textured surfaces, and made a verbal response indicating which surface felt rougher and/or contacted the finger for longer durations. For the first example trials (A1 and A2), the cue directed attention to the roughness differences between the gratings (Roughness). The correct response for A1 was ‘one' since the first surface's ridges were further apart, which feels rougher. The correct response for trial A2 was ‘two' since the second surface of the pair was rougher. For the next example trials (B1 and B2), the cue directed attention to the duration of contact with each grating (Duration). The correct response for B1 was ‘one' since the first surface was longer and therefore contacted the finger for a longer period of time; the correct response for trial B2 was ‘two' since the second surface was longer. For the third example trials (C1 and C2), the cue was neutral and no attribute was specified (Divided). The subjects had to divide their attention between the two features in order to determine which of the two surfaces in a pair was in contact longer or was rougher. The correct response for C1 was ‘one' since the first surface was both rougher and contacted the finger longer; the correct response for trial C2 was ‘two' since the second surface of the pair was longer. Figure 1. View largeDownload slide Schematic representation of surfaces and task conditions for three examples of paired gratings (see Materials and Methods). The subject received an attention-directing cue, was stimulated by a pair of textured surfaces, and made a verbal response indicating which surface felt rougher and/or contacted the finger for longer durations. For the first example trials (A1 and A2), the cue directed attention to the roughness differences between the gratings (Roughness). The correct response for A1 was ‘one' since the first surface's ridges were further apart, which feels rougher. The correct response for trial A2 was ‘two' since the second surface of the pair was rougher. For the next example trials (B1 and B2), the cue directed attention to the duration of contact with each grating (Duration). The correct response for B1 was ‘one' since the first surface was longer and therefore contacted the finger for a longer period of time; the correct response for trial B2 was ‘two' since the second surface was longer. For the third example trials (C1 and C2), the cue was neutral and no attribute was specified (Divided). The subjects had to divide their attention between the two features in order to determine which of the two surfaces in a pair was in contact longer or was rougher. The correct response for C1 was ‘one' since the first surface was both rougher and contacted the finger longer; the correct response for trial C2 was ‘two' since the second surface of the pair was longer. Figure 2. View largeDownload slide Psychophysical results illustrating percentage correct (means and standard errors) as a function of cue validity (Trial Type). Mean performances for 10 subjects were significantly different on trials when cues correctly selected one surface attribute for attention (Valid) compared to trials in which cues did not specify the attribute (Neutral). Figure 2. View largeDownload slide Psychophysical results illustrating percentage correct (means and standard errors) as a function of cue validity (Trial Type). Mean performances for 10 subjects were significantly different on trials when cues correctly selected one surface attribute for attention (Valid) compared to trials in which cues did not specify the attribute (Neutral). Figure 3. View largeDownload slide PET average difference images displayed in selected sagittal planes. The rightmost column shows the corresponding MP-RAGE atlas representative target image. The color bar is scaled in units of normalized increases in PET