1

The Organization of the Human Cerebral Cortex 1

Estimated by Functional Connectivity 2
3
B.T. Thomas Yeo
1,2,*
, Fenna M. Krienen
1,2,*
, Jorge Sepulcre
1,2,3
, Mert R. Sabuncu
2,4
, 4
Danial Lashkari
4
, Marisa Hollinshead
1,2,3
, Joshua L. Roffman
5
, Jordan W. Smoller
5,6
, 5
Lilla Zllei
2
, Jonathan R. Polimeni
2
, Bruce Fischl
2,4,7
, Hesheng Liu
2
, 6
and Randy L. Buckner
1,2,3,5


7
8
1
Harvard University Department of Psychology, Center for Brain Science, Cambridge, 9
MA;
2
Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, 10
Massachusetts General Hospital, Charlestown, MA;
3
Howard Hughes Medical Institute, 11
Cambridge, MA;
4
Computer Science and Artificial Intelligence Lab, Massachusetts 12
Institute of Technology;
5
Department of Psychiatry, Massachusetts General Hospital, 13
Boston, MA;
6
Center for Human Genetics Research, MGH, Boston, MA;
7
Massachusetts 14
Institute of Technology-Harvard Division of Health Sciences and Technology, 15
Cambridge, MA; *BTTY and FMK contributed equally to this work 16
17
RUNNING HEAD: The Human Cerebral Cortex 18
19
Keywords: prefrontal, parietal, association cortex, fMRI, functional connectivity, default 20
network 21
22
Address correspondence to: 23
Dr. Randy L. Buckner 24
Harvard University 25
52 Oxford Street, Northwest Building, 280.06 26
Cambridge, MA 02138 27
randy_buckner@harvard.edu 28
29
30
31
Articles in PresS. J Neurophysiol (June 8, 2011). doi:10.1152/jn.00338.2011
Copyright 2011 by the American Physiological Society.
Articles in PresS. J Neurophysiol (June 8, 2011). doi:10.1152/jn.00338.2011
Copyright 2011 by the American Physiological Society.

2
Abstract 32
33
Information processing in the cerebral cortex involves interactions among 34
distributed areas. Anatomical connectivity suggests that certain areas form local 35
hierarchical relations such as within the visual system. Other connectivity patterns, 36
particularly among association areas, suggest the presence of large-scale circuits without 37
clear hierarchical relations. Here the organization of networks in the human cerebrum 38
was explored using resting-state functional connectivity MRI (fcMRI). Data from 1000 39
subjects were registered using surface-based alignment. A clustering approach was 40
employed to identify and replicate networks of functionally coupled regions across the 41
cerebral cortex. The results revealed local networks confined to sensory and motor 42
cortices as well as distributed networks of association regions. Within the sensory and 43
motor cortices, functional connectivity followed topographic representations across 44
adjacent areas. In association cortex, the connectivity patterns often showed abrupt 45
transitions between network boundaries. Focused analyses were performed to better 46
understand properties of network connectivity. A canonical sensory-motor pathway 47
involving V1, putative MT+, LIP and FEF was analyzed to explore how interactions 48
might arise within and between networks. Results showed that adjacent regions of the 49
MT+ complex demonstrate differential connectivity consistent with a hierarchical 50
pathway that spans networks. The functional connectivity of parietal and prefrontal 51
association cortices was next explored. Distinct connectivity profiles of neighboring 52
regions suggest they participate in distributed networks that, while showing evidence for 53
interactions, are embedded within largely parallel, interdigitated circuits. We conclude by 54
discussing the organization of these large-scale cerebral networks in relation to monkey 55
anatomy and their potential evolutionary expansion in humans to support cognition. 56
57
58

3
Introduction 59
60
Complex behaviors are subserved by distributed systems of brain areas (Felleman 61
and Van Essen 1991; Goldman-Rakic 1988; Mesulam 1990). The organization of these 62
systems can be studied in non-human animals using invasive techniques including 63
histology, anatomic tract tracing, electrophysiology, and lesion methods. The 64
organization of brain systems in the human has been inferred by comparing 65
cytoarchitectonically-defined homologies between species, and by noting similarities in 66
neuropsychological deficits following accidental brain injury to deficits present in animal 67
ablation studies. General agreement has emerged from these comparisons that the basic 68
organization of brain systems is similar across mammalian species. However, there is 69
also evidence that the human cerebral cortex, particularly association cortex, is not 70
simply a scaled version of other species. 71
The German anatomist Korbinian Brodmann (1909) first emphasized that areas 72
comprising the human inferior parietal lobule do not have clear homologues in the 73
monkey -- an observation that continues to motivate contemporary debates (Orban et al. 74
2004). Gross differences are also observed in the human brain when it is compared to our 75
evolutionarily closest relatives. For example, the human brain is triple the size of modern 76
great apes but motor and visual cortices are about the same absolute size (Blinkov and 77
Glezer 1968; Frahm et al. 1984). This observation suggests that expansion of the human 78
cerebrum disproportionately involves areas beyond those subserving basic sensory and 79
motor functions. In a recent analysis of cortical expansion based on 23 homologous areas 80
between the macaque and human, Van Essen and colleagues noted that the greatest 81
growth occurs in regions distributed across frontal, parietal, and temporal association 82
cortices (Van Essen and Dierker 2007; Hill et al. 2010). Preuss (2004) came to a similar 83
conclusion in a detailed review of comparative anatomy. Thus, in addition to expecting 84
the human brain to show broadly similar organizational properties with other well-studied 85
species, expansion and perhaps elaboration of association networks is also expected. 86
In this paper we report results of a comprehensive analysis of networks within the 87
human cerebral cortex using intrinsic functional connectivity MRI (fcMRI). The analysis 88
was based on 1000 young adults who contributed uniformly collected MRI data. The data 89

4
were brought into a common surface coordinate system to help preserve the surface 90
topology of the cortical mantle. Analyses were motivated by two goals. First, we sought 91
to provide reference maps that are a current best estimate of the organization of the 92
human cerebral cortex as measured by functional connectivity. Second, we wanted to 93
better understand how patterns of functional connectivity might give rise to the 94
organizational properties that underlie distributed brain systems. Particular focus was 95
placed on parietal and frontal association cortices. The foundations for the present work 96
come from traditional anatomical studies of cortical organization. 97
98
Organizational properties of the cerebral cortex in the non-human primate 99
Distributed brain systems are organized to facilitate both serial and parallel 100
processing (Felleman and Van Essen 1991; Mesulam 1998). The concept of serial 101
hierarchies is embedded within early ideas about brain organization. For example, 102
William James (1890) proposed that principles governing the reflex arc extend to the 103
cerebral hemispheres. He hypothesized that excitement of sensory systems propagates 104
upwards from lower to higher cerebral centers governing ideas, then to centers 105
producing (or inhibiting) movements. Hubel and Wiesel (1962) formally proposed the 106
concept of serial processing across a hierarchy in cat visual cortex based on their 107
observations of increasingly complex receptive field properties from the lateral geniculate 108
nucleus (LGN) to the simple and complex cells of V1. Based on studies of corticocortical 109
connections in the macaque, Pandya and Kuypers (1969) and Jones and Powell (1970) 110
suggested that hierarchical processing across sensory systems converges on transmodal 111
association areas. 112
The discovery of widespread connections among multiple cortical areas, as well 113
as extensive feedback projections from higher to lower sensory areas, suggested strictly 114
serial processing is not the only organizational scheme in the cerebral cortex. Instead, it 115
was proposed that hierarchical processing exists in a distributed fashion that can be 116
inferred from the laminar distribution of anatomical connectivity (Friedman 1983; 117
Maunsell and Van Essen 1983; Rockland and Pandya 1979). The comprehensive meta- 118
analysis of corticocortical connections in the macaque monkey by Felleman and Van 119
Essen (1991) provided strong evidence that unimodal and heteromodal areas in both the 120

5
visual and somatomotor systems are organized into separate distributed hierarchies (also 121
see Ungerleider and Desimone 1986; Van Essen et al. 1992). Some projections between 122
areas are organized as feedforward (ascending) projections, others as feedback 123
(descending) projections, and still others as lateral projections. For example, consistent 124
with serial processing, the primary visual area (V1) sends forward connections to and 125
receives feedback connections from V2 in a topographic fashion that connects the 126
corresponding receptive field representation in each area. In contrast to strictly serial 127
processing, these unimodal sensory cortical areas (V1 and V2) both project to higher 128
sensory areas. Lateral projections between areas are also common (e.g., CIT and STPp). 129
It becomes considerably more difficult to make inferences about the organization 130
of circuits involving association cortex. Historically, of the four criteria function, 131
cytoarchitecture, connectivity and topography used to define cortical areas and thereby 132
constrain models of organization, topography (e.g., retinotopy) and function are difficult 133
to discern in heteromodal association areas. Cytoarchitecture and connectivity thus 134
become especially valuable for inferring brain circuit organization beyond the sensory 135
and motor systems. However, as noted by Felleman and Van Essen (1991), the number of 136
violated constraints to hierarchical connectivity increases in the progression from early 137
sensory cortex up to association cortex (red lines near the top of the visual hierarchy in 138
Fig. 4 of Felleman and Van Essen 1991). 139
This raises the interesting possibility that the association areas may not follow as 140
rigid a hierarchical organization as canonical sensory and motor areas. Violations of strict 141
hierarchical arrangements are apparent in the visual system as noted above, but violations 142
and alternative connectivity patterns become common in association areas. For example, 143
paired tracer injections in association areas 7a and 46 lead to interdigitating columnar 144
patterns of terminations in some areas and complementary (feedforward and feedback) 145
patterns in other areas (Selemon and Goldman-Rakic 1988). 146
While recognizing that convergence and integration of pathways occurs in the 147
association cortex, Goldman-Rakic (1988) emphasized that primate association cortex is 148
organized into parallel distributed networks (see also Mesulam 1981). There are two key 149
features to her proposed organization that depart from hierarchical organizational models. 150
First, each distributed network consists of association areas spanning frontal, parietal, 151

6
temporal and cingulate cortices. Networks are densely interconnected, such that two areas 152
in the parietal and frontal cortices belonging to the same network are not just 153
anatomically connected to each other, but they are also both connected with other 154
components of the same network (Selemon and Goldman-Rakic 1988). Second, multiple 155
distributed networks exist adjacent to each other: adjacent areas in the parietal cortex 156
belonging to separate networks are differentially connected to adjacent areas of 157
corresponding networks in the frontal, temporal and cingulate cortices (Cavada and 158
Goldman-Rakic 1989a; 1989b). 159
The possibility of parallel distributed circuits will be an important consideration 160
in our analysis of fcMRI networks in the human, particularly within association cortices. 161
An intriguing possibility is that the majority of the human cerebral cortex involves 162
multiple parallel circuits that are interdigitated throughout association cortex, such that 163
each cortical lobe contains components of multiple association networks. That is, the 164
expansion of the cerebral association cortex in humans relative to the macaque may 165
preferentially involve networks organized in the form outlined by Goldman-Rakic (1988) 166
and anticipated by others (e.g., Mesulam 1981). To explore this possibility, analyses will 167
focus both on evidence for hierarchical relations across regions as well as evidence for 168
distributed networks that are interdigitated throughout association cortex. 169
170
Insights into the organization of the cerebral cortex revealed through neuroimaging 171
Noninvasive neuroimaging methods including positron emission tomography 172
(PET; Raichle 1987) and functional MRI (fMRI; Kwong et al. 1992; Ogawa et al. 1992) 173
allow functional response properties to be measured in the human cerebral cortex. The 174
measures are indirect, reflecting blood flow and oxygenation changes that are coupled to 175
neural activity through incompletely understood mechanisms (Logothetis 2008), and the 176
methods are presently limited to a spatial resolution of a few mm (e.g., Engel et al. 1997). 177
Neuroimaging approaches have nonetheless been extremely valuable for providing 178
insights into cortical organization. In some cases it has been possible to directly map the 179
topography within (and borders between) cortical areas (Engel et al. 1994; Sereno et al. 180
1995). More generally, differential response properties between regions are the source of 181
information about cortical mapping. For example, the increase in the complexity of 182

7
receptive field properties measured from primary to secondary sensory areas in visual 183
(Wandell et al. 2007), somatosensory (Iwamura 1998), and auditory (Wessinger et al. 184
2001) cortices suggest that serial hierarchical processing exists in human sensory cortex. 185
Neuroimaging studies of a wide range of cognitive tasks reveal simultaneous 186
activation in multiple regions in the parietal, frontal, temporal and cingulate cortices, 187
suggesting distributed systems of brain areas are involved in cognition. However, it is 188
difficult to assess the organization of these distributed systems based solely on task 189
activity because these cognitive tasks likely tap into multiple, overlapping processes, 190
some of which reflect the operation of distributed systems and others which reflect 191
distinct processing demands of the tasks (see Mesulam 1990 and Posner et al. 1988 for 192
relevant discussion). For these reasons, methods that can measure connectivity may 193
provide novel insights into the organization of distributed brain systems. 194
195
Functional connectivity and diffusion MRI provide tools to explore cortical organization 196
Diffusion MRI (dMRI) and fcMRI have recently emerged as promising tools for 197
mapping the connectivity of the human brain, each with distinct strengths and 198
weaknesses. dMRI measures the diffusion of water thus allowing direct non-invasive 199
mapping of white matter pathways (Basser et al. 1994). However, dMRI is presently 200
limited to resolving major fiber tracts. By contrast, fcMRI measures intrinsic functional 201
correlations between brain regions (Biswal et al. 1995) and is sensitive to coupling 202
between distributed as well as adjacent brain areas (e.g., see Sepulcre et al. 2010 for 203
discussion). While not a direct measure of anatomical connectivity, the functional 204
couplings detected by fcMRI are sufficiently constrained by anatomy to provide insights 205
into properties of circuit organization (for reviews, see Fox and Raichle 2007; Van Dijk 206
et al. 2010). When describing these correlations, we use the term functional connectivity 207
as coined by Karl Friston (1994) to denote temporal correlations between remote 208
neurophysiological events for which the causal relation is undetermined. 209
There are important limitations of fcMRI including sensitivity to indirect 210
anatomical connectivity and functional coupling that changes in response to recent 211
experience and the current task being engaged (Buckner 2010). For these reasons, some 212
discussions of fcMRI have emphasized that intrinsic activity measured by fcMRI reflects 213

8
the prior history of activity through brain systems and not simply static anatomic 214
connectivity (Power et al. 2010). fcMRI also does not presently provide information 215
about whether connections are feedforward (ascending) or feedback (descending). These 216
limitations constrain how analyses are conducted and results can be interpreted. 217
Directly relevant to the present study, prior investigations using fcMRI provide 218
estimates of large-scale cortical networks that have generally (but not in all details) 219
converged across a variety of analytic approaches, including seed-based fcMRI (Biswal 220
et al. 1995), independent component analysis (Beckmann and Smith 2004; Smith et al. 221
2009), clustering (Bellec et al. 2010; Golland et al. 2007) and graph theory (Dosenbach et 222
al. 2007). Because of uncertainties regarding their relation to underlying anatomic brain 223
systems, networks identified using fcMRI have often been labeled based on their 224
relations to task-based functional networks. Some of these networks, such as the default 225
network (Greicius et al. 2003) and dorsal attention system (Fox et al. 2006), have been 226
proposed to be related to anatomical tracing and task-based fMRI in the macaque 227
(Buckner et al. 2008; Saleem et al. 2008; Vincent et al. 2007). 228
Motivated by the usefulness of connectivity in establishing the organization of the 229
cerebral cortex in non-human primates, this study analyzed fcMRI data from 1000 230
subjects with two main goals. First, the analyses sought to provide reference maps that 231
are a current best estimate of the organization of human cortical networks as measured by 232
functional connectivity. Second, by using the power of a large data sample to 233
quantitatively measure functional connectivity strength among many regions, the study 234
explored the patterns of corticocortical functional coupling that give rise to these 235
networks. 236
237
Methods 238
239
Overview 240
The present study explored the organization of large-scale distributed networks in 241
the human cerebral cortex using resting-state fcMRI. The main analyses were based on a 242
core dataset of 1000 healthy, young adults whose fMRI data were acquired using the 243
same MRI sequence on the same hardware (3 Tesla field strength, 12-channel receive coil 244

9
array). For several analyses the data were divided into Discovery (n = 500) and 245
Replication (n = 500) data samples to test for reliability and for unbiased quantification of 246
functional connectivity patterns. Additional supplementary datasets were used to address 247
specific questions that arose during analysis. A first supplementary fMRI dataset (n = 16) 248
contrasted different passive tasks engaged during resting-state functional data acquisition. 249
A second supplementary fMRI dataset (n = 4) consisted of data acquired during visual 250
stimulation optimized to define retinotopic boundaries of early visual areas (Hinds et al. 251
2009; Polimeni et al. 2005). A final supplementary dataset used human histological data 252
to define a range of cytoarchitectonic areas including human V1 (Amunts et al. 2000, 253
Fischl et al. 2008) and the putative homologue to macaque MT+ (Malikovic et al. 2007, 254
Yeo et al. 2010b). All data (fMRI and histological) were brought into a common surface 255
coordinate system based on the cortical surface as reconstructed from each participants 256
structural anatomy. Data analyses began by examining broad properties of cortical 257
network organization and progressed to quantify the detailed patterns of functional 258
connectivity within and between networks. 259
In the first set of analyses, a clustering algorithm was used to parcellate the 260
cerebral cortex into networks of functionally coupled regions. Parcellations were 261
examined for a coarse solution that organized the cortex into 7 networks as well as a finer 262
solution that identified 17 networks. As the results will reveal, the estimated networks 263
were consistent across the Discovery and Replication data samples and confirmed by 264
region-based fcMRI analyses. The full dataset was used to construct a best-estimate 265
parcellation of the human cerebral cortex to serve as a reference for future studies. 266
The second set of analyses explored the coupling of regions that fell within 267
sensory and motor pathways. Since these areas are relatively well understood in both 268
humans and macaques, they provide the opportunity to evaluate the utility and limitations 269
of functional connectivity methods. Analyses examined quantitative coupling properties 270
between individual regions that were within the same network as well as coupling 271
properties between networks focusing on a sensory-motor pathway that is the putative 272
homologue of the well-studied system in the monkey involving MT+, parietal regions at 273
or near LIP, and premotor regions at or near FEF. 274

10
The final set of analyses characterized the organization of distributed networks in 275
higher-order association cortex. The connectivity patterns of regions within frontal and 276
parietal association cortices were quantified. These analyses involved constructing a 277
series of small seed regions across frontal and parietal cortices and examining functional 278
connectivity strength to multiple regions distributed throughout the cerebral cortex, 279
allowing the fingerprint of functional coupling to be identified for each region. For 280
these analyses, regions were always defined in the Discovery data sample or some other 281
source, such as histology, and functional connectivity quantified in the independent 282
Replication data sample. 283
284
Participants 285
Paid participants were clinically healthy, native English speaking young adults 286
with normal or corrected-to-normal vision (ages 18 to 35). Subjects were excluded if their 287
fMRI signal-to-noise ratio (SNR) was low (< 100; see below), artifacts were detected in 288
the MR data, their self-reported health information indicated the presence of any prior 289
neurological or psychiatric condition, or they were taking any psychoactive medications. 290
The core dataset consisted of 1000 individuals imaged during eyes open rest (EOR) and 291
was divided into two independent samples (each n = 500; labeled the Discovery and 292
Replication samples). Age and gender were matched for the Discovery (mean age = 21.3 293
yr, 42.6% male) and Replication (mean age = 21.3 yr, 42.8% male) datasets. These data 294
are new data presented for the first time in this study and were acquired as part of a 295
collaborative effort across multiple local laboratories all acquiring data on matched MRI 296
scanners (at Harvard and at the Massachusetts General Hospital). Participants provided 297
written informed consent in accordance with guidelines set by institutional review boards 298
of Harvard University or Partners Healthcare. 299
Two smaller supplementary datasets were also analyzed. The Task Effect dataset 300
(n = 16, mean age = 21.1 yr, 25.0% male) consisted of fMRI data collected under 301
different passive conditions (eyes closed rest, ECR; EOR; and fixation, FIX) and was 302
analyzed previously (Van Dijk et al. 2010). The Visuotopic dataset (n = 4; mean age = 303
34.5 yr, 100% male) consisted of previously published visuotopic data (Hinds et al. 2009; 304
Polimeni et al. 2005). 305

11
306
MRI data acquisition 307
All data were collected on matched 3T Tim Trio scanners (Siemens, Erlangen, 308
Germany) using a 12-channel phased-array head coil except for the Visuotopic dataset, 309
which was acquired on a custom-built four-channel phased-array surface coil. A software 310
upgrade (VB15 to VB17) occurred on all scanners during the study. Validation studies 311
that acquired structural and functional data on the same individuals before and after the 312
upgrade could not detect an effect of the upgrade. The functional imaging data were 313
acquired using a gradient-echo echo-planar imaging (EPI) sequence sensitive to blood 314
oxygenation level-dependent (BOLD) contrast. Whole-brain coverage including the 315
entire cerebellum was achieved with slices aligned to the anterior-commissure posterior- 316
commissure (AC-PC) plane using an automated alignment procedure ensuring 317
consistency between subjects (van der Kouwe et al. 2005). Structural data included a 318
high-resolution multi-echo T1-weighted magnetization-prepared gradient-echo image 319
(multi-echo MP-RAGE; van der Kouwe et al. 2008). 320
For the core dataset, subjects were instructed to remain still, stay awake and to 321
keep their eyes open. EPI parameters were as follows: TR = 3000 ms, TE = 30 ms, FA = 322
85, 3 x 3 x 3 mm voxels, FOV = 216 and 47 axial slices collected with interleaved 323
acquisition and no gap between slices. Each functional run lasted 6.2 min (124 time 324
points). One or two runs were acquired per subject (average of 1.7 runs). Parameters for 325
the structural scan (multi-echo MP-RAGE; van der Kouwe et al. 2008) were as follows: 326
TR = 2200 ms, TI = 1100 ms, TE=1.54 ms for image 1 to 7.01 ms for image 4, FA = 7, 327
1.2 x 1.2 x 1.2 mm voxels and FOV = 230. The multi-echo MP-RAGE allows increased 328
contrast through weighted-averaging of the four derived images. 329
For the Task Effect dataset, subjects were instructed to remain still with their eyes 330
open (EOR; two runs), eyes closed (ECR; two runs) or to passively fixate a centrally 331
presented crosshair (FIX; two runs). For details see Van Dijk et al (2010). The order of 332
rest conditions was counterbalanced across subjects. EPI parameters were as follows: TR 333
= 3000 ms; TE = 30 ms; FA = 90; 3 x 3 x 3 mm voxels; FOV = 288 and 43 slices 334
collected with interleaved acquisition and no gap between slices. Each functional run 335
lasted 7.15 min (143 time points). Parameters for structural scans (MP-RAGE) were as 336

12
follows: TR = 2530 ms, TI = 1100 ms, TE = 3.44 ms, FA = 7, 1 x 1 x 1 mm voxels and 337
FOV = 256. 338
The Visuotopic dataset was collected using a custom-built four-channel phased- 339
array surface coil placed at the back of the head. During each functional run, subjects 340
were presented with one of four visual stimuli: a clockwise rotating wedge, a 341
counterclockwise rotating wedge, an expanding ring, or a contracting ring (DeYoe et al. 342
1996; Engel et al. 1994; Sereno et al. 1995). Because the four-channel surface coil 343
provided only partial brain coverage, structural data for these four subjects were collected 344
separately on a 1.5T Allegra scanner (Siemens, Erlangen, Germany). Further details of 345
the acquisition and data processing protocol can be found elsewhere (Hinds et al. 2009; 346
Polimeni et al. 2005). 347
Except where noted, the description of data processing and analysis below applies 348
to the whole-brain data (the core dataset of 1000 subjects and the Task Effect dataset) and 349
not the Visuotopic data. 350
351
Functional MRI data preprocessing 352
The fMRI data were preprocessed with a series of steps common to fMRI 353
analyses. Preprocessing involved (1) discarding the first four volumes of each run to 354
allow for T1-equilibration effects, (2) compensating for slice-acquisition-dependent time 355
shifts per volume with SPM2 (Wellcome Department of Cognitive Neurology, London, 356
UK), and (3) correcting for head motion using rigid body translation and rotation with the 357
FSL package (Jenkinson et al. 2002; Smith et al. 2004). 358
The data underwent further processing using procedures adapted from Biswal et 359
al. (1995) and optimized for fcMRI analysis (Fox et al. 2005; Van Dijk et al. 2010; 360
Vincent et al. 2006). Briefly, a low-pass temporal filter removed constant offsets and 361
linear trends over each run while retaining frequencies below 0.08 Hz. Sources of 362
spurious variance, along with their temporal derivatives, were removed through linear 363
regression including: (1) six parameters obtained by correction for rigid body head 364
motion, (2) the signal averaged over the whole brain, (3) the signal averaged over the 365
ventricles, and (4) the signal averaged over the deep cerebral white matter. This 366
regression procedure minimized signal contributions of non-neuronal origin including 367

13
respiration-induced signal fluctuations (Van Dijk et al. 2010). Unlike previously 368
established fcMRI preprocessing procedures, no spatial smoothing of the resting-state 369
data occurred up to this point of the preprocessing stream. 370
371
Structural MRI data preprocessing and functional-structural data alignment 372
The structural data were processed using the FreeSurfer 373
(http://surfer.nmr.mgh.harvard.edu) version 4.5.0 software package. The FreeSurfer 374
software package constitutes a suite of automated algorithms for reconstructing accurate 375
surface mesh representations of the cortex from individual subjects T1 images (Fig. 1B- 376
C) and the overlay of fMRI on the surfaces for group analysis (Fig. 1D-E). Briefly, the 377
cortical surface extraction process (Fig. 1B-C) involved (1) correcting for intensity 378
variations due to MR inhomogeneities (Dale et al. 1999), (2) removing extra-cerebral 379
voxels through skull-stripping (Sgonne et al. 2004), (3) segmenting cortical gray and 380
white matter voxels based on the intensity difference and geometric structure of the gray- 381
white interface (Dale et al. 1999), (4) computing cutting planes to disconnect the two 382
hemispheres and subcortical structures (Dale et al. 1999), (5) filling the interior holes of 383
the segmentation using a connected-component analysis (Dale et al. 1999), (6) 384
tessellating a triangular mesh over the gray-white boundary of each hemispheric volume 385
and deforming the mesh to produce a smooth representation of the gray/white interface 386
and pial surface (Dale et al. 1999), and (7) correcting topological defects in the surface so 387
that the mesh achieves a spherical topology (Fischl et al. 2001; Sgonne et al. 2007). 388
389
------------------------------------------------------------------ 390
Insert Figure 1 About here 391
------------------------------------------------------------------ 392
393
After segmentation of the cortical surface, spatial correspondences among the 394
subjects cortical folding patterns were established via the use of a spherical coordinate 395
system (Fig. 1D-E). Briefly, the process involved (1) inflating each subjects surface 396
mesh into a sphere while minimizing geometric distortion of the original cortical surface 397

14
as measured by geodesic distances among surface vertices and ensuring the inflation 398
constituted a one-to-one mapping, and (2) computing a smooth, invertible deformation of 399
the resulting spherical mesh to a common spherical coordinate system that aligned the 400
cortical folding patterns across subjects (Fischl et al. 1999a; Fischl et al. 1999b). 401
Once the common spherical coordinate system was established, the structural and 402
functional images were aligned (Fig. 1A-B) using boundary-based registration (Greve 403
and Fischl 2009) that is provided as part of FreeSurfers companion package, FsFast 404
(http://surfer.nmr.mgh.harvard.edu/fswiki/FsFast). The preprocessed resting state fMRI 405
data were then propagated to the common spherical coordinate system via sampling from 406
the middle of the cortical ribbon in a single interpolation step (Fig. 1A-E). The choice of 407
sampling fMRI data from the middle of the cortical ribbon was motivated by the desire to 408
reduce the blurring of fMRI signal across sulci or gyri and also by a recent study on the 409
point-spread function of fMRI (Polimeni et al. 2010). The study showed that large 410
draining vessels on the pial surface increased BOLD signal close to the pial surface but 411
reduced spatial specificity of the hemodynamic response. Sampling fMRI data from the 412
middle of the cortical ribbon therefore represented a trade-off between spatial specificity 413
and signal sensitivity. Since our fMRI voxels were relatively large (3mm), we were not as 414
concerned about laminar bias in the functional connectivity analysis. 415
The cerebral cortex is a thin sheet, with common organizational features along its 416
radial axis. Along the dimensions parallel to this sheet is a mosaic of cortical areas that 417
differ in architecture, connectivity, function, and/or topographic organization (Felleman 418
and Van Essen 1991; Kaas 1987). The spherical representation of the cortex therefore 419
affords a more accurate alignment of the cortical folding pattern and has the consequence 420
of improving cytoarchitectural (Fischl et al. 2008; Hinds et al. 2008; Yeo et al. 2010a) 421
and functional (Fischl et al. 1999b; Van Essen 2005) correspondences across subjects 422
compared with 3D volumetric registration even though cortical folds do not completely 423
predict cytoarchitecture or function (Rajkowska and Goldman-Rakic 1995; Thirion et al. 424
2007; Yeo et al. 2010b). The acquisition resolution and inherent limitations of the BOLD 425
signal also provided restrictions on achievable resolution. 426

15
A 6mm FWHM (full width at half maximum) smoothing kernel was applied to 427
the fMRI data in the surface space and the data were downsampled to a 4mm mesh
1
. 428
Smoothing after the fMRI data were projected onto the surface helped to minimize the 429
blurring of fMRI signal across sulci or gyri. Since our algorithms are not perfectly 430
accurate, any registration or segmentation errors will likely cause blurring of fMRI signal 431
across sulci or gyri. Consequently, we did not expect to eliminate the blurring issues 432
completely, which is important to keep in mind when interpreting the results. The steps 433
taken could only minimize the problem. 434
The processing of the Visuotopic dataset was broadly similar except that older 435
versions of FreeSurfer and FsFast were used for the processing and so manual 436
interventions were required to correct the T2* to T1 registration. Details of the processing 437
can be found elsewhere (Hinds et al. 2009; Polimeni et al. 2005). 438
439
Quality control 440
Visual inspection of the registered data suggested that accurate representation of 441
the cortical surface was extracted for each subject and that structural and functional 442
image registration was successful. Figure 2 shows the results of cortical surface 443
extraction from the T1 images and T2* to T1 registration of 3 randomly chosen subjects. 444
These examples represent typical subjects. Note that functional data distortion remains in 445
areas prone to susceptibility artifacts including anterior prefrontal regions, regions near 446
lateral temporal cortex, and orbital frontal cortex. 447
448
------------------------------------------------------------------ 449
Insert Figure 2 About here 450
------------------------------------------------------------------ 451

1
It is not possible to generate a high resolution uniform mesh on the sphere. However,
one can work with approximately uniform spherical meshes at different spatial
resolutions by starting with a regular icosahedron mesh consisting of 20 equal faces and
12 vertices and iteratively subdividing each mesh triangle into four smaller triangles.
Here each cortical hemisphere is represented by a subdivided icosahedron mesh with
20480 faces and 10242 vertices, where neighboring pairs of vertices are on average 3.8
mm apart (max = 4.1 mm, min = 3.4 mm).

