Spontaneously emerging patterns in human visual cortex and their functional connectivity are linked to the patterns evoked by visual stimuli

DoHyun Kim; Tomer Livne; Nicholas V Metcalf; Maurizio Corbetta; Gordon L Shulman

doi:10.1152/jn.00630.2019

. 2020 Sep 23;124(5):1343–1363. doi: 10.1152/jn.00630.2019

Spontaneously emerging patterns in human visual cortex and their functional connectivity are linked to the patterns evoked by visual stimuli

DoHyun Kim ¹, Tomer Livne ^2,³, Nicholas V Metcalf ³, Maurizio Corbetta ^3,^4,^5,^6,^7,⁸, Gordon L Shulman ^3,^4,^✉

PMCID: PMC8356777 PMID: 32965156

Abstract

The function of spontaneous brain activity is an important issue in neuroscience. Here we test the hypothesis that patterns of spontaneous activity code representational patterns evoked by stimuli. We compared in human visual cortex multivertex patterns of spontaneous activity to patterns evoked by ecological visual stimuli (faces, bodies, scenes) and low-level visual features (e.g., phase-scrambled faces). Specifically, we identified regions that preferred particular stimulus categories during localizer scans (e.g., extrastriate body area for bodies), measured multivertex patterns for each category during event-related task scans, and then correlated over vertices these stimulus-evoked patterns to the pattern measured on each frame of resting-state scans. The mean correlation coefficient was essentially zero for all regions/stimulus categories, indicating that resting multivertex patterns were not biased toward particular stimulus-evoked patterns. However, the spread of correlation coefficients between stimulus-evoked and resting patterns, positive and negative, was significantly greater for the preferred stimulus category of an ROI. The relationship between spontaneous and stimulus-evoked multivertex patterns also governed the temporal correlation or functional connectivity of patterns of spontaneous activity between individual regions (pattern-based functional connectivity). Resting multivertex patterns related to an object category fluctuated preferentially between ROIs preferring the same category, and fluctuations of the pattern for a category (e.g., body) within its preferred ROIs were largely uncorrelated with fluctuations of the pattern for a disparate category (e.g., scene) within its preferred ROIs. These results support the proposal that spontaneous multivertex activity patterns are linked to stimulus-evoked patterns, consistent with a representational function for spontaneous activity.

NEW & NOTEWORTHY Spontaneous brain activity was once thought to reflect only noise, but evidence of strong spatiotemporal regularities has motivated a search for functional explanations. Here we show that the spatial pattern of spontaneous activity in human high-level and early visual cortex is related to the spatial patterns evoked by stimuli. Moreover, these patterns partly govern spontaneous spatiotemporal interactions between regions, so-called functional connectivity. These results support the hypothesis that spontaneous activity serves a representational function.

Keywords: BOLD fMRI, functional connectivity, multivariate pattern analysis (MVPA), resting state, spontaneous activity

INTRODUCTION

Spontaneous neural activity is observed throughout the brain, yet its function remains mysterious. An important clue, however, comes from work that has uncovered striking similarities between spontaneous activity and the activity evoked by a task (Arcaro et al. 2015; Berkes et al. 2011; Cole et al. 2016; Fiser et al. 2004; Fox et al. 2007; Heinzle et al. 2011; Kenet et al. 2003; Omer et al. 2019; Raemaekers et al. 2014; Ryu and Lee 2018; Tavor et al. 2016; Tsodyks et al. 1999). For example, the temporal correlation of spontaneous activity between brain regions [functional connectivity (FC)] closely resembles the spatial topography of task-evoked activity (Biswal et al. 1995; de Pasquale et al. 2010; Greicius et al. 2003; He et al. 2008; Power et al. 2011; Yeo et al. 2011), links distributed brain regions into functional networks, and can be used to predict task activation (Cole et al. 2016; Osher et al. 2019; Tavor et al. 2016).

The remarkable spatiotemporal regularities of spontaneous activity and widespread findings that abnormalities in interregional correlations of spontaneous activity in humans are associated with neurological and psychiatric disorders (e.g., Grefkes and Fink 2014; Northoff and Duncan 2016) have motivated a search for functional explanations. One hypothesis is that spontaneous activity has a role in the synaptic homeostasis of structural connections (Deco et al. 2013). Another idea is that fluctuations of spontaneous activity between regions constitute a spatiotemporal prior that facilitates the recruitment of task circuitries during behavior (Petersen and Sporns 2015; Raichle 2011). A third hypothesis is that spontaneous activity has a role in representing information about external (Fiser et al. 2010) and internal states (Harmelech and Malach 2013). Genetically determined circuitries generate spontaneous activity that is shaped in the course of development by experience through Hebbian statistical learning (Berkes et al. 2011). Conversely, spatial and temporal patterns of spontaneous activity constrain task-evoked patterns. As a result of this cyclic process, both spontaneous and task-evoked activity code similar representations of internal and external states (Fiser et al. 2010). The same process determines the spontaneous interactions between regions, which reflect connectivity patterns that are coded as synaptic efficacies in cortical networks (Harmelech and Malach 2013; Strappini et al. 2019). Figure 1 illustrates schematically the representation hypothesis of spontaneous brain activity.

Fig. 1. — The putative cyclic interplay between brain activity evoked by real-world experiences and resting-state activity.

