Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Jun 19.
Published in final edited form as: Neuron. 2013 Jun 19;78(6):1116–1126. doi: 10.1016/j.neuron.2013.04.023

The relationship of anatomical and functional connectivity to resting state connectivity in primate somatosensory cortex

Zheng Wang 1,*, Li Min Chen 4,5,*, László Négyessy 2,3,*, Robert M Friedman 1, Arabinda Mishra 4,5, John C Gore 4,5,†, Anna W Roe 1,†
PMCID: PMC3723346  NIHMSID: NIHMS470732  PMID: 23791200

SUMMARY

Studies of resting state activity in the brain have provoked critical questions about the brain’s functional organization but, its biological basis is not clear. Specifically, the relationships between interregional correlations in resting state measures of activity, neuronal functional connectivity and anatomical connectivity are much debated. To investigate these relationships, we have examined both anatomical and steady state functional connectivity within the hand representation of primary somatosensory cortex (areas 3b and 1) in anesthetized squirrel monkeys. The comparison of three data sets (fMRI, electrophysiological, anatomical) indicate two primary axes of information flow within SI: prominent interdigit interactions within area 3b and predominantly homotopic interactions between area 3b and area 1. These data support a strikingly close relationship between baseline functional connectivity and anatomical connections. This study is also the first to extend findings derived from large-scale cortical networks to the realm of local mm-scale networks.

INTRODUCTION

Understanding the spatiotemporal modulation of ongoing activity in the ‘resting’ (unstimulated) brain is critical for interpretation of what is ‘signal’ and what is ‘noise’ in neural function (Gusnard and Raichle, 2001; Fox and Raichle, 2007; Vincent et al., 2007). These baseline hemodynamic signatures impact highly the interpretation of activated functional networks associated with different sensory, attentive, or cognitive states (Grecius et al., 2009; Honey et al 2009; Biswal et al., 2010; Deco et al., 2011; Smith et al., 2009). The link between resting state metrics and anatomical connectivity has largely been supported by modeling of areal correlations with known inter-areal connection patterns (Fox et al., 2005, 2006; Honey et al., 2007; Vincent et al., 2007; Luczak et al., 2009; Scholvinck et al., 2010; Deco and Jirsa, 2012); however, this relationship has not been examined directly with studies of anatomical connectivity (Matsui et al., 2011). The neuronal basis of resting state is also uncertain. Although hemodynamic based networks have been associated with widespread low frequency correlations in local field potentials (Arieli et al., 1996; Cohen and Kohn, 2011, Kenet et al., 2003), there is little evidence relating resting state connectivity with underlying neuronal connectivity. Moreover, as resting state studies have focused on broad cortico-cortical networks, little attention has been paid to resting state connectivity patterns at finer local cortical scales.

In this study, we seek to establish the relationship between anatomical connectivity, functional neuronal connectivity, and local resting state connectivity patterns revealed by fMRI. Our testbed for this study is the connectivity pattern of digit tip representations in somatosensory cortex (areas 3b and 1) of squirrel monkeys, an area central to manual behavior in monkeys and amenable to study with fMRI, electrophysiological, and anatomical connectivity techniques. This multimodal approach aims to establish an understanding of local (at the mm-scale) baseline networks revealed by resting state connectivity and, furthermore, provide evidence to support a local to global hierarchy of resting states within the brain.

RESULTS

Resting state connectivity

Resting state functional connectivity patterns of digit tip representations in primary somatosensory cortex (SI) were examined in eleven squirrel monkeys (one case shown in Figs. 1AB,D–F). Under isoflurane anesthesia, BOLD maps of digit activation (Fig. 1B) and the resting state acquisitions (static probe touching digit tip skin, Figs. 1D–F) were recorded using a 9.4T Varian MRI spectrometer. Seed locations (open blue squares, Fig. 1A–B) were selected based on optical imaging and/or electrophysiological maps of areas 3b and 1 (Fig. 1A); using surface vasculature as landmarks (arrowheads), these maps were readily coregistered to the maps acquired by MRI (cf. Chen et al., 2007). To examine functional connectivity, a seed (Fig. 1D, solid blue square) was placed at the voxel that responded strongly to D2 tip tactile stimulation in area 3b and voxels with significantly correlated resting state activity (r > 0.7) were identified (Fig. 1D). Significant correlated voxels were revealed within area 3b spanning adjacent digits (D1: white dots, D2: yellow dots, D3: red dots, D4: green dots). The voxels with highest significance in area 3b fell within the seed digit tip or adjacent digit tip representations. Within area 1, the largest number of significantly correlated voxels was more restricted along the mediolateral axis and fell largely within representations matching the seed voxel reflected across the area 3b/1 border (D2/D3 digit tip region). The observed connectivity pattern revealed predominantly digit-matched connectivity between areas 3b and 1 and within area 3b broad connectivity impinging on other digits. These results were not dependent on the specific seed chosen within a locale (Supp Figure 1).

Fig. 1.

Fig. 1

Open in a new tab

Resting state fMRI connectivity within SI cortex of squirrel monkeys. One case is shown in A–B, D–F; population data in C and G–I. (A) Electrophysiological map of digit region. Colored dots: see legend, digits 1–4, palm. Dotted lines: estimated inter-areal borders between areas 3a, 3b, and 1, and between hand/face. White arrows: the central and lateral sulci. Blue boxes: area 3a, area 3b and area 3b face seed regions. White arrowheads: vessel markers used for alignment with image in F (pink arrowheads). (B) BOLD activation in response to vibrotactile stimulation of D2 tip. Activated voxels are observed in areas 3a, 3b, and 1. Correlation maps were thresholded at r > 0.7 with a peak correlation value of 0.9. (C) Boxplot of correlation coefficient values between area 3b and area 1 (3b-1), area 3b and area 3a (3b-3a), area 3a and area 1 (3a-1), and with control locations (3b-ctl, 1-ctl, 3a-ctl). (D–F) BOLD correlation maps in resting state. Seed voxels (solid blue boxes) were placed in the digit regions in area 3b (C), area 3a (D), and face regions in area 3b (E) for voxel-wise correlation analysis. (G–I) Cross-animal (or population) correlation maps of seeds in area 3b (G), 3a (H), and face (I) regions. Correlations are summary of 18 runs (i.e. each map based on 18 seeds, seeds overlay D2, D3, or D4 digit tip) conducted in 10 animals. To average across animals, seed voxels were used to align all the cross-animal images. Correlation maps are centered on the seed region. Because seed location is relative to the imaging field of view across animals, there are some deviations in spatial location between average correlation map and individual cases.

