Four-dimensional maps of the human PNAS PLUS

Pietro Avanzinia,b,1, Rouhollah O. Abdollahia,c, Ivana Sartorid, Fausto Caruanae, Veronica Pellicciad, Giuseppe Casacelid, Roberto Maid, Giorgio Lo Russod, Giacomo Rizzolattia,b,e,1, and Guy A. Orbana

aDipartimento di Neuroscienze, University of Parma, 43125 Parma, Italy; bIstituto di Neuroscienze, Consiglio nazionale delle Ricerche, 43125 Parma, Italy; cCognitive Neuroscience Section, Institute of Neuroscience and Medicine, Research Center Jülich, 52428 Julich, Germany; dCentro per la chirurgia dell’Epilessia “Claudio Munari,” Ospedale Ca’Granda-Niguarda, 20162 Milan, Italy; and eBrain Center for Social and Motor Cognition, Istituto Italiano di Tecnologia, 43125 Parma, Italy

Contributed by Giacomo Rizzolatti, February 5, 2016 (sent for review November 21, 2015; reviewed by Riitta Hari, Jon H. Kaas, Philippe Kahane, and Karl Zilles) A fine-grained description of the spatiotemporal dynamics of human (ii) the electrocorticogram (ECoG), obtained from subdural elec- brain activity is a major goal of neuroscientific research. Limitations in trode grids, covering regions of cortex. The latter technique suffers spatial and temporal resolution of available noninvasive recording from three disadvantages: (i) the technique only samples from and imaging techniques have hindered so far the acquisition of cortical gyri, missing cortical regions buried in sulci (7); (ii)the precise, comprehensive four-dimensional maps of human neural technique is affected by volume conduction (8); and (iii)the activity. The present study combines anatomical and functional data technique does not record directly from gray matter, because pia from intracerebral recordings of nearly 100 patients, to generate and arachnoid mater lie in between the electrodes and the cortex, highly resolved four-dimensional maps of human cortical processing reducing the amplitude and making it more difficult to extract high- of nonpainful somatosensory stimuli. These maps indicate that the frequency activity from the recorded signal. Both approaches suffer human somatosensory system devoted to the hand encompasses a from the so-called sparse-sampling problem (i.e., the limited and widespread network covering more than 10% of the cortical surface uneven coverage of the patients’ brain because the positioning of the of both hemispheres. This network includes phasic components, electrodes in any given patient is dictated by clinical criteria), leading centered on primary somatosensory cortex and neighboring motor, to difficulties in carrying out analyses at the population level. NEUROSCIENCE premotor, and inferior parietal regions, and tonic components, The primary aim of the present study is to show how ana- centered on opercular and insular areas, and involving human parietal tomical and functional data recorded from a large number of rostroventral area and ventral medial-superior-temporal area. The patients could be combined to generate highly resolved four- technique described opens new avenues for investigating the neural dimensional maps of human cortical processing. To demonstrate basis of all levels of cortical processing in humans. the feasibility of computing such maps, indicating consistently active cortical nodes and the nodes’ time course at a millisecond intracranial recordings | stereo-EEG | median nerve | | scale, we leveraged the advantages of sEEG to investigate so- temporal dynamics matosensory processing following electrical stimulation of the median nerve in nearly 100 patients. fMRI studies have consis- detailed description of the spatiotemporal dynamics of hu- tently reported activation of the primary somatosensory complex Aman brain activity is a major goal of neuroscientific re- (SI), secondary somatosensory complex (SII), and insula in response search. However, it has been impossible so far to attain both high to transient nonpainful stimuli (9) but less so of supplementary spatial and temporal resolution using the available noninvasive recording and imaging techniques. Hence, a precise and com- Significance prehensive four-dimensional cartography of human neural activity has not yet been obtained. High spatial resolution, provided by neuroimaging techniques such as functional magnetic resonance Here, we show how anatomical and functional data recorded imaging (fMRI), is crucial for highlighting the topographical orga- from patients undergoing stereo-EEG can be combined to nization of specific areas (e.g., somatotopy of sensorimotor areas) generate highly resolved four-dimensional maps of human as well as identifying the nodes of brain networks endowed with cortical processing. We used this technique, which provides specific functional properties (1). It is not sufficient, however, to spatial maps of the active cortical nodes at a millisecond scale, know which nodes are active; information is also needed about to depict the somatosensory processing following electrical the local dynamics of the nodes, as well as the relative timing of stimulation of the median nerve in nearly 100 patients. The re- their activity, to fully understand functions (2, 3). sults showed that human somatosensory system encompasses Even if the temporal resolution of electroencephalography a widespread cortical network including a phasic component, (EEG) and magnetoencephalography (MEG) allowed one to centered on primary somatosensory cortex and neighboring observe the intra- and interareal dynamics, to date such re- motor, premotor, and inferior parietal regions, as well as a tonic cordings remain too poor in localization power (1–2 cm) (3, 4). component, centered on the opercular and insular areas, lasting Combining EEG and fMRI has been suggested as a solution, more than 200 ms. using EEG to determine the temporal dynamics within and be- Author contributions: P.A., I.S., G.R., and G.A.O. designed research; I.S., V.P., G.C., R.M., tween the areas identified with fMRI (5). However, the disparate and G.L.R. performed research; P.A., R.O.A., and F.C. analyzed data; and P.A., G.R., and nature of the two signals recorded (hemodynamic for fMRI, G.A.O. wrote the paper. electrical for EEG) creates discrepancies in the results that prevent Reviewers: R.H., School of Science, Aalto University; J.H.K., Vanderbilt University; P.K., precise matching of these methods (3). INSERM U836 and Université Joseph Fourier; and K.Z., Research Center Julich. Invasive intracranial EEG offers a unique opportunity to observe The authors declare no conflict of interest. human brain activity with an unparalleled combination of spatial Freely available online through the PNAS open access option. and temporal resolution. Depending on the electrodes used, two 1To whom correspondence may be addressed. Email: [email protected] or kinds of recordings can be made: (i) intraparenchymal recordings, [email protected]. also called stereo-EEG (sEEG) (6), obtained using stereotactically This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. inserted needle-like electrodes with multiple recording leads; and 1073/pnas.1601889113/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1601889113 PNAS Early Edition | 1of8 Downloaded by guest on October 2, 2021 motor area (SMA) (10), anterior (11), superior Statistical analysis revealed that 1,146 of the leads exploring gray parietal lobule (SPL) and (IPL) (12), and matter presented a significant broadband gamma power (50- to (13). In addition to providing sensitive detection 150-Hz) increase (see Fig. 1C for an example of single lead analysis of responsive regions, sEEG responses also reveal time courses, on gamma-band time course) in response to the median nerve which could be relevant to debates concerning cortical dynamics, stimulation (489 in the left hemisphere, 657 in the right; Fig. S4). such as the parallel vs. serial operation of SII relative to SI (14, 15). By localizing each lead, we computed the overall responsiveness Here, we show for the first time, to our knowledge, that compre- (in % of responsive leads) maps for both hemispheres (Fig. 2), in hensive four-dimensional maps of human cortical somatosensory principle comparable with neuroimaging data in that timing is not processing can be computed from high-frequency broadband considered. High responsiveness values were found in SI. In par- gamma activity, thus surpassing previous subdural and EEG/MEG ticular, the highest proportion of responsive leads was present bi- maps in spatiotemporal precision. laterally in areas 3a (left 81%, right 94%) and 3b (left 95%, right 87%). Lower values characterized areas 1 (left 60%, right 60%) Results and 2 (left 55%, right 64%). These results are in agreement with Recordings were obtained from 17,009 leads in 99 patients, of previous MEG results (16). Opercular (OP) and insular regions which 11,983 (5,678 in the left hemisphere and 6,305 in the right) were active in both hemispheres, peaking at 50% in OP1 (17). were localized in the cortical gray matter according to the ana- Responsive regions included mainly OP1, OP2, and the long gyri of (LgI), with both LgI active in the right hemisphere tomical reconstruction procedure (Fig. 1 and Figs. S1–S3). Of and the posterior LgI active in the left hemisphere. OP3, OP4, as note, 48.8% of cortical leads (2,733 in the left hemisphere and well as short gyri of the insula (SgI) were poorly responsive (Fig. 2 3,112 in the right) were localized in sulcal areas [i.e., cortical and Fig. S4). OP1 and OP4 [corresponding to areas S2 (secondary regions characterized by a negative folding index (Materials and Methods somatosensory area) and PV (parietal ventral area), as defined by )], many of which are not accessible to ECoG recordings. Disbrow et al. (18)] are the main parts of classically defined SII. The sampling density maps computed for the two hemispheres High proportions of responsive leads were observed in the D (Fig. 1 and Figs. S2 and S3) demonstrate the extensive coverage motor system, including the primary motor area (maximum of cortical sheet, with peaks bilaterally in the middle temporal values around 85%), large sectors of dorsal and ventral premotor 2 gyrus and in the mesial temporal region (32 leads/cm in the left, cortex, and SMA. Anterior was reliably re- 2 48 leads/cm in the right hemisphere). Only the frontal and oc- sponsive in both hemispheres, approaching 60% in its most cipital tips of the hemispheres as well as the cortical crowns were rostral extent. Further posterior parietal cortex activations were poorly represented because of the obligatory orthogonal insertion biased in favor of the left hemisphere in and of electrodes and to the anatomical and vascular constraints (pres- in a region located between OP4 and area 44 resembling parietal ence of frontal bone sinus and superior sagittal sinus, respectively). rostroventral (PR) area, described by Disbrow et al. (19) in the