difference counts (10 counts = 1% change in blood flow). Foci identified in ‘passive stimulation/counting' minus ‘eyes-closed-rest' subtractions are shown on the left in column that is labeled ‘control.' Letter-tagged red arrows point to correspondingly labeled foci listed in Table 1. The middle column labeled ‘attention' shows averaged data from all attention conditions (attention scan minus ‘passive stimulation/counting'). Number-tagged red arrows point to foci listed in Table 2. Corresponding average MP-RAGE images for the same sections are shown in the right column. All PET subtraction images warped to fit the MP-RAGE images (see text). Abbreviations: CeS, central sulcus; PoCeS, postcentral sulcus; PreCeS, precentral sulcus. Figure 3. View largeDownload slide PET average difference images displayed in selected sagittal planes. The rightmost column shows the corresponding MP-RAGE atlas representative target image. The color bar is scaled in units of normalized increases in PET difference counts (10 counts = 1% change in blood flow). Foci identified in ‘passive stimulation/counting' minus ‘eyes-closed-rest' subtractions are shown on the left in column that is labeled ‘control.' Letter-tagged red arrows point to correspondingly labeled foci listed in Table 1. The middle column labeled ‘attention' shows averaged data from all attention conditions (attention scan minus ‘passive stimulation/counting'). Number-tagged red arrows point to foci listed in Table 2. Corresponding average MP-RAGE images for the same sections are shown in the right column. All PET subtraction images warped to fit the MP-RAGE images (see text). Abbreviations: CeS, central sulcus; PoCeS, postcentral sulcus; PreCeS, precentral sulcus. Figure 4. View largeDownload slide PET data show rCBF increases on selected, interpolated coronal slices through portions of the postcentral gyrus and parietal operculum that contain S1 and S2 representations. The same reconstruction conventions described for Figure 3 apply here. The stereotaxic coordinates of the correspondingly labeled foci are in Table 2. Figure 4. View largeDownload slide PET data show rCBF increases on selected, interpolated coronal slices through portions of the postcentral gyrus and parietal operculum that contain S1 and S2 representations. The same reconstruction conventions described for Figure 3 apply here. The stereotaxic coordinates of the correspondingly labeled foci are in Table 2. Figure 5. View largeDownload slide Average PET counts (n = 9) measured in 11 mm diameter spherical volumes centered on foci 2 (S2) and 6 (S1) for each task condition. The error bars represent SEM (missing for tasks where the standard errors were too small to appear with the scaling factor used to create the graph). Tasks: eyes-closed-rest = ecr; passive tactile stimulation = pass; passive tactile stimulation and counting = pass/counting; all selective and divided attention tasks = attention. Task effects were significant for both foci and, within each cortical location, blood flows were significantly greater during the attention tasks. For S1, blood flows during the passive stimulation/counting task were significantly lower than during ecr and passive stimulation controls. Note significantly greater attention task effect in S2 in comparison to S1. Figure 5. View largeDownload slide Average PET counts (n = 9) measured in 11 mm diameter spherical volumes centered on foci 2 (S2) and 6 (S1) for each task condition. The error bars represent SEM (missing for tasks where the standard errors were too small to appear with the scaling factor used to create the graph). Tasks: eyes-closed-rest = ecr; passive tactile stimulation = pass; passive tactile stimulation and counting = pass/counting; all selective and divided attention tasks = attention. Task effects were