To examine the specificity of this resulting fMRI correlation pattern, we generated two other resting state functional connectivity maps with seeds placed in area 3a (Fig. 1E) and in the face area of area 3b (Fig. 1F). These two seeds produced connectivity patterns quite distinct from the area 3b digit tip seed. The area 3a seed revealed strong functional connectivity with area 1 but not with area 3b, consistent with known anatomical connectivity (Huffman and Krubitzer, 2001; Jones et al., 1978). The seed in area 3b face area exhibited intrinsic connectivity within area 3b but little significant correlated activity elsewhere, also consistent with known anatomy (cf. Manger et al., 1997). This suggests that different seed locations reveal different connectivity patterns that parallel underlying anatomy.

Such patterns were observed across animals, as illustrated in summary resting state connectivity maps (Fig. 1G–I) and in the population pair-wise correlation coefficients (Fig. 1C). Cross animal correlation maps were generated by averaging the correlation r values on a voxel-by-voxel basis across 21 runs obtained from 10 animals. As imaging planes are tangential to the surface of the cortex, seed voxels (located in area 3b or area 3a D2, D3, or D4 digit tip locations or in control area 3b face location) were used to align images across animals. Despite the normal spatial variations of seed location relative to landmarks such as the central sulcus, consistent patterns were observed across cases. For the area 3b seed (Fig. 1G), there was clearly a focus of high correlation near the seed in area 3b and a secondary peak corresponding to area 1. The greater mediolateral extent of strong connectivity within area 3b (Fig. 1G, threshholded red region in area 3b) than within area 1 (Fig. 1G, theshholded red region in area 1) was consistent with that observed in individual cases. The population area 3a digit seed (Fig. 1H) revealed a large focus in area 3a and a second in area 1. The population face seed in area 3b revealed a single large region of high correlation (Fig. 1I, double star). These differences were further quantified, as shown in the boxplot of seed pair correlation coefficients in Figure 1C. Robustness of these patterns across studies are also bolstered by the high reproducibility across days (Fig. 2A–C and 2D–E) and across runs (Fig. 2G–I and 2J–L). Thus, across cases resting state connectivity exhibited similar patterns and were demonstrably not random in nature.

Fig. 2.

Fig. 2

Open in a new tab

Reproducibility of resting state connectivity patterns in SI cortex of squirrel monkeys. Resting state connectivity patterns to seeds placed in area 3b (A, D, G, J), area 3a (B, E, H, K) and a control face location (C, F, I, L) across different days (A–C: Day 1; D–F: Day 2) in monkey SM-M and across different runs (G–I: Run 1; J–L: Run 2) in monkey SM-P. Conventions as in Figure 1D–F. Scale bar: 2 mm.

Having established area-specific resting state connectivity patterns, we further evaluated whether digit-specific connectivity could be revealed. Figure 3A illustrates the correlation r values for all voxels within the imaged field of view for one case with a seed in area 3b D3. This reveals a range of r values with the highest (red) occurring over the matching digit location in area 1, weaker r values with other voxels within area 3b and with non-matching digits in area 1 (green). The weakest r values were between area 3b digit and control locations (blue). These findings were consistent across runs (Figure 3B). Voxel-voxel correlations between face seeds in area 3b produced highest values within local face voxels and not with digit voxels in area 3b, area 3a, or area 1/2 (Figure 3C, 3D). These differences in r values are not a by product of SNR variation, as the region of interest has relatively even intensity (Supp Figure 2). Furthermore, as shown in Fig. 4, seeds placed in voxels overlying other digits produced digit specific connectivity patterns that shifted in appropriate topographic fashion (A: D2 seed in area 3b with D2 in area 1, blue dotted line; B: D3 seed in area 3b with D3 in area 1, pink dotted line; C: D4 seed in area 3b with D4 in area 1, yellow dotted line; D: face seed correlates only with nearby face voxels).

Fig. 3.

Fig. 3

Open in a new tab

3D illustration of the cross-run reproducibility of local correlation patterns of digits and control regions. (A–B) 3D plot of correlation profiles of D3 in two separate runs. Underneath contour maps are thresholded 2D correlation patterns. (C–D) 3D plot of correlation profiles and contour maps of thresholded 2D correlation patterns of control face seeds in two separate runs. Corresponding thresholding values are indicated by color bar and numbers right next to the 2D contour maps. (E) Boxplots of ROI based, pair-wise correlation coefficient (r value) distributions among different ROI groups. *: p<0.001, paired-t-test.

Fig. 4.

Fig. 4

Open in a new tab

Seeds in different digits in area 3b produce corresponding shift in resting state connectivity patterns. (A–C) Thresholded (r = 0.6) 2D correlation patterns to seeds placed in D2, D3 and D4. (D) Correlation pattern of a seed in an area 3b control face region. (E) Digit maps based on electrophysiology. Different color patches represent different digits in both area 3b (left) and area 1 (right). Dotted color lines indicate the location of the seed of different digits in area 3b. Dotted black lines represent the estimated inter-areal border and hand-face border.

Across 3 runs in one animal (high resolution data from one animal), we found on average different connectivity r values between different cortical loci: the strongest occurred between matching digits in area 3b and area 1 (Fig. 3E: 2), intermediate values between different digits within area 3b (Figure 3E: 3), lower values between non-corresponding digits in area 3b and area 1 (Fig. 3E: 4, and lowest values between area 3b digit and control areas (Fig. 3E: 1). These statistically significant digit-specific differences between seed pairs (p < 0.001) suggest that one can evaluate fine structure in anatomical connectivity using resting state connectivity patterns.