Fig. 1. Anatomical and functional data analysis. (A) Axial MR slice of a patient, after the coregistration with the postimplantation CT, on which the entire Y′ electrode implanted in the left hemisphere is visible. The reconstruction of the Y′ electrode is shown on the right side, with the leads numbered 1–18 from the tip, for a total distance of 50 mm. Leads are represented by sets of voxels, colored red if located in cortical gray matter and green otherwise. (B) Midthickness surface of the fs_LR brain template with all leads located in gray matter of the left hemisphere (from 49 patients) indicated as black dots. For four nodes, circles with 0.5 cm (thick line) and 1 cm (thin line) of radius represent the masks. (C, Upper) Average time frequency plot (100 trials) in response to median nerve stimulation in one lead of one patient. (C, Lower) Time course of the average gamma power (50–150 Hz) for the same channel, reported as z scores based on the prestimulus interval. Black asterisks indicate time bins with gamma power significantly exceeding baseline (P < 0.001). (D) Sampling density of the left hemisphere computed from data in B. The color scale is expressed in the number of recording leads per cm2. Cytoarchitectonic areas (1–4, 6, OP1–4, and 44) and anatomically defined areas (long and short gyri of insula) are indicated by white lines, and functionally defined regions are indicated in blue [confidence ellipses of phAIP, DIPSA, and DIPSM from Jastorff et al. (78)] or green [MT cluster from Abdollahi et al. (72)].

2of8 | www.pnas.org/cgi/doi/10.1073/pnas.1601889113 Avanzini et al. Downloaded by guest on October 2, 2021 PNAS PLUS

Fig. 2. Overall responsiveness maps. Overall responsiveness (responsive leads as a percentage of total explored leads per disk) maps for the left (A) and right (B) hemispheres. Only surface nodes with values exceeding 10% were shown. The same conventions as in Fig. 1 are used.

monkey. This area may correspond to OP6 of Amunts and co- optimal number of clusters (average silhouette value: 0.448). The workers (20, 21). Interestingly, a weak but reliable responsiveness centroids obtained for each cluster (Fig. 3B) depicted three syn- (around 20%) was found in right middle temporal (MT) cluster chronous patterns (strong-, middle-, and weak-phasic), charac- (22), extending dorsally into (MTG). terized by a steep rise after stimulus delivery, a peak in the third

This activation partially overlaps with hOc5 (23). time bin (20–30 ms) and a return to baseline level within 50 ms NEUROSCIENCE To rule out the strong responsiveness of motor and premotor from stimulation. A prolonged cluster (green line) exhibited a areas being due to the intensity of stimulation, just above motor similar time course in the rising phase, but persisted two time bins threshold, we computed overall responsiveness for a second set (20 ms) longer than phasic clusters. A completely different pattern of recordings, obtained in 70 patients, in which the intensity was was shown by a tonic cluster (blue line) lasting more than 200 ms, 20% below the motor threshold. No difference in the propor- whose peak was consistently lower and later than all other clusters. tions of active leads (overall responsiveness >10%) of sensory Quality of the clustering was indicated by the small percentage of (BA2) and motor (BA4) areas were found between the supra- negative silhouette values (less than 5% of leads, most of them and subthreshold data in the left hemisphere (χ2 = 3.97; not belonging to middle-phasic and prolonged clusters, whose peak significant). A significant difference (after correction for two amplitudes were very similar). These leads were excluded from comparisons) was found in the right hemisphere (χ2 = 14.75; P < subsequent cluster-related analyses. 0.01), but the proportion of active leads decreased more in BA2 The relative responsiveness (percentage of responsive leads than in BA4 (Table S1) with subthreshold stimulation. belonging to a cluster) maps showed a remarkable topographical Clustering (k means) the time course of gamma band activity distribution across the two hemispheres. As shown in Fig. 3 A–C, (Table S2) over all responsive leads indicated five to be the the strong phasic cluster (see red palette) covered the middle

Fig. 3. Clustering of time courses. (Center) Time courses (centroid ± SD) of the five clusters, as indicated (Inset). (Left and Right) Relative responsiveness (leads belonging to one cluster as a percentage of total number of responsive leads per disk) maps of left (L) and right (R) hemispheres (middle portion)for strong-phasic (yellow red), delayed-phasic (green), and tonic (blue) clusters. Only nodes with values exceeding 20% are shown. The same conventionsasin Fig. 1 are used.