16
452
Visualization 453
While all subsequent analyses were performed in FreeSurfer surface space, for the 454
purpose of visualization, all maps were transformed and displayed on the inflated PALS 455
cortical surfaces using Caret software (Van Essen 2004; 2005; Van Essen and Dierker 456
2007). In addition, this study also transformed and visualized the estimated networks in 457
FMRIB Software Library (FSL) MNI152 space (Smith et al. 2004). The mapping 458
between FSL MNI152 volumetric space and FreeSurfer surface space is detailed in our 459
companion study (Buckner et al. submitted). 460
461
Signal-to-noise ratio (SNR) maps 462
Signal loss and distortion (susceptibility artifacts) occur as a result of magnetic 463
field inhomogeneities. Field inhomogeneities are particularly pronounced in regions 464
where the brain is adjacent to air, causing signal loss and distortion in T2*-dependent 465
(BOLD) images (Ojemann et al. 1997). To estimate the effects of susceptibility artifacts 466
in the present data, the signal-to-noise ratio (SNR) of the motion corrected fMRI time 467
series was computed for each voxel in subjects native volumetric space by averaging the 468
signal intensity across the whole run and dividing it by the standard deviation over time. 469
SNR was also used as exclusionary criteria. If the SNR for the whole brain (mean SNR 470
over all voxels within the brain mask) was <100 for an fMRI run, the subject was 471
excluded. Thus, all 1000 subjects contributed data with SNR > 100 for each fMRI run. 472
For subjects with two runs, the SNR was averaged across the runs. The SNR was then 473
projected to FreeSurfer surface space, averaged across the 1000 subjects from the core 474
dataset, and displayed in Caret PALS space (Fig. 3). As expected, low SNR is present in 475
the anterior portion of the inferior and medial temporal lobe, as well as in the orbital 476
frontal cortex. There is also clear spatial variation in the SNR across the cortical mantle, 477
which is important to keep in mind when interpreting the results, such as the absence of a 478
cortical region of low SNR from a network. 479
480
------------------------------------------------------------------ 481

17
Insert Figure 3 About here 482
------------------------------------------------------------------ 483
484
Clustering 485
We applied a clustering approach to define the boundaries of functionally distinct 486
cortical regions and their relations to regions distributed throughout the cerebral cortex 487
(forming networks). Distinguishing neighboring cortical regions by their pattern of 488
connectivity has a long history in both non-human primate (e.g., Cavada and Goldman- 489
Rakic 1989a; Goldman-Rakic 1988; Passingham et al. 2002) and human (e.g., Cohen et 490
al. 2008; Johansen-Berg et al. 2004; Nelson et al. 2010) research. Here we began our 491
analyses by defining cortical networks to be sets of cortical regions with similar profiles 492
of corticocortical functional connectivity. The idea follows the empirical finding that in 493
primates, regions of association cortex that are anatomically connected tend to have 494
similar patterns of anatomical connectivity to other cortical and subcortical regions, thus 495
forming a densely connected distributed network (Goldman-Rakic 1988). Note that this 496
assumption about the organizational properties of corticocortical connectivity is probably 497
neither a characteristic of all cortical regions nor a full characterization of the 498
connectivity pattern of any cortical region. As will be shown, the procedure identified 499
functionally coupled networks that could be verified with seed-based regional analyses 500
that made no assumptions about the connectivity patterns. 501
For this initial analysis, we defined the connectivity profile of a cortical region to 502
be its functional coupling to 1175 region of interest (ROI) vertices. The 1175 ROI 503
vertices were uniformly sampled in FreeSurfer surface space (shown in Caret PALS 504
space in Fig. 4) and consisted of single vertices spaced about 16mm apart. For each 505
subject, we computed the Pearsons product moment correlation between the fMRI time 506
series at each spatial location (18715 vertices) and the 1175 ROI vertices. Each spatial 507
location is therefore characterized by its functional coupling to the 1175 ROI vertices. 508
We binarized the 18715 x 1175 matrix of correlations for each subject by keeping the top 509
10% of the correlations and averaged the binarized matrices independently across each 510
group of 500 subjects in the Discovery and Replication samples. If a subject had two 511

18
runs, we averaged the correlation matrices across the two runs before binarization. As 512
will be shown in the results, binarization of the correlation matrix leads to significantly 513
better clustering results, although the algorithm appears robust to the particular choice of 514
threshold. Visual inspection of the connectivity profiles (not shown) suggested that the 515
1175 ROI vertices were sufficiently dense to capture spatial variation in corticocortical 516
connectivity given the limits of our acquisition procedures. 517
518
------------------------------------------------------------------ 519
Insert Figure 4 About here 520
------------------------------------------------------------------ 521
522
A clustering algorithm was then applied separately to the Discovery and 523
Replication samples to estimate networks of cortical regions with similar connectivity 524
profiles. The two independent datasets thus allowed exploration of the reliability of 525
estimated networks. The idea behind clustering can be illustrated with a toy example. 526
Figure 5A shows hypothetical points scattered in a structured fashion on a two- 527
dimensional canvas. Clustering aims to recover this structure by dividing the points into 528
different groups so that points within a group are physically close, such as shown in 529
Figure 5B. 530
The clustering algorithm employed in this study modeled the data with a von 531
Mises-Fisher distribution (Lashkari et al. 2010). More specifically, the data were modeled 532
as 18715 points on an 1174-dimensional unit hypersphere embedded in an 1175- 533
dimensional Euclidean space, where distances between points were measured by their 534
geodesic distance on the hypersphere. Like the toy example, clustering aims to group 535
vertices that are close together in this non-Euclidean canvas (i.e., have similar 536
connectivity profiles) into the same cluster or network. Measuring distances between 537
points by their geodesic distance is equivalent to defining the similarity between two 538
correlation profiles to be the correlation between the correlation profiles. By using 539
correlation as a measure of similarity, differences in correlation strength were normalized 540
among points so that regions are clustered together based on their connectivity profiles 541

19
(rather than their strengths of connectivity). In theory, this should mitigate the effects of 542
spatial variation in SNR (Fig. 3). 543
The algorithm operated by randomly assigning the 18715 points to different 544
groups and then iteratively reassigning the group memberships of points to maximize the 545
agreement of connectivity profiles among points of the same group. More details of the 546
clustering algorithm can be found elsewhere (Lashkari et al. 2010). 547
548
------------------------------------------------------------------ 549
Insert Figure 5 About here 550
------------------------------------------------------------------ 551
552
Stability analysis 553
A drawback of most clustering approaches is that one must choose the number of 554
clusters a priori. In this instance, the question is how many clusters cerebral networks 555
are needed to correctly parcellate the cortex? We do not have an answer to this 556
question or know if there is a single correct answer given that the cerebral cortex 557
possesses complex patterns of diverging and converging connections among areas. As 558
such, none of our conclusions will depend on a strong assumption that there is a single 559
correct solution to parcellating the cortex. Nonetheless, we sought a principled approach 560
to identify parcellation solutions that captured significant portions of the correlation 561
structure among cortical regions. 562
One popular method for estimating the number of clusters is by analyzing the 563
stability of the clustering algorithm (Ben-Hur et al. 2002; Lange et al. 2004; also see Fig. 564
5). We employed two variations of the stability analysis on the full set of 1000 subjects. 565
Both variations estimated the same numbers of clusters. The first variation involved 566
(repeatedly and randomly) dividing the ROIs into two groups and measuring the 567
reproducibility of the clustering algorithms results when applied separately to the two 568
groups of ROIs. The second variation involved (repeatedly and randomly) dividing the 569
18715 vertices into two groups and applying the clustering algorithm separately to the 570
two groups of vertices. The model parameters learned from clustering one group of 571

20
vertices was then used to predict the clustering results of the second group of vertices. 572
The agreement between the prediction and clustering results of the second group 573
measured the generalization power of the clustering results (Fig. 6). Further details of the 574
stability analysis can be found elsewhere (Lange et al. 2004). 575
The stability analyses (Fig. 6) suggested 7 and 17 networks were appropriate 576
starting points for parcellating the cortex. As the results will reveal, these parcellation 577
solutions were excellent for capturing significant components of the regional variation 578
that could be replicated across datasets and independently revealed by seed-based 579
analyses. However, the focus on 7- and 17-network solutions should not be taken to 580
imply that meaningful properties are absent in alternative parcellation schemes. By 581
focusing on both a relatively coarse solution (7-network) and fine-resolution solution (17- 582
network), we were able to survey the solution space broadly. 583
584
------------------------------------------------------------------ 585
Insert Figure 6 About here 586
------------------------------------------------------------------ 587
588
Parcellation maps 589
Parcellation maps of the cerebral cortex were generated for both 7-network and 590
17-network solutions for the Discovery sample and replicated in the Replication sample. 591
The reliability analysis was conducted to illustrate the stability of the topographic 592
boundaries that the solutions converged upon. In this regard, a powerful feature of 593
analyzing large data samples is that the analyses are able to detect the presence of stable 594
cerebral networks and also to establish the boundaries of regions with a high degree of 595
confidence, including contiguous regions that may be part of distinct networks. As a final 596
step the full sample (N = 1000) was used to compute parcellations that represent our best 597
estimates of the networks (Figs. 11 and 13). 598
599
Confidence maps 600

21
A useful visualization of the cortical parcellation is to look at the confidence of 601
each spatial location belonging to its assigned network. We used the silhouette measure 602
(Rousseeuw 1987) from the clustering literature for this purpose (Figs. 8 and 10). The 603
silhouette of a data point (spatial location in our case) measures the similarity (correlation 604
in our case) of the data point to other data points of the same cluster (network in our case) 605
compared with data points belonging to the next closest cluster. The resulting silhouette 606
at each spatial location lies between -1 and 1, so that a larger value indicates higher 607
confidence of the spatial location in belonging to its assigned network. A negative value 608
indicates that the connectivity profile at the spatial location is on average closer to the 609
next closest cluster than to its assigned cluster. A negative value is therefore unlikely, but 610
is still possible because the clustering cost function is not equivalent to the silhouette 611
measure. 612
613
Correlation maps and correlations between regions 614
Large-scale cortical networks can be reliably estimated. To better understand the 615
meaning of the networks resolved by the clustering technique, all salient results were 616
followed up with focused analyses using seed-based regional analysis. The coordinates 617
for all seed regions used in these analyses can be found in Tables 1-5. For these analyses, 618
group-averaged functional connectivity maps were used to inspect the validity of 619
clustering results and to visualize differences in connectivity patterns of regions in 620
sensory and association cortices. 621
Each region consisted of single surface vertex (~4 mm x 4 mm) but should be 622
considered spatially more extensive because of the spatial smoothing and intersubject 623
averaging. Correlation maps were obtained by computing the Pearsons product moment 624
correlation between the regions preprocessed resting fMRI time course and the time 625
courses of all other vertices across the cortical mantle. To obtain a group-averaged 626
correlation z-map, the correlation map of each subject in the group was converted to 627
individual subject z-map using Fishers r-to-z transformation and then averaged across all 628
subjects in the group. The Fishers r-to-z transformation increases normality of the 629
distribution of correlations in the sample. For subjects with multiple runs, the individual 630
subject z-maps were first averaged across the runs before submitting to the group 631

22
average. An inverse Fishers r-to-z transformation was then applied to the group- 632
averaged correlation z-map yielding a group-averaged correlation map. 633
To quantify functional connectivity among regions, Fishers r-to-z transformed 634
correlations were computed among the regions for each subject within a group. For 635
several targeted, a priori analyses, classical statistical tests, including t-tests (e.g., see 636
Figs. 20 and 21) and ANOVA (e.g., see Figs. 22 and 27), were performed on the z- 637
transformed correlations using Matlab 7.4 (Mathworks, Natick, USA) or SPSS 18.0 638
(IBM, Armonk, USA). All tests survive Bonferroni correction for multiple comparisons. 639
640
Selecting regions for functional connectivity analysis 641
Throughout the analyses, seed regions for functional connectivity were selected 642
using different criteria depending on the purpose of the analysis. In all cases, if a 643
particular dataset was used for selecting the region (e.g., Discovery sample), functional 644
connectivity was always computed with a different dataset (e.g., Replication sample), 645
thus providing an unbiased measurement of correlation strength. We detail the method 646
used for region selection in the results as they are implemented for each particular 647
analysis in the Results section. The following procedures describe the general strategies 648
adopted. 649
First, when testing for seed-based confirmation of resolved networks, the 650
estimated network boundaries and confidence maps of the Discovery sample were used to 651
derive regional vertices to be tested in the Replication sample (e.g., see Fig. 16). Regions 652
were chosen for (1) maximal spatial coverage of estimated networks, (2) avoiding 653
network boundaries, and/or (3) their confidence in network assignments. We also defined 654
new regions based on the correlation maps from the Discovery sample. For example, new 655
regions might be chosen to be at or near the peaks of the correlation maps. 656
Second, for some analyses we utilized task-based fMRI to select regions. For 657
example, visuotopic and functional characteristics revealed using fMRI can be used to 658
estimate visual areas in the human (Hadjikhani et al. 1998; Sereno et al. 1995). The Caret 659
software database provides estimated locations of multiple visual areas that were mapped 660
into Caret PALS space using surface-based registration of an individual case in 661
Hadjikhani et al (1998), although the foveal and peripheral extents of these areas are 662

23
likely to be underestimated (Van Essen 2004). Landmark-based surface registration 663
between FreeSurfer and Caret PALS allowed us to utilize these fMRI-defined visuotopic 664
regions for guiding our selection of regions in FreeSurfer surface space (e.g., V3A in Fig. 665
25). In addition, we also considered peak activation coordinates reported in fMRI 666
literature (e.g., see Fig. 27). When the peak coordinates were reported in MNI space, we 667
projected the coordinates to FreeSurfer surface space. The mapping between MNI152 668
volumetric space and FreeSurfer surface space is detailed in our companion study 669
(Buckner et al. submitted). In cases where the peak activation coordinates were reported 670
in the atlas space of Talairach and Tournoux (1988), the coordinates were first mapped to 671
FSL MNI152 space (Lancaster et al. 2007) before being projected to FreeSurfer surface 672
space. 673
Third, probabilistic histological maps in FreeSurfer surface space allowed for the 674
selection of regions within histologically-defined areas (e.g., see Fig. 22). Postmortem 675
human brains of fifteen subjects with no history of neurologic or psychiatric diseases 676
were processed and analyzed (Amunts et al. 1999; Schleicher et al. 1999; Schormann and 677
Zilles 1998). The histological sections were aligned to postmortem MR volume of the 678
same brain using nonlinear warping (Schormann and Zilles 1998) to build an undistorted 679
3-dimensional histological volume. Cytoarchitectonic areas, including V1 (Amunts et al. 680
2000) and hOc5/MT+ (Malikovic et al. 2007) were segmented using observer- 681
independent criteria (Schleicher et al. 1999). The MR volumes were segmented to 682
separate white matter from other tissue classes, and the segmentation was used to 683
generate topologically correct and geometrically accurate surface representations of the 684
cerebral cortex using FreeSurfer (Fischl et al. 2008). The cortical surfaces of the fifteen 685
subjects were registered to FreeSurfer surface space and the histological areas were 686
sampled onto the surface space. While there were fifteen subjects, each cytoarchitectonic 687
area was only analyzed in at most ten subjects. 688
Prior work has demonstrated good across-subject alignment of lower order 689
cortical areas in the surface coordinate system, with average misregistration errors as 690
small as 2-3 mm for V1 (Fischl et al. 2008; Hinds et al. 2008; Yeo et al. 2010a), which is 691
around the spatial resolution of the present fMRI data. For higher-order regions such as 692
BA44, BA45 and hOc5/MT+, intersubject agreement is worse, with average 693

24
misalignment errors in the order of 6-12 mm (Fischl et al. 2008; Yeo et al. 2010a; 694
2010b), but still an improvement from standard volumetric alignment (Amunts et al. 695
1999). In the case of histologically defined hOc5, considered to be putative MT+ 696
(Malikovic et al. 2007), we were able to verify (not shown) that the probabilistic map of 697
hOc5 in FreeSurfer surface space (Yeo et al. 2010b) was consistent with that of MT+ 698
defined in Caret PALS space (Van Essen 2004) and peak MT+ coordinates reported in 699
fMRI attention literature (Shulman et al. 1999). Certain anatomical landmarks were also 700
useful in the selection of regional vertices. For example, the calcarine fissure was used as 701
a guide to select regions in the lower and upper visual field representations as well as in 702
the central and peripheral visual field representations within V1 (e.g., see Fig. 22). 703
704
Comparison of network boundaries with cytoarchitectonic areas 705
In addition to their utility for selecting regions, the probabilistic histological maps 706
were useful in relating the estimated network boundaries to human cytoarchitectonic 707
areas. Because the Statistical Parametric Mapping (SPM) Anatomy toolbox contained a 708
more complete set of probabilistic histological maps of the same subjects in MNI Colin27 709
volumetric space (Eickhoff et al. 2005), we projected these probability maps to 710
FreeSurfer surface space by establishing spatial correspondence between Colin27 and 711
FreeSurfer surface space using the same procedure as that used for mapping between 712
MNI152 volumetric space and FreeSurfer surface space (Buckner et al. submitted). 713
Cytoarchitectonic areas common to both datasets were those of the primary motor cortex 714
(areas 4a and 4p; Geyer et al. 1996), premotor cortex (area 6; Geyer 2004), primary 715
somatosensory cortex (areas 3, 2 and 1; Geyer et al. 1999), early visual cortex (areas 17 716
and 18; Amunts et al. 2000), hOc5/MT+ (Malikovic et al. 2007) and BA44/45 (Amunts et 717
al. 1999). Consistent with previous discussion about surface-based versus volume-based 718
registration, we found Eickhoffs probabilistic maps in FreeSurfer space to be more 719
diffuse than the maps obtained from the purely surface-based approach, implying poorer 720
intersubject alignment. Consequently, for cytoarchitectonic areas common to both 721
datasets, the surface-based probabilistic maps were used (Fischl et al. 2008; Yeo et al. 722
2010b) (e.g., see Fig. 22). We were also able to verify reasonable overlap (not shown) 723
between the projected Eickhoffs maps and the purely surface-based probabilistic maps 724

25
for areas common to both datasets, substantiating the validity of the mapping between the 725
Colin27 space and FreeSurfer surface space. 726
727
Effect of resting condition on functional connectivity 728
For certain analyses, it was important to check that findings were not the result of 729
overt eye movements that might shift edges and visual boundaries in and out of the 730
central field. The core dataset (N = 1000) employed an eyes open rest condition because 731
it is comparable to visual fixation in terms of signal strength (Van Dijk et al. 2010) but 732
can be acquired in studies that do not employ a setup for visual display. To examine the 733
effects of the task employed during the resting-state, the effect of condition was analyzed 734
for certain key analyses (e.g., for analyses that quantified the functional connectivity 735
strengths among visual regions). As the results will show, the type of resting condition 736
(EOR, ECR or FIX) is not a significant factor influencing our results (e.g., see Fig. 17). 737
738
Visuotopic fMRI data 739
The analyses of the visual cortex involved the visuotopic organization of the V1- 740
V3 complex. The fMRI data were analyzed in the native subjects volumetric space and 741
the results sampled onto FreeSurfer surface space and averaged across subjects. The 742
details of the analysis, which provided eccentricity estimates of the visual representation 743
in the V1-V3 complex, are described elsewhere (Hinds et al. 2009). A 1mm smoothing 744
kernel was applied to the averaged eccentricity estimate in FreeSurfer surface space. 745
Because of the limited range of visual angle that could be stimulated in the MRI scanner, 746
and because fixational eye movements that occur during visual stimulation prevent stable 747
stimulation of the fovea, the eccentricity estimates did not cover the representation of the 748
periphery or of the center of visual field within the V1-V3 complex (Hinds et al. 2009; 749
Polimeni et al. 2005), but were sufficient for our analyses. 750
751
Distribution of parcellations and raw data 752
A primary result of this study is the parcellation of cortical networks and the 753
estimation of boundaries of regions within the networks. The parcellations of parietal and 754
prefrontal cortices, in particular, represent demarcations of complex topographical 755

26
regions that have been perplexing to understand in relation to task-based functional 756
neuroimaging studies. We have uploaded the parcellations in Caret PALS surface space 757
into the Surface Management System Database (SumsDB) for open use (Dickson et al. 758
2001) (link under construction). The parcellations in FreeSurfer surface space can be 759
downloaded from http://surfer.nmr.mgh.harvard.edu/fswiki. The raw fMRI data from the 760
1000 subjects will be made openly available to researchers using the procedures 761
established by the OASIS data releases (Marcus et al. 2007; 2010) and the 1000 762
Functional Connectomes Project (Biswal et al. 2010). 763
764
Results 765
766
Estimates of cerebral networks are reliable 767
The cerebral cortex was parcellated into multiple networks using clustering. The 768
parcellations resulted in networks that involved primarily adjacent areas (e.g., visual 769
cortex) and networks that involved areas widely distributed throughout the cortex (e.g., 770
heteromodal association cortex). Figure 7 shows the 7-network estimates for the 771
Discovery and Replication samples. A total of 97.4% of the vertices were assigned to the 772
same networks across both datasets. Varying the particular choice of binarization 773
threshold (ranging from 5% to 15%) and smoothing (ranging from no smoothing to 6mm 774
FWHM) minimally affected the results (not shown). Figure 8 shows the confidence 775
(silhouette) value for each vertex with respect to its assigned network for the 7-network 776
estimate. Regions close to the boundaries between networks were less confident in their 777
assignment. Spatial variation within individual components of the estimated networks 778
was also observed beyond the boundary regions. Often these low-confidence regions 779
anticipated further fractionation of the networks into smaller subnetworks that emerged 780
when larger numbers of networks were allowed (e.g., compare the lateral prefrontal 781
extent of the orange network in Fig. 7 with its confidence map in Fig. 8, and then note 782
subsequent fractionation of this region in Fig. 9). Figure 9 shows the 17-network 783
estimates for the Discovery and Replication data samples, and Figure 10 shows the 784
confidence map for the Discovery dataset. For the 17-network estimate, 97.0% of the 785
vertices were assigned to the same networks across both datasets. 786

27
787
------------------------------------------------------------------ 788
Insert Figures 7, 8, 9 and 10 About here 789
------------------------------------------------------------------ 790
791
Estimates of cerebral networks from 1000 subjects 792
To provide the best estimates of the cerebral cortical networks, clustering was 793
performed on the full sample of 1000 subjects. Figures 11 and 13 show the 7- and 17- 794
network parcellation estimates. Several results are notable. A salient feature of the 795
estimated networks is the separation of the early sensory and late motor cortices (blue and 796
purple) from association cortex, consistent with the observation that early sensory and 797
late motor regions exhibit dense local anatomical connectivity in primates (Felleman and 798
Van Essen 1991; Jones et al. 1978; Markov et al. 2010) and preferential local functional 799
coupling in humans (Sepulcre et al. 2010). Sensory and motor cortices, whose functional 800
connectivity networks were preferentially local, comprised only 35% of the cerebral 801
mantle and were the exception in terms of network structure. 802
The majority of the human cerebral cortex is comprised of multiple, distinct 803
networks of association areas. The association networks in the 7-network estimate 804
converged and extended upon networks previously described in the resting-state 805
literature, including those referred to as the dorsal and ventral attention (green and violet 806
respectively; Fox et al. 2006), the frontoparietal control (orange; Dosenbach et al. 2007; 807
Vincent et al. 2008) and the default (red; Buckner et al. 2008; Greicius et al. 2003) 808
networks (Fig. 12). We also note that the 7-network parcellation of the parietal cortex is 809
similar to those proposed using seed-based approaches (Vincent et al. 2008) and using the 810
areal boundary detection method (Cohen et al. 2008; Nelson et al. 2010). The 811
convergence of multiple different analysis approaches suggests that the parcellation is 812
intrinsic to the resting-state data rather than an artifact of the algorithm used. 813
Generally, the 17-network estimate fractionated the 7-network estimate into 814
smaller subnetworks. Some aspects of the fractionation, such as the emergence of a 815
parahippocampal-retrosplenial-lateral parietal network, are anticipated by other studies 816

28
using hierarchical-clustering techniques (e.g., Andrews-Hanna et al. 2010). Other aspects 817
of the fractionation were unexpected, such as the emergence of subnetworks within the 818
visual and motor cortices that did not respect areal boundaries but rather appear to align 819
with topographic organization. In the following sections, we quantify and further explore 820
the patterns of functional connectivity that give rise to these networks. 821
822
------------------------------------------------------------------ 823
Insert Figures 11, 12 and 13 About here 824
------------------------------------------------------------------ 825
826
A cautionary note about potential artifacts 827
Before exploring the estimated networks in more detail, it is important to point 828
out aspects of the data that are difficult to interpret because of potential fMRI signal 829
blurring across gyri and signal loss associated with susceptibility (Ojemann et al. 1997). 830
Figure 14A illustrates one example. Somatomotor, auditory and posterior insular cortices 831
are correlated within a single network (also see Fig. 11). The ventral portion of 832
somatomotor cortex is clustered with the auditory cortex in the 17-network parcellation 833
(Fig. 13). While non-human primate tracing studies suggest auditory and somatomotor 834
cortices are connected via multiple areas within the insular cortex (Disbrow et al. 2003; 835
Friedman et al. 1986; Mesulam and Mufson 1982; Mufson and Mesulam 1982), an 836
equally likely explanation for the observed correlation is blurring of the BOLD signal 837
across the Sylvian fissure (Fig. 14A). We could not find a way, in these data, to resolve 838
whether the coupling was an artifact of limited resolution or a true, coupled network. 839
Figure 14B illustrates a second example of how fMRI signal blurring might affect 840
the interpretation of the results. In this case, the primary somatosensory cortex (S1) and 841
primary motor cortex (M1) are clustered within the same network (Fig. 11). While there 842
is anatomical evidence of direct connectivity between S1 and M1 in the macaque (Jones 843
et al. 1978; Pons and Kaas 1986), we are unable to resolve whether the coupling was an 844
artifact of limited resolution due to the close proximity of M1 and S1 in volumetric space. 845