Anatomical connectivity

To examine whether the resting state pattern of area 3b parallels anatomical connectivity, we made single focal anatomical tracer injections (BDA) into functionally identified digit tip locations in three other squirrel monkeys. Since squirrel monkeys are small (600–900 g) animals and can only undergo a limited number (1–3) of procedures involving recovery from surgery or anesthesia, we have collected different datasets in separate animals. Electrophysiological mapping revealed the orderly digit topography in area 3b and area 1 (Fig. 5A). Consistent with our previous studies (Chen et al., 2001; Friedman et al., 2004), optical imaging of cortical activation in response to stimulation of single digit tips revealed two activation spots, one in area 3b and one in area 1 (response to D2 stimulation shown in Fig. 5B). A focal injection of BDA confined to the single digit tip representation (< 500 μm in diameter) (Fig. 5C) was made in the D2 tip location; and the resulting cellular label was reconstructed (Fig. 5D, Fig. 6). The injection resulted in heavy labeling of cells (orange and yellow) near the injection site in area 3b as well as patchy label distant (green and blue) from the injection site in the hand area in area 3b (Fig. 5D, see also Fig. 6). These included adjacent digit locations within 3b in distal D1, D3, and D4. Heavy label was also observed in area 1 predominantly in a D2/D3 region with heavy focus in the tip representation zone (Fig. 5D). Consistent with reciprocal connectivity patterns in somatosensory cortex, BDA labeled axonal terminal patches (Fig. 6) were also observed to share a similar pattern of connectivity (Negyessy et al., 2013). Thus, the labeling in this case suggests topographically widespread inputs from other digit locations within area 3b, and relatively mediolaterally restricted inputs from area 1, from largely topographically matched locations. This differential intra- vs. inter-areal pattern of labeling was also seen in two other cases (Figs. 5E–F, see also Fig. 6). Thus, anatomical connections were characterized by two primary axes of information flow (broad intra-areal, curved red arrows in 5D–F, and comparatively focused inter-areal connectivity, straight red arrows in 5D–F). This pattern was consistent with the strong digit matched resting state connectivity between area 3b and area 1, the weaker but distinct connectivity between different digits within 3b, and the even weaker connectivity between non-matching digits between area 3b and area 1 (Fig. 3E).

Fig. 5.

Fig. 5

Open in a new tab

Anatomical connectivity of digits revealed by injections in area 3b. Following optical imaging and electrophysiological mapping, in each case, a focal injection of BDA was made into a digit tip location and resulting label distribution was reconstructed. (A) Electrophysiological mapping of digit region in areas 3b and 1. Same field of view as in B–D. (B) T-map of optical image obtained in response to vibrotactile stimulation of D2, revealing activation in area 3b and in area 1 (white arrows). Red pixels: p < 0.01. (C) Injection of BDA into D2 location in area 3b. Arrows: sites of strong BDA label in area 3b and in area 1; lowest arrow is injection site. Label in area 1 is not visible at this magnification (see panel D and Fig. 6). Same arrow positions are shown in A–D for reference. Small circles: mark vessel lumen, used for alignment with blood vessel image. (D) Resulting cellular densities (see Methods) from injection in C. Color scale shown at left; high to low cell density: orange, yellow, green, blue. (E–F) A second (E) and third (F) case illustrating BDA label distribution following injection of D2 tip in area 3b. Each case reveals mediolateral intra-areal axis of label and largely digit-matched inter-areal label. Scalebar: 1 mm. m: medial, p: posterior, applies to all.

Fig. 6.

Fig. 6

Open in a new tab

Photomicrographs showing anterograde and retrograde BDA labeling. (A–C) Inverse bright field micrographs providing dark field-like view of BDA labeling following injections in D2 tip of area 3b in three squirrel monkeys. Arrows identify D2t injection locations. High density labeling at projection sites (white patches, shown by the arrows) is seen in area 1 as well as in other locations in area 3b (arrowheads). Approximate depths of the sections are indicated in the upper right corner. A1/A2: areas 1 and 2; CS: central sulcus; D2t and D4t: distal finger pad of the 2nd and 4th fingers. Rostral is top; lateral is left. Box in C: region from which D–F are taken. (D–F) Distributions of labeled cells, processes, and bouton-like structures seen at high power under bright field were reconstructed. (D) Patchy labeling contained numerous retrogradely labeled cells and a dense meshwork of anterogradely labeled axons. Note the details of the dendritic branches of the retrogradely labeled neurons. (E) BDA labeled long range axonal fibers. (F) High power microscopic image shows anterogradely labeled axon terminal like structures. Scale bars represent: 1 mm in A–C; 50 μm in D,E and 25 μm in F.

Functional connectivity

These patterns of connectivity within area 3b and between area 3b and 1 also were supported by electrophysiological recordings of steady state neuron-neuron interactions in four other squirrel monkeys. Following optical imaging and electrophysiology mapping (Fig. 7A–B), on separate electrodes, single units were isolated in the digit tip representations (D2, D3, and D4 tips) of area 3b and area 1. Area 3b-area 1 (A3b-A1) pairs were either same-digit or adjacent-digit pairs; area 3b-area 3b (A3b-A3b) pairs were all adjacent-digit pairs. To achieve sufficient spike numbers, we elevated spontaneous activity by placing a static tactile probe on the digit tips (raster plots and histograms shown in Supp Fig. 3A); this is a common method of increasing spike rate for functional connectivity studies that does not affect the nature of the neuronal correlation in primates (Kaas et al., 1990; cf Bruno and Sakmann, 2006; Hung et al., 2010; Luczak et al., 2009; similar results were obtained with true, i.e no probe, spontaneous conditions, see Supp Fig. 3DE). Such baseline activity of each neuron was collected and neuronal interaction assessed with cross correlation methods (see Supplementary Methods).

Fig. 7.

Fig. 7

Open in a new tab

Recording in squirrel monkey primary somatosensory cortex. (A) Electrophysiological mapping in 3b and 1 overlaid on blood vessel map. Inset: Color code for distal, middle, and proximal phalanges on each digit. Star shows localization of the tracer injection used for anatomical studies. Scale bar: 1 mm. M: medial, P: posterior. (B) Left panel: Intrinsic signal optical image of the same cortical field of view in response to tactile stimulation of distal D2 pad. Two activation spots are revealed, one in area 3b and one in area 1 (red arrows). Right panel: Outline of D2 and D4 activations overlaid on the blood vessel image. Red dashed lines indicate approximate areal borders. (C–F) Interaction characteristics of the CCGs of neuronal pairs across the population. Population correlograms for (C) inter-areal interactions and (D) intra-areal interactions. Standard deviations indicated by colored shaded areas. Both shuffled and simulated spike trains produced flat cross correlograms. (E,F) Histograms of asymmetry indices of (E) inter-areal and (F) intra-areal correlograms. For inter-areal asymmetries, both positive (feedforward, solid lines) and negative (feedback, dotted lines) are plotted. As there is no directionality in interdigit interactions in area 3b, all intra-areal asymmetries are defined as positive. Distributions are significantly shifted to the right, indicating presence of directionality in functional connectivity (Wilcoxon signed-rank test, same digit pairs: p < 0.001, n = 160; different digit pairs p < 0.001, n = 153).