Avanzini et al. PNAS Early Edition | 3of8 Downloaded by guest on October 2, 2021 strip of primary somatosensory areas (peaks of 81% in left 3b and 70% in right 3a), primary motor area and anterior part of inferior parietal cortex, corresponding to the caudal sectors of PF, PFt, and PFm (24, 25). The prolonged cluster (green palette) was re- stricted to a small sector of dorsal premotor area bilaterally (both reaching 33%), area 4 in the right hemisphere and SMA in the left hemisphere. The tonic cluster (see blue palette) was highly specific for secondary somatosensory areas (OP1 and OP2) and long gyri of insular cortex (peaks around 80% and 90%). Also, activity in human PR, ventral premotor cortex (bilaterally), and MT cluster (right hemisphere) exhibited a tonic time course. Middle- and weak-phasic clusters (Fig. S5) showed less focal distributions, in- stead spreading centrifugally from the with the middle-phasic cluster confined to closer areas and the weak- phasic cluster to more distant ones. Of note, these two clusters largely avoided the regions belonging to the three focal clusters. The region of interest (ROI) analysis confirmed that all four primary somatosensory areas are characterized by a strong-phasic time course, with areas 3a, 3b, and 1 similar in terms of peak amplitude (Fig. 4A). A repeated-measures ANOVA returned significant main effects for both time and ROIs, and a significant ROI-by-time interaction [F(3,177) = 2.787; P < 0.0001]. However, post hoc comparisons did not reveal any consistent difference across ROIs. The same analysis conducted on areas of the motor system (Fig. 4B) showed significant main effects for both time and ROIs, and a significant ROI-by-time interaction [F(4,236) = 6.545; P < 0.0001]. Interestingly, post hoc analysis identified three time intervals differentiating the motor ROIs: (i) around the peak re- sponse (synchronous for all ROIs), the medial premotor area shows significantly lower power values than area 4, putative hu- man anterior intraparietal area (phAIP), and dorsal premotor area (PMd); (ii) in the 50- to 70-ms interval, phAIP shows no sustained power, in contrast to all other ROIs, with PMd pre- senting the highest values; (iii) between 100 and 200 ms, power in the ventral premotor area (PMv) and the medial premotor area significantly exceeded that of PMd, probably reflecting the pres- ence of tonic cluster activity (Fig. 3). A repeated-measures ANOVA run on secondary somatosensory areas returned a significant ROI- by-time interaction [F(2,118) = 4.367; P < 0.0001]. Post hoc analysis showed a significant difference between OP1 and long gyri of insula, and between OP1 and other OPs combined in the 20- to 30-ms period, with OP1 presenting a stronger phasic response relative to other ROIs. In a late interval (100–200 ms), however, gamma power recorded from LgI exceeded that in opercular areas. The motor temporal pattern was tested also for the sub- threshold dataset, with the specific aim of validating that the sustained behavior of PMd between 50 and 70 ms was unrelated to thumb twitch induced by the median nerve stimulation. The Fig. 4. ROI analysis. Average (±SE) time course of leads in somatosensory ROI-by-time interaction again reached significance [F(4,236) = areas (A), motor regions (B), and opercular/insular areas (C). Areas are listed 1.294; P < 0.005] with subthreshold stimulation. Post hoc anal- (Insets). Black marks under the curves indicate significant post hoc compar- isons (P < 0.002). Number of responsive leads for each graph: 34 in area 1, 54 ysis, comparing mean gamma power across ROIs in the 50- to in area 2, 21 in area 3a, 50 in area 3b (A); 68 in area 4, 157 in PMd, 45 in 70-ms time window confirmed the prolonged activation in motor medial premotor areas, 42 in PMv, and 28 in phAIP (B); and 133 in OP1, 54 in and premotor areas relative to phAIP (Fig. S6). other OPs, and 38 in LgI (C). Clustering and ROI analyses have provided valuable insights into the dynamics of hand somatosensory processing. Both analyses have limitations because the first lumps large numbers plitude maps over time. Because this operation is performed for of leads together, and the second blends leads selected on a each time-bin, a movie that links together the entire sequence vi- priori anatomical criteria. To provide data-driven evidence, we sualizes the 4-dimensional dynamics of the cortical activity in left applied map computation directly to gamma power amplitude at and right hemispheres (Movies S1–S4). By inspecting four frames each time bin. In this way, we constructed a temporal sequence of such movie (Fig. 5) one can easily see: (i) the prevalence of the of maps, one every 10 ms, depicting the distribution of average strong phasic response in area 3b, 3a and 1 at 20–30 ms (Fig. 5A); amplitude (considering only responsive leads, see Fig. S7 A–C) (ii) the marked involvement of posterior part of dorsal premotor over cortex. To complement this information, we time-resolved the cortex visible around 40–50 ms after stimulation (Fig. 5B); (iii)the overall responsiveness maps (Fig. 2) to achieve time dependent dominance of premotor cortex and OP1 at 60–70 ms after stimu- responsiveness maps (Fig. S7 D–F). Both datasets contain distinct lation, when primary sensory areas are already on the decrease and independent information, related to variations in time of (Fig. 5C); and, finally, (iv) the restriction of late activity (more than strength and reliability of response. Hence, we combined them by 100 ms following the onset of the electrical stimulation) to OP1, means of a pointwise multiplication, thus obtaining weighted am- OP2 and LgI.

4of8 | www.pnas.org/cgi/doi/10.1073/pnas.1601889113 Avanzini et al. Downloaded by guest on October 2, 2021 PNAS PLUS

Fig. 5. Weighted amplitude maps over time. Four weighted amplitude maps of the left hemisphere at time bins indicated. The color range is dynamically NEUROSCIENCE adjusted from frame to frame. Note how the activity peaks in primary sensory areas in first frame, spreads to dorsal premotor area (second frame), remains active longer (third frame), whereas residual activity after 100 ms is recorded mostly from parietal (OP1) and long gyri of insula (last frame). The same conventions as in Fig. 1 are used.