significant for both foci and, within each cortical location, blood flows were significantly greater during the attention tasks. For S1, blood flows during the passive stimulation/counting task were significantly lower than during ecr and passive stimulation controls. Note significantly greater attention task effect in S2 in comparison to S1. References Allport A (1989) Visual attention. In: Foundations of cognitive science (Posner MI, ed.), pp. 631-682. Cambridge, MA: MIT Press. Google Scholar Andersen R ( 1989) Visual and eye movement functions of the posterior parietal cortex. Annu Rev Neurosci  12: 377–403. Google Scholar Andersen RA ( 1997) Multimodal integration for the representation of space in the posterior parietal cortex. Phil Trans R Soc Lond B Biol Sci  352: 1421–1428. Google Scholar Büchel C, Josephs O, Rees G, Turner R, Frith C, Friston K ( 1998) The functional anatomy of attention to visual motion. A functional MRI Study. Brain  121: 1281–1294. Google Scholar Buckner R, Tulving E (1995) Neuroimaging studies of memory: theory and recent PET results. In: Handbook of neuropsychology (Boller F, Grafman J, eds), pp. 439–466. Amsterdam: Elsevier. Google Scholar Buckner R, Petersen S, Raichle M ( 1995) Dissociation of human prefrontal cortex areas across different speech production tasks and gender groups. J Neurophysiol  74: 2163–2173. Google Scholar Buckner RL, Petersen SE, Ojemann JG, Miezin FM, Squire LR, Raichle ME ( 1995) Functional anatomical studies of explicit and implicit memory retrieval tasks. J Neurosci  15: 12–29. Google Scholar Buckner RL, Raichle ME, Miezin FM, Petersen SE ( 1996) Functional anatomic studies of memory retrieval for auditory words and visual pictures. J Neurosci  16: 6219–6235. Google Scholar Burton H, Sinclair RJ ( 1990) Second somatosensory cortical area in macaque monkeys. I. Neuronal responses to controlled, punctate indentations of glabrous skin on the hand. Brain Res  520: 262–71. Google Scholar Burton H, Videen TO, Raichle ME ( 1993) Tactile vibration activated foci in insular and parietal opercular cortex studied with positron emission tomography: mapping the second somatosensory area in humans. Somatosens Mot Res  10: 297–308. Google Scholar Burton H, MacLeod A-MK, Videen T, Raichle M ( 1997) Multiple foci in parietal and frontal cortex activated by rubbing embossed grating patterns across fingerpads: a positron emission tomography study in humans. Cereb Cortex  7: 3–17. Google Scholar Burton H, Sinclair R, Hong S-Y, Pruett JJ, Whang K ( 1997) Tactile-spatial and cross-modal attention effects in the second somatosensory and 7b cortical areas of rhesus monkeys. Somatosens Mot Res  14: 237–267. Google Scholar Burton H, Sinclair R, Kuo J ( 1997) Responses in macaque monkey primary somatosensory (SI) and motor cortex during a tactile attention task. Soc Neurosci Abstr  23: 1801. Google Scholar Carlson M, Burton H ( 1988) Recovery of tactile function after damage to primary or secondary somatic sensory cortex in infant Macaca mulatta. J Neurosci  8: 833–859. Google Scholar Chapman CE ( 1994) Active versus passive touch: factors influencing the transmission of somatosensory signals to primary somatosensory cortex. Can J Physiol Pharmacol  72: 558–570. Google Scholar Coghill RC, Talbot JD, Evans AC, Meyer E, Gjedde A, Bushnell MC, Duncan GH ( 1994) Distributed processing of pain and vibration by the human brain. J Neurosci  14: 4095–108. Google Scholar Colby CL, Duhamel JR, Goldberg ME (1993) The analysis of visual space by the lateral intraparietal area of the monkey: the role of extraretinal signals. [Review]. In: Progress in brain research (Hicks TP, Molotchnikoff S, Ono T, eds), pp. 307–316. New York: Elsevier. Google Scholar Collins DL, Neelin P, Peters TM, Evans AC ( 1994) Automatic 3D inter-subject registration of MR