29
As an example of uncertainty occurring near regions of MR susceptibility, Figure 846
14C illustrates a cream-colored network of regions in the temporal pole and orbital 847
frontal cortex (also see Fig. 11). While there is anatomical evidence from the primate 848
tracing literature supporting the existence of this network (Carmichael and Price 1995; 849
Kondo et al. 2003; Moran et al. 1987), the spatial distortion and signal loss caused by MR 850
susceptibility creates uncertainty in the location of the network boundaries. 851
852
------------------------------------------------------------------ 853
Insert Figure 14 About here 854
------------------------------------------------------------------ 855
856
Sensory and motor cortices exhibit topographically-specific functional connectivity 857
The sensory and motor cortices were clustered separately from the association 858
cortex in the 7-network estimate (Fig. 11). This result by itself suggests that sensory and 859
motor cortices distinguish themselves from distributed networks of association areas. 860
Within the sensory and motor networks, there were a number of further observations. Of 861
most interest, the 17-network parcellation fractionated the sensory and motor cortices into 862
subnetworks (Fig. 13). Specifically, early visual regions formed two distinct subnetworks 863
that did not respect areal boundaries. Somatomotor cortex was similarly fractionated 864
along its lateral extent. Our hypothesis is that these fractionations reflect topographic 865
organization topographic representation of visual space in the visual regions and 866
topographic representation of body space in the somatomotor regions. 867
Visual topography. The visual network in the 7-network estimate (Fig. 11) was 868
fractionated into two separate subnetworks (purple and bright red) in the 17-network 869
estimate (Fig. 13). The boundary between the two visual subnetworks cut perpendicularly 870
across the calcarine fissure, suggesting the division of the early visual areas into central 871
and peripheral components. To evaluate this possibility, Figure 15 overlays the 872
eccentricity estimates of the Visuotopic dataset over the boundaries of the two separate 873
visual subnetworks. The early visual areas were divided into two subnetworks along an 874
isoeccentricity line of approximately 4
o
. We refer to these two subnetworks as central 875

30
and peripheral, although an immediate question arises as to whether the division of 876
lower visual areas into central and peripheral components extends to higher visual areas. 877
The particular traveling wave paradigm used in the Visuotopic dataset was designed to 878
estimate visual eccentricity within the V1-V3 complex and is therefore unreliable outside 879
the complex (Wandell et al. 2007). To gain further clarification on visual areas outside 880
the V1-V3 complex, we inspected the boundaries of the 17-network parcellation overlaid 881
on the map of approximate human visual areas provided by Van Essen (2004). The 882
boundary between the central and peripheral representations continues through the 883
extrastriate visual areas consistent with the possibility that the division of lower visual 884
areas into central and peripheral components generally applies to the extrastriate cortex 885
with certain caveats that will be taken up in the Discussion. 886
887
------------------------------------------------------------------ 888
Insert Figure 15 About here 889
------------------------------------------------------------------ 890
891
To assess the validity of the clustering analysis of visual cortex, six seed regions 892
were selected from the Discovery sample (Table 1), and their fcMRI maps were 893
computed using the Replication sample. The regions were selected to include V1 and V3 894
regions that fell within the central and peripheral representations. Using the calcarine 895
fissure, histological V1 estimates (Amunts et al. 2000; Fischl et al. 2008) and the network 896
boundaries as landmarks, two regions labeled V1
c
and V1
p
were selected to correspond to 897
central and peripheral V1 respectively. Using the probabilistic histological maps of the 898
SPM Anatomy toolbox (Eickhoff et al. 2005), two regions labeled V3
cv
and V3
pv
were 899
selected at or near visual area V3v (Rottschy et al. 2007; Wilms et al. 2010) 900
corresponding to the central and peripheral representations respectively. The two 901
remaining regions were selected from the extrastriate regions of the central and peripheral 902
visual subnetworks. In all cases, the confidence (silhouette) map of the 17-network 903
estimate from the Discovery sample (Fig. 10) was used as a guide. 904

31
The fcMRI maps of the six seed regions confirm the network demarcations (Fig. 905
16). Compared with the other seed regions of the central visual system, V1
c
demonstrated 906
weaker correlation to the extrastriate regions of the central visual system. This is reflected 907
by the lower confidence of V1
c
in its network assignment as shown in Table 1. Consistent 908
with this observation, as the number of networks in the clustering analysis was increased 909
(not shown), the central V1 region separated from the extrastriate component of the 910
central visual system. 911
912
------------------------------------------------------------------ 913
Insert Table 1 and Figure 16 About here 914
------------------------------------------------------------------ 915
916
To quantify the dissociation between the central and peripheral representations 917
within the visual subnetworks, Figure 17A shows polar plots of the correlation of V1
p
918
and V1
c
with five regions in the Replication sample. Since the Discovery and Replication 919
samples consist of resting data collected under the eyes open rest condition, the patterns 920
of functional connectivity were also quantified for the Task Effect dataset (Fig. 17B). The 921
results revealed that the central and peripheral V1 seed regions displayed distinct patterns 922
of functional connectivity that generalized across multiple data acquisition conditions 923
including eyes closed rest and visual fixation. 924
925
------------------------------------------------------------------ 926
Insert Figure 17 About here 927
------------------------------------------------------------------ 928
929
Although there are differences in visual field properties, such as magnification 930
factors and receptive field sizes, between the central and peripheral regions of the V1-V3 931
complex, these differences vary smoothly from central to peripheral vision within an area 932
(Balasubramanian et al. 2002; Dow et al. 1981; Van Essen et al. 1984). To explore this 933

32
further, functional connectivity was examined among seed regions spanning the 934
eccentricity axes of V1 and V3v (Fig. 18). Results did not suggest a sharp transition in 935
functional connectivity between V1 and V3v seed regions moving from central to 936
peripheral representations. The resulting division of the V1-V3 complex along the 937
isoeccentricity line of 4
o
was therefore likely driven by the functional connectivity of 938
visual regions outside the V1-V3 complex or may be an artifact of the small number of 939
networks being mandated by the analyses. 940
941
------------------------------------------------------------------ 942
Insert Figure 18 About here 943
------------------------------------------------------------------ 944
945
Somatomotor topography. The 7-network parcellation estimate clustered the 946
somatomotor cortex into a single network (the blue network in Fig. 11). Figure 19 shows 947
the boundaries of the 7-network estimate overlaid on the probabilistic histological maps 948
of areas 6 (Geyer 2004), 2 (Grefkes et al. 2001) and 5L (Scheperjans et al. 2008a; 2008b). 949
The histological estimates of areas 1, 3 and 4 (Geyer et al. 1996; Geyer et al. 1999) are 950
sandwiched between areas 2 and 6 and are not shown. Recent work (Amunts et al. 2010) 951
has delineated three additional premotor areas anterior to the ventral half of area 6 shown 952
in Figure 19; the ventral half of area 6 in Figure 19 is therefore an underestimation. Based 953
on these areal references, the somatomotor network likely includes MI (area 4) and 954
caudal premotor area 6, SI (areas 3, 1 and 2) and most, if not all of early somatosensory 955
area 5L. The somatomotor network also includes a small portion of the mid-cingulate 956
sulcus and possibly area 5M (not shown; Scheperjans et al. 2008a; 2008b). 957
958
------------------------------------------------------------------ 959
Insert Figure 19 About here 960
------------------------------------------------------------------ 961
962

33
The 17-network parcellation divided the somatomotor strip into dorsal and ventral 963
subnetworks across the axis that represents body space (Fig. 13). To investigate this 964
division, the parcellations were compared to activation maps of 24 subjects who were 965
instructed to move their tongue, hand or foot in response to a visual cue (for a detailed 966
explanation of this dataset, see Buckner et al. submitted). As shown in Figure 20A, the 967
boundary between the dorsal and ventral somatomotor subnetworks was roughly 968
positioned between the hand and tongue representations. To quantify this observation, 969
seed regions were selected from the left hemisphere hand, foot and tongue activation 970
maps and verified to fall within the probabilistic histological map of area 4 (Fischl et al. 971
2008; Geyer et al. 1996). Figure 20B shows pairwise correlations computed among the 972
hand, foot and tongue regions averaged over 1000 subjects (Hand coordinates: -41, -20, 973
62; Foot coordinates: -6, -26, 76; Tongue coordinates: -55, -4, 26). The hand-foot 974
correlation was significantly higher than the hand-tongue correlation (p < .001) and foot- 975
tongue correlation (p < .001). Thus, like the visual system, functional coupling forms 976
networks within the somatomotor system that reflect topographic organization. 977
There were two further results that must be interpreted cautiously because of 978
volumetric signal blurring. The 7- and 17-network parcellations of somatomotor cortex 979
included both the precentral and postcentral representations of body space thus forming 980
networks that spanned areas (M1 and S1). This observation is difficult to interpret 981
because precentral and postcentral gyri abut each other in volumetric space (see Fig. 982
14B). Similarly, the ventral somatomotor network in the 17-network parcellation 983
included parts of the insula and auditory cortex, perhaps reflecting a polysynaptic circuit 984
of functional coupling linked to speech movements and hearing ones own voice. We do 985
not interpret this observation because of their volumetric proximity (see Fig. 14A). 986
987
------------------------------------------------------------------ 988
Insert Figure 20 About here 989
------------------------------------------------------------------ 990
991
Asymmetry of functional coupling varies across somatomotor topography 992

34
Analysis of asymmetries in functional coupling is beyond the scope of this paper. 993
However, we observed an interesting variation in the asymmetry of functional coupling 994
that reinforces the observation of functional differences along the somatomotor body 995
representation. Specifically, pairwise correlations between homotopic pairs of hand, foot 996
and tongue representations (Hand coordinates: 41, -20, 62; Foot coordinates: 6, -26, 997
76; Tongue coordinates: 55, -4, 26) were measured between the hemispheres (Fig. 21A). 998
The homotopic hand correlation was significantly weaker than that of the homotopic foot 999
(p < .001) and tongue (p < .001) correlations. We are unable to rule out the possibility 1000
that the higher correlation between homotopic tongue regions is an artifact of subjects 1001
moving their tongues during scanning. 1002
1003
------------------------------------------------------------------ 1004
Insert Figure 21 About here 1005
------------------------------------------------------------------ 1006
1007
To explore the somatotopy of interhemispheric fcMRI beyond the hand, foot and 1008
tongue representations, we estimated the sequence of vertices lying in the shortest path 1009
connecting the left tongue region with the left hand region and those lying in the shortest 1010
path connecting the left hand region with the left foot region. For each left hemisphere 1011
motor vertex, we found the corresponding right hemisphere vertex that was maximally 1012
correlated with it in the Discovery sample. Figure 21B plots the correlation between the 1013
left hemisphere vertices, arranged ventral to dorsal, and the corresponding right 1014
hemisphere vertices in the Replication data sample. Defining the maximally correlated 1015
right hemisphere vertices in the Discovery sample avoided bias in the correlation values 1016
and the issue of picking truly homotopic representations in either hemisphere. Consistent 1017
with Figure 21A, the maximal tongue correlation was higher than that of the foot, which 1018
was in turn higher than that of the hand. The region in between the hand and foot 1019
representations, possibly corresponding to the trunk representation, also displayed higher 1020
maximal interhemispheric correlation than those of the hand or foot representations. 1021

35
The observed somatotopy of interhemispheric fcMRI is consistent with non- 1022
human primate studies that have shown that the representations of midline structures in 1023
S1 and M1, such as the face and trunk, have denser callosal connections than those of 1024
distal limbs, such as the hand and foot (Gould et al. 1986; Jones and Wise 1977; 1025
Killackey et al. 1983). 1026
1027
Hierarchical processing within a canonical sensory-motor pathway 1028
Parcellation of the cerebral cortex into distinct networks will not capture 1029
information about interactions between regions that fall across separate networks. This is 1030
particularly problematic because the canonical system-level description of cortical 1031
processing involves interactions across hierarchical pathways of sensory and motor areas. 1032
The above analyses leave open the question of how the distinct networks interact as is 1033
expected for sensory-motor pathways. 1034
To explore this question, we selected the canonical sensory-motor pathway that 1035
extends from primary visual cortex to the precentral motor regions (including the putative 1036
homologue of FEF) via the motion-sensitive MT+ complex and posterior parietal cortex 1037
at or near putative human LIP. In the macaque literature, this pathway has been 1038
extensively studied in relation to sensory-guided decisions resulting in eye movements 1039
and associated processes linked to spatial attention (e.g., Andersen and Buneo 2002; 1040
Colby and Goldberg 1999; Gold and Shadlen 2007; Shadlen and Newsome 2001). In the 1041
human literature, this pathway has been studied both in relation to spatially-directed 1042
movements and also in relation to spatial attention, with components of the pathway 1043
sometimes referred to as the dorsal attention system or network (Corbetta and Shulman 1044
2002). Most critically, the anatomic relations between the areas within the pathway have 1045
been extensively explored beginning with the seminal work of Maunsell and Van Essen 1046
(1983). 1047
Early visual cortex. Analysis focused initially on the bottom of the sensory-motor 1048
pathway by investigating functional connectivity between V1 and putative human MT+
2
. 1049

2
In humans, the MT+ complex is used to denote the putative human homologue of
macaque area MT and neighboring visual areas that are sensitive to motion stimuli
(DeYoe et al. 1996). Here we are able to constrain the location of MT+ using surface

36
For this analysis, two V1 regions were selected in dorsal (V1
cd
) and ventral (V1
cv
) central 1050
V1 corresponding to the lower and upper visual field representations. Two additional V1 1051
regions were selected in dorsal (V1
pd
) and ventral (V1
pv
) peripheral V1 corresponding to 1052
the lower and upper visual field representations. Two MT+ seed regions (MT+
d
, MT+
v
) 1053
were selected following the dorsoventral extent of the surface-based probabilistic 1054
histological map of MT+ (Malikovic et al. 2007; Yeo et al. 2010b). The MT+
d
and MT+
v
1055
seed regions likely correspond to peripheral and central visual field representations 1056
respectively (Huk et al. 2002; Maunsell and Van Essen 1987) and were analyzed to 1057
illustrate differential connectivity within MT+. Because of the over-representation of the 1058
lower visual field within macaque MT (Maunsell and Van Essen 1987), the distinct MT+ 1059
seed regions might also represent the lower visual field, although human fMRI studies 1060
have so far failed to yield evidence of this representation bias within MT+ (Amano et al. 1061
2009; Kolster et al. 2010; Tootell et al. 1995). In the following analyses, MT+ was either 1062
analyzed as two regions using the extreme seed regions (MT+
d
and MT+
v
) or as a single 1063
region using the center seed region (MT+) for analyses that did not require the 1064
complexities of topographic distinctions. A single aMT+ (anterior MT+) region was 1065
chosen anterior to and outside the histological MT+
3
. The regions are shown in Figures 1066
22A and 24; their coordinates are reported in Table 2. 1067
1068
------------------------------------------------------------------ 1069
Insert Table 2 and Figure 22 About here 1070
------------------------------------------------------------------ 1071
1072

based histological maps of hOc5 which is thought to be the cytoarchitectonic correlate of
the human MT+ complex (Malikovic et al. 2007; Yeo et al. 2010b).
3
Multiple seed regions are used throughout the paper. The abbreviation of the names of
these seed regions obeyed the following convention: the suffix following a region
indicates relative spatial location within the region, while the prefix preceding a region
indicates relative spatial location outside the region. Therefore, MT+
d
is a seed region
within the dorsal aspect of the MT+ complex, while aMT+ is a seed region anterior to and
outside the MT+ complex.

37
Figure 22B quantifies the differential functional coupling of the four V1 regions 1073
with MT+ and aMT+ within the Replication sample. V1 showed strong functional 1074
coupling to MT+ and significantly weaker coupling to aMT+ (p < .001 for both MT+
v
1075
and MT+
d
). The dorsal and ventral MT+ regions also demonstrated differential functional 1076
coupling with the four V1 regions, confirmed by a 2x2x2 ANOVA including eccentricity 1077
(central or peripheral), polar angle (upper or lower visual field), and MT+ region (dorsal 1078
or ventral). Both MT+ regions were more strongly coupled with the lower visual field V1 1079
seed regions than the upper visual field V1 seed regions (p < .001). Furthermore, dorsal 1080
MT+ was more strongly coupled with peripheral V1 than central V1 (p < .001), while 1081
ventral MT+ was more strongly coupled with central V1 than peripheral V1 (p < .05). 1082
These results suggest the topographic pathways between V1 and the MT+ complex are 1083
largely consistent with the visual field representations of the two areas. 1084
Since V1 is a relatively large structure compared with MT+, to ensure the results 1085
were robust to the particular choice of the V1 seed regions, Figure 23 shows fcMRI maps 1086
of the aMT+, MT+
d
and MT+
v
regions computed using the Replication dataset. By all 1087
accounts, V1 is functionally coupled to MT+ while aMT+ shows minimal coupling, 1088
leading to their separation into distinct networks in the cortical parcellation. 1089
1090
------------------------------------------------------------------ 1091
Insert Figure 23 About here 1092
------------------------------------------------------------------ 1093
1094
Visual association and parietal association cortices. Moving up the sensory- 1095
motor pathway, the functional connectivity of MT+ and aMT+ with parietal and frontal 1096
cortices was next examined. Figure 24 shows the fcMRI maps of MT+ and aMT+ seed 1097
regions computed using the Replication sample with views focusing on the parietal and 1098
lateral frontal cortices. In the parietal lobe, both MT+ and aMT+ demonstrated 1099
correlation with the SPL and IPS, but little or no correlation with the inferior parietal lobe 1100
(IPL). In the frontal cortex, correlations were mostly limited to the precentral sulcus and 1101
gyrus. Compared with the MT+ seed region, the aMT+ seed region demonstrated 1102

38
stronger correlation with the parietal and frontal cortices. The MT+ and aMT+ seed 1103
regions were also maximally correlated with different parts of the parietal and frontal 1104
cortices, suggesting differential influence on nearby regions of sensory-association 1105
cortex. 1106
1107
------------------------------------------------------------------ 1108
Insert Figure 24 About here 1109
------------------------------------------------------------------ 1110
1111
To quantify the pattern of differential functional coupling, we computed the 1112
correlation of the MT+ and aMT+ seed regions with four visual, four parietal and two 1113
frontal cortical regions (Fig. 25A). Two of the visual cortex regions, V1
cd
and V1
pd
, were 1114
previously used in the V1-MT+ connectivity analysis. V3A and V4 were selected based 1115
on their high correlation with MT+ in the Discovery sample and named using the 1116
approximate map of human visual areas (Van Essen 2004) as reference. The four parietal 1117
regions were chosen at or near the IPS. Three of the parietal regions (IPS2, SPL7A, 1118
SPL7P) were selected based on a meta-analysis of fMRI literature of tasks that reportedly 1119
activate the human homologues of macaque areas AIP (anterior intraparietal), LIP (lateral 1120
intraparietal) and PIP (posterior intraparietal) respectively. The final parietal region, 1121
IPS3
m
, was selected using the Discovery sample to be physically between IPS2 and 1122
SPL7A. The parietal regions were labeled based on their proximity to the probabilistic 1123
histological maps. In particular, IPS2 is at or near area hIP2 at the anterior most part of 1124
IPS (Choi et al. 2006), while SPL7A and SPL7P are at or near areas 7A and 7P of the 1125
SPL (Scheperjans et al. 2008a; 2008b). Finally, IPS3
m
is on the medial wall of IPS at or 1126
near area hIP3 (Scheperjans et al. 2008a; 2008b). The dorsal frontal region, putatively 1127
FEF, was selected based on the meta-analysis of fMRI literature of saccade tasks. The 1128
ventral frontal region PrC
v
(precentral ventral) was selected based on its high correlation 1129
with aMT+ in the Discovery sample. The locations of the visual, parietal and frontal 1130
regions are reported in Table 2. Table 3 summarizes the set of fMRI studies used to 1131
derive the coordinates of IPS2, SPL7A, SPL7P and FEF. 1132

39
1133
------------------------------------------------------------------ 1134
Insert Table 3 and Figure 25 About here 1135
------------------------------------------------------------------ 1136
1137
Figure 25B shows the polar plots of the regional correlations using the 1138
Replication sample. To ensure that the results were not the result of overt eye 1139
movements, the polar plots were also replicated using the Task Effect dataset in Figures 1140
25C and 25D. In all cases, MT+ had significantly stronger correlation with early visual 1141
cortex as compared with aMT+. In contrast, aMT+ had significantly stronger correlation 1142
with the parietal and frontal regions compared with MT+. MT+ and aMT+ are strongly 1143
coupled to one another (r = 0.30 in the Replication sample). Thus, consideration of 1144
functional coupling between these regions in detail suggests that early visual areas are 1145
coupled to regions at or near MT+, which are in turn directly (or indirectly through 1146
intermediate regions, such as aMT+) coupled to parietal and frontal regions associated 1147
with sensory-motor integration. The differential coupling of MT+ and aMT+ resulted in 1148
their inclusion into distinct cortical networks. 1149
Frontoparietal interactions. We next considered the differential functional 1150
coupling of FEF and PrC
v
with the parietal cortex. Figure 26 shows the fcMRI maps of 1151
FEF and PrC
v
computed using the Replication sample. While both FEF and PrC
v
1152
demonstrated strong correlation with the SPL and IPS, differences in correlation patterns 1153
also emerged. Specifically, PrC
v
was strongly correlated with more ventral portions of 1154
rostral SPL and IPS, while FEF was strongly correlated with caudal SPL and IPS. To 1155
quantify this phenomenon, we computed the correlation of FEF and PrC
v
with five 1156
parietal regions at or near SPL and IPS. Four of the parietal regions (IPS2, IPS3
m
, 1157
SPL7A, SPL7P) were the same as those used in the previous analysis. Using the 1158
Discovery sample, a fifth parietal region was selected on the lateral wall of rostral IPS, 1159
within what is often termed the frontoparietal control network (orange; Fig. 11). Since the 1160
region was located at or near the histological map of hIP1 (Choi et al. 2006) projected to 1161

40
FreeSurfer space, we labeled the seed IPS1. The regions are displayed in Figure 27A and 1162
their coordinates reported in Table 2. 1163
1164
------------------------------------------------------------------ 1165
Insert Figures 26 and 27 About here 1166
------------------------------------------------------------------ 1167
1168
Figure 27B shows the correlation of FEF and PrC
v
with the five parietal regions 1169
computed in the Replication sample, arranged in rostral (lateral) to caudal (medial) order. 1170
The rostrolateral IPS regions (IPS1, IPS2, IPS3
m
) were more strongly correlated with 1171
PrC
v
than FEF, while the mediocaudal SPL regions (SPL7A, SPL7P) were more strongly 1172
correlated with FEF than PrC
v
. A 2x6 ANOVA including frontal regions (FEF, PrC
v
) and 1173
parietal regions (IPS1, IPS2, IPS3
m
, SPL7A, SPL7P) found a significant interaction (p < 1174
.001). In particular, FEF was more strongly correlated with the three IPS seeds (all p < 1175
.001) while PrC
v
was more strongly correlated with SPL7A and SPL7P (both p < .001). 1176
These results reveal differential coupling between distinct parietal and frontal regions, 1177
likely reflecting multiple pathways involved in sensory-motor integration (Kurata 1991; 1178
Rizzolatti et al. 1998; Tann-Garipy et al. 2002). 1179
Hierarchical analysis. Assuming a visual hierarchy similar to that proposed by 1180
Maunsell and Van Essen (1983) exists in human cerebral cortex, V1 is expected to be 1181
near the bottom of the hierarchy and FEF to be near the top (also see Felleman and Van 1182
Essen 1991; Ungerleider and Desimone 1986). While there are uncertainties as to the 1183
exact homologies of our MT+ and aMT+ seed regions, they are likely to be in the middle 1184
of the hierarchy with aMT+ higher than MT+. In our analysis, we found that V1 and 1185
MT+, which are close together in the hierarchy, have stronger correlation than V1 and 1186
aMT+, which are farther apart in the hierarchy (Figs. 22 and 25). We also found that 1187
aMT+ and FEF, which are closer in the hierarchy, have stronger correlation than MT+ 1188
and FEF, which are farther apart in the hierarchy (Figs. 24 and 25). 1189
These results suggest that correlation between regions in the visual hierarchy may 1190
provide some insight into the organization of the pathways. To explore this hypothesis, 1191

41
we examined the arrangement of six seed regions (V1
pd
, V3A, MT+, aMT+, SPL7A, 1192
FEF) discussed earlier based on the assumption that stronger correlations are consistent 1193
with closer positioning in the processing hierarchy. By examining alternative 1194
arrangements of regions, the analysis was able to investigate which particular 1195
arrangement was most consistent with the functional connectivity pattern. Figure 28 1196
illustrates this approach. Figure 29 shows the two best hierarchical arrangements of the 1197
six regions when seeking a five-level hierarchy using the Replication sample. In both 1198
cases, the relative ordering of the regions within the functional hierarchy agrees well with 1199
the proposed macaque visual hierarchy of Maunsell and Van Essen (1983). By contrast, 1200
alternative arrangements of the visual areas lead to large proportions of violated 1201
constraints (Figs. 29D and 29E). 1202
Control analyses using the Discovery set shows that the fcMRI strengths of all 1203
pairs of the six seed regions are inversely correlated (r = -0.78, p < .001) with respect to 1204
the hierarchical distance between the regions in the hierarchy as defined by Felleman and 1205
Van Essen (1991). By contrast, the fcMRI strengths of all pairs of the six seed regions are 1206
not correlated (p = 0.23) with respect to the physical distance between the seed regions. 1207
These results suggest that spatial proximity does not explain these correlations; rather, the 1208
pattern of correlations reflects relative positions of regions within the hierarchy. 1209
These findings converge on two main results. First, there are functional 1210
interactions between networks but the interactions are selective. V1 shows minimal 1211
coupling with association regions within parietal cortex but it does couple with MT+. The 1212
between-network interactions also do not typically violate the network organizations 1213
revealed earlier by the clustering approaches: regions show stronger functional coupling 1214
to other regions within their network as contrast to regions outside their network. Second, 1215
when examined together, the pattern of correlations between regions in different 1216
networks is consistent with a hierarchical organization by which sensory information 1217
could influence regions associated with motor control (Fig. 29). 1218
1219
------------------------------------------------------------------ 1220
Insert Figures 28 and 29 About here 1221

42
------------------------------------------------------------------ 1222
1223
Parietal and prefrontal association cortices possess multiple regions with distinct 1224
connectivity profiles 1225
The analyses above describe separate sensory and motor networks, and address 1226
the question of how sensory and motor networks might interact through intermediate 1227
areas such as aMT+. However, these networks make up only a minority of the human 1228
cerebral cortex. Networks of widely distributed regions comprise the majority of the 1229
human cerebral cortex as illustrated by the violet, orange, and red networks in the 7- 1230
network solution (Fig. 11). The network complexity expands when the 17-network 1231
parcellation is considered (Fig. 13). These association networks will be the remaining 1232
focus of the results. First, we examined the topography of parietal and frontal association 1233
cortices to quantitatively illustrate that multiple neighboring regions possess markedly 1234
different functional connectivity fingerprints. Next we explored the relations between 1235
distributed regions to demonstrate that the association cortex is comprised of multiple, 1236
interdigitated association networks. 1237
Parietal association cortex. The basic approach employed to characterize the 1238
connectivity properties of parietal cortex was to extend the strategy applied in Figure 25 1239
to construct polar plots that quantify coupling with multiple, distributed cortical regions. 1240
Following the work of Passingham et al. (2002), we refer to these plots as functional 1241
connectivity fingerprints with the idea that the graphical representation will facilitate the 1242
visualization of similarities in connectivity properties between regions as well as any 1243
differences that make them unique. The plots are quantitative in that eccentricity displays 1244
the strength of functional coupling between each seed region and a single target region. 1245
For this analysis, multiple regions within parietal cortex were selected to cover 1246
the IPL and immediately adjacent portions of the SPL. Regions were further selected to 1247
survey distinct networks as revealed by the clustering maps. Some of the regions were 1248
used previously in the sensory-motor pathway analysis (IPS2, SPL7A). Other regions 1249
(PF, PGa, IPS3
l
, PGp
d
and PGp
v
) were selected using the Discovery dataset and then 1250
labeled according to the probabilistic histological maps. In particular, the PF and PGa 1251
seed regions were at or near human inferior parietal areas PF and PGa, respectively 1252