Discussion Responsiveness Maps Document the Wide Extent of the Cortical In the present study, by localizing intracerebral electrodes in Somatosensory Network. The overall responsiveness (Fig. 2) and individual hemispheres, assessing the gamma broadband activity weighted amplitude maps (Fig. 5) reveal the cortical network of all recording leads, and mapping the results to a common involved in processing somatosensory stimuli and document its template, we obtained a comprehensive four-dimensional car- time course. This network encompasses several areas the in- tography of the cortical processing of median nerve stimulation. volvement of which remained controversial so far. Indeed, beside Beyond its neuroscientific relevance, this knowledge may allow a primary somatosensory cortices and SII/insula region (39), we more careful interpretation of the sEEG activity and a better revealed the activation of motor and premotor cortex (13), SMA design of the sEEG implantations based on the ictal clinical (10, 40), superior and inferior parietal lobules (12, 41, 42). Thus, semiology. despite the very stringent response criteria we used, computed This result was achieved thanks to the group analysis of in- maps indicate a widespread network (over 10% of the surface in tracranial recordings from a large number of patients. Similar both hemispheres) devoted to processing median nerve input. This attempts to solve the sparse-sampling problem have been re- result is even more notable given that the present report is limited cently made using ECoG (26–31). However, although useful for to stimulation of only one of the three nerves innervating the hand. investigating activity from areas near the crowns of gyri, such In addition, two additional regions were identified as responding recordings are blind to neuronal activity from deep regions (7). to median nerve stimulation. The first is located in both hemi- It is worth noting that half of the leads recorded in the present spheres in the same sector of parieto-ventral region that has been study by sEEG were localized in sulci, including major sulci for called human PR (19). This area appears to be the counterpart of somatosensory processing like the sylvian and the rolandic fis- monkey PR, which is connected to IPL, area 4, and premotor sures. Previous sEEG studies (32–35) reported maps derived cortex (19, 43). Disbrow et al. (18) reported that human PR from population analyses, but the limited number of patients responded to passive tactile stimulation, although in only 60% of prevented them from obtaining the full picture of cortical tested participants. Investigating the temporal relationship be- processing. tween areas S2 and PR using MEG, Hinkley et al. (44) concluded High-frequency broadband gamma activity (50–150 Hz) was that both S2 and PR contribute to the somatomotor integration extracted from sEEG as an index of cortical activity, being needed for manual exploration and object discrimination, with S2 spatially and functionally more specific than other frequency activity preceding that of PR. The tonic time course observed in bands and presumably reflecting population spiking activity human PR in the present study supports the involvement of this (36, 37), and thus allowing networks to be investigated at mil- region in these discriminative functions. lisecond time-scales (33). A comparison between gamma time The second site was revealed by leads exploring the ventral course and ERP responses from the same leads is presented in medial–superior–temporal area (pMSTv), part of the MT cluster Figs. S8 and S9. To obtain a precise spatial localization for (22), and neighboring posterior MTG in the right hemisphere continuous maps, we minimized the circular kernel size by using (Fig. 2B). This region is classically considered involved in high- the geodesic distance (38) and logistic weighting function. Of order , whereas its contribution to somatosensory note, masks centered at 1.5-cm geodesic distance were com- processing is controversial. van Kemenade et al. (45) reported pletely independent, having no shared nodes. that different neuronal populations in human middle temporal