volumetric data in standardized Talairach space. J Comput Assist Tomogr  18: 192–205. Google Scholar Corbetta M ( 1998) Frontoparietal cortical networks for directing attention and the eye to visual locations: identical, independent, or overlapping neural systems? Proc Natl Acad Sci USA  95: 831–838. Google Scholar Corbetta M, Miezin FM, Dobmeyer S, Shulman GL, Petersen SE ( 1991) Selective and divided attention during visual discriminations of shape, color, and speed: functional anatomy by positron emission tomography. J Neurosci  11: 2383–2402. Google Scholar Corbetta M, Miezin FM, Shulman GL, Petersen SE ( 1993) A PET study of visuospatial attention. J Neurosci  13: 1202–1226. Google Scholar Coull JT ( 1998) Neural correlates of attention and arousal: insights from electrophysiology, functional neuroimaging and psychopharmacology. Prog Neurobiol  55: 343–361. Google Scholar Critchley M ( 1949) The phenomenon of tactile inattention with special reference to parietal lesions. Brain  72: 538–561. Google Scholar Damasio H (1995) Human brain anatomy in computerized images. New York: Oxford University Press. Google Scholar Derbyshire S, Vogt B, Jones A ( 1998) Pain and Stroop interference tasks activate separate processing modules in anterior cingulate cortex. Exp Brain Res  118: 52–60. Google Scholar Drevets WC, Burton H, Videen TO, Snyder AZ, Simpson J Jr, Raichle ME ( 1995) Blood flow changes in human somatosensory cortex during anticipated stimulation. Nature  373: 249–252. Google Scholar Fiez JA, Raichle ME, Miezen FM, Petersen SE, Tallal P, W.F. K{???} ( 1995) PET studies of auditory and phonological processing: effects of stimulus characteristics and task demands. J Cogn Neurosci  7: 357–375. Google Scholar Forss N, Hari R, Salmelin R, Ahonen A, Hamalainen M, Kajola M, Knuutila J, Simola J ( 1994) Activation of the human posterior parietal cortex by median nerve stimulation. Exp Brain Res  99: 309–315. Google Scholar Fox PT, Mintun MA, Raichle ME, Herscovitch P ( 1984) A noninvasive approach to quantitative functional brain mapping with H215O and positron emission tomography. J Cereb Blood Flow Metab  4: 329–333. Google Scholar Fox PT, Burton H, Raichle ME ( 1987) Mapping human somatosensory cortex with positron emission tomography. J Neurosurg  67: 34–43. Google Scholar Fox PT, Mintun MA, Reiman EM, Raichle ME ( 1988) Enhanced detection of focal brain responses using intersubject averaging and change-distribution analysis of subtracted PET images. J Cereb Blood Flow Metab  8: 642–653. Google Scholar Friston K, Holmes A, Worsley K, Poline J, Frith C, Frackowiak R ( 1995) Statistical parametric maps in functional imaging: a general linear approach. Hum Brain Map  2: 189–210. Google Scholar Frith C, Dolan RJ ( 1997) Brain mechanisms associated with top-down processes in perception. Phil Trans R Soc Lond B Biol Sci  352: 1221–1230. Google Scholar Garcha HS, Ettlinger G ( 1978) The effects of unilateral or bilateral removals of the second somatosensory cortex (area SII): a profound tactile disorder in monkeys. Cortex  14: 319–326. Google Scholar Garcha HS, Ettlinger G, Maccabe JJ ( 1982) Unilateral removal of the second somatosensory projection cortex in the monkey: evidence for cerebral predominance? Brain  105: 787–810. Google Scholar Hari R, Reinikainen K, Kaukoranta E, Hämäläinen M, Ilmoniemi R, Penttinen A, Salminen J, Teszner D ( 1984) Somatosensory evoked cerebral magnetic fields from SI and SII in man. Electroenceph Clin Neurophysiol  57: 254–263. Google Scholar Haxby JV, Grady CL, Ungerleider LG, Horwitz B ( 1991) Mapping the functional neuroanatomy of the intact human brain with brain work imaging. Neuropsychologia  29: 539–55. Google Scholar Haxby JV, Horwitz B, Ungerleider LG, Maisog JM, Pietrini P, Grady CL ( 1994) The functional organization of human