43
(Caspers et al. 2006; 2008), while the IPS3
l
seed region was on the lateral lip of the IPS 1253
at or near human hIPS3 (Scheperjans et al. 2008a; 2008b). PGp
d
and PGp
v
corresponded 1254
to the dorsal and ventral aspects of area PGp in the IPL (Caspers et al. 2006; 2008). One 1255
additional seed region at or near the temporoparietal junction (TPJ) was selected based on 1256
the peak coordinates reported in a human fMRI study of mental state inference (Saxe and 1257
Powell 2006), which accords well with the predicted mean peak in a recent meta-analysis 1258
on the same topic (Van Overwalle and Baetens 2009). 1259
Twenty-two cortical target regions were distributed over the lateral and medial 1260
aspects of the cerebral cortex such that multiple regions covered each association 1261
network. Four of the cortical regions (aMT+, MT+, FEF and PrC
v
) were used previously 1262
in the sensory-motor pathway analysis. Many labels for the frontal seed regions reflect 1263
gross anatomical landmarks as atlas-based human histological references were not 1264
available; subscripts denote relative positions. It is worth noting that target region 6vr+ 1265
was selected to be anterior to the map of premotor area 6 (Geyer 2003) and labeled to 1266
reflect a new delineation of the ventral precentral sulcus (Amunts et al. 2010). Target 1267
region PFC
v
fell within the histological map of BA45 (Amunts et al. 1999). The 1268
coordinates of the parietal seed regions and cortical target regions are reported in Table 4. 1269
1270
------------------------------------------------------------------ 1271
Insert Table 4 and Figure 30 About here 1272
------------------------------------------------------------------ 1273
1274
The results are illustrated in Figure 30. Parietal association cortex is comprised of 1275
multiple regions that can show similarities but often display markedly different functional 1276
connectivity fingerprints. Visual inspection of the plots allows appreciation of the 1277
resemblance of correlation patterns for parietal seed regions that cause them to fall within 1278
common networks. For example, the PGp
d
, PGp
v
, and TPJ seed regions all fall within the 1279
default network as defined in the extant neuroimaging literature. Their connectivity 1280
fingerprints are largely defined by their correlations with other association and limbic 1281
regions, and near complete absence of correlations with visual regions including MT+ 1282

44
and aMT+. This general fingerprint pattern can be directly contrasted with parietal 1283
regions such as those represented by IPS2 and SPL7A that are associated with distributed 1284
cortical regions linked to sensory and motor function (e.g., strong correlations with MT+ 1285
and aMT+). These distinct parietal regions (IPS2 and SPL7A) are likely at or near the 1286
putative human homologue to macaque AIP and LIP and fall within the functional 1287
network that has been discussed as the dorsal attention network in the human literature 1288
and described in detail above in Figures 22 to 29. 1289
There are also subtle differences across regions that fall within the same broad 1290
networks. For example, while parietal seed regions TPJ, PGp
v
and PGp
d
are strongly 1291
correlated with posterior cingulate cortex (PCC) and the precuneus (pCun), PGp
v
has 1292
comparatively stronger correlation to retrosplenial cortex (RSP) and parahippocampal 1293
complex (PHC). In contrast, TPJ has the strongest correlation with posterior superior 1294
temporal sulcus (STS
p
) and lower correlation with RSP and PHC. PGp
d
has the strongest 1295
correlation with medial prefrontal cortex (PFC
m
). These differential correlations account 1296
for the fractionation of the posterior IPL into three components in the 17-network 1297
estimate (Fig. 13) and may be important to the functional properties of the region. 1298
Frontal association cortex. Analysis of frontal association cortex applied the 1299
same approach as for parietal association cortex. The 8 frontal seed regions selected for 1300
analysis were among the cortical targets in the parietal plots discussed previously. The 22 1301
cortical target regions selected for analysis were distributed throughout the cortex and 1302
selected so that multiple target regions covered each association network. Eleven regions 1303
were carried forward from the previous analysis of the parietal cortex (e.g., IPS2, SPL7A, 1304
FEF, PrC
v
, aMT+ and MT+) and these were arranged in the same spatial locations in the 1305
polar plots (Cinga clockwise to pCun). The remaining 11 cortical target regions were 1306
unique to analysis of frontal cortex but are arranged in the plots so that cortical target 1307
regions at the same location in the polar plots belong to the same network in the 7- 1308
network estimate. The arrangement of the cortical target regions therefore facilitates the 1309
comparison of polar plots across the parietal and frontal figures. The coordinates of the 1310
frontal seed regions and cortical target regions are reported in Table 4. 1311
1312
------------------------------------------------------------------ 1313

45
Insert Figure 31 About here 1314
------------------------------------------------------------------ 1315
1316
Figure 31 shows the polar plots of correlations for the 8 frontal seed regions with 1317
the 22 target regions in parietal, temporal, frontal and cingulate cortices. Many of the 1318
same properties observed for parietal association cortex were again apparent. Figure 31 1319
demonstrates that frontal seed regions in the same network generally share similar 1320
functional connectivity fingerprints that are distinct from those of neighboring seed 1321
regions belonging to a different network. A simple posterior-to-anterior hierarchy is not 1322
present as might be expected from some models of frontal hierarchy based on anatomical 1323
connectivity (e.g., Petrides 2005). Rather, regions with similar functional connectivity 1324
fingerprints appeared at posterior and anterior locations as evidenced by PFC
da
, which is 1325
considerably rostral to most of its partner regions, and PFC
dp
which is considerably 1326
caudal to its partner regions. 1327
The arrangement of the cortical target regions also allows a key feature of 1328
connectivity to be appreciated when frontal Figure 31 is directly compared to parietal 1329
Figure 30: parietal and frontal seed regions from the same network, for instance PG
pd
and 1330
PFC
m
, have similar functional connectivity fingerprints. These similarities suggest that 1331
association cortex is made of parallel interdigitated networks of regions consistent with 1332
suggestions based on double-labeling analyses of monkey association cortex (Goldman- 1333
Rakic 1988; Selemon and Goldman-Rakic 1988). In our final analysis, we explored this 1334
possibility directly by comparing functional connectivity maps for multiple seed regions 1335
distributed across each association network. 1336
1337
Association cortex is comprised of multiple, interdigitated large-scale networks 1338
The parcellations derived from clustering suggest that there are multiple, large- 1339
scale networks interdigitated throughout association cortex. Thus, components of the 1340
green, violet, orange and red networks in the 7-network estimates are spatially adjacent in 1341
the parietal, temporal and frontal cortices (Fig. 11). A similar organization is observed in 1342
the 17-network estimate (Fig. 13). However, such an impression could be an artifact of 1343

46
the winner-take-all implementation of clustering that assigned each cortical vertex to a 1344
single network. The complex functional connectivity fingerprints demonstrated that 1345
regions within the same networks showed similar functional connectivity fingerprints 1346
even when distributed across the cortex, consistent with their inclusion in common 1347
networks, but also revealed differences between regions that suggest subtler 1348
organizational properties. In our final analysis, we sought to explore patterns of 1349
functional connectivity using an exclusively seed-based approach in a comprehensive 1350
manner to contrast connectivity patterns for multiple regions embedded within the same 1351
networks. 1352
For this analysis, six left hemisphere seed regions from each of the four major 1353
association networks were analyzed beginning with the canonical sensory-motor network 1354
(known as the dorsal attention network in human neuroimaging literature). The ventral 1355
attention, frontoparietal control and default networks were subsequently explored. These 1356
four networks were identified in the 7-network estimate based on the Discovery dataset. 1357
Where possible we used the same seed regions as used by the other analyses in this paper. 1358
For this reason, the selected regions were not always the most confident in terms of 1359
network assignment as assessed by their silhouette plots or other means. Table 5 reports 1360
the coordinates of the seed regions. 1361
The left hemisphere fcMRI maps of the seed regions were computed using the 1362
Replication dataset and illustrated in Figures 32 to 35. Each seed region is functionally 1363
coupled primarily to regions within the same network, thus largely confirming the 7- 1364
network estimate and the existence of multiple, large-scale distributed networks in human 1365
association cortex. The network patterns were even reproduced when target regions were 1366
isolated in relation to adjacent cortical regions (e.g., Fig. 33A). Thus, this verifies that the 1367
networks identified via clustering capture the predominant functional coupling patterns 1368
within association cortex. Note that this result is not obligated the maps derived from 1369
the analysis of seed regions are not constrained to respect the borders of the networks 1370
defined by clustering. Furthermore, while the broad patterns confirmed the boundaries of 1371
the networks as expected, several exceptions were also observed. 1372
1373
------------------------------------------------------------------ 1374

47
Insert Table 5 and Figures 32, 33, 34 and 35 About here 1375
------------------------------------------------------------------ 1376
1377
Close inspection of the maps (Figs. 32 to 35) reveals evidence for crosstalk 1378
among the networks, consistent with the earlier analyses of the canonical sensory-motor 1379
pathway. For example, Figure 32C/D suggests interaction between the dorsal attention 1380
network and regions within the visual system, possibly related to the previously described 1381
hierarchical flow of information within the sensory-motor pathway (Fig. 29). In contrast, 1382
Figure 32F suggests functional coupling between the dorsal attention and the frontal 1383
components of the frontoparietal control and ventral attention systems. 1384
A further observation is that the fcMRI maps of seed regions chosen from the 1385
same network anticipate fractionation of the networks in ways also suggested by the 1386
analysis of connectivity fingerprints. For example, the PHC seed region is strongly 1387
coupled with the RSP, but not with the full extent of the posterior cingulate (Fig. 35E). It 1388
is also strongly coupled to the IPL but not the anterior IPL. This suggests that the 7- 1389
network estimate is likely insufficient to capture the full complexity of the functional 1390
couplings across different brain regions. As predicted by the fcMRI map of the PHC 1391
seed, the 17-network estimate differentiates the posterior IPL, RSP and the posterior PHC 1392
into an IPL-RSP-PHC system (dark blue in Fig. 13) distinct from the other regions of the 1393
default network. 1394
These collective results thus illustrate patterns of connectivity that are largely 1395
captured by a relatively small number of interdigitated, large-scale networks but also 1396
reveal more detailed properties that suggest crosstalk and fractionation within the major 1397
networks. 1398
1399
Discussion 1400
1401
The present results suggest association cortex comprises the majority of the human 1402
cerebral mantle and is made up of multiple, interdigitated association networks. The 1403
properties of association networks were found to be quite different from that of sensory 1404

48
and motor cortices. Sensory and motor areas are embedded within cerebral networks that 1405
are organized in a topographic fashion forming preferentially local networks meaning 1406
that adjacent areas tend to show strong functional coupling with one another. Hierarchical 1407
relations progress from early visual areas to premotor areas. By contrast, multiple 1408
association networks involve areas distributed throughout the cortex, always including 1409
discrete regions within prefrontal, parietal, temporal and midline cortices. These 1410
distributed association networks are interdigitated in a manner that yields complex zones, 1411
particularly in parietal and prefrontal association cortices. Within these zones, nearby 1412
regions possess markedly different connectivity patterns that can be explained by their 1413
being embedded within distinct association networks. In the following sections we 1414
explore the details of these patterns and discuss what the connectivity patterns suggest 1415
about how the human cerebral cortex may have expanded during evolution. 1416
1417
The cerebral cortex comprises multiple, distinct functionally coupled networks 1418
A primary result of our analyses is the cerebral maps depicted in Figures 11 and 1419
13. Figure 11 displays a coarse parcellation of the cerebral cortex into 7 functionally 1420
coupled networks, and Figure 13 displays a finer parcellation into 17 networks. These 1421
maps represent our current best estimate of the organization of the human cerebral cortex 1422
based on fcMRI. The large number of contributing subjects (N = 1000) and surface-based 1423
alignment procedures helped to detect topographical details with considerable 1424
confidence. However, these maps are still limited in resolution owing to between-subject 1425
averaging and the acquisition resolution of the data, and will presumably be improved in 1426
the future. At their present resolution, they have been surprisingly informative for 1427
providing insights into cerebral organization and, as will be illustrated by our companion 1428
paper, they also provide a basis to explore the organization of subcortical structures 1429
(Buckner et al. submitted). 1430
Two broad properties are immediately apparent when examining the maps, which 1431
will be the focus of discussion. First, sensorimotor and visual cortices form their own 1432
networks in the coarse parcellation that are fractionated into subnetworks in the finer 1433
parcellation. The fractionations do not follow simple divisions such as between 1434
somatosensory and motor areas along the central sulcus, or between early (e.g., V1) and 1435

49
late (e.g., V3) retinotopic visual areas in the occipital cortex. We will discuss hypotheses 1436
about what these fractionations might represent in upcoming sections. Second, 1437
association cortex comprises multiple, interdigitated networks that are distributed 1438
throughout cortex. Multiple aspects of these networks, in particular the properties of the 1439
coarse networks displayed in Figure 11, have previously been described using seed-based 1440
(e.g., Biswal et al. 1995; Greicius et al. 2003; Fox et al. 2006; Vincent et al. 2006) and 1441
independent component analysis (e.g., Beckmann et al. 2005; Damoiseaux et al. 2006; De 1442
Luca et al. 2006) approaches. Many aspects of the organization, especially within the 1443
higher resolution parcellation (Fig. 13), are novel. We will discuss the details of the 1444
organization of association cortex in upcoming sections with particular focus on zones of 1445
parietal and prefrontal cortices that represent nexuses of convergence among multiple, 1446
distinct networks. 1447
1448
Visual cortex displays local, topographically organized functional coupling 1449
Early visual areas were strongly functionally coupled to one another and only 1450
minimally correlated to regions outside of visual cortex. Examining the coupling 1451
properties more closely revealed evidence for an intrinsic gradient within the early visual 1452
cortex that likely corresponds to the transition from the central to peripheral visual field 1453
representations (Figs. 15 to 17). The division is not absolute (Fig. 18) but appears as an 1454
abrupt division into central and peripheral networks when assignments into distinct 1455
networks are forced, providing a convenient way to map the topographical organization 1456
across visual areas. Inspection of the boundaries of the functionally coupled networks on 1457
visual areas suggests that the division of lower visual areas into central and peripheral 1458
components might extend to higher visual areas. The ventral boundary of the central and 1459
peripheral visual systems continued outside the V1-V3 complex and divided V4v in two. 1460
Since the eccentricity representation of the V1-V3 complex is continuous through V4v 1461
(Brewer et al. 2005; Hadjikhani et al. 1998), it is likely that V4v was also divided into 1462
central and peripheral components. Anterior to V4v, at least two hemifield maps have 1463
been found, all of which have a distinctly large foveal representation and respond 1464
strongly to central visual stimuli throughout their extent (Brewer et al. 2005; Wandell et 1465

50
al. 2005), and are therefore consistent with the inclusion of V8 (Hadjikhani et al. 1998) 1466
within the central visual system. 1467
However, there were inconsistencies as might be expected because of limited 1468
resolution and the inherent complexity of the visual cortex (see Wandell et al. 2007 for a 1469
review). Multiple hemifield maps have been found in the extrastriate cortex lateral to the 1470
V1-V3 complex (DeYoe et al. 1996; Larsson and Heeger 2006; Tootell and Hadjikhani 1471
2001), such as regions that are part of the object-selective lateral occipital complex 1472
(LOC; Malach et al. 1995) and motion-selective area MT+ (Huk et al. 2002; Malikovic et 1473
al. 2007; Tootell and Taylor 1995). The inclusion of these entire visual regions within the 1474
central visual system suggests a violation of the central-peripheral division in these 1475
areas, although we note that the ventral visual stream (of which LOC is a part) is 1476
dominated by signals from the fovea (Baizer et al. 1991). The dorsal boundary of the 1477
central and peripheral visual networks continued outside the V1-V3 complex and cut 1478
through the hemifield maps dorsal and lateral to V3v, possibly including V3A, V3B and 1479
V7 (Larsson and Heeger 2006; Swisher et al. 2007; Tootell et al. 1998). Because these 1480
regions have separate foveal representation from the V1-V3 complex (see Fig. 1 in 1481
Larsson and Heeger 2006) and because of the complex trajectory of the boundary through 1482
these regions, we were unable to judge whether the central-peripheral division applied to 1483
these regions. 1484
The central-peripheral functional coupling of the human visual system is 1485
consistent with many aspects of known anatomy. Anatomical studies in non-human 1486
primates have shown that the topography of connections between visual areas generally 1487
respects the visual field representation (Cragg 1969; Maunsell and Van Essen 1983; Van 1488
Essen and Zeki 1978; Zeki 1969). For example, a region of V1 that responds strongly to a 1489
particular eccentricity and polar angle will project to the region of V2 that responds 1490
strongly to the same eccentricity and polar angle, presumably because of feedforward 1491
input (Van Essen and Zeki 1978). Because of the larger receptive field size of neurons in 1492
V2, the degenerated target area tends to be larger than that of the source lesion. Our result 1493
of higher functional connectivity strength between V1c and V3c (as well as between V1p 1494
and V3p) than between V1c and V1p may reflect the topography of anatomical 1495
connections within the visual system. 1496

51
If anatomical connectivity respects visuotopic representation and functional 1497
connectivity is constrained by anatomical connectivity, a question then arises as to why 1498
the clustering analysis divided the visual areas (especially the lower visual areas) into 1499
central and peripheral regions rather than into upper and lower visual fields? A split along 1500
the eccentricity axis is supported by proposals that higher visual areas are organized 1501
according to central versus peripheral field bias (Baizer et al. 1991; Grill-Spector and 1502
Malach 2004; Levy et al. 2001), although such an organization has also been disputed 1503
(Larsson and Heeger 2006; Wandell et al. 2005). A more mundane and likely reason is 1504
that eccentricity representation runs in parallel across multiple visual areas in contrast 1505
with angular representation that alternates in visual field sign across multiple visual areas 1506
(see Fig. 1 in Larsson and Heeger 2006; Sereno et al. 1995). Because this study utilized 1507
smoothing and intersubject averaging to boost SNR, any topography in functional 1508
connectivity that respects the high spatial frequency of angular representation is likely to 1509
be washed out. The resulting limitation in spatial resolution might also explain why 1510
visual regions dominated by foveal signals are grouped entirely within the central visual 1511
system even though they possess quarterfield or hemifield representations spanning both 1512
central and peripheral vision. 1513
Together these observations suggest that functional connectivity of the visual 1514
cortex followed the topographic organization of these areas up to a point. Certain 1515
observations, such as the lack of functional connectivity topography reflecting angular 1516
visual representation, could not be reconciled with prior anatomical observations. These 1517
differences may reflect the limitations of our approach, true differences between fcMRI 1518
and anatomical connectivity or novel connectivity findings that have yet to be revealed in 1519
anatomical studies. We will discuss these limitations and ambiguities later. 1520
1521
Somatomotor cortex displays topographical organization 1522
Estimates of functional connectivity networks grouped multiple somatosensory 1523
and motor areas into a single functionally coupled network for the low-resolution 1524
estimate of cortical organization (Fig. 11) and into a dorsal-ventral division for the high- 1525

52
resolution estimate (Fig. 13)
4
. Like visual cortex, the somatomotor network was 1526
characterized by its strong functional coupling to nearby areas but absence of functional 1527
coupling to distributed regions across the cortex (barring regions across the insular 1528
cortex). Even when regional analyses were explored that did not constrain the topography 1529
to form separate networks, no evidence was found for functional coupling to distributed 1530
cortical regions. Figure 19 shows that the motor component of the somatomotor network 1531
included MI (area 4) and caudal premotor area 6, while the somatosensory component 1532
included SI and most, if not all of early somatosensory area 5L. The somatomotor 1533
network also included a small portion of the mid-cingulate sulcus and possibly part of 1534
area 5M. Within the somatomotor cortex of non-human primates, areas 1 to 6 are 1535
densely, but not fully, connected (see meta-analysis in Felleman and Van Essen 1991) by 1536
association fibers that enter the white matter and re-enter the cortex (Jones et al. 1978). 1537
For example, these association fibers connect area 2 with areas 1, 3, 4, 5 and 6 (Jones et 1538
al. 1978; Pons and Kaas 1986), as well as area 4 with areas 2 and 6 (Matelli et al. 1986; 1539
Matelli et al. 1998; Muakkassa and Strick 1979), but not areas 1 and 3 (Jones et al. 1978; 1540
Pons and Kaas 1986). Jones et al. (1978) also reported non-specific local connections 1541
between adjacent primary somatosensory and motor areas (e.g., areas 3 and 4). 1542
The non-human primate premotor area 6 has been subdivided into rostral and 1543
caudal areas that may provide further insight into the functional connectivity patterns 1544
(Barbas and Pandya 1987; Matelli et al. 1985; 1991; Zilles et al. 1995). The caudal 1545
premotor areas are densely and topographically connected to M1, but not to the prefrontal 1546
cortex. In contrast, the rostral premotor areas are connected to prefrontal cortex but not 1547
M1 (Barbas and Pandya 1987; Luppino et al. 1993; Matelli et al. 1986). The caudal 1548
premotor areas are also more densely connected to each other than the adjacent rostral 1549
premotor areas. While there are competing hypotheses over the exact homology between 1550
primate and human premotor areas (Geyer 2004; Petrides 2005; Rizzolatti et al. 1998), 1551
converging studies have suggested a rostrocaudal subdivision of human premotor cortex 1552
(see chapter 4.4 of Geyer 2004 for a review), which might reflect the absorption of caudal 1553

4
Auditory cortex was also functionally coupled to the ventral somatomotor network.
However, as illustrated in Figure 14, this may be an artifact of signal bleeding across the
insula in volumetric space. For this reason, while the finding may reflect a meaningful
functional interaction, we do not discuss it further in the present paper.

53
human area 6 into the somatomotor network as revealed by functional coupling. Despite 1554
evidence that the early somatosensory and late motor fields are densely integrated, we 1555
must be cautious of the possibility of fMRI signal blurring across the central sulcus, 1556
which may cause an overestimation of functional coupling between the parallel strips of 1557
SI, MI and premotor area 6. 1558
Of most interest, the analysis also revealed a dorsoventral division of the 1559
somatomotor strip, which was confirmed by targeted regional analyses (Fig. 20). In non- 1560
human primates, the connections between the different somatosensory-motor areas are 1561
generally topographic, so that for example, the hand region of area 3 projects to the hand 1562
region of areas 1 and 2 (Jones et al. 1978; Pons and Kaas 1986). In addition, the 1563
somatomotor association fibers have been reported to terminate in mediolaterally oriented 1564
strips (see for example Fig. 4 in Jones et al. 1978; Jones and Wise 1977), although others 1565
have raised questions concerning the strip-like nature of callosal connections (Gould et 1566
al. 1986; Killackey et al. 1983). Our observations of fcMRI correlations within the human 1567
somatomotor cortex, whose anteroposterior extents were shorter than their mediolateral 1568
extents, were therefore consistent with the observations of ipsilateral strip-like anatomical 1569
connections. One possibility is that the dorsoventral boundary revealed by our clustering 1570
analysis might correspond to the boundary of the face and body representations. 1571
1572
Evidence for a prototype distributed cortical pathway 1573
The results discussed above reveal interesting organizational properties of local 1574
networks of areas within sensorimotor and visual cortices. In considering the organization 1575
of the cerebral cortex, an immediate question arises as to how information in the sensory 1576
systems might propagate to influence motor representations? Functional connectivity is 1577
limited in its ability to provide insight into this question, but some aspects of the results 1578
are informative and consistent with anatomical and physiological studies of sensory- 1579
motor pathways. 1580
The canonical sensory-motor pathway that has been studied in the monkey is the 1581
pathway that includes retinotopic visual cortex, the MT+ complex, parietal area LIP, and 1582
the FEF (e.g., Andersen et al. 1990; Colby and Goldberg 1999; Gold and Shadlen 2007; 1583
Shadlen and Newsome 2001). The basic idea is that incoming visual information 1584

54
propagates from early visual areas to MT+, which provides constraints on decision 1585
processes that arise from interactions with LIP and FEF. FEF, in turn, interacts with 1586
motor regions to generate motor signals. Hierarchical anatomic connections support such 1587
a pathway (Felleman and Van Essen 2001; Maunsell and Van Essen 1983). The question 1588
here is whether the human functional connectivity results also reveal evidence for such a 1589
pathway and, if so, what properties emerge? 1590
Of particular interest are the quantitative results presented in Figures 22 to 27 that 1591
reveal interactions between distinct networks. While primary visual cortex is largely 1592
absent functional coupling to association or premotor cortices, it is topographically 1593
coupled to MT+ (Fig. 22), which is in turn strongly coupled to aMT+ as well as modestly 1594
coupled to parietal regions including SPL7A (Figs. 24 and 25). aMT+ is functionally 1595
coupled to PrC
v
and FEF (Fig. 25) completing the pathway. While the organization of the 1596
human MT+ complex is still unresolved, human MT+ is thought to include the human 1597
homologues of macaque MT, MST and FST (Amano et al. 2009; Huk et al. 2002; Kolster 1598
et al. 2010). Based on the relative location of aMT+ with respect to MT+, aMT+ might 1599
correspond to macaque MST/FST (Maunsell and Van Essen 1983) or TEO/PIT 1600
(Felleman and Van Essen 1991; von Bonin and Bailey 1947). Our observation that aMT+ 1601
is less correlated with V1 than MT+ (Figs. 22 and 23) is therefore consistent with 1602
multiple studies showing that macaque TEO/PIT or MST/FST is less densely connected 1603
with V1 compared with MT (Distler et al. 1993; Felleman and Van Essen 1991; Markov 1604
et al. 2010). Although there are uncertainties to using functional connectivity to infer 1605
hierarchical arrangements among areas, if one uses the simple assumption that the more 1606
strongly two areas are functionally coupled together, the closer they are to one another in 1607
a processing hierarchy, a sensory-motor processing hierarchy emerges that is consistent 1608
with the extant literature (Figs. 28 and 29). 1609
Analysis of interactions among networks illustrates a number of further points. 1610
Most critically, the hierarchy across networks reveals how a distributed network might 1611
serve as a bridge between sensory and motor networks that themselves possess 1612
preferentially local interactions (resulting in their parcellation into their own networks in 1613
Figs. 11 and 13). One speculation is that the distributed cortical network illustrated as the 1614
green network in Figure 11 -- and known within the neuroimaging literature as the dorsal 1615

55
attention system -- is the prototype distributed cortical network. It possesses multiple 1616
properties that are common to all association networks. Specifically, its component 1617
regions are distributed throughout temporal, parietal, and frontal cortices and show strong 1618
functional couplings between all pairs of regions (Fig. 32). We speculate that this 1619
network is a prototype because it is more strongly functionally coupled to extrastriate 1620
sensory regions and premotor regions than the remaining association networks that will 1621
be discussed later and also because it is well represented in the macaque. The presence of 1622
similar networks in humans and macaques suggests that they are homologous, and thus 1623
were present in the last common ancestor, which lived about 25-30 million years ago 1624
(Pilbeam and Young 2004). As will be discussed in the next section, the remaining 1625
distributed association networks, which in the human represent the majority of 1626
association cortex, display the same general organization but appear to have lost direct 1627
functional coupling to sensory and motor regions, at least insofar as measured by intrinsic 1628
functional connectivity. 1629
Another important feature of this hierarchy is apparent when one considers finer 1630
details of the functional coupling patterns. Within the broad hierarchy, there is evidence 1631
for specialization indicative of parallel hierarchical arrangements. This is perhaps best 1632
illustrated by the differential functional coupling of premotor regions FEF and PrC
v
with 1633
the multiple regions localized around the superior parietal lobule (Figs. 26 and 27). 1634
Anatomical tracing work in the macaque has suggested two segregated sensory-motor 1635
pathways from parietal cortex to dorsal and ventral aspects of frontal cortex, including 1636
premotor cortex (Kurata 1991; Rizzolatti et al. 1998; Tann-Garipy et al. 2002) and the 1637
frontal eye fields (Petrides and Pandya 2006; Stanton et al. 1995). These frontoparietal 1638
connections have a dorsomedial to ventrolateral axis: dorsal portions of caudal frontal 1639
cortex are preferentially connected to medial and dorsal parts of parietal cortex including 1640
the superior parietal lobule and medial parietal cortex, while caudal ventral frontal cortex 1641
preferentially communicates with lateral and ventral aspects of parietal cortex and largely 1642
lacks connections with dorsal and medial parietal areas. Our demonstration that PrC
v
was 1643
strongly correlated with more ventral portions of rostral SPL and IPS, while FEF was 1644
strongly correlated with caudal SPL and IPS (Figs. 26 and 27) is consistent with 1645
descriptions of frontoparietal sensory-motor circuits in the macaque. Thus, evidence for 1646