Avanzini et al. PNAS Early Edition | 5of8 Downloaded by guest on October 2, 2021 (hMT)+/V5 process tactile and visual motion information. In clinical grounds, using seizure semiology, scalp-EEG, and neuroimaging as contrast, Jiang et al. (46) reported only a weak (2% of voxels) re- guide. Patients were fully informed regarding the electrode implantation and sEEG recordings. The present study received the approval of the Ethics sponse to tactile stimulation. Given the high sensitivity of in- ’ tracranial recordings, we were able to show that indeed somatosensory Committee of Ospedale Ca Granda-Niguarda (ID 939-2.12.2013) and in- formed consent was obtained. Intracerebral recordings were performed information reaches pMSTv, suggesting a possible integration be- according to sEEG methodology to define the cerebral structures involved in tween tactile and visual information (47). If pMSTv functions in the onset and propagation of seizure activity (65, 66). No seizure occurred, the same capacity as its monkey counterpart, this integration may no alterations in the sleep/wake cycle were observed, and no additional serve to control tracking of moving objects not only with eyes, as pharmacological treatments were applied during the 24 h before the ex- described by Newsome et al. (48), but also with arms (49). perimental recording. Neurological examination was unremarkable in all cases; in particular, no motor or sensory deficit was found in any patient. Four-Dimensional Maps Provide Previously Unidentified Information About Interareal Dynamics. There is an ongoing debate of whether Electrode Implantation. Most implantations were unilateral, because clinical SI and SII operate serially or in parallel (14, 15). Both views are evidence generally indicates the hemisphere generating the seizures. Only compatible with anatomical findings demonstrating that SI is 8 of the 99 patients were implanted bilaterally, resulting in a total of – 107 implanted hemispheres. A number of depth electrodes (range: 9–19; connected to SII via cortico cortical connections (50, 51) and average: 13) were implanted in different regions of the hemisphere using that various thalamic nuclei project in parallel to SI and SII stereotactic coordinates. Each cylindrical electrode had a diameter of 0.8 mm (52). Neuroimaging studies addressed this issue using causal and consisted of eight to eighteen 2-mm-long contacts (leads), spaced modeling approach, but with mixed results: Kalberlah et al. 1.5 mm apart (DIXI Medical, Besancon, France). (53) and Khoshnejad et al. (54) supported serial processing, Immediately after the implantation, cone-beam computed tomography whereas Chung et al. (55) and Liang et al. (56) the parallel (CBCT) was obtained with the O-arm scanner (Medtronic) and registered to model. Even intracranial studies yielded conflicting results. preimplantation 3D T1-weighted MR images. Subsequently, multimodal Barba et al. (57) reported a negative ERP peaking at 30 ms in scenes were built with the 3D Slicer software package (67), and the exact parietal operculum following the first negative component oc- position of each lead was determined, at the single patient level, looking at multiplanar reconstructions (68). Following clinical conventions, all leads are curring at 20 ms waveform, thus favoring the serial interpretation. identified by a letter corresponding to the electrode shaft, followed by a On the contrary, Karhu and Tesche (58) observed synchronous number starting from the tip of the electrode. neuronal population activity in contralateral SII area 20–30 ms after stimulation, coinciding in time with the first responses generated Median Nerve Stimulation. The day after the implantation, patients were in SI, thus supporting the parallel hypothesis. admitted to the neurology ward, to undergo clinical and neuropsychological In the present study, the phasic component in OP1 coincided in tests to functionally characterize the recording leads. To map leads involved time with the phasic activity in SI, thus favoring the view of a in processing hand somatosensory information, a median nerve stimulation parallel propagation of sensory information to the two areas. test was administered to the patients lying in bed with the eyes closed. The However, because of the 10-ms time bin of our analysis, this con- median nerve opposite to the recorded hemisphere was stimulated at the clusion should be taken with caution. Interestingly, our findings wrist, using 100 constant-current pulses (0.2-ms duration) at 1 Hz. The in- tensity and exact site of stimulation were varied until an observable thumb distinguished different patterns of activation in peri-sylvian areas, twitch was obtained. The motor threshold in our sample ranged from 3.2 to with OP1 presenting the strongest phasic component followed by a 5.8 mA. The stimulation intensity was set at 10% above the motor threshold. tonic pattern, and OP2, OP3, and OP4 exhibiting mainly tonic As a control, most of the patients (n = 70) were also tested with a stimulation activity (Fig. 4C). The relative timing of phasic and tonic activities intensity 20% below the motor threshold. in operculoinsular areas is consistent with previous MEG data (59, 60). Hence, OP1 appears to play the role of a hub dispatching Anatomical Reconstruction of Electrodes. The aim of this computational stage information to neighboring opercular cortices (61). To settle this was to precisely locate the recording leads in individual cortical surfaces using point, further studies are needed using for example functional the multimodal explorations performed in each patient, and import these connectivity techniques like cortico–cortical evoked potentials (62). locations into a common template. Each patient underwent a structural MRI (Achieva; Philips Medical Systems) before electrodes implantation. The T1 Finally, our data showed that insular cortex was consistently images were segmented using FreeSurfer software (69) to identify the pial activated by median nerve stimulation and presented a markedly and the white matter surfaces for the native mesh of each patient. The tonic time course (Fig. 3). This activation was limited to the quality of segmentation was verified by visual inspection of the resulting posterior part of insula, anatomically corresponding to the in- surfaces. The midthickness surface, i.e., the average surface lying between sular long gyri. This finding is in agreement with the metaanalysis the pial and white matter surfaces, was extracted. Moreover, starting from by Kurth et al. (63), which demonstrated that this part of the the pial surface, the sulci pattern was extracted through a procedure that insula is a specific sector related to somatosensory processing. evaluates the normalized geometric depth of each point [folding index 70)]. Our finding of a tonic activity of this sector may be related to the The more negative this parameter, the deeper the point is buried in a sulcus. involvement of posterior insula in the integration of somato- After the implantation, each patient underwent a volumetric brain CT, to locate precisely the recording leads using the artifacts generated on CT sensory inputs with other sensory modalities (64). images (Fig. 1B). The MRI and CT datasets were coregistered [FMRIB’s linear image registration tool (FLIRT), 6 df, mutual information (71)] to get the Materials and Methods anatomical brain data and the implanted electrodes into the same co- Participants. Stereo-EEG data were collected from 99 patients (52 male, 47 ordinate space. Thresholding of the CT signal intensity was used to segre- female) suffering from drug-resistant focal epilepsy. Only patients present- gate the recording leads (Fig. 1B) and to reconstruct their position in the CT ing with no anatomical alterations (n = 78) or with small abnormalities volume. During this process, the knowledge of the number of implanted outside of the sensorimotor areas (n = 21), as evident on MR, were included. electrodes, the number of leads on each electrode, and the electrodes’ entry Fifteen of the patients with positive MR showed alterations of the temporal and target points were used as constraints to ensure a reliable reconstruction. lobe [4 hippocampal sclerosis, 4 minimal periventricular nodular hetero- Subtracting white-matter from pial surfaces yielded a ribbon surface topia, 2 cavernomas of temporal pole, 2 focal cortical dysplasia (FCDII), and 3 corresponding to the gray matter thickness. Because the spatial resolution of post ischemic injuries]. Four patients presented alteration of the posterior CT images exceeded that of the MRI imaging and the CT images were reduced (no overlap with active ROIs reported in the study), of whom, to binary information (1: presence of a recording lead; 0: absence of it), the two patients had FCDII, one had cavernoma, and one had post-ischemic ribbon was oversampled at a resolution of 0.4 mm in all directions. The in- injury. Finally, one patient presented a FCDII of the , and one tersection of CT images with the ribbon surface detected the contact points patient had an anterior periventricular nodular heterotopia. located in the cortical gray matter (Fig. 1B, red dots). Thus, at this stage, a These patients were stereotactically implanted with intracerebral elec- lead located in the gray matter is represented by the number of 0.064-mm3 trodes as part of their presurgical evaluation, at the Centro per la chirurgia voxels it shares with the gray matter. For those leads in the gray matter, the dell’Epilessia “Claudio Munari.” Implantation sites were selected solely on centroid of the artifact was computed and projected onto the nearest nodes

6of8 | www.pnas.org/cgi/doi/10.1073/pnas.1601889113 Avanzini et al. Downloaded by guest on October 2, 2021 of the midthickness native surface, thus obtaining a single surface node and the secondary somatosensory areas (OP1, OP2, OP3, OP4, and LgI). These PNAS PLUS representing the projection of a lead located in gray matter. ROIs were defined either by cytoarchitectonical or functional criteria (see Finally, the individual midthickness surface was resampled to match the Continuous Maps). For each of these groups, we performed a two-way re- number of nodes (163.842) of the template (Fs-LR-average) and coregistered peated-measure ANOVA with time and area as factors. Significant interac- with this template using Freesurfer_to_fs_LR pipeline (brainvis.wustl.edu/wiki/ tions were explored in post hoc analysis with planned-comparisons t tests index.php/Caret:Download). In this step, single-lead representations were en- (P < 0.001 to account for the 50 time bins). hanced by arbitrarily adding six nodes (on the hexagonal mesh) surrounding the original single nodes on the native surfaces, to ensure that all leads in the Continuous Maps. To provide a continuous view of the topographic pattern of individual surfaces were maintained on the template surface. To quantify the active leads, we built a circular mask based on the geodesic distance between goodness of fit between the native meshes and the template, we computed two cortical points (76) (i.e., the minimum pathway within the gray matter two local indices [i.e., deformation and distortion, evaluated for each node of connecting the source and the target nodes). For each cortical node, we defined the mesh (72)], indexing the shrinkage and translation of each node during the the nodes within a 1-cm geodesic distance from the original node and transformation. These indices were used to calculate the surface area of nodes weighted the contribution of each node by a sigmoid function. Node weight in a group taking into account individual cortical surface area. was defined as a logistic function with unitary amplitude, a steepness of 2 and a In conclusion, the recording leads located in the gray matter are repre- midpoint at 7.5 mm. As a result, each node of the cortical mesh was associated sented in three different formats in the reconstruction procedure: as the with a collection of surrounding nodes (averaging 806 and 813 nodes for the number of 0.064-mm3 native voxels and as a fixed number (7) of nodes in right and left hemispheres), with all nodes within 5 mm of the origin maximally the native and in the template meshes. Whereas the first format provides the precise location in the cortical depth, the latter two provide precise in- weighted, whereas those between 5 and 10 mm were gradually reduced in formation about location in the cortical surface (at midthickness). weight, to avoid edge effects. By this approach, we computed five different functional variables: SEEG Data Recording and Processing. For each implanted patient, the initial i) Cortical sampling density [i.e., number of explored leads per cm2, using the recording procedure included the selection of an intracranial reference, fixed number (7) of nodes per lead and the average surface of a disk]. which was chosen by clinicians using both anatomical and functional criteria. ii) Overall responsiveness [i.e., the number of responsive nodes (Nr) as a The reference was computed as the average of two adjacent leads both percent of the number of explored nodes (Ne) within a disk]. Given the exploring white matter. These leads were selected time-by-time because they fixed number of nodes representing a lead (7), this variable is equivalent did not present any response to standard clinical stimulations, including to the number of responsive leads as a percent of the number of ex- somatosensory (median, tibial, and trigeminal nerves), visual (flash), and plored leads within a disk. This time-independent variable provides an ’ acoustical (click) stimulations. Nor did the leads electrical stimulation evoke overall picture of the cortical responsiveness, ranging from 0% to 100%, any sensory and/or motor behavior. The sEEG trace was recorded with a directly comparable to neuroimaging studies. Data were thresholded at