We observed robust intra-areal and inter-areal synchronized neuronal firing in this cortical network. Typical of cortical correlograms (CCGs), peaks tended to be broad (approximately 50 ms in width at half height) and centered on zero (individual examples shown in Supp Fig. 3B, blue lines). Both shuffled and simulated spike trains produced flat cross correlograms (Supp Fig. 3B, red lines). Population correlograms are shown in Fig. 7C and 7D. We examined whether there were any differences in connectivity between same-digit and adjacent-digit pairs. Of the A3b-A1 pairs that exhibited significant correlogram peaks (A3b/A1 same digit: 50.5%, 160/317; A3b/A1 adjacent digit: 48.7%, 153/314), we found the stronger interactions (as measured by both peak size and peak area) in the A3b-A1 same digit than adjacent digit population (Fig. 7C, blue: A3b-A1 same digit, red: A3b-A1 adjacent digit, p < 0.001; grand mean peak values same-digit: 0.080, adjacent-digit: 0.065). Frequency histograms of the peak strengths, together with their empirical cumulative distributions further corroborated the peak size differences (Supp Fig. 3C). In sum, these comparisons revealed that, as a population, same digit interactions were stronger than different-digit interactions between area 3b and 1, consistent with the relative robustness of same-digit interaction revealed anatomically and with resting state fMRI.

CCGs also revealed robust intra-areal interactions in area 3b (all recorded between adjacent D2–D3 or D3–D4 digit pairs). Although somewhat weaker than inter-areal interactions (grand mean peak 0.058, Wilcoxon rank sum test, p < 0.05), significant interactions were observed in roughly half of the pairs (area 3b/3b: 45.6%, 63 out of 138 pairs). This is illustrated in the population correlogram in Fig. 7D. Thus, in an area which traditionally has been considered highly topographic, these data underscore surprisingly prominent inter-digit interactions in area 3b, and is consistent with the intra-areal connectivity patterns revealed by anatomy and resting state fMRI.

We next examined whether there was any evidence for directionality in the population of interactions between area 3b and area 1. Consistent with previous studies on cortical neuronal interactions (Steinmetz et al., 2000; Roe and Ts’o, 1999), most neuronal interactions in SI were centered on zero; that is, the bulk of CCGs had peaks occurring at zero. However, we noticed that there were often asymmetries in the overall shape of CCGs, indicating that of the neuronal interactions that were directional, more were in the positive than negative direction (i.e. more feedforward than feedback interactions). To quantify these impressions across the population, for each CCG, we computed an asymmetry index [ASI = (R − L)/(R + L), where R and L are the number of interactions to the right and to the left of zero]. This index ranges from −1 to 1, with larger numbers indicating greater asymmetry, where a value of 0.33 indicates the distribution to the right of zero is twice that to the left of zero. This index indicates the directionality of the population of coincidences within a CCG and is not the same as peak position.

For both same-digit (Fig. 7E, blue) and adjacent-digit (Fig. 7E, red) populations of A3b-A1 pairs, the distributions of ASI of individual CCGs were significantly shifted to the right (Wilcoxon signed-rank tests: p ≪ 0.001, same digit: median value = 0.07, n = 160 pairs; adjacent-digit: median = 0.06, n = 153), suggesting an overall feed-forward direction from area 3b to area 1. There were no significant differences in ASI distribution between same-digit (blue) and adjacent-digit (red) inter-areal pairs (Fig. 7E, p > 0.1). Thus, although the strongest interactions appear to be due to common input (i.e. correlograms are centered on zero), for coincidences slightly weaker in strength (i.e. away from 0), more occur with positive than negative latency. This population bias is consistent with a predominance of feedforward interactions.

We also examined directionality in the intra-areal A3b-A3b population. All of these pairings were between adjacent digits. For all 3b-3b pairs, we defined all asymmetries as positive (biased to the right, since there is no expected difference between e.g. D2–D3 vs D3-D2) and combined all pairs into a single histogram (Figure 7F). We found that the ASI distributions exhibited a strong positive bias (p ≪ 0.001, n = 63 pairs of A3b-A3b, median value 0.20). What is interesting here is that we did not obtain symmetric peaks, which suggests that 3b-3b interactions are less likely to be due to common input and that a large portion of the interactions are directional (from one digit to the adjacent digit). Furthermore, the fact that this intra-areal asymmetry is so prominent, significantly more so than that between 3b-1 (7E, p ≪ 0.001), suggests a strong lateral flow of intra-areal information. In summary, these neuronal interactions are consistent with and extend the interpretation of anatomical and resting state connectivity patterns.

DISCUSSION

The functional connectivity patterns within and between areas 3b and 1 are consistent with the strongly mediolateral and anteroposterior axes of anatomical labeling and resting state connectivity patterns. Previous studies have suggested that global resting state connectivity is anchored by anatomical connectivity. For the first time, our study establishes an anatomical basis for resting state networks at a local mm-based scale. Furthermore, while previous studies have linked resting state networks to broad-based (<0.1 Hz) functional connectivity, no study has related resting state networks to functional interactions at the single neuron level. We suggest this fine scale spatial and temporal interaction comprises one level of a local to global multiscale hierarchy in resting brain states.

Relationship of functional interactions to anatomical connections

Figure 8 summarizes the common resting state interactions found across the BOLD-based, anatomical, and neuronal connectivity datasets. All three datasets reveal a strong same-digit interaction between area 3b and area 1 (Fig. 8, straight red arrow from area 3b to area 1) and all three datasets reveal interdigit interaction within area 3b (Fig. 8, curved red arrows). Thus, these two prominent interaction types underie two axes of information flow: an antero-posterior axis between area 3b and 1 and a mediolateral axis within area 3b. In addition, there are weaker interactions present between area 3b and 1 that are not digit-specific (Fig. 8, thin straight arrows).

Fig. 8.

Fig. 8

Open in a new tab

Summary schematic of common resting state interactions between area 3b and area 1 found across BOLD-based, anatomical, and neuronal connectivity datasets. All three datasets reveal a strong same-digit interaction between area 3b and area 1 (straight red arrow from area 3b to area 1) and all three datasets reveal interdigit interaction within area 3b (curved red arrows). Thus, there are two axes of information flow: an antero-posterior axis between area 3b and 1 and a mediolateral axis within area 3b. In addition, there are weaker interactions present between area 3b and 1 that are not digit-specific (thin straight arrows).

Inter-areal interactions

The asymmetry of the A3b-A1 CCGs indicate a feedforward bias in steady-state interactions (Fig. 8, straight red arrow from area 3b to area 1). For inter-areal interactions, we observed a significantly greater interaction strength for same-digit (Fig. 8, heavy red arrow) than adjacent-digit interactions (Fig. 8, thinner red arrows). We suggest that this is consistent with the density of anatomical connectivity. That is, since anatomical connections are more robust for same-digit locations in 3b and 1, these would underlie the most direct and strongest interactions. Those between different digits may be mediated by a smaller proportion of direct anatomical connections or by indirect interactions between digits within area 1, resulting in weaker overall functional interactions.