extrastriate cortex: a PET–rCBF study of selective attention to faces and locations. J Neurosci  14: 6336–6353. Google Scholar Holmes P (1994) Statistical issues in functional brain mapping. Ph.D. thesis, University of Glasgow (www.fil.ion.ucl.ac.ak/spm/papers/ APH_thesis). Google Scholar Hsiao SS, Johnson KO, Twombly IA ( 1993) Roughness coding in the somatosensory system. Acta Psychol  84: 53–67. Google Scholar Hsiao SS, O'Shaughnesy DM, Johnson KO ( 1993) Effects of selective attention on spatial form processing in monkey primary and secondary somatosensory cortex. J Neurophysiol  70: 444–447. Google Scholar Hunton DL, Miezin FM, Buckner RL, van Mier HI, Raichle ME, Petersen SE ( 1996) An assessment of functional-anatomical variability in neuroimaging studies. Hum Brain Map  4: 122–139. Google Scholar Huttunen J, Homberg V ( 1991) Modification of cortical somatosensory evoked potentials during tactile exploration and simple active and passive movements. Electroenceph Clin Neurophysiol  81: 216–23. Google Scholar Huttunen J, Wikström H, Korvenoja A, Seppäläinen AM, Aronen H, Ilmoniemi RJ ( 1996) Significance of the second somatosensory cortex in sensorimotor integration: enhancement of sensory responses during finger movements. NeuroReport  7: 1009–1012. Google Scholar Janer KW, Pardo JV ( 1991) Deficits in selective attention following bilateral anterior cingulotomy. J Cogn Neurosci  3: 231–241. Google Scholar Johannsen P, Jakobsen J, Bruhn P, Hansen SB, Gee A, Stødkilde-Jørgensen H, Gjedde A ( 1997) Cortical sites of sustained and divided attention in normal elderly humans. NeuroImage  6: 145–155. Google Scholar Kertzman C, Schwarz U, Zeffiro TA, Hallett M ( 1997) The role of posterior parietal cortex in visually guided reaching movements in humans. Exp Brain Res  114: 170–183. Google Scholar LaMotte RH, Mountcastle VB ( 1979) Disorders in somesthesis following lesions of parietal lobe. J Neurophysiol  42: 400–19. Google Scholar Lancaster J, Glass T, Lankipalli B, Downs H, Mayberg H, Fox P ( 1995) A modality-independent approach to spatial normalization of tomographic images of the human brain. Hum Brain Map  3: 209–223. Google Scholar Laplane D, Degos JD, Baulac M, Gray F ( 1981) Bilateral infarction of the anterior cingulate gyri and of the fornices. Report of a case. J Neurol Sci  51: 289–300. Google Scholar Ledberg A, O'Sullivan BT, Kinomura S, Roland PE ( 1995) Somatosensory activations of the parietal operculum of man. A PET study. Eur J Neurosci  7: 1934–1941. Google Scholar Lin W, Kuppusamy K, Haacke EM, Burton H ( 1996) Functional MRI in human somatosensory cortex activated by touching textured surfaces. J Magn Reson Imag  6: 565–572. Google Scholar MacLeod AK, Buckner RL, Miezin FM, Petersen SE, Raichle ME ( 1998) Right anterior prefrontal cortex activation during semantic monitoring and working memory. NeuroImage  7: 41–48. Google Scholar Mesulam M-M ( 1981) A cortical network for directed attention and unilateral neglect. Ann Neurol  10: 309–325. Google Scholar Mesulam M-M ( 1990) Large-scale neurocognitive networks and distributed processing for attention, language, and memory. Ann Neurol  28: 597–613. Google Scholar Meyer E, Ferguson S, Zatorre R, Alivisatos B, Marrett S, Evans A, Hakim A ( 1991) Attention modulates somatosensory cerebral blood flow response to vibrotactile stimulation measured by positron emission tomography. Ann Neurol  29: 440–443. Google Scholar Mintun MA, Fox PT, Raichle ME ( 1989) A highly accurate method of localizing regions of neuronal activation in the human brain with positron emission tomography. J Cereb Blood Flow Metab  9: 96–103. Google Scholar Moscovitch M, Behrmann M ( 1994) Coding of spatial information in the somatosensory system: evidence from patients with neglect following parietal lobe damage. J Cogn Neurosci  