Neurofax EEG-1100 (Nihon Kohden System) at 1-kHz sampling rate. 10% so as to exclude the contribution of sparsely responsive regions. NEUROSCIENCE Clinicians visually inspected recordings to verify for ictal epileptic dis- iii) Relative responsiveness [i.e., the number of nodes (leads) exhibiting a charges (IEDs) during the stimulation protocol. In 87 patients, the ROIs were specific temporal response pattern in percent of the number of respon- devoid of any IED. In the remaining 12 patients, sparse IEDs were recorded sive nodes (leads) within a disk]. This variable indexes the degree to during the stimulation protocol. In these cases, however, false-positive re- which an area responds with a specific temporal pattern, information sponses are very unlikely because none of the patients included showed reflex to which neuroimaging studies are completely blind. Results were IEDs, and thus IEDs were not synchronized to the stimulus. This absence was thresholded at a 1/N value, where N represents the number of clusters, confirmed by visual inspection of the quality of the data averaged. to show only areas in which proportions of single clusters exceed chance. The recordings from all leads in the gray matter were filtered (band-pass: iv) The response amplitude for each time bin (i.e., the average z score – 0.015 500 Hz; notch: 50 Hz) to avoid aliasing effects and decomposed into across all responsive nodes in the disk). This variable indexes the – ’ time frequency plots using complex Morlet s wavelet decomposition. Power strength of a response, regardless of how reliably it occurs within a disk. in the gamma (50- to 150-Hz) frequency band was extracted in a window No minimum threshold was set for the z score, but results were masked extending from 100 ms before to 500 ms after the electrical stimulation, and for overall responsiveness exceeding 10%. subdivided into 60 nonoverlapping 10-ms bins. Following previous in- v) The weighted amplitude for each time bin, i.e., a combination of re- tracranial studies (35, 73, 74), gamma power was estimated for 10 adjacent sponse amplitude and overall responsiveness, depicting the variation nonoverlapping 10-Hz frequency bands. over the cortical surface of the response amplitude weighted by To identify the leads responsive to median nerve stimulation, the gamma the responsiveness: band power in each poststimulus bin was compared with baseline using a t       test. Significance was Bonferroni corrected for 50 comparisons (P = 0.001), Nr overall ΣA response ΣA weighted = , and to decrease the false-positive ratio, only leads with significant gamma Ne responsiveness * Nr amplitude Ne amplitude increases in at least three time bins were designated as responsive. To nor- malize data across patients and leads, power in poststimulus bins was where Ne is the number of explored nodes (leads) and Nr is the number of transformed into z scores relative to the prestimulus interval. responsive nodes (leads) in a disk. Computation was restricted to regions A k mean clustering was applied to the gamma band time course of the whose overall responsiveness was higher than 10%. responsive leads to group them according to the time course of the response, These maps were plotted using CARET software (77) (www.nitrc.org/projects/ independently of their location. Nineteen clustering procedures were performed caret) and directly compared with retinotopic regions in the occipital cortex imposing increasing numbers of clusters (3–20) and computing silhouette values (72) to confidence ellipses of rostral intraparietal sulcus (78), the most rostral of (75) to evaluate clustering validity. The optimal clustering was defined by the which corresponds to phAIP, and to cytoarchitectonic sensorimotor regions maximal average silhouette value, and leads with negative silhouette values (79), opercular areas (17), and area 44 (80). The subdivision between dorsal and were not considered in the evaluation of the optimal clustering using a re- ventral premotor areas was made according to Tomassini et al. (81). peated-measurement ANOVA with time and cluster as factors. t tests (P < 0.001 to account for 50 time bins) were used to evaluate post hoc comparisons. ACKNOWLEDGMENTS. The authors thank Dr. S. Raiguel for revising the Finally, to compare cortical areas, leads were grouped by their location in a English version of the manuscript. This study was supported by European ROI-based analysis to evaluate differences among primary sensory areas (3a, Research Council (ERC) “Cogsystem” project no. 250013 (to G.R.) and ERC 3b, 1, and 2), motor areas (4, medial premotor area, PMd, PMv, and phAIP), “Parietalaction” project no. 323606 (to G.A.O.).

1. Van Essen DC (2013) Cartography and connectomes. Neuron 80(3):775–790. 6. Cardinale F, Lo Russo G (2013) Stereo-electroencephalography safety and effective- 2. Kopell NJ, Gritton HJ, Whittington MA, Kramer MA (2014) Beyond the connectome: ness: Some more reasons in favor of epilepsy surgery. Epilepsia 54(8):1505–1506. The dynome. Neuron 83(6):1319–1328. 7. Noy N, et al. (2015) Intracranial recordings reveal transient response dynamics during 3. Hari R, Parkkonen L (2015) The brain timewise: How timing shapes and supports brain information maintenance in human cerebral cortex. Hum Brain Mapp 36(10):3988–4003. function. Philos Trans R Soc Lond B Biol Sci 370(1668):pii: 20140170. 8. Whitmer D, Worrell G, Stead M, Lee IK, Makeig S (2010) Utility of independent 4. Burle B, et al. (2015) Spatial and temporal resolutions of EEG: Is it really black and component analysis for interpretation of intracranial EEG. Front Hum Neurosci 4:184. white? A scalp current density view. Int J Psychophysiol 97(3):210–220. 9. Ferretti A, et al. (2007) Cortical brain responses during passive nonpainful median 5. Debener S, Ullsperger M, Siegel M, Engel AK (2006) Single-trial EEG-fMRI reveals the nerve stimulation at low frequencies (0.5-4 Hz): An fMRI study. Hum Brain Mapp dynamics of cognitive function. Trends Cogn Sci 10(12):558–563. 28(7):645–653.