56
specialization of subpathways is present in this canonical sensory-motor pathway. Taking 1647
the speculation that this canonical pathway represents the prototype distributed 1648
association network, it is tempting to wonder whether the interdigitated association 1649
networks that comprise the remaining human association cortex are evolutionary 1650
expansions of this basic prototype with multiple, interdigitated pathways that have 1651
become nearly completely differentiated. 1652
1653
Association cortices are nexuses of regions with distinct connectivity fingerprints. 1654
Association cortex, in particular parietal cortex near the inferior parietal lobule, 1655
has been challenging to characterize. Human association cortex shows disproportionate 1656
expansion in relation both to macaque and great apes (Preuss 2004; Van Essen and 1657
Dierker 2007; Hill et al. 2010). In a few instances, it is an open question as to whether 1658
homologies should be expected (Orban et al. 2006). For example, in his seminal work in 1659
1909, Korbinian Brodmann noted that the human inferior parietal lobule included two 1660
cytoarchitectonic areas that are absent in the monkey (areas 39 and 40). The possibility 1661
that these association areas are vastly expanded in hominid evolution, or are even novel 1662
areas altogether, figured prominently in the classic description of disconnection 1663
syndromes by Geschwind (1965). For these reasons anatomic connectivity in the monkey 1664
cannot uniformly be presumed to apply to the human. Adding further complication, 1665
nearby regions of association cortex are often active across quite distinct forms of tasks 1666
suggesting functional diversity (e.g., Culham and Kanwisher 2001). As extreme examples 1667
of functional diversity, parietal regions near the superior parietal lobule, including those 1668
discussed in the preceding section, respond during sensory-motor decision tasks 1669
(Corbetta and Shulman 2002). Regions near the temporoparietal junction respond during 1670
social tasks that require participants to infer what others are thinking (Saxe 2006; Van 1671
Overwalle and Baetens 2009), and regions within the caudal portion of the inferior 1672
parietal lobule respond during episodic remembering (Cabeza et al. 2008; Vilberg and 1673
Rugg 2008; Wagner et al. 2005). 1674
Our results demonstrate that parietal association cortex includes multiple nearby 1675
regions that possess markedly different connectivity profiles that parallel similar 1676
distinctions in prefrontal cortex. The results in parietal cortex are anticipated by both 1677

57
anatomic studies in the monkey and prior studies using human functional connectivity. 1678
Specifically, the caudal portion of macaque 7a, labeled Opt by Pandya and colleagues 1679
(Pandya and Seltzer 1982), has connections to the parahippocampal cortex, retrosplenial 1680
cortex, and posterior cingulate (Seltzer and Pandya 1994; Andersen et al. 1990; Suzuki et 1681
al. 1994; Lavenex et al. 2002). Nearby areas, such as LIP, are preferentially connected to 1682
visual association cortex and premotor areas leading Andersen et al. (1990) to note that 1683
area 7a appears to be very different from other visual areas in the inferior parietal lobule 1684
in that it is the only area that connects to some of the highest centers of the brain. 1685
Examination of functional connectivity of parietal association cortex in the human has 1686
also revealed notable diversity. Vincent et al. (2006) illustrated that neighboring parietal 1687
regions are functionally coupled to distinct sensory-motor and limbic circuits. In later 1688
studies, parietal association cortex was found to possess between three (Vincent et al. 1689
2008) and four (Nelson et al. 2010) distinct zones distinguished by their functional 1690
connectivity profiles (see also Sestieri et al. 2011). 1691
Differential functional coupling across parietal and prefrontal regions are 1692
displayed in Figures 30 and 31. Drawing from Passingham and colleagues (2002), we 1693
refer to these regional connectivity profiles as fingerprints because they illustrate the 1694
connectivity patterns across regions that make them distinct. In examining the many 1695
fingerprints, several principles emerge that provide insight into cortical organization. 1696
First, nearby regions can show abrupt transitions in their connectivity fingerprints. The 1697
transitions from SPL7A to IPS3
l
and from IPS3
l
to PGp
d
are such examples. SPL7A, 1698
which may be at or near the human homologue to macaque LIP, is functionally coupled 1699
to extrastriate (MT+ and aMT+) and premotor (FEF and PrC
v
) regions. IPS3
l
shows a 1700
fundamentally distinct connectivity fingerprint with coupling to prefrontal regions that 1701
are within dorsolateral prefrontal cortex (PFC
l
and PFC
lp
) and amodal posterior temporal 1702
association regions (ITG). IPS3
l
is absent coupling to sensory or motor regions. PGp
d
is 1703
distinguished by prominent functional coupling to limbic regions including posterior 1704
cingulate (PCC), retrosplenial cortex (RSP), and the medial temporal lobe (PHC). This 1705
tripartite division separates the major parietal zones that form the dorsal attention, 1706
frontoparietal control, and default networks discussed extensively in the human 1707
neuroimaging literature (green, orange, and red networks in Fig. 11). 1708

58
Second, within these broad divisions there are further distinctions that demarcate 1709
more subtle regional differences. Across these regions, most connectivity properties are 1710
shared but there are also key differences, forming connectional families of regions (see 1711
Passingham et al. 2002 for discussion). These differences within connectional families 1712
lead to the finer parcellation observed in Figure 13 and are likely of functional 1713
importance. For example, the TPJ and PGp
v
possess similar connectivity fingerprints, and 1714
both fall within the broader network that is globally called the default network. However, 1715
the TPJ is preferentially coupled to medial prefrontal regions (PFC
dm
), the posterior 1716
cingulate cortex (PCC) and precuneus (pCun), and the superior temporal sulcus (STS and 1717
STS
p
). While PGp
v
possesses a broadly similar fingerprint, it is also prominently coupled 1718
to regions associated with the medial temporal lobe memory system (RSP and PHC). 1719
This is of particular interest because both of these parietal association regions have been 1720
proposed to be involved with higher mental functions linked to internal mentation and 1721
social cognition (see Buckner et al. 2008 for review), but functional distinctions have also 1722
been noted (e.g., Rosenbaum et al. 2007). 1723
The final organization principle that emerges is that association regions belonging 1724
to the same connectional families can always be found widely distributed across the 1725
cortical mantle. Contrasting the fingerprints of regions within prefrontal cortex (Fig. 31) 1726
with those falling within parietal cortex (Fig. 30) illustrates this last principle. PFC
dm
and 1727
TPJ are prime examples. These regions possess nearly the same functional connectivity 1728
fingerprints that differ from nearby regions within their own lobes. This accounts for why 1729
these distributed regions form such tightly coupled functional connectivity networks and 1730
suggests that association cortex might be best conceptualized as a series of interdigitated, 1731
distributed networks. 1732
1733
Association cortex is comprised of multiple, interdigitated large-scale circuits 1734
The majority of the human cerebral cortex is made up of multiple large-scale 1735
networks that include functionally coupled regions distributed across the brain. Such 1736
organization is apparent in prior studies (e.g., Beckmann et al. 2005; Damoiseaux et al. 1737
2006; De Luca et al. 2006; Fox et al. 2006; Greicius et al. 2003; 2004; Vincent et al. 1738
2008) and is evident in all of the analyses presented here. By analyzing the complete 1739

59
topography of the cortical surface, we were further able to illustrate that the multiple, 1740
distributed networks are interdigitated with one another forming complex convergence 1741
zones in parietal and prefrontal association cortices (Figs. 32 to 35). What do these 1742
patterns suggest about the organization of the cerebral cortex? 1743
At the broadest level, these observations emphasize the need to adopt network 1744
approaches when exploring cerebral function. By network approach, we refer to the idea 1745
that the relevant functional unit may be the interconnected network itself, as suggested by 1746
Mesulam (1981; 1986) and Goldman-Rakic (1988). Mesulam (1981) proposed a network 1747
approach as an alternative to centrist approaches to localization, in which complex 1748
functions relied on specific cortical areas exclusively devoted to that function. While 1749
recognizing that functional specializations exist among areas of the same network, the 1750
network framework emphasizes that functions arise as emergent properties of these 1751
reciprocally connected systems of brain areas (see Fig. 4 of Mesulam 1981; 1990). In 1752
other words, distributed systems of areas, spanning different cortical lobes and 1753
subcortical structures, form functional units via their dense interconnections, giving rise 1754
to complex behavior. This is a quite different organization than is evident in sensory 1755
cortex, which is characterized by dense connectivity among local areas. The two 1756
frameworks are not entirely different because all areas are presumably embedded within 1757
systems of interacting brain areas; however, the focus in many theories is on processing 1758
specialization within areas and processing hierarchies among neighboring areas. The 1759
present analyses suggest that functional unit of interest may be the distributed network 1760
itself. 1761
Goldman-Rakic (1988) offered three specific anatomic observations that provide 1762
evidence for distributed cerebral networks. First, prefrontal and parietal areas that are 1763
directly connected to one another also tend to have convergent projections to additional 1764
temporal and limbic areas (see Fig. 3 of Goldman-Rakic 1988). Second, interconnected 1765
association areas are tied together by common thalamic connections. And third, arguing 1766
against a typical hierarchical model of cortical organization, interconnected association 1767
areas tend not to have laminar projection patterns with clear feedforward and feedback 1768
relations (also see Felleman and Van Essen 1991). We suspect that the distributed 1769
networks that comprise the majority of human association cortex are networks of this 1770

60
type highly interconnected, without strong hierarchical relations among areas, and 1771
integrated into a common functional unit of some form. This does not mean that areas 1772
within prefrontal and parietal cortices are making the same contributions to the network, 1773
but the emphasis does shift to asking how the separate networks make distinct functional 1774
contributions rather than asking how different areas within prefrontal or parietal cortices 1775
may be locally differentiated. That is, association cortex is characterized by multiple 1776
modules (Bullmore and Sporns 2009) that are each comprised of nodes distributed widely 1777
across the cortex. 1778
Of further interest, the distributed cerebral networks converge on regions of 1779
association cortex that are late to develop in terms of myelination (see Fig. 3 of Catani 1780
and ffytche 2005) and cortical surface area (Hill et al. 2010), and are expanded in the 1781
human brain relative to the modern macaque brain (Van Essen and Dieker 2007). For 1782
these reasons, it is likely that distributed association networks have been under strong 1783
selective pressure to expand in recent hominid evolution. The network parcellations 1784
presented in Figures 11 and 13 provide a current best estimate of the organization of the 1785
interdigitated networks that comprise human association cortex. 1786
1787
Caveats and limitations 1788
Measuring functional connectivity is not the same as directly measuring 1789
anatomical connectivity. Functional connectivity is constrained by anatomic connectivity 1790
(e.g., Johnston et al. 2008; Honey et al. 2009) but those constraints are not restricted to 1791
monosynaptic connectivity. For example, functional coupling is present between the two 1792
hemispheres for striate cortex in the macaque (Vincent et al. 2007) while 1793
interhemispheric connections are absent except at the border of V1 (Van Essen et al. 1794
1982). The pervasiveness of functional coupling is both a weakness and strength of the 1795
technique as it allows one to map large-scale polysynaptic circuits, but leaves ambiguities 1796
and uncertainties, which will require follow-up by other methods such as diffusion 1797
imaging techniques and detailed examination of homologies to non-human primates. 1798
Cerebrocerebellar circuits are a place that perhaps best illustrates that functional 1799
connectivity is constrained by anatomy but is more pervasive than monosynaptic 1800
connectivity. Cerebrocerebellar circuits which are exclusively polysynaptic (Evarts and 1801

61
Thach 1968; Kemp and Powell 1971; Strick 1985) demonstrate functional coupling that 1802
tracks the contralateral organization of anatomic projections and is topographically 1803
specific as described in our companion paper (Buckner et al. submitted). Thus, functional 1804
connectivity is informative but should not be considered a direct measure of anatomic 1805
connectivity. Functional connectivity also appears sensitive to other factors, including 1806
recent experience and the state of the subject during scanning (Fox and Raichle 2007; 1807
Moeller et al. 2009; Buckner 2010). An assumption made in the present paper is that the 1808
dominant contribution to the measured correlations reflect stable properties of cortical 1809
architecture an assumption we believe is warranted but nonetheless needs to be made 1810
explicit as boundary conditions and violations of this assumption may emerge. 1811
A further limitation of the present work is resolution. Even at the relatively finer 1812
resolution of the 17-network estimate, discrete components of a given network likely 1813
span multiple cytoarchitecturally distinct cortical areas. The association networks 1814
described in this paper are at a coarser resolution than the networks inferred by Goldman- 1815
Rakic (1988) and others in the macaque. Our limited data resolution, as a result of voxel 1816
size, smoothing and intersubject averaging including potential errors in surface-based 1817
alignment, may miss important features of the cortical topography and the results should 1818
be interpreted accordingly. Future, high resolution studies of individuals may provide 1819
better estimates of cortical topography. 1820
A final limitation is our use of clustering to parcellate the cerebral cortex. The 1821
assumption made by the clustering approach is that each vertex belongs to a single 1822
network. Our seed-based analyses (Figs. 32-35) suggest that this is a reasonable 1823
beginning point for analysis even though crosstalk exists between networks. However, 1824
there are specific places where the parcellation results might be particularly sensitive to 1825
inaccuracies, including the characterization of cortical regions that serve as putative hubs 1826
of communication between networks (Buckner et al. 2009; Hagmann et al. 2008; 1827
Mesulam 1998). We also note that the 17-network estimate does not cleanly fractionate 1828
individual networks within the 7-network estimate, implying that cortical networks are 1829
not spatially organized in a strictly hierarchical fashion
5
, like that suggested in the toy 1830

5
Here hierarchy refers to the spatial organization of the networks rather than the
concept of hierarchical processing pathways.

62
example (Fig. 5). Our efforts to enforce strict subdivisions of coarser networks by using 1831
hierarchical clustering (not shown) failed to provide stable results across the Discovery 1832
and Replication datasets. Recent advances in graph theoretic clustering approaches are 1833
promising in providing the possibility for regions to belong to multiple networks or 1834
communities (e,g., Ahn et al. 2010). 1835
1836
Conclusions 1837
Different regions of the cerebral cortex display distinct characteristics. Functional 1838
connectivity of retinotopic visual areas display dense local functional coupling that is 1839
organized across areas in a fashion that respects functional topography. Association 1840
cortex is comprised of multiple, interdigitated large-scale networks that, while exhibiting 1841
crosstalk, possess predominantly parallel organization. The map of these cerebral 1842
networks is provided as a reference for future functional characterization and 1843
confirmation by complementary approaches that can directly visualize anatomic 1844
connectivity. 1845
1846
1847

63
Acknowledgements 1848
1849
We thank Elizabeth Hemphill, Leah Bakst, and Sara Rubenstein for assistance 1850
with data collection, Karl Zilles and Katrin Amunts for the human histological data, 1851
Donna Dierker and David Van Essen for help with the Caret software, Doug Greve for 1852
help with the FreeSurfer/FsFast software, Simon Eickhoff for help with the SPM 1853
Anatomy toolbox, Avram Holmes, Todd Pruess, Roger Tootell, Justin Vincent and Justin 1854
Baker for useful discussion, the Harvard Center for Brain Science Neuroimaging Core 1855
and the Athinoula A. Martinos Center for imaging support, and the Harvard 1856
Neuroinformatics Research Group (Gabriele Fariello, Timothy OKeefe, and Victor 1857
Petrov). FMK was supported by fellowships from the Department of Defense, an Ashford 1858
Graduate Fellowship in the Sciences and the Sackler Scholars Program in Psychobiology. 1859
This work was supported by NIH Grants (AG021910, P41RR14074, K08MH067966, 1860
P41RR14075, U24RR021382, AG022381, RC1AT005728-01, R01NS052585-01, 1861
1R21NS072652-01, 1S10RR023401, 1S10RR019, 1S10RR023043, 1R01NS070963), the 1862
Ellison Medical Foundation, the Howard Hughes Medical Institute and the Simons 1863
Foundation. 1864
1865
1866

64
Table 1: Locations of Visual Cortex Seed Regions 1867
1868
Seed Region Coordinates Confidence
V3
pv
(Rottschy et al. 2007;
Wilms et al. 2010)
-12, -67, -3 0.74
ExP -3, -74, 23 0.81
V1
p
(Amunts et al. 2000;
Fischl et al. 2008)
-16, -74, 7 0.78
V3
cv
(Rottschy et al. 2007;
Wilms et al. 2010)
-23, -91, -15 0.67
ExC -32, -89, -1 0.75
V1
c
(Amunts et al. 2000;
Fischl et al. 2008)
-13, -100, -8 0.48
1869
Notes: V1
c
and V1
p
seed regions are selected from the central and peripheral regions of V1. 1870
V3
cv
and V3
pv
seed regions are selected from the central and peripheral regions at or near 1871
V3v. ExC and ExP are selected from the purple-colored central and bright red-colored 1872
peripheral visual systems respectively. The confidence of the seeds in their network 1873
assignment is computed from the Replication dataset. Coordinates reflect the approximate 1874
center location based on the atlas space of the Montreal Neurological Institute (MNI). 1875
1876

65
Table 2: Locations of Seed Regions Utilized in the Sensory-Motor Pathway Analysis 1877
1878
Name Coordinates
MT+
d

-44, -72, 8
MT+ -45, -72, 3
MT+
v

-45, -79, -1
aMT+ -51, -64, -2
V1
pd

-18, -70, 8
V1
pv

-8, -63, 6
V1
cd

-8, -95, 3
V1
cv

-8, -92, -5
V3A -17, -92, 20
V4 -22, -65, -9
FEF -26, -6, 48
PrC
v

-50, 6, 30
IPS2 -40, -37, 42
IPS3
m

-31, -48, 46
SPL7A -28, -61, 60
SPL7P -14, -68,64
IPS1 -46, -49, 51
1879
Notes: Four V1 seed regions were selected using the calcarine fissure, histological V1 1880
estimates (Amunts et al. 2000; Fischl et al. 2008) and the 17-network boundary estimate in 1881
the Discovery dataset. V1
pd
and V1
pv
were from the dorsal and ventral parts of peripheral 1882
V1, likely representing the ventral and dorsal peripheral visual fields respectively. V1
cd
1883
and V1
cv
were from the dorsal and ventral parts of central V1, likely representing the 1884
ventral and dorsal central visual fields respectively. MT+
d
, MT+ and MT+
v
seed regions 1885
were selected from the dorsal, center and ventral parts of the histological map of the MT+ 1886
complex, so that MT+
d
and MT+
v
likely represented peripheral and central visual fields 1887
respectively. Based on the over-representation of the lower visual field within macaque 1888
MT (Maunsell and Van Essen 1987), it is possible that the three MT+ seed regions might 1889
also represent the lower visual field. aMT+ was selected to be anterior and outside the 1890
MT+ histological map. V3A and V4 were selected based on their high correlation with 1891
MT+ in the Discovery sample and named using the approximate map of human visual 1892
areas (Van Essen 2004) as reference. FEF and PrC
v
were selected from the caudal frontal 1893

66
cortex, while IPS2, IPS3
m
, SPL7A, SPL7P and IPS1 were at or near IPS. FEF, IPS2, 1894
SPL7A and SPL7P were derived from the meta-analysis of fMRI studies (see Table 4). 1895
The PrC
v
region was selected based on its correlation with aMT+. IPS3
m
was chosen 1896
spatially between IPS2 and SPL7A. IPS1 was chosen on the lateral wall of rostral IPS. 1897
IPS2, IPS3
m
, SPL7A, SPL7P and IPS1 were named using the probabilistic histological 1898
maps of the parietal cortex as reference (Choi et al. 2006; Scheperjans et al. 2008a; 1899
2008b). Coordinates reflect the approximate center location based on the MNI atlas space. 1900
1901

67
Table 3: Summary of fMRI Meta-analysis to Obtain Coordinates for IPS2, SPL7A, SPL7P 1902
and FEF 1903
1904
Name Coordinates Putative Macaque Homology Literature
IPS2 -40, -37, 42 AIP (anterior intraparietal) Binkofski et al. 1999; Binkofski
et al. 1998; Culham et al. 2003;
Grefkes et al. 2002; Jncke et al.
2001; Shikata et al. 2001;
Shikata et al. 2003
SPL7A -28, -61, 60 LIP (lateral intraparietal) Hagler et al. 2007; Heide et al.
2001; Koyama et al. 2004; Luna
et al. 1998; Medendorp et al.
2003; Sereno et al. 2001 (as
reported in Table 1 of Hagler et
al. 2007); Shulman et al. 2003
SPL7P -14, -68, 64 PIP (posterior intraparietal) Faillenot et al. 2001; Shikata et
al. 2001; Shikata et al. 2003;
Taira et al. 2001
FEF -26, -6, 48 FEF (frontal eye fields) Connolly et al. 2000; Connolly
et al. 2002; Corbetta et al. 1998;
Heide et al. 2001; Koyama et al.
2004; Luna et al. 1998; Perry
and Zeki 2000
1905
Notes: For papers that only report right hemisphere coordinates, left hemisphere 1906
coordinates were obtained by reflection across the midline. Coordinates reflect the 1907
approximate center location based on the MNI atlas. For IPS2, tasks generally involved 1908
perception of 3D objects. Studies reporting responses at or near SPL7A used saccadic eye 1909
movement tasks. Studies reporting responses at or near SPL7P involved tasks requiring 1910
surface/orientation discrimination. FEF coordinates come from studies of saccadic eye 1911
movements. 1912
1913
1914
1915

68
Table 4: Locations of Seed Regions Used for Analysis of Connectivity Fingerprints (Figs. 1916
30 and 31) 1917
1918
Parietal Cortex Coordinates
PGa -52, -50, 49
IPS2 -40, -37, 42
SPL7A -28, -61, 60
IPS3
l

-35, -56, 42
PGp
d

-49, -63, 45
PGp
v

-49, -69, 28
TPJ -51, -57, 27
PF -60, -37, 38
5Ci -16 -32 39
PCC -3, -49, 25
pCun -10, -57, 35
Frontal Cortex
PFC
la

-41, 55, 4
PFC
l

-38,33,16
PFC
da

-31,39,30
PFC
lp

-45, 29, 32
PFC
dp

-44,15,48
PrC
v

-50, 6, 30
FEF -26, -6, 48
PFC
y

-55,24,13
6vr+ -55, 6, 11
PFC
m

-7, 46, -2
PFC
dm

-4, 49, 32
PFC
mp

-5, 22, 47
Cing
a

-10, 13, 40
Cing
m

-5, 2, 29
Occipital/Temporal Cortex
STS
a

-49, 5, -26
STS -55, -10 -16
STS
p

-49, -34, -4
ITG -59, -53, -14
aMT+ -51,-64,-2
MT+ -45, -72, 3
Limbic Cortex
PHC -25, -31, -20
RSP -7, -50, 7

69
Notes: Where possible, seed regions were carried forward from previous analyses (i.e., 1919
IPS2, SPL7A, FEF, PrC
v
, aMT+ and MT+; see Tables 2 & 3). Seed regions PF and PGa 1920
were at or near human inferior parietal area PF and PGa, respectively (Caspers et al. 1921
2006; 2008). The IPS3
l
seed region was selected to be on the lateral lip of the IPS at or 1922
near human hIPS3, while the 5Ci seed region corresponded to the human posterior 1923
cingulate sulcal region 5Ci (Scheperjans et al. 2008a; 2008b). PG
pd
and PG
pv
1924
corresponded to the dorsal and ventral aspects of area PGp in the inferior parietal lobule 1925
(Caspers et al. 2006; 2008). The TPJ seed region was selected based on the peak 1926
coordinates reported in a human fMRI study of mental state inference (Saxe and Powell 1927
2006) which accords well with the predicted mean peak in a recent meta-analysis on the 1928
same topic (Van Overwalle and Baetens 2009). Seed region 6vr+ was selected to be 1929
anterior to the map of premotor area 6 (Geyer 2003) and labeled to reflect a new 1930
cytoarchitectonic delineation within the ventral precentral region (Amunts et al. 2010). 1931
Seed region PFC
v
fell within the histological map of BA45 (Amunts et al. 1999; Fischl et 1932
al. 2008). The remaining seed regions were selected to provide comprehensive coverage 1933
of the lateral and medial aspects of the cortical surface and such that each association 1934
network was represented by multiple targets. Labels for the remaining seed regions were 1935
determined according to gross anatomical landmarks and given subscripts to denote 1936
relative positions: Cing = cingulate sulcus; pCun = precuneus; ITG = inferior temporal 1937
gyrus; PCC = posterior cingulate cortex; PFC = prefrontal cortex; PHC = 1938
parahippocampal cortex; RSP = retrosplenial cortex; STS = superior temporal sulcus. 1939
Subscripts a = anterior; p = posterior, l = lateral, m = medial, d = dorsal and v = ventral. 1940
Coordinates reflect the approximate center location based on the MNI atlas. 1941
1942

70
Table 5: Locations of Seed Regions Used for Analysis of Parallel Networks in Association 1943
Cortex 1944
1945
Name Coordinates Confidence
Dorsal Attention A -22, -8, 54 0.54
Dorsal Attention B -34, -38, 44 0.53
Dorsal Attention C -18, -69, 51 0.46
Dorsal Attention D -51, -64, -2 0.55
Dorsal Attention E -8, -63, 57 0.32
Dorsal Attention F -49, 3, 34 0.49

Ventral Attention A -31, 39, 30 0.49
Ventral Attention B -54, -36, 27 0.63
Ventral Attention C -60, -59, 11 0.27
Ventral Attention D -5, 15, 32 0.65
Ventral Attention E -8, -24, 39 0.57
Ventral Attention F -31, 11, 8 0.67

Control A -40, 50, 7 0.52
Control B -43, -50, 46 0.51
Control C -57, -54, -9 0.25
Control D -5, 22, 47 0.43
Control E -6, 4, 29 0.27
Control F -4, -76, 45 0.25

Default A -27, 23, 48 0.46
Default B -41, -60, 29 0.63
Default C -64, -20, -9 0.61
Default D -7, 49, 18 0.60
Default E -25, -32, -18 0.22
Default F -7, -52, 26 0.61
1946
Notes: The confidence of the seed regions in their network assignment is computed from 1947
the Replication dataset. Coordinates reflect the approximate center location based on the 1948
MNI atlas. 1949
1950

71
Figure Legends 1951
1952
Figure 1 Surface coordinate system for fMRI analysis. For each subject, the T2* images 1953
yielding BOLD-contrast fMRI data (A) were registered to the T1-weighted structural data 1954
(B). The cortical gray-white and pial surfaces were estimated from the structural data. 1955
The red lines show the estimated gray-white surface (A, B). Pial surface is shown in C. 1956
The gray-white surface was inflated into a sphere (D). The inflated spheres were then 1957
aligned across subjects using surface-based registration of the cortical folding pattern, 1958
resulting in a common spherical coordinate system (E). BOLD data of individual subjects 1959
(A) can then be projected onto the spherical coordinate system (E) in a single 1960
transformation step to reduce artifacts due to multiple interpolations. 1961
1962
Figure 2 Examples of intrasubject surface extraction and registration of structural- 1963
functional images. Examples of extracted cortical gray-white surfaces (red lines) are 1964
overlaid on T2* and T1 images of three random subjects in their native T1 space. 1965
Imperfections are apparent in BOLD data especially in regions of susceptibility artifact 1966
(e.g., orbital frontal cortex). 1967
1968
Figure 3 Signal-to-noise ratio (SNR) maps of the functional data from the full sample (N 1969
= 1000). The mean estimate of the BOLD fMRI data SNR is illustrated for multiple 1970
views of the left hemisphere in Caret PALS space. A = anterior, P = posterior, D = dorsal 1971
and V = ventral. 1972
1973
Figure 4 Cortical regions utilized in constructing functional connectivity profiles. A total 1974
of 1175 regions are sampled uniformly on the surface-based representations of the left 1975
and right hemispheres within the FreeSurfer surface coordinate system and shown here in 1976
Caret PALS space, where each red patch represents the location of a single regional 1977
vertex. Each vertex in the surface coordinate system is characterized by its profile of 1978
functional connectivity to the 1175 regions. The visually non-uniform distribution of the 1979
regions in Caret PALS space is due to the non-linear deformation from FreeSurfer space 1980