6: 151–155. Google Scholar Mugler JP, Brookeman JR ( 1990) Three-dimensional magnetization-prepared rapid gradient-echo imaging (3D MP RAGE). Magn Reson Med  15: 152–157. Google Scholar Nobre AC, Sebestyen GN, Gitelman DR, Mesulam MM, Frackowiak RSJ, Frith CD ( 1997) Functional localization of the system for visuospatial attention using positron emission tomography. Brain  120: 515–533. Google Scholar Pardo JV, Fox PT, Raichle ME ( 1991) Localization of a human system for sustained attention by positron emission tomography. Nature  349: 61–64. Google Scholar Pardo JV, Pardo PJ, Janer KW, Raichle ME ( 1990) The anterior cingulate cortex mediates processing selection in the Stroop attentional conflict paradigm. Proc Natl Acad Sci USA  87: 256–9. Google Scholar Petersen SE, Fox PT, Posner MI, Mintun MA, Raichle ME ( 1989) Positron emission tomographic studies of the processing of single words. J Cogn Neurosci  1: 153–170. Google Scholar Picard N, Strick P ( 1996) Motor areas of the medial wall: a review of their location and functional activation. Cereb Cortex  6: 342–353. Google Scholar Pons TP, Garraghty PE, Friedman DP, Mishkin M ( 1987) Physiological evidence for serial processing in somatosensory cortex. Science  237: 417–420. Google Scholar Pons TP, Garraghty PE, Cusick CG, Kaas JH ( 1985) The somatotopic organization of area 2 in macaque monkeys. J Comp Neurol  241: 445–66. Google Scholar Poranen A, Hyvärinen J ( 1982) Effects of attention on multiunit responses to vibration in the somatosensory regions of the monkey's brain. Electroenceph Clin Neurophysiol  53: 525–37. Google Scholar Posner MI (1986) Chronometric explorations of mind. New York: Oxford University Press. Google Scholar Posner MI, Petersen SE ( 1990) The attention system of the human brain. Annu Rev Neurosci  13: 25–42. Google Scholar Posner MI, Driver J ( 1992) The neurobiology of selective attention [Review]. Curr Opin Neurobiol  2: 165–169. Google Scholar Posner MI, Dehaene S ( 1994) Attentional networks [Review]. Trends Neurosci  17: 75–79. Google Scholar Posner MI, Petersen SE, Fox PT, Raichle ME ( 1988) Localization of cognitive operations in the human brain. Science  240: 1627–1631. Google Scholar Posner MI, Walker JA, Friedrech FJ, Rafal RD ( 1984) Effects of parietal injury on covert orienting of attention. J Neurosci  4: 1863–1874. Google Scholar Raczkowski D, Kalat JW, Nebes R ( 1974) Reliability and validity of some handedness questionnaire items. Neuropsychologia  6: 43–47. Google Scholar Rademacher J, Caviness VS Jr, Steinmetz H, Galaburda AM ( 1993) Topographical variation of the human primary cortices: implications for neuroimaging, brain mapping, and neurobiology. Cereb Cortex  3: 313–329. Google Scholar Robinson CJ, Burton H ( 1980) Somatotopographic organization in the second somatosensory area of M. fascicularis. J Comp Neurol  192: 43–67. Google Scholar Roland PE ( 1981) Somatotopical tuning of postcentral gyrus during focal attention in man. A regional cerebral blood flow study. J Neurophysiol  46: 744–754. Google Scholar Roland PE ( 1982) Cortical regulation of selective attention in man. A regional cerebral blood flow study. J Neurophysiol  48: 1059–1078. Google Scholar Roland PE ( 1987) Somatosensory detection of microgeometry, macrogeometry and kinesthesia after localized lesions of the cerebral hemispheres in man. Brain Res  434: 43–94. Google Scholar Roland PE, Geyer S, Amunts K, Schormann T, Schleicher A, Malikovic A, Zilles K ( 1997) Cytoarchitectural maps of the human brain in standard anatomical space. Hum Brain Map  5: 222–227. Google Scholar Roland PE, O'Sullivan B, Kawashima R ( 1998) Shape and roughness activate different somatosensory areas in the human brain. Proc Natl Acad Sci USA  95: 3295–3300. Google Scholar Seitz RJ, Roland PE ( 1992) Vibratory