Intra-areal interactions

Contrary to the traditional view that area 3b neurons have receptive fields confined to single digits, an increasing number of reports in anesthetized and awake monkeys suggest a significant level of interdigit integration of tactile input (Reed et al., 2008; Chen et al., 2003; Lipton et al., 2010). The prevalent interdigit interactions found in this study (Fig. 8, curved red arrows) are consistent with the proposal that such interdigit interactions are mediated by intra-areal anatomical connections. Indeed, not only are interdigit interactions prevalent, they occur with significant peak asymmetry, potentially implicating the role of intrinsic horizontal connections within areas. Although it is difficult to infer specific circuitry from cross correlation studies, the presence of prominent asymmetry in 3b-3b interactions is suggestive that, in addition to common input, intrinsic horizontal connections within 3b may contribute strongly to intra-areal interdigit interactions.

Another possible source of the comparatively greater intra-areal (than inter-areal) asymmetry may relate to the fact that intrinsic connections are considered slowly conducting (0.05–0.5 m/s, resulting in longer latencies and more asymmetric correlograms), while inter-areal interactions are considered fast-conducting (3–20 m/s, resulting in shorter latencies and less asymmetric correlograms) (Bringuier et al., 1999; Girard et al., 2001). Consistent with this, a large proportion of axons coursing from area 3b to area 1 are apparently myelinated fibers, while those within area 3b or area 1 are unmyelinated axons (not shown).

In sum, the functional correlations observed within area 3b and between area 3b and 1 are consistent with the observed anatomical connections. While functional interactions assessed by CCGs may be due to either direct or indirect connectivity, the anatomical connectivity would contribute strongly to the observed functional biases. Under steady-state conditions, the asymmetry in functional interactions suggests a prominent bias of information flow from area 3b to area 1, especially for same digit interactions. Intra-areal interactions comprise a prominent orthogonal direction of information flow. These findings add to our understanding of the relative strengths of interaction and the overall direction of information flow within SI.

This view of steady-state functional connectivity patterns in SI will be relevant for interpreting data obtained under conditions of tactile stimulation and manual behavior (cf. Hung et al., 2007, 2010). These connection patterns suggest that intra-areal and inter-areal connections mediate distinct functional transformations, and may play differential roles in manual behaviors requiring digit-specific integration vs. interdigit coordination (e.g. multifinger tasks and exploration) (Johansson and Flanagan, 2009; Keysers et al., 2010).

Implications for resting state networks

The concept that baseline functional correlations are based in anatomical connectivity is relevant to the large ‘resting state’ literature. Although the exact relationship between anatomical connectivity and functional connectivity remains elusive at multiple levels, there is consensus that baseline functional connectivity does reflect to some extent anatomical connectivity patterns (e.g. (Vincent et al., 2007; Honey et al., 2009); for review, see (Deco and Corbetta, 2011; Behrens and Spons, 2011). Largely based on analyses of BOLD signals collected in fMRI studies, this literature suggests that functional circuits in the baseline state have inherent biases in their interactions within brain networks. External sensory stimulation then interacts with this baseline state, resulting in various network modulations (e.g. switching between or selecting amongst different cortical networks or otherwise ‘pushing’ the network into an alternative state). Comparisons between such functionally defined connectivity networks in the resting and activated states have further emphasized the notion that activated connectivities arise from anatomically based connectional specificity (Matsui et al., 2011).

Our data also support this view. Typical of CCGs, peaks are centered on zero indicating a predominance of common input, arising either from thalamic inputs or other cortical sources. This pedestal of common input is accompanied by a prominent, feedforward direction of information flow, as indicated by the strength of inter-areal interactions (Fig. 7C) and the predominantly positive asymmetry indices (Fig. 7E). In visual cortex, V1–V2 interactions are on average stronger than V1-V1 interactions, reflecting the larger degree of spatial integration in V2 and concomitant larger network size (Hung et al., 2010; Livingstone and Hubel, 1984; Ts’o and Gilbert, 1988; Roe and Ts’o, 1999). In somatosensory cortex, greater inter-areal integration may also be expected due to the larger receptive field sizes in area 1 than area 3b. Steady-state intrinsic interactions within area 3b and within area 1 may also provide a baseline configuration upon which sensory stimuli or other active states are further elaborated (Reed et al., 2008; cf Steinmetz et al., 2000). Thus, sensory stimulation may further enhance the pre-existing biases, producing in SI a strongly feedforward direction of information flow in the stimulated state. Such hypotheses have been supported by studies of macroscale networks. This study now extends these ideas to the local microscale network.

EXPERIMENTAL PROCEDURES

Animal preparation

Eighteen squirrel monkeys (fMRI: 11 monkeys, anatomy: 3 monkeys; electrophysiology: 4 monkeys) were anesthetized with ketamine hydrochloride (10 mg/kg)/atropine (0.05 mg/kg) and maintained with isoflurane anesthesia (0.8–1.1%) delivered in a 70:30 O2/N2O mixture. All procedures were in compliance with and approved by the Institutional Animal Care and Use Committee of Vanderbilt University.

MRI methods and data analysis

All MRI scans were performed on a 9.4-T Varian Inova spectrometer (Varian Medical Systems, Palo Alto, CA, USA) using a 3-cm surface coil. T2-weighted oblique structural images (TE: 16; TR: 200 ms) at 78×78 ×1000 μm3 resolution were acquired and coregistered with fMRI maps and with blood vessel maps. Functional MRI data acquired from the same slices using a gradient echo planar sequence (TE: 16 ms; TR: 1.5 s) at voxel sizes of 575×575×2000 μm3 (and for one case at 275x275x2000 μm3) were reconstructed and imported into Matlab (Mathworks, Natick, MA, USA) for analysis. Within each imaging session, both tactile stimulus-driven and resting state BOLD images were acquired. Eighteen sets of resting state fMRI data were acquired from eleven anesthetized squirrel monkeys. Determination of seed voxels in area 3b and 3a in each animal were based on stimulus-driven fMRI activation maps and the available electrophysiology maps for each animal. Voxel-wise correlation was calculated and then thresholded at r>=0.7 for display.