Avanzini et al. PNAS Early Edition | 7of8 Downloaded by guest on October 2, 2021 10. Manganotti P, et al. (2009) Steady-state activation in somatosensory cortex after changes 46. Jiang F, Beauchamp MS, Fine I (2015) Re-examining overlap between tactile and visual in stimulus rate during median nerve stimulation. Magn Reson Imaging 27(9):1175–1186. motion responses within hMT+ and STS. Neuroimage 119:187–196. 11. Arienzo D, et al. (2006) Somatotopy of anterior cingulate cortex (ACC) and supple- 47. Ricciardi E, Bonino D, Pellegrini S, Pietrini P (2014) Mind the blind brain to understand mentary motor area (SMA) for electric stimulation of the median and tibial nerves: An the sighted one! Is there a supramodal cortical functional architecture? Neurosci fMRI study. Neuroimage 33(2):700–705. Biobehav Rev 41:64–77. 12. Ruben J, et al. (2001) Somatotopic organization of human secondary somatosensory 48. Newsome WT, Wurtz RH, Komatsu H (1988) Relation of cortical areas MT and MST to cortex. Cereb Cortex 11(5):463–473. pursuit eye movements. II. Differentiation of retinal from extraretinal inputs. 13. Boakye M, Huckins SC, Szeverenyi NM, Taskey BI, Hodge CJ, Jr (2000) Functional J Neurophysiol 60(2):604–620. magnetic resonance imaging of somatosensory cortex activity produced by electrical 49. Ilg UJ, Schumann S (2007) Primate area MST-l is involved in the generation of goal- stimulation of the median nerve or tactile stimulation of the index finger. directed eye and hand movements. J Neurophysiol 97(1):761–771. J Neurosurg 93(5):774–783. 50. Jones EG, Powell TP (1969) Connexions of the somatic of the rhesus 14. Rowe MJ, Turman AB, Murray GM, Zhang HQ (1996) Parallel organization of so- monkey. I. Ipsilateral cortical connexions. Brain 92(3):477–502. matosensory cortical areas I and II for tactile processing. Clin Exp Pharmacol Physiol 51. Jones EG, Powell TP (1969) Connexions of the somatic sensory cortex of the rhesus 23(10-11):931–938. monkey. II. Contralateral cortical connexions. Brain 92(4):717–730. 15. Forss N, Hietanen M, Salonen O, Hari R (1999) Modified activation of somatosensory 52. Almeida TF, Roizenblatt S, Tufik S (2004) Afferent pathways: A neuroanatomical cortical network in patients with right-hemisphere stroke. Brain 122(Pt 10):1889–1899. review. Brain Res 1000(1-2):40–56. 16. Kaukoranta E, Hämäläinen M, Sarvas J, Hari R (1986) Mixed and sensory nerve stim- 53. Kalberlah C, Villringer A, Pleger B (2013) Dynamic causal modeling suggests serial ulations activate different cytoarchitectonic areas in the human primary somato- processing of tactile vibratory stimuli in the human somatosensory cortex–an fMRI sensory cortex SI. Neuromagnetic recordings and statistical considerations. Exp Brain study. Neuroimage 74:164–171. – Res 63(1):60 66. 54. Khoshnejad M, Piché M, Saleh S, Duncan G, Rainville P (2014) Serial processing in 17. Eickhoff SB, Grefkes C, Zilles K, Fink GR (2007) The somatotopic organization of cy- primary and secondary somatosensory cortex: A DCM analysis of human fMRI data in – toarchitectonic areas on the human parietal operculum. Cereb Cortex 17(8):1800 1811. response to innocuous and noxious electrical stimulation. Neurosci Lett 577:83–88. 18. Disbrow E, Roberts T, Krubitzer L (2000) Somatotopic organization of cortical fields in 55. Chung YG, et al. (2014) Intra- and inter-hemispheric effective connectivity in the – the of Homo sapiens: Evidence for SII and PV. JCompNeurol418(1):1 21. human somatosensory cortex during pressure stimulation. BMC Neurosci 15:43. 19. Disbrow E, Litinas E, Recanzone GH, Padberg J, Krubitzer L (2003) Cortical connections 56. Liang M, Mouraux A, Iannetti GD (2011) Parallel processing of nociceptive and non- of the second somatosensory area and the parietal ventral area in macaque monkeys. nociceptive somatosensory information in the human primary and secondary so- – J Comp Neurol 462(4):382 399. matosensory cortices: Evidence from dynamic causal modeling of functional magnetic ’ 20. Amunts K, et al. (2010) Broca s region: Novel organizational principles and multiple resonance imaging data. J Neurosci 31(24):8976–8985. receptor mapping. PLoS Biol 8(9):e1000489. 57. Barba C, Frot M, Mauguière F (2002) Early secondary somatosensory area (SII) SEPs. 21. Amunts K, Zilles K (2012) Architecture and organizational principles of Broca’s region. Data from intracerebral recordings in humans. Clin Neurophysiol 113(11):1778–1786. Trends Cogn Sci 16(8):418–426. 58. Karhu J, Tesche CD (1999) Simultaneous early processing of sensory input in human 22. Kolster H, Peeters R, Orban GA (2010) The retinotopic organization of the human primary (SI) and secondary (SII) somatosensory cortices. JNeurophysiol81(5):2017–2025. middle temporal area MT/V5 and its cortical neighbors. J Neurosci 30(29):9801–9820. 59. Hari R, et al. (1993) Functional organization of the human first and second somato- 23. Malikovic A, et al. (2007) Cytoarchitectonic analysis of the human extrastriate cortex sensory cortices: A neuromagnetic study. Eur J Neurosci 5(6):724–734. in the region of V5/MT+: A probabilistic, stereotaxic map of area hOc5. Cereb Cortex 60. Simões C, Jensen O, Parkkonen L, Hari R (2003) Phase locking between human primary 17(3):562–574. and secondary somatosensory cortices. Proc Natl Acad Sci USA 100(5):2691–2694. 24. Caspers S, et al. (2006) The human inferior parietal cortex: Cytoarchitectonic parcel- 61. Mazzola L, Faillenot I, Barral F-G, Mauguière F, Peyron R (2012) Spatial segregation lation and interindividual variability. Neuroimage 33(2):430–448. of somato-sensory and pain activations in the human operculo-insular cortex. 25. Caspers S, et al. (2008) The human inferior parietal lobule in stereotaxic space. Brain Neuroimage 60(1):409–418. Struct Funct 212(6):481–495. 62. David O, et al. (2013) Probabilistic functional tractography of the human cortex. 26. Miller KJ, et al. (2007) Real-time functional brain mapping using electrocorticography. Neuroimage 80:307–317. Neuroimage 37(2):504–507. 