72
to Caret PALS space. This image thus also serves to illustrate the subtle differences 1981
between the two surface space coordinate systems. 1982
1983
Figure 5 Toy example illustrating clustering. (A) Hypothetical points are scattered in a 1984
structured fashion on a two-dimensional canvas. Clustering aims to recover the 1985
underlying structure. (B) Example solutions for M = 2, 3, 4 or 5 clusters are shown. The 1986
solutions for M = 2 or 5 clusters agree with visual assessment of the underlying structure 1987
and are therefore useful representations. On the other hand, seeking 3 or 4 clusters does 1988
not lead to satisfying solutions because solutions are ambiguous. For example, the M = 3 1989
solution is not unique in the sense that an equally good alternate solution is for one 1990
group of points in the red cluster to be grouped with the orange cluster. Seeking M = 3 or 1991
4 clusters is therefore unstable in the sense that different random initializations of the 1992
clustering algorithm lead to different equally good solutions. In the present study we 1993
employed a stability analysis to estimate the numbers of clusters and also examined both 1994
a relatively coarse solution (7-network) and fine-resolution solution (17-network) so as to 1995
survey the solution space broadly (see Fig. 6). 1996
1997
Figure 6 7 and 17 networks can be stably estimated. Instability of the clustering 1998
algorithm is plotted as a function of the number of estimated networks for the vertex- 1999
resampling variant of the stability analysis applied to 1000 subjects. The clustering 2000
algorithm is less stable with increasing number of estimated networks, which is an 2001
expected property since the number of estimated networks enlarges the solution space 2002
(and thus complexity) of the clustering problem. The local minima of the graphs (marked 2003
with asterisks) indicate the number of networks that can be stably estimated by the 2004
clustering algorithm. The stability analysis suggests that 7, 10, 12 or 17 networks can be 2005
stably estimated. Resampling the ROIs yields almost identical results and is not shown. 2006
Here we focus on the 7- and 17-network estimates to provide a broad survey of the 2007
solution space. 2008
2009

73
Figure 7 Discovery and Replication of a 7-network cortical parcellation. The 7-network 2010
estimates are highly consistent across the Discovery (n = 500) and Replication (n = 500) 2011
datasets. 97.4% of the vertices were assigned to the same network. 2012
2013
Figure 8 Confidence of 7-network estimate in Discovery dataset. Confidence (silhouette) 2014
value for each vertex with respect to its assigned network is shown for the Discovery 2015
dataset. Regions close to the boundaries between networks were less confident of their 2016
assignment, although we also observed structured spatial variation within individual 2017
components of the estimated networks, such as lateral prefrontal cortex, which 2018
foreshadows its division in the 17-network estimate (Fig. 10). 2019
2020
Figure 9 Discovery and Replication of a 17-network cortical parcellation. The 17- 2021
network estimates are highly consistent across the Discovery (n = 500) and Replication (n 2022
= 500) datasets. A total of 97.0% of the vertices were assigned to the same network. 2023
2024
Figure 10 Confidence of 17-network estimate in Discovery dataset. Confidence 2025
(silhouette) value for each vertex with respect to its assigned network is shown for the 2026
Discovery dataset. Again, regions close to the boundaries between networks were less 2027
confident of their assignment, while structured spatial variation were observed within 2028
individual components of the estimated networks. 2029
2030
Figure 11 A coarse (7-network) parcellation of the human cerebral cortex based on 1000 2031
subjects. To provide the best estimates of the 7 cortical networks, clustering was 2032
performed on the fMRI data of the full 1000 subjects. A salient feature is the separation 2033
of the early sensory and late motor cortices (blue and purple) from the association cortex. 2034
The association networks converged and extended upon networks previously described in 2035
the resting-state literature, including the dorsal attention, ventral attention, frontoparietal 2036
control and the default networks. 2037
2038
Figure 12 Table of colors assigned to networks in the 7-network estimate. Common 2039
names associated with each network in the neuroimaging literature are included in the 2040

74
brackets. This should not be taken to mean that our estimated networks correspond 2041
exactly to those in the literature or that the networks code solely for functions associated 2042
with their assigned name. For example, components of the green network are implicated 2043
in spatial attentional tasks (Corbetta and Shulman 2002) and so we refer to the green 2044
network as the dorsal attention system. However, from the macaque literature, 2045
components of the green network have also been described in relation to the hierarchical 2046
processing of sensory information (e.g., Felleman and Van Essen 1991) and recruited in 2047
tasks involving sensory-guided decisions (e.g., Andersen and Buneo 2002). 2048
2049
Figure 13 A fine-resolution (17-network) parcellation of the human cerebral cortex based 2050
on 1000 subjects. To provide the best estimates of the 17 cortical networks, clustering 2051
was performed on the fMRI data of the full 1000 subjects. The 17-network estimate 2052
fractionated the 7-network into smaller networks. Some aspects of the fractionations have 2053
been previously noted in other studies. 2054
2055
Figure 14 Uncertain observations due to limited data resolution and MR susceptibility. 2056
When interpreting the clustering results, potential artifacts and uncertainties must be 2057
considered. Because of the close proximity of the somatomotor and auditory cortices (A) 2058
and the close proximity of the pre- and post-central gyri (B), we are unable to resolve 2059
whether the clustering of the somatomotor and auditory cortices (A) and the clustering of 2060
the primary somatosensory and primary motor cortices (B) are due to the result of fMRI 2061
blurring across sulci or a true, coupled network of distributed areas as predicted by 2062
macaque tracing studies. (C) The orbital frontal-temporopolar network (cream color) 2063
consists of temporopolar and orbital frontal regions that are affected by MR 2064
susceptibility. Since MR susceptibility spatially distorts the MR signal and reduces SNR, 2065
there is uncertainty in the exact boundary of the orbital frontal-temporopolar network and 2066
the true extent of the network is probably underestimated. 2067
2068
Figure 15 Eccentricity estimates quantify the division of the early visual cortex into 2069
central and peripheral systems. Eccentricity estimates in the early visual areas of 4 2070
subjects were averaged and overlaid on the boundaries (in black) of the 17-network 2071

75
estimate. The boundary between areas 18 and 19 estimated from the Histological dataset 2072
is overlaid in green. The 17-network estimate divides the early visual areas along an 2073
isoeccentricity line of approximately 4
o
. Note that the eccentricity estimates are not 2074
reliable outside the V1-V3 complex. 2075
2076
Figure 16 Evidence that the fractionation of the visual system reflects fcMRI topography 2077
within the visual cortex. Six left hemisphere seed regions were picked from the 2078
Discovery dataset: V1
c
and V1
p
correspond to central and peripheral visual field 2079
representation within V1 respectively; V3
cv
and V3
pv
correspond to central and peripheral 2080
V3v respectively; ExC and ExP correspond to two seed regions within the extrastriate 2081
visual cortex in the estimated locations of the central and peripheral visual fields (purple 2082
and bright red in center panel). The six seed regions are illustrated in the center panel and 2083
their coordinate locations are reported in Table 1. Their left hemisphere fcMRI maps 2084
were computed using the Replication dataset and arranged around the center panel. Note 2085
that the central visual seed regions are selectively correlated with the central visual 2086
representation, while the peripheral visual seed regions are selectively correlated with the 2087
peripheral visual representation. 2088
2089
Figure 17 Quantification of fcMRI topography within the visual cortex and 2090
independence of the topography from task condition. (A) Quantification measures of 2091
functional connectivity strength are plotted in polar form for V1
c
(central V1) and V1
p
2092
(peripheral V1) seed regions for the Replication dataset. Note that V1 refers to V1
c
for 2093
the V1
p
polar plot (blue) and V1
p
for the V1
c
polar plot (red). Coordinate locations for all 2094
six seed regions (V1
c
, V1
p
, V3
cv
, V3
pv
, ExC, ExP) are reported in Table 1. (B) Polar plots 2095
of (A) replicated with the Task Effects dataset (EOR = eyes open rest; ECR = eyes closed 2096
rest; FIX = fixation) to ensure that the results obtained using the EOR Replication dataset 2097
were not due to overt eye movements that might shift edges and visual boundaries in and 2098
out of the central field. Left figure shows V1
p
polar plot. Right figure shows V1
c
polar 2099
plot. The polar plots quantify the differential functional coupling of central and peripheral 2100
V1 with higher visual areas. The polar scales range from r = -0.1 (center) to r = 0.7 (outer 2101
boundary) in 0.2-step increments. 2102

76
2103
Figure 18 V1 and V3 functional correlations display a smooth transition from the central 2104
to peripheral representations. Correlation of two series of seed regions spanning the 2105
eccentricity axes of V1 and V3v is shown for the full 1000 subjects. V1 seed regions of 2106
low eccentricity are strongly correlated with V3 seed regions of low eccentricity. V1 seed 2107
regions of high eccentricity are strongly correlated with V3 seed regions of high 2108
eccentricity. There is a gradual transition in functional connectivity strength between the 2109
central to peripheral representations. 2110
2111
Figure 19 7-network boundaries on probabilistic maps of areas 6, 2 and 5L. Boundaries 2112
of 7-network estimate based on the full 1000 subjects are overlaid on the surface-based 2113
probabilistic histological maps of areas 6, 2 and 5L. The somatomotor network includes 2114
most, if not all of areas 2 and 5L, but only caudal half of area 6. 2115
2116
Figure 20 Evidence that the fractionation of the somatomotor cortex reflects fcMRI 2117
topography within the somatosensory and motor cortex. (A) Average fMRI activation 2118
maps of 24 subjects instructed to move their tongue (blue), right hand (red) or right foot 2119
(green) across separate conditions. Black lines correspond to boundaries of the 17- 2120
network estimate. The dorsoventral split of the somatomotor network occurs spatially 2121
between the tongue and hand activations. (B) Quantification of correlation strength 2122
between the left hemisphere tongue, hand and foot seed regions selected from the 2123
activation maps. Hand coordinates = -41, -20, 62; Foot coordinates = -6, -26, 76; Tongue 2124
coordinates: -55, -4, 26. Hand-foot correlation is significantly higher than hand-tongue 2125
correlation, which is in turn significantly higher than foot-tongue correlation. 2126
2127
Figure 21 Evidence that the interhemispheric fcMRI of homotopic regions within the 2128
primary motor cortex is topographically organized. (A) Correlation strength of left 2129
hemisphere tongue, hand and foot seed regions with corresponding contralateral seed 2130
regions averaged over all 1000 subjects. Right hemisphere vertices were obtained by 2131
reflection across the midline. Hand coordinates = 41, -20, 62; Foot coordinates = 6, - 2132
26, 76; Tongue coordinates = 55, -4, 26. The tongue representation has the strongest 2133

77
interhemispheric correlation, followed by the foot and then the hand. (B) Plot of 2134
interhemispheric correlation along the ventral (tongue) to dorsal (foot) extent of motor 2135
cortex. Maximal interhemispheric correlation is highest near the tongue representation 2136
and also peaks between the hand and foot representations, possibly corresponding to the 2137
trunk representation. 2138
2139
Figure 22 Functional connectivity between MT+ and V1 is topographically organized. 2140
(A) Two MT+ seed regions, MT+
d
and MT+
v
, were selected in the dorsal and ventral 2141
portions of the histological MT+ estimate. aMT+ (anterior MT+) seed region was 2142
selected anterior to histological MT+. Four V1 seed regions were selected using the 2143
histological V1 estimate: V1
cd
and V1
cv
were selected in dorsal and ventral central V1; 2144
V1
pd
and V1
pd
were selected in dorsal and ventral peripheral V1. Coordinate locations of 2145
seed regions are reported in Table 3. (B) Correlation strength of aMT+ and MT+ seed 2146
regions with V1 in the Replication dataset. There are four observations to be noted: (1) 2147
V1-aMT+ correlation is weaker than V1-MT+ correlation, (2) MT+ correlation with the 2148
lower visual field is stronger than the upper visual field, (3) MT+
d
correlation with 2149
peripheral V1 is stronger than central V1 and (4) MT+
v
correlation with central V1 is 2150
stronger than peripheral V1. 2151
2152
Figure 23 Functional connectivity maps of MT+ reveal topographic organization. 2153
Functional connectivity maps of aMT+, MT+
v
and MT+
d
are computed using the 2154
Replication dataset and shown with views focusing on V1. V1 showed little or no 2155
correlation to aMT+, but strong correlations with both MT+ seeds. In both MT+ fcMRI 2156
maps, there is stronger correlation with dorsal V1 (lower visual field) than ventral V1 2157
(upper visual field). There is also increasing correlation with central V1 as we proceed 2158
from MT+
d
to MT+
v
. Yellow line denotes areal boundary of V1. 2159
2160
Figure 24 aMT+ and MT+ demonstrate differential functional connectivity with parietal 2161
and frontal cortices. Functional connectivity maps of aMT+ and MT+ seed regions 2162
computed using the Replication dataset. Coordinate locations of the regions are reported 2163
in Table 2. MT+ and aMT+ are more strongly correlated with SPL and IPS than with 2164

78
IPL. Correlation with frontal cortex is mostly limited to precentral sulcus and gyrus. 2165
aMT+ demonstrates stronger overall correlation with parietal and frontal cortices, 2166
compared with MT+. MT+ and aMT+ are maximally correlated with different parts of 2167
parietal and frontal cortices. 2168
2169
Figure 25 aMT+ and MT+ functional connectivity patterns generalize across task 2170
conditions. (A) Four visual, four parietal and two frontal seed regions were used to 2171
quantify the functional coupling of aMT+ and MT+ to distributed cortical regions. 2172
Coordinate locations of the seed regions are reported in Table 3 and were chosen either 2173
using the Discovery dataset or meta-analysis of fMRI studies (Table 4). (B) Polar plots of 2174
MT+ (blue) and aMT+ (red) connectivity with the visual, parietal and frontal seed 2175
regions are computed using the Replication dataset. MT+ is more strongly correlated with 2176
visual cortex as compared with parietal and frontal cortices. The converse is true for 2177
aMT+. (C, D) Polar plots of MT+ (blue) and aMT+ (red) connectivity replicated in the 2178
Task Effects dataset demonstrate that the functional coupling differences generalize 2179
across multiple data acquisition conditions. The polar scales range from r = -0.1 (center) 2180
to r = 0.6 (outer boundary) in 0.35-step increments. 2181
2182
Figure 26 Differential connectivity of dorsal and ventral caudal frontal cortex with SPL 2183
and IPS. Functional connectivity maps of FEF and PrC
v
are computed using the 2184
Replication dataset and shown with view focusing on parietal cortex. Both FEF and PrC
v
2185
demonstrate strong correlation with SPL and IPS. PrC
v
is more strongly correlated with 2186
ventral portions of rostral SPL and IPS, while FEF is more strongly correlated with 2187
caudal SPL and IPS. 2188
2189
Figure 27 Evidence for segregated pathways linking caudal frontal cortex with SPL and 2190
IPS. (A) Five parietal seed regions were selected along the rostrocaudal extent of SPL 2191
and IPS. Two frontal seed regions, FEF and PrC
v
, were selected in dorsal and ventral 2192
precentral sulcus. All seed regions were selected using the Discovery dataset or meta- 2193
analysis of fMRI studies (Table 4). The coordinate locations are reported in Table 3. (B) 2194
Functional connectivity of FEF and PrC
v
with the five parietal seed regions, arranged in 2195

79
rostral (lateral) to caudal (medial) order, from the Replication dataset. Rostrolateral IPS 2196
seed regions (IPS1, IPS2, IPS3
m
) were more strongly correlated with PrC
v
than FEF, 2197
while the mediocaudal SPL seed regions (SPL7A, SPL7P) were more strongly correlated 2198
with FEF than PrC
v
. 2199
2200
Figure 28 Examples of satisfied and violated constraints in estimating the functional 2201
hierarchy of cortical regions based on fcMRI. A functional hierarchy is estimated based 2202
on the assumption that regions closer in a hierarchy have stronger correlation. (A) Five 2203
cortical regions are arranged in a four-level hierarchy whose functional connectivity 2204
strengths satisfy both hierarchical and lateral constraints. (B) Identical arrangement of 2205
five cortical regions in a four-level hierarchy with different functional connectivity 2206
strengths that violate both hierarchical and lateral constraints. Thick lines correspond to 2207
strong correlations. Thin lines correspond to weak correlations. (i) Regions a and c 2208
are farther apart that regions a and b. In example (A), correlation of regions a and 2209
c is weaker than correlation of regions a and b, so a hierarchical constraint is 2210
satisfied. In example (B), correlation of regions a and c is stronger than correlation of 2211
regions a and b, so a hierarchical constraint is violated. (ii) Regions c and d are 2212
on the same hierarchical level. In example (A), correlation of regions c and e is 2213
approximately the same as the correlation of regions d and e, so a lateral constraint is 2214
satisfied. In example (B), correlation of regions c and e is stronger than the 2215
correlation of regions d and e, so a lateral constraint is violated. In the context of 2216
hierarchy estimation in this paper, we consider two correlations within 0.2 of each other 2217
to be approximately the same when assessing lateral constraints. Given the pairwise 2218
correlations of a set of seed regions and a known number of levels in the hierarchy, we 2219
can seek the best hierarchical arrangement of the seed regions with the following local 2220
optimization procedure: (1) randomly arrange the seed regions into a hierarchy, (2) 2221
consider swapping a pair of seed regions or shifting a single seed region to a different 2222
hierarchical level without changing the number of levels in the hierarchy such that the 2223
proportion of violated constraints is maximally decreased, (3) terminate when no further 2224
improvement in the proportion of violated constraints can be achieved and (4) repeat the 2225
preceding steps 20 times picking the solution with the least proportion of violated 2226

80
constraints. The best solution obtained using this optimization procedure is the same as a 2227
brute-force search over all possible hierarchical arrangements of the seed regions. We 2228
note that we are generally unable to infer the number of levels in the functional hierarchy, 2229
since the number of possible constraints can be drastically different for hierarchies with 2230
different number of levels, and so the proportion of violated or satisfied constraints is not 2231
comparable across hierarchies with different levels. In practice however, the solution 2232
space is robust; for example, the best solution for a 5-level hierarchy typically differs 2233
from the best solution for a 4-level hierarchy by the collapsing of regions from two 2234
adjacent levels into one level. Uncovering the true hierarchical structure in the macaque 2235
visual hierarchy based on anatomical connectivity has also proved to be problematic 2236
(Hilgetag et al. 1996). 2237
2238
Figure 29 Functional connectivity estimates of the hierarchical organization of a 2239
canonical sensory-motor pathway. (A) Six seed regions are arranged into a five-level 2240
functional hierarchy using the Replication dataset. (B, C) Two best hierarchical 2241
arrangements of the seed regions as measured by the proportion of violated hierarchical 2242
and lateral constraints. A violation occurred when the ordering placed more strongly 2243
correlated regions farther apart in the hierarchy than more weakly correlated regions (see 2244
Fig. 28). (D, E) Two poor hierarchical arrangements of the seed regions as measured by 2245
the proportion of violated hierarchical and lateral constraints. Relative ordering of the 2246
seed regions (A, B) within the functional hierarchy agrees well with the proposed 2247
macaque visual hierarchy (see text). 2248
2249
Figure 30 Adjacent parietal regions exhibit distinct functional connectivity fingerprints. 2250
Correlations of eight parietal seed regions (center panel) with 22 cortical target regions 2251
(top panel) from the Replication dataset and displayed as polar plots. Colors represent the 2252
7-network segmentation (from Fig. 11). The coordinate locations are reported in Table 4. 2253
Parietal seed regions that belong to the same network (e.g., TPJ, PGp
v
and PGp
d
) have 2254
generally similar functional connectivity fingerprints that are distinct from other parietal 2255
seed regions. Close inspection of the polar plots reveals distinct connectivity fingerprints 2256
even for parietal regions within the same network, some of which anticipate the further 2257

81
fractionation of the parietal cortex in the 17-network estimate (Fig. 13). Note that the 2258
cortical targets from Cinga to pCun on the left side of the polar plots are the same as that 2259
of the frontal polar plots (Fig. 31) to allow for comparison across the two sets of polar 2260
plots. The remaining cortical targets are different across the two sets of polar plots but are 2261
arranged so that cortical targets at the same location in the polar plots belong to the same 2262
network in the 7-network estimate. The polar scales range from r = -0.4 (center) to r = 2263
0.5 (outer boundary) in 0.3-step increments. 2264
2265
Figure 31 Adjacent frontal regions exhibit distinct functional connectivity fingerprints. 2266
The format and plotting are the same as Figure 30 with regions tailored for exploration of 2267
frontal cortex. The coordinate locations are reported in Table 4. The polar scales range 2268
from r = -0.4 (center) to r = 0.5 (outer boundary) in 0.3-step increments. 2269
2270
Figure 32 Functional connectivity for regions within the canonical distributed cortical 2271
network. This network is often called the dorsal attention network. The five seed regions 2272
are displayed in the center overlaid on top of the 7-network parcellation from Figure 11. 2273
The coordinate locations are reported in Table 5. Each panel displays the functional 2274
connectivity map for one of the six seed regions for the Replication dataset overlaid on a 2275
surface that shows the 7-network boundaries (in light black lines) as reference. Each seed 2276
region displays functional coupling with all of the regions of the distributed network. 2277
However, there are important differences between regions. In particular the regions near 2278
aMT+ (panel D) and SPL7A (panel C) show strong functional coupling with earlier 2279
visual areas. The region at or near the putative homologue of FEF (panel A) shows 2280
minimal functional coupling with earlier visual areas but does show strong coupling with 2281
midline motor regions (see midline section of panel A). The color scale (bottom) shows 2282
the plotted correlation range for the maps. 2283
2284
Figure 33 Functional connectivity for distributed regions within a second large-scale 2285
association network. This network is often called the ventral attention network. The 2286
format and plotting are the same as Figure 32 and coordinate locations are reported in 2287
Table 5. Each seed region is functionally coupled mostly with regions within the same 2288

82
network revealing that each component of the network recapitulates the others. There is a 2289
general absence of crosstalk between networks except for local correlation around the 2290
seed regions. 2291
2292
Figure 34 Functional connectivity for distributed regions within a third large-scale 2293
association network. This network is often called the frontoparietal control network. The 2294
format and plotting are the same as Figure 32 and coordinate locations are reported in 2295
Table 5. In addition to a general absence of crosstalk between networks this network also 2296
shows no functional coupling to sensory and motor regions. Rather its topography reveals 2297
a distributed network that is interdigitated with the networks illustrated in Figures 33 and 2298
35. 2299
2300
Figure 35 Functional connectivity for distributed regions within a fourth large-scale 2301
association network. This network makes up the prominent components of the network 2302
often called the default network. The format and plotting are the same as Figure 32 and 2303
coordinate locations are reported in Table 5. Each seed region is functionally coupled 2304
mostly with regions within the same network again revealing that each component of the 2305
network recapitulates components the remaining network, with some exceptions. For 2306
example, the seed region in the parahippocampal cortex (panel E) shows functional 2307
connectivity with the retrosplenial cortex, ventral medial prefrontal cortex, and a specific 2308
region with the caudal IPL. These patterns of functional connectivity lead to the 2309
fractionation into subnetworks as illustrated in Figure 13. The fractionation of this 2310
particular network is largely to subdivide the broader network rather than to display 2311
functional coupling with regions in distinct networks. 2312
2313
2314
2315
2316

83
References 2317
2318
Ahn Y-Y, Bagrow JP, Lehmann S. Link communities reveal multiscale complexity in 2319
networks. Nature 466: 761-764, 2010. 2320
Amano K, Wandell BA, Dumoulin SO. Visual field maps, population receptive field 2321
sizes, and visual field coverage in the human MT+ complex. J Neurophysiol 102: 2322
2704-2718, 2009. 2323
Amunts K, Lenzen M, Friederici AD, Schleicher A, Morosan P, Palomero-Gallagher 2324
N, Zilles K. Broca's region: Novel organizational principles and multiple receptor 2325
mapping. PLoS Biol 8: e1000489, 2010. 2326
Amunts K, Malikovic A, Mohlberg H, Schormann T, Zilles K. Brodmann's areas 17 2327
and 18 brought into stereotaxic space - Where and how variable? Neuroimage 11: 66- 2328
84, 2000. 2329
Amunts K, Schleicher A, Brgel U, Mohlberg H, Uylings HBM, Zilles K. Broca's 2330
region revisited: Cytoarchitecture and intersubject variability. J Comp Neurol 412: 2331
319-341, 1999. 2332
Andersen RA, Buneo CA. Intentional maps in posterior parietal cortex. Annu Rev 2333
Neurosci 25: 189-220, 2002. 2334
Andersen RA, Asanuma C, Essick G, Siegel RM. Corticocortical connections of 2335
anatomically and physiologically defined subdivisions within the inferior parietal 2336
lobule. J Comp Neurol 296: 65-113, 1990. 2337
Andrews-Hanna JR, Reidler JS, Sepulcre J, Poulin R, Buckner RL. Functional- 2338
anatomic fractionation of the brain's default network. Neuron 65: 550-562, 2010. 2339
Baizer JS, Ungerleider LG, Desimone R. Organization of visual inputs to the inferior 2340
temporal and posterior parietal cortex in macaques. J Neurosci 11: 168-190, 1991. 2341
Balasubramanian M, Polimeni J, Schwartz EL. The V1-V2-V3 complex: 2342
Quasiconformal dipole maps in primate striate and extra-striate cortex. Neural 2343
Networks 15: 1157-1163, 2002. 2344
Barbas H, Pandya D. Architecture and frontal cortical connections of the premotor 2345
cortex (area 6) in the rhesus monkey. J Comp Neurol 256: 211-228, 1987. 2346

84
Basser PJ, Mattiello J, LeBihan D. Estimation of the effective self-diffusion tensor 2347
from the NMR spin echo. J Magn Reson B 103: 247-254, 1994. 2348
Beckmann CF, DeLuca M, Devlin JT, Smith SM. Investigations into resting-state 2349
connectivity using independent component analysis. Philos Trans R Soc Lond B Biol 2350
Sci 360: 1001-1013, 2005. 2351
Beckmann CF, Smith SM. Probabilistic independent component analysis for functional 2352
magnetic resonance imaging. IEEE Trans Med Imaging 23: 137-152, 2004. 2353
Bellec P, Rosa-Neto P, Lyttelton OC, Benali H, Evans AC. Multi-level bootstrap 2354
analysis of stable clusters in resting-state fMRI. Neuroimage 51: 1126-1139, 2010. 2355
Ben-Hur A, Elisseeff A, Guyon I. A stability based method for discovering structure in 2356
clustered data. Pacific Symposium on Biocomputing: 6-17, 2002. 2357
Binkofski F, Buccino G, Posse S, Seitz RJ, Rizzolatti G, Freund H-J. A fronto- 2358
parietal circuit for object manipulation in man: Evidence from an fMRI-study. Eur J 2359
Neurosci 11: 3276-3286, 1999. 2360
Binkofski F, Dohle C, Posse S, Stephan KM, Hefter H, Seitz RJ, Freund H-J. Human 2361
anterior intraparietal area subserves prehension: a combined lesion and functional 2362
MRI activation study. Neurology 50: 1253-1259, 1998. 2363
Biswal B, Yetkin FZ, Haughton VM, Hyde JS. Functional connectivity in the motor 2364
cortex of resting human brain using echo-planar MRI. Magn Reson Med 34: 537-541, 2365
1995. 2366
Biswal BB, Mennes M, Zuo X-N, Gohel S, Kelly C, Smith SM, Beckmann CF, 2367
Adelstein JS, Buckner RL, Colcombe S, Dogonowski A-M, Ernst M, Fair D, 2368
Hampson M, Hoptman MJ, Hyde JS, Kiviniemi VJ, Ktter R, Li S-J, Lin C-P, 2369
Lowe MJ, Mackay C, Madden DJ, Madsen KH, Margulies DS, Mayberg HS, 2370
McMahon K, Monk CS, Mostofsky SH, Nagel BJ, Pekar JJ, Peltier SJ, Petersen 2371
SE, Riedl V, Rombouts SARB, Rypma B, Schlaggar BL, Schmidt S, Seidler RD, 2372
Siegle GJ, Sorg C, Teng G-J, Veijola J, Villringer A, Walter M, Wang L, Weng 2373
X-C, Whitfield-Gabrieli S, Williamson P, Windischberger C, Zang Y-F, Zhang 2374
H-Y, Castellanos FX, Milham MP. Toward discovery science of human brain 2375
function. Proc Natl Acad Sci USA 107: 4734-4739, 2010. 2376