Stimulus protocol

Fingers were secured and tactile stimulation (8 Hz vibrotactile stimulation) of fingerpads was delivered with a piezoceramic device. In fMRI runs, stimuli were presented in 24 sec on and off blocks, 300 imaging volumes per run. For electrophysiological recordings, to achieve sufficient spike numbers, the stimulus probe remained in contact with the skin with a constant displacement, thereby achieving a steady state firing level. The data length of each steady state epoch was 650 ms long and collected in sessions of 100 ~ 300 trials; these trials were randomly interleaved with single- and dual-site stimulation of the digits.

Electrophysiological recording and analysis

Single units were isolated online and sorted (Plexon Inc.). Spike synchrony was measured by simultaneous recordings of single units isolated on separate electrodes. Three types of area 3b (A3b) and area 1 (A1) unit pairs were collected: A3b-A3b, A3b-A1 same-digit, and A3b-A1 adjacent-digit. All A3b-A3b pairs were from adjacent digits. The temporal resolution of spikes was 1 ms and response histograms were constructed with 5-ms time bins. 100–300 trials (repetitions) were collected. Joint PSTH were generated. The level of synchrony above or below chance was computed by subtracting the shift-predictor correlogram from the raw correlogram (Aertsen et al., 1989; Brody, 1999ab). Cross-correlograms (CCGs) and their 95% confidence intervals were computed using a 500 ms window, ± 250 ms around a lag of 0 ms. CCG peaks were counted as significant if two consecutive values exceeded the confidence intervals within ± 50 ms lag (Cohen et al., 2010). CCGs were normalized for differences in firing rate (Brody, 1999ab) and shuffle corrected (Perkel, 1969). Additionally, we further assessed significance of correlation by synthesizing thousands of artificial spike trains based on recorded spike times (random permutation approach) and calculating deviation from this baseline distribution. The correlation strength (CS) ( Takeuchi et al., 2011) was defined as CS = R + L, where R and L indicate the summed bins on the right and left sides of each CCG within ±50 ms from the center bin (0 ms). An asymmetry index (ASI) was defined as ASI = (R − L)/(R + L). A peak weighted to the right suggests prevalence of feedforward interaction, one weighted to the left suggests prevalence of feedback interaction, and one with equal left and right weights suggests common inputs or recurrent connections. For population comparisons, the nonparametric Wilcoxon test (Kruskal-Wallis test for group comparison) was used to determine significant differences (p < 0.05) between the cumulative distributions of peak-correlation coefficients, the correlation strength (CS) and the asymmetry index (AI).

Tracer injection and anatomical reconstruction

Focal injections of tracer were made in digit tip locations in area 3b and area 1, as determined by optical imaging and electrophysiological recording. We injected through glass micropipettes with tip inner diameter of 15–20 μm a 1:1 mixture of 10% biotinylated dextrans via iontophoresis (3 μA, 7 seconds on/off cycle, 20 minutes) at depth of 400 μm. After 10–20 days survival, animals were given an overdose of Pentobarbitol (100 mg/kg) and perfused transcardially with fixative. The cortical region of interest was removed, flattened and cut at 50 μm. The most superficial layer containing surface vasculature was used to align sections with optical images (e.g. Kaskan et al., 2009). Maps of the BDA label were made using Neurolucida® (MicroBrightField Europe, E.K. Magdeburg, Germany) and an Olympus microscope equipped with a motorized stage. Density measurements of the retrograde labeling were performed by Voronoi tessellation (http://mathworld.wolfram.com/VoronoiDiagram.html), an algorithm that generates areas inversely related to the density of the BDA labeled neurons. Two dimensional density plots were then computed by averaging the logarithm of the Voronoi areas and color-coding the density values by 2 units of standard deviation (sd) above and below the average (Negyessy et al., 2013). Labeling around the injection site of 250–300 μm diameter was omitted from analyses.

Supplementary Material

01

Supp Fig. 1. Correlation pattern not dependent upon specific seed voxel within a locale. (A–C) Correlation patterns of three neighboring seed voxels placed at the peak voxel (red arrow, B), and one voxel above (A) and one below (C) in area 3b (left columns). Images in right column show the correlation contour maps thresholded at r values of 0.6 and 0.8. (D–F) Correlation patterns of three neighboring seed voxels placed at the peak voxel (red arrow, B), and one voxel above (A) and one below (C) in area 1 (left columns). Images in right column shows the correlation contour maps thresholded at r values of 0.6 and 0.8.

Supp Fig. 2. Uniform intensity field. To illustrate uniformity of intensity in our imaged maps, we show a representative EPI intensity map (average SNR is around 180) along with corresponding T2* weighted structural image below. This shows that the SNR is reasonably even within the field of view. EPI intensity map of the SI region (A) and its corresponding high-resolution T2* weighted structural image (B).

Supp Fig. 3. Electrophysiological recordings and analysis. (A) Steady state firing of neurons in A3b and A1. Top: raster plots of units on D2 (red) and D3 (yellow) in A3b and A1. Bottom: PSTHs of this steady state activity. (B) Example correlograms. F1: A3b-A1 same digit. F2: A3b-A1 adjacent digit. F3: A3b-A3b. (C) Frequency histograms of correlation strength (area under peak) for A3b-A1 network. Blue: same digit pairs. Red: adjacent digit pairs. In each plot, left ordinate represents the count in each histogram bin, and right ordinate shows the cumulative distributions. Shaded area: 95% confidence range. (D) Cross correlograms of spontaneous (no probe on digit) spike trains. A total of 186 pairs of CCGs from A3b-A1 same-digit (grand mean peak: 0.0417), 188 pairs from A3b-A1 adjacent-digit (grand mean peak: 0.0348) and 162 pairs from A3b-A3b adjacent-digit (grand mean peak: 0.0275) were obtained from two monkeys. (E) Corresponding simulated CCG curves. Grand mean CCGs are plotted as thick black line. In comparison to the unit data with digit indentation, we obtained qualitatively similar correlograms under spontaneous conditions.

HIGHLIGHTS.