stimulation increases and decreases the regional cerebral blood flow and oxidative metabolism: a positron emission tomography (PET) study. Acta Neurol Scand  86: 60–67. Google Scholar Shulman GL, Corbetta M, Buckner RL, Raichle ME, Fiez JA, Miezin FM, Petersen SE ( 1997) Top-down modulation of early sensory cortex. Cereb Cortex  7: 193–206. Google Scholar Sinclair RJ, Burton H ( 1993) Neuronal activity in the second somatosensory cortex of monkeys (Macaca mulatta) during active touch of gratings. J Neurophysiol  70: 331–350. Google Scholar Snyder AZ (1996) Difference image vs. ratio image error function forms in PET–PET realignment. In: Quantification of brain function using PET (Bailey D, Jones T, eds), pp. 131–137. San Diego: Academic Press. Google Scholar Squire LR, Ojemann JG, Miezin FM, Petersen SE, Videen TO, Raichle ME ( 1992) Activation of the hippocampus in normal humans: a functional anatomical study of memory. Proc Natl Acad Sci USA  89: 1837–1841. Google Scholar Talairach J, Tournoux P (1988) Co-planar stereotaxic atlas of the human brain. Stuttgart: Thieme. Google Scholar Vandenberghe R, Dupont P, De Bruyn B, Bormans G, Michiels J, Mortelmans L, Orban G ( 1996) The influence of stimulus location on the brain activation pattern in detection and orientation discrimination. Brain  119: 1263–1276. Google Scholar Vogt BA, Finch DM, Olson CR ( 1992) Functional heterogeneity in cingulate cortex: the anterior executive and posterior evaluative regions. Cereb Cortex  2: 435–43. Google Scholar Vogt B, Derbyshire S, Jones A ( 1996) Pain processing in four regions of human cingulate cortex localized with co-registered PET and MR imaging. Eur J Neurosci  8: 1461–1473. Google Scholar Watson RT, Heilman KM, Cauthen JC, King FA ( 1973) Neglect after cingulectomy. Neurology  23: 1003–1007. Google Scholar Whalen PJ, Bush G, McNally RJ, Wilhelm S, McInerney SC, Jenike MA, Rauch SL ( 1998) The emotional counting Stroop paradigm: a functional magnetic resonance imaging probe of the anterior cingulate affective division [in process citation]. Biol Psychiat  44: 1219–1228. Google Scholar White L, Andrews T, Hulette C, Richards A, Groelle M, Paydarfar J, Purves D ( 1997) Structure of the human sensorimotor system. I: morphology and cytoarchitecture of the central sulcus. Cereb Cortex  7: 18–30. Google Scholar Woldorff MG, Gallen CC, Hampson SA, Hillyard SA, Pantev C, Sobel D, Bloom FE ( 1993) Modulation of early sensory processing in human auditory cortex during auditory selective attention. Proc Natl Acad Sci USA  90: 8722–8726. Google Scholar Woodruff PW, Benson RR, Bandettini PA, Kwong KK, Howard RJ, Talavage T, Belliveau J, Rosen BR ( 1996) Modulation of auditory and visual cortex by selective attention is modality-dependent. NeuroReport  7: 1909–1913. Google Scholar Woolsey CN, Erickson TC, Gilson WE ( 1979) Localization in somatic sensory and motor areas of human cerebral cortex as determined by direct recording of evoked potentials and electrical stimulation. J Neurosurg  51: 476–506. Google Scholar Zilles K, Schleicher A, Langemann C, Amunts K, Morosan P, Palomero-Gallagher N, Schormann T, Mohlberg H, Bürgel U, Steinmetz H, Schlaug G, Roland P ( 1997) Quantitative analysis of sulci in the human cerebral cortex: development, regional heterogeneity, gender difference, asymmetry, intersubject variability and cortical architecture. Hum Brain Map  5: 218–221. Google Scholar © Oxford University Press
TI - Tactile Attention Tasks Enhance Activation in Somatosensory Regions of Parietal Cortex: A Positron Emission Tomography Study
JF - Cerebral Cortex
DO - 10.1093/cercor/9.7.662
DA - 1999-10-01
UR - https://www.deepdyve.com/lp/oxford-university-press/tactile-attention-tasks-enhance-activation-in-somatosensory-regions-of-DPJjVlXnqV
SP - 662
EP - 674
VL - 9
IS - 7
DP - DeepDyve
ER -