85
Blinkov SM, Glezer II. The Human Brain in Figures and Tables. New York: Basic 2377
Books 1968. 2378
Brewer AA, Liu J, Wade AR, Wandell BA. Visual field maps and stimulus selectivity 2379
in human ventral occipital cortex. Nat Neurosci 8: 1102-1109, 2005. 2380
Brodmann K. Localization in the Cerebral Cortex. Translated by Garey LJ. New York: 2381
Springer 1909/2006. 2382
Buckner RL. Human functional connectivity: New tools, unresolved questions. Proc 2383
Natl Acad Sci USA 107: 10769-10770, 2010. 2384
Buckner RL, Andrews-Hanna JR, Schacter DL. The brain's default network: 2385
Anatomy, function, and relevance to disease. Ann N Y Acad Sci 1124: 1-38, 2008. 2386
Buckner RL, Krienen FM, Castellanos A, Diaz JC, Yeo BTT. The organization of the 2387
human cerebellum estimated by functional connectivity. J Neurophysiol, submitted 2388
Buckner RL, Sepulcre J, Talukdar T, Krienen FM, Liu H, Hedden T, Andrews- 2389
Hanna JR, Sperling RA, Johnson KA. Cortical hubs revealed by intrinsic 2390
functional connectivity: Mapping, assessment of stability, and relation to Alzheimer's 2391
disease. J Neurosci 29: 1860-1873, 2009. 2392
Bullmore E, Sporns O. Complex brain networks: Graph theoretical analysis of structural 2393
and functional systems. Nat Rev Neurosci 10:186-198, 2009. 2394
Cabeza R, Ciaramelli E, Olson IR, Moscovitch M. The parietal cortex and episodic 2395
memory: An attentional account. Nat Rev Neurosci 9: 613-625, 2008. 2396
Carmichael ST, Price JL. Limbic connections of the orbital and medial prefrontal 2397
cortex in macaque monkeys. J Comp Neurol 363: 615-641, 1995. 2398
Caspers S, Eickhoff SB, Geyer S, Scheperjans F, Mohlberg H, Zilles K, Amunts K. 2399
The human inferior parietal lobule in stereotaxic space. Brain Struct Funct 212: 481- 2400
495, 2008. 2401
Caspers S, Geyer S, Schleicher A, Mohlberg H, Amunts K, Zilles K. The human 2402
inferior parietal cortex: Cytoarchitectonic parcellation and interindividual variability. 2403
Neuroimage 33: 430-448, 2006. 2404
Catani M, ffytche DH. The rises and falls of disconnection syndromes. Brain 128:2224- 2405
2239, 2005. 2406

86
Cavada C, Goldman-Rakic PS. Posterior parietal cortex in rhesus monkey: II. Evidence 2407
for segregated corticocortical networks linking sensory and limbic areas with the 2408
frontal lobe. J Comp Neurol 287: 422-445, 1989a. 2409
Cavada C, Goldman-Rakic PS. Posterior parietal cortex in rhesus monkey: I. 2410
Parcellation of areas based on distinctive limbic and sensory corticocortical 2411
connections. J Comp Neurol 287: 393-421, 1989b. 2412
Choi H-J, Zilles K, Mohlberg H, Schleicher A, Fink GR, Armstrong E, Amunts K. 2413
Cytoarchitectonic identification and probabilistic mapping of two distinct areas 2414
within the anterior ventral bank of the human intraparietal sulcus. J Comp Neurol 2415
495: 53-69, 2006. 2416
Cohen AL, Fair DA, Dosenbach NUF, Miezin FM, Dierker D, Van Essen DC, 2417
Schlaggar BL, Petersen SE. Defining functional areas in individual human brains 2418
using resting functional connectivity MRI. Neuroimage 41: 45-57, 2008. 2419
Colby CL, Goldberg ME. Space and attention in parietal cortex. Annu Rev Neurosci 2420
22:319-349, 1999. 2421
Connolly JD, Goodale MA, Desouza JF, Menon RS, Vilis T. A comparison of 2422
frontoparietal fMRI activation during anti-saccades and anti-pointing. J Neurophysiol 2423
84: 1645-1655, 2000. 2424
Connolly JD, Goodale MA, Menon RS, Munoz DP. Human fMRI evidence for the 2425
neural correlates of preparatory set. Nat Neurosci 5: 1345-1352, 2002. 2426
Corbetta M, Akbudak E, Conturo TE, Snyder AZ, Ollinger JM, Drury HA, 2427
Linenweber MR, Petersen SE, Raichle ME, Van Essen DC, Shulman GL. A 2428
common network of functional areas for attention and eye movements. Neuron 21: 2429
761-773, 1998. 2430
Corbetta M, Shulman GL. Control of goal-directed and stimulus-driven attention in the 2431
brain. Nat Rev Neurosci 3: 201-215, 2002. 2432
Cragg BG. The topography of the afferent projections in the circumstriate visual cortex 2433
of the monkey studied by the Nauta method. Vision Res 9: 733-747, 1969. 2434
Culham JC, Danckert SL, DeSouza JFX, Gati JS, Menon RS, Goodale MA. Visually 2435
guided grasping produces fMRI activation in dorsal but not ventral stream brain 2436
areas. Exp Brain Res 153: 180-189, 2003. 2437

87
Culham JC, Kanwisher NG. Neuroimaging of cognitive functions in human parietal 2438
cortex. Curr Opin Neurobiol 11: 157-163, 2001. 2439
Dale AM, Fischl B, Sereno MI. Cortical surface-based analysis - I. Segmentation and 2440
surface reconstruction. Neuroimage 9: 179-194, 1999. 2441
Damoiseaux JS, Rombouts SARB, Barkhof F, Scheltens P, Stam CJ, Smith SM, 2442
Beckmann CF. Consistent resting-state networks across healthy subjects. Proc Natl 2443
Acad Sci USA 103: 13848-13853, 2006. 2444
De Luca M, Beckmann CF, De Stefano N, Matthews PM, Smith SM. fMRI resting 2445
state networks define distinct modes of long-distance interactions in the human brain. 2446
Neuroimage 29: 1359-1367, 2006. 2447
DeYoe EA, Carman GJ, Bandettini P, Glickman S, Wieser J, Cox R, Miller D, Neitz 2448
J. Mapping striate and extrastriate visual areas in human cerebral cortex. Proc Natl 2449
Acad Sci USA 93: 2382-2386, 1996. 2450
Dickson J, Drury H, Van Essen DC. 'The surface management system' (SuMS) 2451
database: A surface-based database to aid cortical surface reconstruction, 2452
visualization and analysis. Philos Trans R Soc Lond B Biol Sci 356: 1277-1292, 2001. 2453
Disbrow E, Litinas E, Recanzone GH, Padberg J, Krubitzer L. Cortical connections 2454
of the second somatosensory area and the parietal ventral area in macaque monkeys. J 2455
Comp Neurol 462: 382-399, 2003. 2456
Distler C, Boussaoud D, Desimone R, Ungerleider LG. Cortical connections of inferior 2457
temporal area TEO in macaque monkeys. J Comp Neurol 334: 125-150, 1993. 2458
Dosenbach NUF, Fair DA, Miezin FM, Cohen AL, Wenger KK, Dosenbach RAT, 2459
Fox MD, Snyder AZ, Vincent JL, Raichle ME, Schlaggar BL, Petersen SE. 2460
Distinct brain networks for adaptive and stable task control in humans. Proc Natl 2461
Acad Sci USA 104: 11073-11078, 2007. 2462
Dow BM, Snyder AZ, Vautin RG, Bauer R. Magnification factor and receptive field 2463
size in foveal striate cortex of the monkey. Exp Brain Res 44: 213-228, 1981. 2464
Eickhoff SB, Stephan KE, Mohlberg H, Grefkes C, Fink GR, Amunts K, Zilles K. A 2465
new SPM toolbox for combining probabilistic cytoarchitectonic maps and functional 2466
imaging data. Neuroimage 25: 1325-1335, 2005. 2467

88
Engel SA, Glover GH, Wandell BA. Retinotopic organization in human visual cortex 2468
and the spatial precision of functional MRI. Cereb Cortex 7: 181-192, 1997. 2469
Engel SA, Rumelhart DE, Wandell BA, Lee AT, Glover GH, Chichilnisky EJ, 2470
Shadlen MN. fMRI of human visual cortex. Nature 369: 525, 1994. 2471
Evarts EV, Thach WT. Motor mechanisms of the CNS: Cerebrocerebellar 2472
interrelations. Ann Rev Physiol 31:451-498, 1969. 2473
Faillenot I, Sunaert S, Van Hecke P, Orban GA. Orientation discrimination of objects 2474
and gratings compared: An fMRI study. Eur J Neurosci 13: 585-596, 2001. 2475
Felleman DJ, Van Essen DC. Distributed hierarchical processing in the primate cerebral 2476
cortex. Cereb Cortex 1: 1-47, 1991. 2477
Fischl B, Liu A, Dale AM. Automated manifold surgery: Constructing geometrically 2478
accurate and topologically correct models of the human cerebral cortex. IEEE Trans 2479
Med Imaging 20: 70-80, 2001. 2480
Fischl B, Rajendran N, Busa E, Augustinack J, Hinds O, Yeo BTT, Mohlberg H, 2481
Amunts K, Zilles K. Cortical folding patterns and predicting cytoarchitecture. Cereb 2482
Cortex 18: 1973-1980, 2008. 2483
Fischl B, Sereno MI, Dale AM. Cortical surface-based analysis - II: Inflation, flattening, 2484
and a surface-based coordinate system. Neuroimage 9: 195-207, 1999a. 2485
Fischl B, Sereno MI, Tootell RBH, Dale AM. High-resolution intersubject averaging 2486
and a coordinate system for the cortical surface. Hum Brain Mapp 8: 272-284, 1999b. 2487
Fox MD, Snyder AZ, Vincent JL, Corbetta M, Van Essen DC, Raichle ME. The 2488
human brain is intrinsically organized into dynamic, anticorrelated functional 2489
networks. Proc Natl Acad Sci USA 102: 9673-9678, 2005. 2490
Fox MD, Corbetta M, Snyder A, Vincent J, Raichle ME. Spontaneous neuronal 2491
activity distinguishes human dorsal and ventral attention systems. Proc Natl Acad Sci 2492
USA 103: 10046, 2006. 2493
Fox MD, Raichle ME. Spontaneous fluctuations in brain activity observed with 2494
functional magnetic resonance imaging. Nat Rev Neurosci 8: 700-711, 2007. 2495
Frahm HD, Stephan H, Baron G. Comparison of brain structure volumes in Insectivora 2496
and Primates: V. Area striata (AS). J. Hirnforsch 25:537-577, 1984. 2497

89
Friedman DP, Murray EA, O'Neill JB, Mishkin M. Cortical connections of the 2498
somatosensory fields of the lateral sulcus of macaques: Evidence for a corticolimbic 2499
pathway for touch. J Comp Neurol 252: 323-347, 1986. 2500
Friedman DP. Laminar patterns of termination of cortico-cortical afferents in the 2501
somatosensory system. Brain Res 273: 147-151, 1983. 2502
Friston KJ. Functional and effective connectivity in neuroimaging: A synthesis. Hum 2503
Brain Mapp 2: 56-78, 1994. 2504
Geschwind N. Disconnexion syndromes in animals and man. I. Brain 88:237-294, 1965. 2505
Geyer S. The microstructural border between the motor and the cognitive domain in the 2506
human cerebral cortex. Adv Anat Embryol Cell Biol 174: 1-89, 2004. 2507
Geyer S, Ledberg A, Schleicher A, Kinomura S, Schormann T, Brgel U, Klingberg 2508
T, Larsson J, Zilles K, Roland PE. Two different areas within the primary motor 2509
cortex of man. Nature 382: 805-807, 1996. 2510
Geyer S, Schleicher A, Zilles K. Areas 3a, 3b, and 1 of human primary somatosensory 2511
cortex 1. Microstructural organization and interindividual variability. Neuroimage 10: 2512
63-83, 1999. 2513
Gold JI, Shadlen MN. The neural basis of decision making. Annu Rev Neurosci 30: 535- 2514
574, 2007. 2515
Goldman-Rakic PS. Topography of cognition: Parallel distributed networks in primate 2516
association cortex. Annu Rev Neurosci 11: 137-156, 1988. 2517
Golland P, Golland Y, Malach R. Detection of spatial activation patterns as 2518
unsupervised segmentation of fMRI data. In Proc. MICCAI: Medical Image 2519
Computing and Computer Assisted Intervention, LNCS 5241:745-753, 2007. 2520
Gould HJ, Cusick CG, Pons TP, Kaas JH. The relationship of corpus callosum 2521
connections to electrical stimulation maps of motor, supplementary motor, and the 2522
frontal eye fields in owl monkeys. J Comp Neurol 247: 297-325, 1986. 2523
Grefkes C, Geyer S, Schormann T, Roland P, Zilles K. Human somatosensory area 2: 2524
Observer-independent cytoarchitectonic mapping, interindividual variability, and 2525
population map. Neuroimage 14: 617-631, 2001. 2526

90
Grefkes C, Weiss PH, Zilles K, Fink GR. Crossmodal processing of object features in 2527
human anterior intraparietal cortex: An fMRI study implies equivalencies between 2528
humans and monkeys. Neuron 35: 173-184, 2002. 2529
Greicius MD, Krasnow B, Reiss AL, Menon V. Functional connectivity in the resting 2530
brain: A network analysis of the default mode hypothesis. Proc Natl Acad Sci USA 2531
100: 253-258, 2003. 2532
Greicius MD, Srivastava G, Reiss AL, Menon V. Default-mode network activity 2533
distinguishes Alzheimer's disease from healthy aging: Evidence from functional MRI. 2534
Proc Natl Acad Sci USA 101: 4637-4642, 2004. 2535
Greve DN, Fischl B. Accurate and robust brain image alignment using boundary-based 2536
registration. Neuroimage 48: 63-72, 2009. 2537
Grill-Spector K, Malach R. The human visual cortex. Annu Rev Neurosci 27: 649-677, 2538
2004. 2539
Hadjikhani N, Liu A, Dale A, Cavanagh P, Tootell RBH. Retinotopy and color 2540
sensitivity in human visual cortical area V8. Nat Neurosci 1: 235-241, 1998. 2541
Hagler DJ, Riecke L, Sereno MI. Parietal and superior frontal visuospatial maps 2542
activated by pointing and saccades. Neuroimage 35: 1562-1577, 2007. 2543
Hagmann P, Cammoun L, Gigandet X, Meuli R, Honey CJ, Wedeen VJ, Sporns O. 2544
Mapping the structural core of human cerebral cortex. PLoS Biol 6: e159, 2008. 2545
Heide W, Binkofski F, Seitz RJ, Posse S, Nitschke MF, Freund H-J, Kmpf D. 2546
Activation of frontoparietal cortices during memorized triple-step sequences of 2547
saccadic eye movements: An fMRI study. Eur J Neurosci 13: 1177-1189, 2001. 2548
Hilgetag C, O'Neill M, Young M. Indeterminate organization of the visual system. 2549
Science 271:776-777, 1996. 2550
Hill J, Inder T, Neil J, Dierker D, Harwell J, Van Essen D. Similar patterns of cortical 2551
expansion during human development and evolution. Proc Natl Acad Sci USA 107: 2552
13135-13140, 2010. 2553
Hinds O, Polimeni JR, Rajendran N, Balasubramanian M, Amunts K, Zilles K, 2554
Schwartz EL, Fischl B, Triantafyllou C. Locating the functional and anatomical 2555
boundaries of human primary visual cortex. Neuroimage 46: 915-922, 2009. 2556

91
Hinds OP, Rajendran N, Polimeni JP, Augustinack JC, Wiggins G, Wald LL, Diana 2557
Rosas HD, Potthast A, Schwartz EL, Fischl B. Accurate prediction of V1 location 2558
from cortical folds in a surface coordinate system. Neuroimage 39: 1585-1599, 2008. 2559
Honey CJ, Sporns O, Cammoun L, Gigandet X, Thiran JP, Meuli R, Hagmann P. 2560
Predicting human resting-state functional connectivity from structural connectivity. 2561
Proc Natl Acad Sci USA 106: 2035-2040, 2009. 2562
Hubel DH, Wiesel TN. Receptive fields, binocular interaction and functional 2563
architecture in the cat's visual cortex. J Physiol 160: 106-154, 1962. 2564
Huk AC, Dougherty RF, Heeger DJ. Retinotopy and functional subdivision of human 2565
areas MT and MST. J Neurosci 22: 7195-7205, 2002. 2566
Iwamura Y. Hierarchical somatosensory processing. Curr Opin Neurobiol 8: 522-528, 2567
1998. 2568
James W. Principles of Psychology, Vol 1. New York: Holt 1890. 2569
Jenkinson M, Bannister P, Brady M, Smith S. Improved optimization for the robust 2570
and accurate linear registration and motion correction of brain images. Neuroimage 2571
17: 825-841, 2002. 2572
Johansen-Berg H, Behrens TEJ, Robson MD, Drobnjak I, Rushworth MFS, Brady 2573
JM, Smith SM, Higham DJ, Matthews PM. Changes in connectivity profiles define 2574
functionally distinct regions in human medial frontal cortex. Proc Natl Acad Sci USA 2575
101: 13335-13340, 2004. 2576
Johnston JM, Vaishnavi SN, Smyth MD, Zhang D, He BJ, Zempel JM, Shimony JS, 2577
Snyder AZ, Raichle ME. Loss of resting interhemispheric functional connectivity 2578
after complete section of the corpus callosum. J Neurosci 28: 6453-6458, 2008. 2579
Jones EG, Coulter JD, Hendry SHC. Intracortical connectivity of architectonic fields in 2580
the somatic sensory, motor and parietal cortex of monkeys. J Comp Neurol 181: 291- 2581
347, 1978. 2582
Jones EG, Powell TPS. An anatomical study of converging sensory pathways within the 2583
cerebral cortex of the monkey. Brain 93: 793-820, 1970. 2584
Jones EG, Wise SP. Size, laminar and columnar distribution of efferent cells in the 2585
sensory-motor cortex of monkeys. J Comp Neurol 175: 391-438, 1977. 2586

92
Kaas JH. The organization of neocortex in mammals: Implications for theories of brain 2587
function. Annu Rev Psychol 38: 129-151, 1987. 2588
Kemp JM, Powell TPS. The connexions of the striatum and globus pallidus: Synthesis 2589
and speculation. Phil Trans R Soc Lond B 262:441-457, 1971. 2590
Killackey HP, Gould HJ, Cusick CG, Pons TP, Kaas JH. The relation of corpus 2591
callosum connections to architectonic fields and body surface maps in sensorimotor 2592
cortex of new and old world monkeys. J Comp Neurol 219: 384-419, 1983. 2593
Kolster H, Peeters R, Orban GA. The retinotopic organization of the human middle 2594
temporal area MT/V5 and its cortical neighbors. J Neurosci 30: 9801-9820, 2010. 2595
Kondo H, Saleem KS, Price JL. Differential connections of the temporal pole with the 2596
orbital and medial prefrontal networks in macaque monkeys. J Comp Neurol 465: 2597
499-523, 2003. 2598
Koyama M, Hasegawa I, Osada T, Adachi Y, Nakahara K, Miyashita Y. Functional 2599
magnetic resonance imaging of macaque monkeys performing visually guided 2600
saccade tasks: Comparison of cortical eye fields with humans. Neuron 41: 795-807, 2601
2004. 2602
Kurata K. Corticocortical inputs to the dorsal and ventral aspects of the premotor cortex 2603
of macaque monkeys. Neurosci Res 12: 263-280, 1991. 2604
Kwong KK, Belliveau JW, Chesler DA, Goldberg IE, Weisskoff RM, Poncelet BP, 2605
Kennedy DN, Hoppel BE, Cohen MS, Turner R, Cheng H-M, Brady T, Rosen 2606
BR. Dynamic magnetic resonance imaging of human brain activity during primary 2607
sensory stimulation. Proc Natl Acad Sci USA 89: 5675-5679, 1992. 2608
Lancaster JL, Tordesillas-Gutierrez D, Martinez M, Salinas F, Evans A, Zilles K, 2609
Mazziotta JC, Fox PT. Bias between MNI and Talairach coordinates analyzed using 2610
the ICBM-152 brain template. Hum Brain Mapp 28: 1194-1205, 2007. 2611
Larsson J, Heeger DJ. Two retinotopic visual areas in human lateral occipital cortex. J 2612
Neurosci 26: 13128-13142, 2006. 2613
Lange T, Roth V, Braun M, Buhmann JM. Stability-based validation of clustering 2614
solutions. Neural Comput 16: 1299-1323, 2004. 2615
Lashkari D, Vul E, Kanwisher N, Golland P. Discovering structure in the space of 2616
fMRI selectivity profiles. Neuroimage 50: 1085-1098, 2010. 2617

93
Lavenex P, Suzuki WA, Amaral DG. Perirhinal and parahippocampal cortices of the 2618
macaque monkey: Projections to the neocortex. J Comp Neurol 447: 394-420, 2002. 2619
Levy I, Hasson U, Avidan G, Hendler T, Malach R. Center-periphery organization of 2620
human object areas. Nat Neurosci 4: 533-539, 2001. 2621
Logothetis NK. What we can do and what we cannot do with fMRI. Nature 453: 869- 2622
878, 2008. 2623
Luna B, Thulborn KR, Strojwas MH, McCurtain BJ, Berman RA, Genovese CR, 2624
Sweeney JA. Dorsal cortical regions subserving visually guided saccades in humans: 2625
an fMRI study. Cereb Cortex 8: 40-47, 1998. 2626
Luppino G, Matelli M, Camarda R, Rizzolatti G. Corticocortical connections of area 2627
F3 (SMA-proper) and area F6 (pre-SMA) in the macaque monkey. J Comp Neurol 2628
338: 114-140, 1993. 2629
Malach R, Reppas JB, Benson RR, Kwong KK, Jiang H, Kennedy WA, Ledden PJ, 2630
Brady TJ, Rosen BR, Tootell RBH. Object-related activity revealed by functional 2631
magnetic resonance imaging in human occipital cortex. Proc Natl Acad Sci USA 92: 2632
8135-8139, 1995. 2633
Malikovic A, Amunts K, Schleicher A, Mohlberg H, Eickhoff SB, Wilms M, 2634
Palomero-Gallagher N, Armstrong E, Zilles K. Cytoarchitectonic analysis of the 2635
human extrastriate cortex in the region of V5/MT+: A probabilistic, stereotaxic map 2636
of area hOc5. Cereb Cortex 17: 562-574, 2007. 2637
Marcus DS, Fotenos AF, Csernansky JG, Morris JC, Buckner RL. Open access 2638
series of imaging studies: Longitudinal MRI data in nondemented and demented older 2639
adults. J Cogn Neurosci 22: 2677-2684, 2010. 2640
Marcus DS, Wang TH, Parker J, Csernansky JG, Morris JC, Buckner RL. Open 2641
access series of imaging studies (OASIS): Cross-sectional MRI data in young, middle 2642
aged, nondemented, and demented older adults. J Cogn Neurosci 19: 1498-1507, 2643
2007. 2644
Markov NT, Misery P, Falchier A, Lamy C, Vezoli J, Quilodran R, Gariel MA, 2645
Giroud P, Ercsey-Ravasz M, Pilaz LJ, Huissoud C, Barone P, Dehay C, 2646
Toroczkai Z, Van Essen DC, Kennedy H, Knoblauch K. Weight consistency 2647
specifies regularities of macaque cortical networks. Cereb Cortex 2010. 2648

94
Matelli M, Camarda R, Glickstein M, Rizzolatti G. Afferent and efferent projections 2649
of the inferior area 6 in the macaque monkey. J Comp Neurol 251: 281-298, 1986. 2650
Matelli M, Govoni P, Galletti C, Kutz DF, Luppino G. Superior area 6 afferents from 2651
the superior parietal lobule in the macaque monkey. J Comp Neurol 402: 327-352, 2652
1998. 2653
Matelli M, Luppino G, Rizzolatti G. Architecture of superior and mesial area 6 and the 2654
adjacent cingulate cortex in the macaque monkey. J Comp Neurol 311: 445-462, 2655
1991. 2656
Matelli M, Luppino G, Rizzolatti G. Patterns of cytochrome oxidase activity in the 2657
frontal agranular cortex of the macaque monkey. Behav Brain Res 18: 125-136, 1985. 2658
Maunsell JHR, Van Essen DC. The connections of the middle temporal visual area 2659
(MT) and their relationship to a cortical hierarchy in the macaque monkey. J Neurosci 2660
3: 2563, 1983. 2661
Maunsell JHR, Van Essen DC. Topographic organization of the middle temporal visual 2662
area in the macaque monkey: Representational biases and the relationship to callosal 2663
connections and myeloarchitectonic boundaries. J Comp Neurol 266: 535-555, 1987. 2664
Medendorp WP, Goltz HC, Vilis T, Crawford JD. Gaze-centered updating of visual 2665
space in human parietal cortex. J Neurosci 23: 6209-6214, 2003. 2666
Mesulam MM. A cortical network for directed attention and unilateral neglect. Ann 2667
Neurol 10: 309-325, 1981. 2668
Mesulam MM. Frontal cortex and behavior. Ann Neurol 19: 320-325, 1986. 2669
Mesulam MM. Large-scale neurocognitive networks and distributed processing for 2670
attention, language, and memory. Ann Neurol 28: 597-613, 1990. 2671
Mesulam MM. From sensation to cognition. Brain 121 (6): 1013-1052, 1998. 2672
Mesulam MM, Mufson EJ. Insula of the old world monkey. III: Efferent cortical output 2673
and comments on function. J Comp Neurol 212: 38-52, 1982. 2674
Moeller S, Nallasamy N, Tsao DY, Freiwald WA. Functional connectivity of the 2675
macaque brain across stimulus and arousal states. J Neurosci 29: 5897-5909, 2009. 2676
Moran MA, Mufson EJ, Mesulam MM. Neural inputs into the temporopolar cortex of 2677
the rhesus monkey. J Comp Neurol 256: 88-103, 1987. 2678

95
Muakkassa KF, Strick PL. Frontal lobe inputs to primate motor cortex: Evidence for 2679
four somatotopically organized 'premotor' areas. Brain Res 177: 176-182, 1979. 2680
Mufson EJ, Mesulam MM. Insula of the old world monkey. II: Afferent cortical input 2681
and comments on the claustrum. J Comp Neurol 212: 23-37, 1982. 2682
Nelson SM, Cohen AL, Power JD, Wig GS, Miezin FM, Wheeler ME, Velanova K, 2683
Donaldson DI, Phillips JS, Schlaggar BL, Petersen SE. A parcellation scheme for 2684
human left lateral parietal cortex. Neuron 67: 156-170, 2010. 2685
Ogawa S, Tank DW, Menon R, Ellermann JM, Kim SG, Merkle H, Ugurbil K. 2686
Intrinsic signal changes accompanying sensory stimulation: Functional brain mapping 2687
with magnetic resonance imaging. Proc Natl Acad Sci USA 89: 5951-5955, 1992. 2688
Ojemann JG, Akbudak E, Snyder AZ, McKinstry RC, Raichle ME, Conturo TE. 2689
Anatomic localization and quantitative analysis of gradient refocused echo-planar 2690
fMRI susceptibility artifacts. NeuroImage 6: 156-167, 1997. 2691
Orban GA, Van Essen D, Vanduffel W. Comparative mapping of higher visual areas in 2692
monkeys and humans. Trends in Cogn Sci 8: 315-324, 2004. 2693
Orban GA, Claeys K, Nelissen K, Smans R, Sunaert S, Todd JT, Wardak C, 2694
Durand J-B, Vanduffel W. Mapping the parietal cortex of human and non-human 2695
primates. Neuropsychologia 44: 2647-2667, 2006. 2696
Pandya DN, Kuypers HG. Cortico-cortical connections in the rhesus monkey. Brain Res 2697
13: 13-36, 1969. 2698
Pandya DN, Seltzer B. Intrinsic connections and architectonics of posterior parietal 2699
cortex in the rhesus monkey. J Comp Neurol 204: 196-210, 1982. 2700
Passingham RE, Stephan KE, Ktter R. The anatomical basis of functional localization 2701
in the cortex. Nat Rev Neurosci 3: 606-616, 2002. 2702
Perry RJ, Zeki S. The neurology of saccades and covert shifts in spatial attention: An 2703
event-related fMRI study. Brain 123: 2273-2288, 2000. 2704
Petrides M. Lateral prefrontal cortex: Architectonic and functional organization. Philos 2705
Trans R Soc Lond B Biol Sci 360: 781-795, 2005. 2706
Petrides M, Pandya DN. Efferent association pathways originating in the caudal 2707
prefrontal cortex in the macaque monkey. J Comp Neurol 498: 227-251, 2006. 2708