  • Anatomical, neural, and resting state connectivity patterns are very similar in SI

  • Intra-areal connections are cross-digit; inter-areal connections are same-digit

  • Resting state patterns reveal functional connectivity at local millimeter scale

Acknowledgments

This study was supported by FIRCA NS059061 (AWR), NIH NS044375 (AWR), NIH NS069909 (LMC), Dana Foundation (LMC), NIH NS078680 (JCG), Vanderbilt Core Grant (P30EY008126), and the Hungarian Scientific Research Fund OTKA NN79366 (LN). The technical assistance of Chang Gu, Yan Yan Chu and Alyssa Zuehl is highly appreciated. We thank the help of Mária Ashaber, Emese Pálfi and Cory Palmer in anatomical data analyses. Hui-Xin Qi for assistance in some electrophysiology mapping experiments, and Baxter Rogers in guidance with fMRI analysis.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Arieli A, Sterkin A, Grinvald A, Aertsen A. Dynamics of ongoing activity: explanation of the large variability in evoked cortical responses. Science. 1996;273:1868–1871. doi: 10.1126/science.273.5283.1868. [DOI] [PubMed] [Google Scholar]
  2. Behrens TE, Sporns O. Human connectomics. Curr Opin Neurobiol. 2011;22:144–153. doi: 10.1016/j.conb.2011.08.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Biswal BB, Mennes M, Zuo XN, Gohel S, Kelly C, Smith SM, Beckmann CF, Adelstein JS, Buckner RL, Colcombe S, et al. Toward discovery science of human brain function. Proc Natl Acad Sci U S A. 2010;107:4734–4739. doi: 10.1073/pnas.0911855107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bringuier V, Chavane F, Glaeser L, Fregnac Y. Horizontal propagation of visual activity in the synaptic integration field of area 17 neurons. Science. 1999;283:695–699. doi: 10.1126/science.283.5402.695. [DOI] [PubMed] [Google Scholar]
  5. Bruno RM, Sakmann B. Cortex is driven by weak but synchronously active thalamocortical synapses. Science. 2006;312:1622–7. doi: 10.1126/science.1124593. [DOI] [PubMed] [Google Scholar]
  6. Chen LM, Friedman RM, Roe AW. Optical imaging of a tactile illusion in area 3b of the primary somatosensory cortex. Science. 2003;302:881–885. doi: 10.1126/science.1087846. [DOI] [PubMed] [Google Scholar]
  7. Chen LM, Turner GH, Friedman RM, Zhang N, Gore JC, Roe AW, Avison MJ. High-resolution maps of real and illusory tactile activation in primary somatosensory cortex in individual monkeys with functional magnetic resonance imaging and optical imaging. J Neurosci. 2007;27:9181–9191. doi: 10.1523/JNEUROSCI.1588-07.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cohen MR, Kohn A. Measuring and interpreting neuronal correlations. Nat Neurosci. 2011;14:811–819. doi: 10.1038/nn.2842. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Deco G, Jirsa VK, McIntosh AR. Emerging concepts for the dynamical organization of resting-state activity in the brain. Nat Rev Neurosci. 2011;12:43–56. doi: 10.1038/nrn2961. [DOI] [PubMed] [Google Scholar]
  10. Deco G, Corbetta M. The dynamical balance of the brain at rest. Neuroscientist. 2011;17:107–123. doi: 10.1177/1073858409354384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Deco G, Jirsa V. Ongoing cortical activity at rest: criticality, multistability and ghost attractors. J Neurosci. 2012;32:3366–3375. doi: 10.1523/JNEUROSCI.2523-11.2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fox MD, Snyder AZ, Vincent JL, Corbetta M, Van Essen DC, Raichle ME. The human brain is intrinsically organized into dynamic, anticorrelated functional networks. Proc Natl Acad Sci U S A. 2005;102:9673–9678. doi: 10.1073/pnas.0504136102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fox MD, Corbetta M, Snyder AZ, Vincent JL, Raichle ME. Spontaneous neuronal activity distinguishes human dorsal and ventral attention systems. Proc Natl Acad Sci U S A. 2006;103:10046–10051. doi: 10.1073/pnas.0604187103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fox MD, Raichle ME. Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging. Nat Rev Neurosci. 2007;8:700–711. doi: 10.1038/nrn2201. [DOI] [PubMed] [Google Scholar]
  15. Friedman RM, Chen LM, Roe AW. Modality maps within primate somatosensory cortex. Proc Natl Acad Sci U S A. 2004;101:12724–12729. doi: 10.1073/pnas.0404884101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Girard P, Hupe JM, Bullier J. Feedforward and feedback connections between areas V1 and V2 of the monkey have similar rapid conduction velocities. J Neurophysiol. 2001;85:1328–1331. doi: 10.1152/jn.2001.85.3.1328. [DOI] [PubMed] [Google Scholar]
  17. Grecius MD, Supekar K, Menon V, Dougherty RF. Resting-state functional connectivity reflects structural connectivity in the default mode network. Cereb Cortex. 2009;19:72–78. doi: 10.1093/cercor/bhn059. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gusnard DA, Raichle ME. Searching for a baseline: functional imaging and the resting human brain. Nat Rev Neurosci. 2001;2:685–694. doi: 10.1038/35094500. [DOI] [PubMed] [Google Scholar]
  19. Honey CJ, Kotter R, Breakspear M, Sporns O. Network structure of cerebral cortex shapes functional connectivity on multiple time scales. Proc Natl Acad Sci U S A. 2007;104:10240–10245. doi: 10.1073/pnas.0701519104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Honey CJ, Sporns O, Cammoun L, Gigandet X, Thiran JP, Meuli R, Hagmann P. Predicting human resting-state functional connectivity from structural connectivity. Proc Natl Acad Sci U S A. 2009;106:2035–2040. doi: 10.1073/pnas.0811168106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Huffman KJ, Krubitzer L. Area 3a: topographic organization and cortical connections in marmoset monkeys. Cereb Cortex. 2001;11:849–867. doi: 10.1093/cercor/11.9.849. [DOI] [PubMed] [Google Scholar]
  22. Hung CP, Ramsden RM, Roe AW. A functional circuitry for edge-induced brightness perception. Nature Neurosci. 2007;10:1185–1190. doi: 10.1038/nn1948. [DOI] [PubMed] [Google Scholar]
  23. Hung CP, Ramsden BM, Roe AW. Inherent Biases in Spontaneous Cortical Dynamics. In: Ding M, Glanzman HD, editors. The Dynamic Brain: an exploration of neuronal variability and its functional significance. Oxford University Press; 2010. pp. 83–103. [Google Scholar]
  24. Johansson RS, Flanagan JR. Coding and use of tactile signals from the fingertips in object manipulation tasks. Nat Rev Neurosci. 2009;10:345–359. doi: 10.1038/nrn2621. [DOI] [PubMed] [Google Scholar]
  25. Jones EG, Coulter JD, Hendry SH. Intracortical connectivity of architectonic fields in the somatic sensory, motor and parietal cortex of monkeys. J Comp Neurol. 1978;181:291–347. doi: 10.1002/cne.901810206. [DOI] [PubMed] [Google Scholar]
  26. Kaas JH, Krubitzer LA, Chino YM, Langston AL, Polley EH, Blair N. Reorganization of retinotopic cortical maps in adult mammals after lesions of the retina. Science. 1990;248:229–231. doi: 10.1126/science.2326637. [DOI] [PubMed] [Google Scholar]
  27. Kenet T, Bibitchkov D, Tsodyks M, Grinvald A, Arieli A. Spontaneously emerging cortical representations of visual attributes. Nature. 2003;425:954–956. doi: 10.1038/nature02078. [DOI] [PubMed] [Google Scholar]
  28. Keysers C, Kaas JH, Gazzola V. Somatosensation in social perception. Nat Rev Neurosci. 2010;11:417–428. doi: 10.1038/nrn2833. [DOI] [PubMed] [Google Scholar]
  29. Lipton ML, Liszewski MC, O’Connell MN, Mills A, Smiley JF, Branch CA, Isler JR, Schroeder CE. Interactions within the hand representation in primary somatosensory cortex of primates. J Neurosci. 2010;30:15895–15903. doi: 10.1523/JNEUROSCI.4765-09.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Livingstone MS, Hubel DH. Specificity of intrinsic connections in primate primary visual cortex. J Neurosci. 1984;4:2830–2835. doi: 10.1523/JNEUROSCI.04-11-02830.1984. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Luczak A, Bartho P, Harris KD. Spontaneous events outline the realm of possible sensory responses in neocortical populations. Neuron. 2009;62:413–425. doi: 10.1016/j.neuron.2009.03.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Manger PR, Woods TM, Munoz A, Jones EG. Hand/face border as a limiting boundary in the body representation in monkey somatosensory cortex. J Neurosci. 1997;17:6338–6351. doi: 10.1523/JNEUROSCI.17-16-06338.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Matsui T, Tamura K, Koyano KW, Takeuchi D, Adachi Y, Osada T, Miyashita Y. Direct comparison of spontaneous functional connectivity and effective connectivity measured by intracortical microstimulation: an fMRI study in macaque monkeys. Cereb Cortex. 2011;10:2348–56. doi: 10.1093/cercor/bhr019. [DOI] [PubMed] [Google Scholar]
  34. Negyessy L, Palfi E, Ashaber M, Palmer C, Balazs J, Friedman RM, Chen LM, Roe AW. Intrinsic horizontal connections process global haptic features in the primary somatosensory cortex: neuroanatomical evidence. J Comp Neurol. 2013 doi: 10.1002/cne.23317. in press. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Reed JL, Pouget P, Qi HX, Zhou Z, Bernard MR, Burish MJ, Haitas J, Bonds AB, Kaas JH. Widespread spatial integration in primary somatosensory cortex. Proc Natl Acad Sci U S A. 2008;105:10233–10237. doi: 10.1073/pnas.0803800105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Roe AW, Ts’o DY. Specificity of color connectivity between primate V1 and V2. J Neurophysiol. 1999;82:2719–2730. doi: 10.1152/jn.1999.82.5.2719. [DOI] [PubMed] [Google Scholar]
  37. Scholvinck ML, Maier A, Ye FQ, Duyn JH, Leopold DA. Neural basis of global resting-state fMRI activity. Proc Natl Acad Sci U S A. 2010;107:10238–10243. doi: 10.1073/pnas.0913110107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Smith SM, Fox PT, Miller KL, Glahn DC, Fox PM, Mackay CE, Filippini N, Watkins KE, Toro R, Laird AR, Beckmann CF. Correspondence of the brain’s functional architecture during activation and rest. Proc Natl Acad Sci U S A. 2009;106:13040–13045. doi: 10.1073/pnas.0905267106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Steinmetz PN, Roy A, Fitzgerald PJ, Hsiao SS, Johnson KO, Niebur E. Attention modulates synchronized neuronal firing in primate somatosensory cortex. Nature. 2000;404:187–190. doi: 10.1038/35004588. [DOI] [PubMed] [Google Scholar]
  40. Ts’o DY, Gilbert CD. The organization of chromatic and spatial interactions in the primate striate cortex. J Neurosci. 1988;8:1712–1727. doi: 10.1523/JNEUROSCI.08-05-01712.1988. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Vincent JL, Patel GH, Fox MD, Snyder AZ, Baker JT, Van Essen DC, Zempel JM, Snyder LH, Corbetta M, Raichle ME. Intrinsic functional architecture in the anaesthetized monkey brain. Nature. 2007;447:83–86. doi: 10.1038/nature05758. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01