63. Kurth F, Zilles K, Fox PT, Laird AR, Eickhoff SB (2010) A link between the systems: 27. Burke JF, et al. (2013) Synchronous and asynchronous theta and gamma activity Functional differentiation and integration within the human insula revealed by meta- during episodic memory formation. J Neurosci 33(1):292–304. analysis. Brain Struct Funct 214(5-6):519–534. 28. Conner CR, Chen G, Pieters TA, Tandon N (2014) Category specific spatial dissociations 64. Mesulam MM (1998) From sensation to cognition. Brain 121(Pt 6):1013–1052. of parallel processes underlying visual naming. Cereb Cortex 24(10):2741–2750. 65. Munari C, et al. (1994) Stereo-electroencephalography methodology: Advantages and 29. Davidesco I, et al. (2013) Spatial and object-based attention modulates broadband high- limits. Acta Neurol Scand Suppl 152(Suppl.c):56–67. frequency responses across the human visual cortical hierarchy. J Neurosci 33(3):1228–1240. 66. Cossu M, et al. (2005) Stereoelectroencephalography in the presurgical evaluation of 30. Dykstra AR, et al. (2012) Individualized localization and cortical surface-based regis- focal epilepsy: A retrospective analysis of 215 procedures. Neurosurgery 57(4):706–718. tration of intracranial electrodes. Neuroimage 59(4):3563–3570. 67. Fedorov A, et al. (2012) 3D Slicer as an image computing platform for the Quanti- 31. Kadipasaoglu CM, et al. (2015) Development of grouped icEEG for the study of tative Imaging Network. Magn Reson Imaging 30(9):1323–1341. cognitive processing. Front Psychol 6:1008. 68. Dale AM, Fischl B, Sereno MI (1999) Cortical surface-based analysis. I. Segmentation 32. Juphard A, et al. (2011) Direct evidence for two different neural mechanisms for reading – familiar and unfamiliar words: An intra-cerebral EEG study. Front Hum Neurosci 5:101. and surface reconstruction. Neuroimage 9(2):179 194. 33. Lachaux J-P, Axmacher N, Mormann F, Halgren E, Crone NE (2012) High-frequency 69. Reuter M, Schmansky NJ, Rosas HD, Fischl B (2012) Within-subject template estimation – neural activity and human cognition: Past, present and possible future of intracranial for unbiased longitudinal image analysis. Neuroimage 61(4):1402 1418. EEG research. Prog Neurobiol 98(3):279–301. 70. Van Essen DC, Drury HA (1997) Structural and functional analyses of human cerebral – 34. Ossandón T, et al. (2012) Efficient “pop-out” visual search elicits sustained broadband cortex using a surface-based atlas. J Neurosci 17(18):7079 7102. γ activity in the dorsal attention network. J Neurosci 32(10):3414–3421. 71. Jenkinson M, Smith S (2001) A global optimisation method for robust affine regis- – 35. Vidal JR, et al. (2010) Category-specific visual responses: An intracranial study com- tration of brain images. Med Image Anal 5(2):143 156. paring gamma, beta, alpha, and ERP response selectivity. Front Hum Neurosci 4:195. 72. Abdollahi RO, et al. (2014) Correspondences between retinotopic areas and myelin – 36. Popivanov ID, Jastorff J, Vanduffel W, Vogels R (2014) Heterogeneous single-unit maps in human . Neuroimage 99:509 524. selectivity in an fMRI-defined body-selective patch. J Neurosci 34(1):95–111. 73. Caruana F, et al. (2014) Human cortical activity evoked by gaze shift observation: An – 37. Ray S, Maunsell JHR (2011) Different origins of gamma rhythm and high-gamma intracranial EEG study. Hum Brain Mapp 35(4):1515 1528. activity in macaque visual cortex. PLoS Biol 9(4):e1000610. 74. Caruana F, Sartori I, Lo Russo G, Avanzini P (2014) Sequencing biological and physical 38. Kadipasaoglu CM, et al. (2014) Surface-based mixed effects multilevel analysis of events affects specific frequency bands within the human premotor cortex: An in- grouped human electrocorticography. Neuroimage 101:215–224. tracerebral EEG study. PLoS One 9(1):e86384. 39. Kaas JH (2004) Somatosensory System. The Human Nervous System, eds Paxinos G, 75. Rousseeuw PJ (1987) Silhouettes: A graphical aid to the interpretation and validation Mai JK (Elsevier Academic, New York), 2nd Ed, pp 1059–1092. of cluster analysis. J Comput Appl Math 20:53–65. 40. Korvenoja A, et al. (1999) Activation of multiple cortical areas in response to so- 76. Knutsen AK, et al. (2010) A new method to measure cortical growth in the developing matosensory stimulation: Combined magnetoencephalographic and functional brain. J Biomech Eng 132(10):101004. magnetic resonance imaging. Hum Brain Mapp 8(1):13–27. 77. Van Essen DC (2012) Cortical cartography and Caret software. Neuroimage 62(2): 41. Forss N, Salmelin R, Hari R (1994) Comparison of somatosensory evoked fields to 757–764. airpuff and electric stimuli. Electroencephalogr Clin Neurophysiol 92(6):510–517. 78. Jastorff J, Begliomini C, Fabbri-Destro M, Rizzolatti G, Orban GA (2010) Coding ob- 42. Gelnar PA, Krauss BR, Szeverenyi NM, Apkarian AV (1998) Fingertip representation in served motor acts: Different organizational principles in the parietal and premotor the human somatosensory cortex: An fMRI study. Neuroimage 7(4 Pt 1):261–283. cortex of humans. J Neurophysiol 104(1):128–140. 43. Padberg J, Disbrow E, Krubitzer L (2005) The organization and connections of ante- 79. Geyer S, Schormann T, Mohlberg H, Zilles K (2000) Areas 3a, 3b, and 1 of human rior and posterior parietal cortex in titi monkeys: Do New World monkeys have an primary somatosensory cortex. Part 2. Spatial normalization to standard anatomical area 2? Cereb Cortex 15(12):1938–1963. space. Neuroimage 11(6 Pt 1):684–696. 44. Hinkley LB, Krubitzer LA, Nagarajan SS, Disbrow EA (2007) Sensorimotor integration 80. Amunts K, et al. (1999) Broca’s region revisited: Cytoarchitecture and intersubject in S2, PV, and parietal rostroventral areas of the human sylvian fissure. J Neurophysiol variability. J Comp Neurol 412(2):319–341. 97(2):1288–1297. 81. Tomassini V, et al. (2007) Diffusion-weighted imaging tractography-based parcella- 45. van Kemenade BM, et al. (2014) Tactile and visual motion direction processing in tion of the human lateral premotor cortex identifies dorsal and ventral subregions hMT+/V5. Neuroimage 84:420–427. with anatomical and functional specializations. J Neurosci 27(38):10259–10269.

8of8 | www.pnas.org/cgi/doi/10.1073/pnas.1601889113 Avanzini et al. Downloaded by guest on October 2, 2021