96
Pilbeam D, Young N. Hominoid evolution: Synthesizing disparate data. Comptes 2709
Rendus Palevol 3:305-321, 2004. 2710
Preuss T. What is it like to be a human? The Cognitive Neurosciences, Third Edition, 2711
MIT Press 5-22, 2004. 2712
Polimeni JR, Fischl B, Greve DN, Wald LL. Laminar analysis of 7T BOLD using an 2713
imposed spatial activation pattern in human V1. Neuroimage 52: 1334-1346, 2010. 2714
Polimeni JR, Hinds OP, Balasubramanian M, van der Kouwe AJW, Wald L,L Dale 2715
AM, Fischl B, Schwartz EL. The human V1V2V3 visuotopic map complex 2716
measured via fMRI at 3 and 7 Tesla. Abstr Soc Neurosci, 2005. 2717
Pons TP, Kaas JH. Corticocortical connections of area 2 of somatosensory cortex in 2718
macaque monkeys: A correlative anatomical and electrophysiological study. J Comp 2719
Neurol 248: 313-335, 1986. 2720
Posner MI, Petersen SE, Fox PT, Raichle ME. Localization of cognitive operations in 2721
the human brain. Science 240: 1627-1631, 1988. 2722
Power JD, Fair DA, Schlaggar BL, Petersen SE. The development of human 2723
functional brain networks. Neuron 67: 735-748, 2010. 2724
Raichle ME. 1987 Circulatory and metabolic correlates of brain function in normal 2725
humans. (In) Handbook of Physiology. The Nervous System V (Mountcastle VB, 2726
Plum F, Geiger SR eds) Bethesda MD, American Physiological Society (pp) 643 2727
674. 2728
Rajkowska G, Goldman-Rakic PS. Cytoarchitectonic definition of prefrontal areas in 2729
the normal human cortex: II. Variability in locations of areas 9 and 46 and 2730
relationship to the Talairach Coordinate System. Cereb Cortex 5: 323-337, 1995. 2731
Rizzolatti G, Luppino G, Matelli M. The organization of the cortical motor system: 2732
New concepts. Electroencephalogr Clin Neurophysiol 106: 283-296, 1998. 2733
Rockland KS, Pandya DN. Laminar origins and terminations of cortical connections of 2734
the occipital lobe in the rhesus monkey. Brain Res 179: 3-20, 1979. 2735
Rosenbaum RS, Stuss DT, Levine B, Tulving E. Theory of mind is independent of 2736
episodic memory. Science 318: 1257, 2007. 2737

97
Rottschy C, Eickhoff SB, Schleicher A, Mohlberg H, Kujovic M, Zilles K, Amunts 2738
K. Ventral visual cortex in humans: Cytoarchitectonic mapping of two extrastriate 2739
areas. Hum Brain Mapp 28: 1045-1059, 2007. 2740
Rousseeuw PJ. Silhouettes: A graphical aid to the interpretation and validation of cluster 2741
analysis. J Comput Appl Math 20: 1987. 2742
Saleem KS, Kondo H, Price JL. Complementary circuits connecting the orbital and 2743
medial prefrontal networks with the temporal, insular, and opercular cortex in the 2744
macaque monkey. J Comp Neurol 506: 659-693, 2008. 2745
Saxe R. Uniquely human social cognition. Curr Opin Neurobiol 16: 235-239, 2006. 2746
Saxe R, Powell LJ. It's the thought that counts: Specific brain regions for one component 2747
of theory of mind. Psychol Sci 17: 692-699, 2006. 2748
Scheperjans F, Eickhoff SB, Hmke L, Mohlberg H, Hermann K, Amunts K, Zilles 2749
K. Probabilistic maps, morphometry, and variability of cytoarchitectonic areas in the 2750
human superior parietal cortex. Cereb Cortex 18: 2141-2157, 2008a. 2751
Scheperjans F, Hermann K, Eickhoff SB, Amunts K, Schleicher A, Zilles K. 2752
Observer-independent cytoarchitectonic mapping of the human superior parietal 2753
cortex. Cereb Cortex 18: 846-867, 2008b. 2754
Schleicher A, Amunts K, Geyer S, Morosan P, Zilles K. Observer-independent method 2755
for microstructural parcellation of cerebral cortex: A quantitative approach to 2756
cytoarchitectonics. Neuroimage 9: 165-177, 1999. 2757
Schormann T, Zilles K. Three-dimensional linear and nonlinear transformations: An 2758
integration of light microscopical and MRI data. Hum Brain Mapp 6: 339-347, 1998. 2759
Sgonne F, Dale AM, Busa E, Glessner M, Salat D, Hahn HK, Fischl B. A hybrid 2760
approach to the skull stripping problem in MRI. Neuroimage 22: 1060-1075, 2004. 2761
Sgonne F, Pacheco J, Fischl B. Geometrically accurate topology-correction of cortical 2762
surfaces using nonseparating loops. IEEE Trans Med Imaging 26: 518-529, 2007. 2763
Selemon LD, Goldman-Rakic PS. Common cortical and subcortical targets of the 2764
dorsolateral prefrontal and posterior parietal cortices in the rhesus monkey: Evidence 2765
for a distributed neural network subserving spatially guided behavior. J Neurosci 8: 2766
4049-4068, 1988. 2767

98
Seltzer B, Pandya DN. Parietal, temporal, and occipital projections to cortex of the 2768
superior temporal sulcus in the rhesus monkey: A retrograde tracer study. J Comp 2769
Neurol 343: 445-463, 1994. 2770
Seltzer B, Pandya DN. Further observations on parieto-temporal connections in the 2771
rhesus monkey. Exp Brain Res 55:301-312, 1984. 2772
Sepulcre J, Liu H, Talukdar T, Martincorena I, Yeo BTT, Buckner RL. The 2773
organization of local and distant functional connectivity in the human brain. PLoS 2774
Comput Biol 6: e1000808, 2010. 2775
Sereno MI, Dale AM, Reppas JB, Kwong KK, Belliveau JW, Brady TJ, Rosen BR, 2776
Tootell RBH. Borders of multiple visual areas in humans revealed by functional 2777
magnetic resonance imaging. Science 268: 889-893, 1995. 2778
Sereno MI, Pitzalis S, Martinez A. Mapping of contralateral space in retinotopic 2779
coordinates by a parietal cortical area in humans. Science 294: 1350-1354, 2001. 2780
Sestieri C, Corbetta M, Romani GL, Shulman GL. Episodic memory retrieval, parietal 2781
cortex, and the default mode network: Functional and topographic analyses. J 2782
Neurosci 31: 4407-4420, 2011. 2783
Shadlen MN, Newsome WT. Neural basis of a perceptual decision in the parietal cortex 2784
(area LIP) of the rhesus monkey. J Neurophysiol 86: 1916-1936, 2001. 2785
Shikata E, Hamzei F, Glauche V, Knab R, Dettmers C, Weiller C, Bchel C. Surface 2786
orientation discrimination activates caudal and anterior intraparietal sulcus in 2787
humans: an event-related fMRI study. J Neurophysiol 85: 1309-1314, 2001. 2788
Shikata E, Hamzei F, Glauche V, Koch M, Weiller C, Binkofski F, Bchel C. 2789
Functional properties and interaction of the anterior and posterior intraparietal areas 2790
in humans. Eur J Neurosci 17: 1105-1110, 2003. 2791
Shulman GL, McAvoy MP, Cowan MC, Astafiev SV, Tansy AP, d'Avossa G, 2792
Corbetta M. Quantitative analysis of attention and detection signals during visual 2793
search. J Neurophysiol 90: 3384-3397, 2003. 2794
Shulman GL, Ollinger JM, Akbudak E, Conturo TE, Snyder AZ, Petersen SE, 2795
Corbetta M. Areas involved in encoding and applying directional expectations to 2796
moving objects. J Neurosci 19: 9480-9496, 1999. 2797

99
Smith SM, Fox PT, Miller KL, Glahn DC, Fox PM, Mackay CE, Filippini N, 2798
Watkins KE, Toro R, Laird AR, Beckmann CF. Correspondence of the brain's 2799
functional architecture during activation and rest. Proc Natl Acad Sci USA 106: 2800
13040-13045, 2009. 2801
Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TEJ, Johansen- 2802
Berg H, Bannister PR, De Luca M, Drobnjak I, Flitney D, Niazy RK, Saunders 2803
J, Vickers J, Zhang Y, De Stefano N, Brady JM, Matthews PM. Advances in 2804
functional and structural MR image analysis and implementation as FSL. Neuroimage 2805
23: S208-S219, 2004. 2806
Stanton GB, Bruce CJ, Goldberg ME. Topography of projections to posterior cortical 2807
areas from the macaque frontal eye fields. J Comp Neurol 353: 291-305, 1995. 2808
Strick PL. How do the basal ganglia and cerebellum gain access to the cortical motor 2809
areas? Behav Brain Res 18:107-123, 1985. 2810
Suzuki WA, Amaral DG. Perirhinal and parahippocampal cortices of the macaque 2811
monkey: Cortical afferents. J Comp Neurol 350: 497-533, 1994. 2812
Swisher JD, Halko MA, Merabet LB, McMains SA, Somers DC. Visual topography 2813
of human intraparietal sulcus. J Neurosci 27: 5326-5337, 2007. 2814
Taira M, Nose I, Inoue K, Tsutsui K. Cortical areas related to attention to 3D surface 2815
structures based on shading: an fMRI study. Neuroimage 14: 959-966, 2001. 2816
Talairach J, Tournoux P. Co-planar Stereotaxic Atlas of the Human Brain: 3- 2817
Dimensional Proportional System: An Approach to Cerebral Imaging. New York: 2818
Thieme Medical Publishers 1988. 2819
Tann-Garipy J, Rouiller EM, Boussaoud D. Parietal inputs to dorsal versus ventral 2820
premotor areas in the macaque monkey: Evidence for largely segregated visuomotor 2821
pathways. Exp Brain Res 145: 91-103, 2002. 2822
Thirion B, Pinel P, Meriaux S, Roche A, Dehaene S, Poline J-B. Analysis of a large 2823
fMRI cohort: Statistical and methodological issues for group analyses. Neuroimage 2824
35: 105-120, 2007. 2825
Tootell RBH, Hadjikhani N. Where is 'dorsal V4' in human visual cortex? Retinotopic, 2826
topographic and functional evidence. Cereb Cortex 11: 298-311, 2001. 2827

100
Tootell RBH, Hadjikhani N, Hall EK, Marrett S, Vanduffel W, Vaughan JT, Dale 2828
AM. The retinotopy of visual spatial attention. Neuron 21: 1409-1422, 1998. 2829
Tootell RBH, Reppas JB, Kwong KK, Malach R, Born RT, Brady TJ, Rosen BR, 2830
Belliveau JW. Functional analysis of human MT and related visual cortical areas 2831
using magnetic resonance imaging. J Neurosci 15: 3215-3230, 1995. 2832
Tootell RBH, Taylor JB. Anatomical evidence for MT and additional cortical visual 2833
areas in humans. Cereb Cortex 5: 39-55, 1995. 2834
Ungerleider LG, Desimone R. Cortical connections of visual area MT in the macaque. J 2835
Comp Neurol 248: 190-222, 1986. 2836
van der Kouwe AJW, Benner T, Fischl B, Schmitt F, Salat DH, Harder M, Sorensen 2837
AG, Dale AM. On-line automatic slice positioning for brain MR imaging. 2838
Neuroimage 27: 222-230, 2005. 2839
van der Kouwe AJW, Benner T, Salat DH, Fischl B. Brain morphometry with 2840
multiecho MPRAGE. Neuroimage 40: 559-569, 2008. 2841
Van Dijk KRA, Hedden T, Venkataraman A, Evans KC, Lazar SW, Buckner RL. 2842
Intrinsic functional connectivity as a tool for human connectomics: Theory, 2843
properties, and optimization. J Neurophysiol 103: 297-321, 2010. 2844
Van Essen DC. Organization of visual areas in macaque and human cerebral cortex. 2845
Visual Neurosci 1: 507-521, 2004. 2846
Van Essen DC. A population-average, landmark-and surface-based (PALS) atlas of 2847
human cerebral cortex. Neuroimage 28: 635-662, 2005. 2848
Van Essen DC, Anderson CH, Felleman DJ. Information processing in the primate 2849
visual system: An integrated systems perspective. Science 255: 419-423, 1992. 2850
Van Essen DC, Dierker DL. Surface-based and probabilistic atlases of primate cerebral 2851
cortex. Neuron 56: 209-225, 2007. 2852
Van Essen DC, Newsome WT, Bixby JL. The pattern of interhemispheric connections 2853
and its relationship to extrastriate visual areas in the macaque monkey. J Neurosci 2: 2854
265-283, 1982 2855
Van Essen DC, Newsome WT, Maunsell JHR. The visual field representation in striate 2856
cortex of the macaque monkey: Asymmetries, anisotropies, and individual variability. 2857
Vision Res 24: 429-448, 1984. 2858

101
Van Essen DC, Zeki SM. The topographic organization of rhesus monkey prestriate 2859
cortex. J Physiol 277: 193-226, 1978. 2860
Vilberg KL, Rugg MD. Memory retrieval and the parietal cortex: A review of evidence 2861
from a dual-process perspective. Neuropsychologia 46:1787-1799, 2008. 2862
Vincent JL, Patel GH, Fox MD, Snyder AZ, Baker JT, Van Essen DC, Zempel JM, 2863
Snyder LH, Corbetta M, Raichle ME. Intrinsic functional architecture in the 2864
anaesthetized monkey brain. Nature 447:83-86, 2007. 2865
Vincent JL, Kahn I, Snyder AZ, Raichle ME, Buckner RL. Evidence for a 2866
frontoparietal control system revealed by intrinsic functional connectivity. J 2867
Neurophysiol 100: 3328-3342, 2008. 2868
Vincent JL, Snyder AZ, Fox MD, Shannon BJ, Andrews JR, Raichle ME, Buckner 2869
RL. Coherent spontaneous activity identifies a hippocampal-parietal memory 2870
network. J Neurophysiol 96: 3517-3531, 2006. 2871
Van Overwalle F, Baetens K. Understanding others' actions and goals by mirror and 2872
mentalizing systems: A meta-analysis. Neuroimage 48: 564-584, 2009. 2873
von Bonin G, Bailey P. The Neocortex of Macaca Mulatta. Urbana: University of 2874
Illinois Press 1947. 2875
Wagner AD, Shannon BJ, Kahn I, Buckner RL. Parietal lobe contributions to episodic 2876
memory retrieval. Trends Cogn Sci 9: 445-453, 2005. 2877
Wandell BA, Brewer AA, Dougherty RF. Visual field map clusters in human cortex. 2878
Philos Trans R Soc Lond B Biol Sci 360: 693-707, 2005. 2879
Wandell BA, Dumoulin SO, Brewer AA. Visual field maps in human cortex. Neuron 2880
56: 366-383, 2007. 2881
Wessinger CM, VanMeter J, Tian B, Van Lare J, Pekar J, Rauschecker JP. 2882
Hierarchical organization of the human auditory cortex revealed by functional 2883
magnetic resonance imaging. J Cogn Neurosci 13: 1-7, 2001. 2884
Wilms M, Eickhoff SB, Hmke L, Rottschy C, Kujovic M, Amunts K, Fink GR. 2885
Comparison of functional and cytoarchitectonic maps of human visual areas V1, V2, 2886
V3d, V3v, and V4(v). Neuroimage 49: 1171-1179, 2010. 2887

102
Yeo BTT, Sabuncu MR, Vercauteren T, Ayache N, Fischl B, Golland P. Spherical 2888
demons: Fast diffeomorphic landmark-free surface registration. IEEE Trans Med 2889
Imaging 29: 650-668, 2010a. 2890
Yeo BTT, Sabuncu MR, Vercauteren T, Holt DJ, Amunts K, Zilles K, Golland P, 2891
Fischl B. Learning task-optimal registration cost functions for localizing 2892
cytoarchitecture and function in the cerebral cortex. IEEE Trans Med Imaging 29: 2893
1424-1441, 2010b. 2894
Zeki SM. Representation of central visual fields in prestriate cortex of monkey. Brain 2895
Res 14: 271-291, 1969. 2896
Zilles K, Schlaug G, Matelli M, Luppino G, Schleicher A, Q M, Dabringhaus A, 2897
Seitz R, Roland PE. Mapping of human and macaque sensorimotor areas by 2898
integrating architectonic, transmitter receptor, MRI and PET data. J Anat 187: 515- 2899
537, 1995. 2900
A
B C D
E
Subject 1
Subject 2
Figure 1
T2*
T1
T2*
T1
R
T2*
T1
S
u
b
j
e
c
t

3

S
u
b
j
e
c
t

1

S
u
b
j
e
c
t

2

Figure 2
D
V
D
V
A
P A
P
Dorsal
Lateral
Medial
Ventral
P A
A P
Parietal Frontal
0 100
Figure 3
Figure 4
A
B
M = 3
M = 5 M = 4
M = 2
Figure 5
Figure 6

Number of Networks
I
n
s
t
a
b
i
l
i
t
y

3 5 7 9 11 13 15 17 19 21 23 25
0
0.05
0.10
0.15
0.20
Discovery Sample (n = 500)
Replication Sample (n = 500)
Figure 7
0 0.80
Figure 8
Discovery Sample (n = 500)
Replication Sample (n = 500)
Figure 9
0 0.70
Figure 10
Left
Right
7-Network Parcellation (N=1000)
Figure 11
Figure 12
Purple (Visual)
Blue (Somatomotor)
Green (Dorsal Attention)
Violet (Ventral Attention)
Cream (Limbic)
Orange (Frontoparietal)
Red (Default)
17-Network Parcellation (N=1000)
Figure 13
Left
Right
13 -23

-10 9

-19 58

Figure 14
6 24
Figure 15
V1
p
V3
pv
ExC

ExP

V3
cv
V1
c
0.25 0.60
V3
cv
V1
c
ExC

ExP

V1
p
V3
pv
Figure 16
B

ECR
EOR
FIX

V1
c
ExC

V3
cv
ExP

V3
pv

V1
p
ExC

V3
cv
ExP

V3
pv

V1
c
V1
p
A

V1

ExC

V3
cv
ExP

V3
pv
Figure 17
0.62 -0.03
V1
c
V1
p
V3
cv
V3
pv
Figure 18
Area 6 Area 2 Area 5L
0 1
Figure 19
A
0
0.1
0.2
0.3
0.4
0
0.1
0.2
0.3
0.4
B
Hand-Foot Hand-Tng Foot-Tng
M
e
a
n

C
o
r
r
e
l
a
t
i
o
n

z
(
r
)

Figure 20
Tongue Hand Foot
0.4
0.6
0.8
1
1 2 3 4 S 6 7 8 9 10 11 12 13 14 1S 16 17 18 19 20 21 22 23 24 2S 26 27 28 29 30 31 32 33 34 3S 36 37 38 39 40 41 42
0
0.2
0.4
0.6
0.8
1
A
B
Tongue Hand Foot
0
0.2
0.4
0.6
0.8
Tng-Tng Hand-Hand Foot-Foot
M
e
a
n

C
o
r
r
e
l
a
t
i
o
n

z
(
r
)

1.0
0.6
0.8
0.4
1.0
M
e
a
n

C
o
r
r
e
l
a
t
i
o
n

z
(
r
)

Figure 21
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
aMT+ MT+
v
MT+
d
aMT+
MT+
d
MT+
v
V1
cd
V1
pd
V1
cv
V1
pv
A

M
e
a
n

C
o
r
r
e
l
a
t
i
o
n

z
(
r
)

0
0.02
0.04
0.06
0.08
0.10
0.12
0.14
Figure 22
aMT+ MT+
d
0.05 0.30
MT+
v
aMT+
MT+
v
MT+
d
Figure 23
0.15 0.50
aMT+ MT+

Figure 24
aMT+ aMT+
MT+

V3A

SPL7P

SPL7A

IPS3
m
IPS2

FEF

PrC
v
MT+
aMT+

V1
cd
V4

V1
pd
PrC
v
FEF

IPS2

IPS3
m
SPL7A

SPL7P

V3A

ECR
EOR
FIX
MT+

aMT+

V1
cdV4
V1
pd
V1
cd
V4

V1
pd
PrC
v
FEF

IPS2

IPS3
m
SPL7A

SPL7P

V3A

V1
cd
V4

V1
pd
PrC
v
FEF

IPS2

IPS3
m
SPL7A

SPL7P

V3A

Figure 25
FEF

0.15 0.50
FEF

PrC
v
PrC
v
Figure 26
IPS3
m
B

0
0.1
0.2
0.3
0.4
0.5
0.6
0
0.1
0.2
0.3
0.4
0.5
0.6
FEF

PrC
v
IPS3
m
IPS1

SPL7P

M
e
a
n

C
o
r
r
e
l
a
t
i
o
n

z
(
r
)

FEF

PrC
v
IPS2

IPS2

IPS1

SPL7A

SPL7A SPL7P
Figure 27
a
b
d c
e
a
b
d c
e
A B
i i
ii ii
i
ii ii
i
Figure 28
FEF

SPL7A

V3A

aMT+

MT+ V1
p
A
C D B E
7.0% 9.8% 37.3% 75.4%
FEF
d
aMT+
SPL7A
MT+
V1
pd
V3A
FEF
d
aMT+
SPL7A
MT+
V1
pd
V3A
FEF
d
aMT+
SPL7A
MT+
V1
pd
V3A
FEF
d
aMT+ SPL7A
MT+
V1
pd
V3A
FEF aMT+
SPL7A
MT+
V3A
V1
pd

FEF
aMT+ SPL7A
MT+
V3A
V1
pd

SPL7A
FEF aMT+
V3A
MT+
V1
pd

SPL7A V3A
FEF
V1
pd

MT+
aMT+
Figure 29

PFC
da
PrC
v
FEF

PFC
l
PFC
dp
PFC
la
STS

STS
p
aMT+

PFC
m

PFC
v
6vr+

PHC
pCun
Cing
a
PFC
mp
PFC
lp
ITG

MT+

RSP

PCC
PFC
dm

PFC
da
PrC
v
FEF

PFC
l
PFC
dp
PFC
la
STS

STS
p
aMT+

PFC
m

PFC
v
6vr+

PHC
pCun
Cing
a
PFC
mp
PFC
lp
ITG

MT+

RSP

PCC
PFC
dm

PFC
da
PrC
v
FEF

PFC
l
PFC
dp
PFC
la
STS

STS
p
aMT+

PFC
m

PFC
v
6vr+

PHC
pCun
Cing
a
PFC
mp
PFC
lp
ITG

MT+

RSP

PCC
PFC
dm

PFC
da
PrC
v
FEF

PFC
l
PFC
dp
PFC
la
STS

STS
p
aMT+

PFC
m

PFC
v
6vr+

PHC
pCun
Cing
a
PFC
mp
PFC
lp
ITG

MT+

RSP

PCC
PFC
dm
PGa IPS2 SPL7A
IPS3
l
PGp
d
PGp
v
TPJ

PF

STS

STS
p
ITG
aMT+

MT+
PFC
la
PFC
v
6vr+

PFC
l
PFC
da
PFC
lp
PFC
dp
PrC
v
FEF

PHC
PFC
m
PFC
dm
PFC
mp
Cing
a
PCC

pCun

RSP

PFC
da
PrC
v
FEF

PFC
l
PFC
dp
PFC
la
STS

STS
p
aMT+

PFC
m

PFC
v
6vr+

PHC
pCun
Cing
a
PFC
mp
PFC
lp
ITG

MT+

RSP

PCC
PFC
dm

PFC
da
PrC
v
FEF

PFC
l
PFC
dp
PFC
la
STS

STS
p
aMT+

PFC
m

PFC
v
6vr+

PHC
pCun
Cing
a
PFC
mp
PFC
lp
ITG

MT+

RSP

PCC
PFC
dm

PFC
da
PrC
v
FEF

PFC
l
PFC
dp
PFC
la
STS

STS
p
aMT+

PFC
m

PFC
v
6vr+

PHC
pCun
Cing
a
PFC
mp
PFC
lp
ITG

MT+

RSP

PCC
PFC
dm

PFC
da
PrC
v
FEF

PFC
l
PFC
dp
PFC
la
STS

STS
p
aMT+

PFC
m

PFC
v
6vr+

PHC
pCun
Cing
a
PFC
mp
PFC
lp
ITG

MT+

RSP

PCC
PFC
dm
IPS2
SPL7A
IPS3
l
PGp
d
PGp
v
TPJ
PF
PGa
Figure 30

STS

STS
p
aMT+

PHC
pCun
Cing
a
PFC
mp
ITG

MT+

RSP

PCC
STS
a
PGa
IPS3
l
Cing
m
PGp
v
PGp
d
TPJ

IPS2

SPL7A

5Ci

PF

STS

STS
p
aMT+

PHC
pCun
Cing
a
PFC
mp
ITG

MT+

RSP

PCC
STS
a
PGa
IPS3
l
Cing
m
PGp
v
PGp
d
TPJ

IPS2

SPL7A

5Ci

PF

STS

STS
p
aMT+

PHC
pCun
Cing
a
PFC
mp
ITG

MT+

RSP

PCC
STS
a
PGa
IPS3
l
Cing
m
PGp
v
PGp
d
TPJ

IPS2

SPL7A

5Ci

PF

STS

STS
p
aMT+

PHC
pCun
Cing
a
PFC
mp
ITG

MT+

RSP

PCC
STS
a
PGa
IPS3
l
Cing
m
PGp
v
PGp
d
TPJ

IPS2

SPL7A

5Ci

PF

STS

STS
p
aMT+

PHC
pCun
Cing
a
PFC
mp
ITG

MT+

RSP

PCC
STS
a
PGa
IPS3
l
Cing
m
PGp
v
PGp
d
TPJ

IPS2

SPL7A

5Ci

PF

PFC
da
PFC
da
PFC
m
PFC
dm
PFC
la
PFC
l
PFC
dp
PrC
v
FEF

STS

STS
p
ITG
aMT+

MT+
PHC
PFC
mp
Cing
a
PCC

pCun

RSP

5Ci
PF
IPS2PGa
IPS3
l
SPL7A
TPJ
PGp
d
PGp
v
STS
a
Cing
m

STS

STS
p
aMT+

PHC
pCun
Cing
a
PFC
mp
ITG

MT+

RSP

PCC
STS
a
PGa
IPS3
l
Cing
m
PGp
v
PGp
d
TPJ

IPS2

SPL7A

5Ci

PF

PFC
dp
STS

STS
p
aMT+

PHC
pCun
Cing
a
PFC
mp
ITG

MT+

RSP

PCC
STS
a
PGa
IPS3
l
Cing
m
PGp
v
PGp
d
TPJ

IPS2

SPL7A

5Ci

PF

PrC
v

STS

STS
p
aMT+

PHC
pCun
Cing
a
PFC
mp
ITG

MT+

RSP

PCC
STS
a
PGa
IPS3
l
Cing
m
PGp
v
PGp
d
TPJ

IPS2

SPL7A

5Ci

PF

FEF

PFC
dm
PFC
m
PFC
la
PFC
l
Figure 31
F
C
E
B
A
A B
F C
D E
D
0.15 0.60
Figure 32
A
D
B
E
F
C
B
F C
D E
A
0.15 0.60
Figure 33
C
A
D
F
E
C
B
B
D
0.15 0.60
A
E
F
Figure 34
B
D E
A
0.15 0.60
A
D
F
E
C
B
C F
Figure 35