Supp Fig. 1. Correlation pattern not dependent upon specific seed voxel within a locale. (A–C) Correlation patterns of three neighboring seed voxels placed at the peak voxel (red arrow, B), and one voxel above (A) and one below (C) in area 3b (left columns). Images in right column show the correlation contour maps thresholded at r values of 0.6 and 0.8. (D–F) Correlation patterns of three neighboring seed voxels placed at the peak voxel (red arrow, B), and one voxel above (A) and one below (C) in area 1 (left columns). Images in right column shows the correlation contour maps thresholded at r values of 0.6 and 0.8.

Supp Fig. 2. Uniform intensity field. To illustrate uniformity of intensity in our imaged maps, we show a representative EPI intensity map (average SNR is around 180) along with corresponding T2* weighted structural image below. This shows that the SNR is reasonably even within the field of view. EPI intensity map of the SI region (A) and its corresponding high-resolution T2* weighted structural image (B).

Supp Fig. 3. Electrophysiological recordings and analysis. (A) Steady state firing of neurons in A3b and A1. Top: raster plots of units on D2 (red) and D3 (yellow) in A3b and A1. Bottom: PSTHs of this steady state activity. (B) Example correlograms. F1: A3b-A1 same digit. F2: A3b-A1 adjacent digit. F3: A3b-A3b. (C) Frequency histograms of correlation strength (area under peak) for A3b-A1 network. Blue: same digit pairs. Red: adjacent digit pairs. In each plot, left ordinate represents the count in each histogram bin, and right ordinate shows the cumulative distributions. Shaded area: 95% confidence range. (D) Cross correlograms of spontaneous (no probe on digit) spike trains. A total of 186 pairs of CCGs from A3b-A1 same-digit (grand mean peak: 0.0417), 188 pairs from A3b-A1 adjacent-digit (grand mean peak: 0.0348) and 162 pairs from A3b-A3b adjacent-digit (grand mean peak: 0.0275) were obtained from two monkeys. (E) Corresponding simulated CCG curves. Grand mean CCGs are plotted as thick black line. In comparison to the unit data with digit indentation, we obtained qualitatively similar correlograms under spontaneous conditions.

NIHMS470732-supplement-01.docx (2.2MB, docx)

RESOURCES