The representation hypothesis has been supported by human and animal work. In animals, imaging of neural activity at a scale that extends across many cortical columns has shown that the macroscale spatial pattern of spontaneous neural activity within a sensory area in an anesthetized animal mirrors the pattern of activity evoked by stimulation of a specific visual feature (Kenet et al. 2003; Omer et al. 2019). In humans, functional magnetic resonance imaging (fMRI) studies of early visual cortex have shown that FC between individual voxels respects the stimulus-evoked selectivity of voxels for polar angle, eccentricity, and low-level stimulus features (Arcaro et al. 2015; Heinzle et al. 2011; Raemaekers et al. 2014; Ryu and Lee 2018). Recent work has also shown that voxelwise resting FC in visual cortex is better approximated by the FC evoked by movies than by more artificial stimuli such as rotating checkerboards or static pictures of stimuli (Strappini et al. 2019; Wilf et al. 2017). Finally, Chen and colleagues (2017) measured the voxelwise resting FC between regions of ventral visual cortex and a second region that was functionally related to a specific visual category (e.g., tools). The multivoxel pattern of voxelwise FC values in a ventral visual region was similar to the multivoxel pattern of activity evoked in that region by exemplars of the category.

Here we further test the representation hypothesis in human visual cortex by determining whether multivertex patterns of resting activity resemble stimulus-evoked patterns of activity. This analysis bears some resemblance to a representational similarity analysis, which is a standard methodology for identifying the representational content of evoked activity in a brain region (Kriegeskorte et al. 2008a). Unlike Chen et al. (2017), we examine the multivertex pattern of resting activity at each timepoint rather than separately analyzing each individual vertex over timepoints in a resting FC analysis.

Specifically, we measured the similarity of resting and stimulus-evoked multivertex patterns in different category specific regions of human visual cortex [e.g., extrastriate body area (EBA), fusiform face area (FFA), and parahippocampal place area (PPA)], for different stimulus categories (e.g., bodies, faces, places). We also examined the similarity of resting and stimulus-evoked patterns in early visual cortex and in category specific regions of visual cortex for the low-level features comprising phase-scrambled and grid-scrambled stimuli. Finally, we examined whether, in the resting state, fluctuations over time in the amplitude of stimulus-evoked patterns were correlated across regions (hereafter termed “pattern-based functional connectivity”). For example, do fluctuations in the amplitude of a resting multivertex pattern of activity in a region that is related to a particular category (e.g., bodies) correlate over time with fluctuations in the amplitude of the multivertex pattern of activity for the same category in a different region? Because this analysis measures the correlated fluctuations of spontaneous patterns of activity across regions, it differs from a standard functional connectivity (FC) analysis, in which the activity of a single voxel is temporally correlated with the activity of other voxels, or the activity averaged across the voxels of a region is temporally correlated with the activity of other voxels or averaged regions. Similar multivertex, pattern-based FC analyses have previously been reported, but only for task-evoked activity, not for resting activity (Anzellotti and Coutanche 2018; Chen et al. 2018; Coutanche and Thompson-Schill 2013).

The representation hypothesis of spontaneous activity makes several specific predictions about the expected multivertex pattern of activity within a region and the interaction of that pattern with the patterns from other regions. First, if spontaneous activity carries information about stimulus categories such as faces, bodies, or scenes, then spontaneous multivertex activity patterns, i.e., patterns of activity observed at rest in the absence of any stimulation, in functionally specialized occipital regions such as the EBA (Downing et al. 2001) should be more related to the activity patterns evoked by a preferred stimulus category (e.g., bodies) than by nonpreferred categories. This predicted relationship between spontaneous and task-evoked multivertex patterns putatively reflects the entrainment of task-evoked patterns into spontaneous activity during development and through experience. Second, if the FC between regions reflects in part correlated fluctuations of the spontaneous representational content of those regions, then the magnitude of resting pattern-based FC between two regions should depend on whether the regions prefer the same visual stimulus category and on whether the tested patterns that are being correlated in the two regions code that preferred category.

MATERIALS AND METHODS

Participants

The study included 16 healthy young adult volunteers (10 female; age 21–35 yr old) with no prior history of neurological or psychiatric disorders. All participants were right-handed native English speakers with normal or corrected-to-normal vision. All participants gave written informed consent to take part in the experiment, and the study was approved by the Institutional Review Board (IRB) of Washington University in St. Louis School of Medicine.

Stimuli

Nine categories of color images subtending 8° × 8° of visual angle were included in event-related “task” fMRI scans. Seven “whole object” categories consisted of images that are encountered in real-world environments: human faces, human bodies, mammals, chairs, tools, scenes, and words. Stimuli, excluding the word category, were obtained from Downing et al. (2006). faces, bodies, and mammals served as animate categories and chairs, tools, and scenes as inanimate categories (Kriegeskorte et al. 2008a). Word stimuli were included for exploratory analyses and results for those stimuli will not be considered in this paper.

Two control stimulus categories were constructed from the above stimuli excluding the word stimuli. A low-level control consisted of phase-scrambled stimuli that preserved the spatial frequency amplitude spectrum of the whole objects images. An intermediate-level control consisted of grid-scrambled stimuli that included basic visual properties of the whole objects images such as line segments and connectors. For the low-level control condition, 2D phase-scrambled images of the exemplars from the six categories were generated by applying the same set of random phases to each two-dimensional frequency component of the original image while keeping the magnitude constant (Watson et al. 2016). Exemplars from all six whole-object categories except real word stimuli were 2D phase scrambled, yielding a total of 144 2D phase-scrambled stimuli. For the intermediate-level control condition, grid-scrambled images of exemplars from the six whole-object categories were generated by subdividing each image into a 10 × 10 grid (each grid is 0.8° × 0.8°) and randomly rearranging the individual grid segments.

Color images of exemplars from seven categories were included in blocked-design localizer scans: human faces, human bodies, objects (chairs and tools), scenes, words, false font character strings, and phase-scrambled images. The categories for the localizer scans differed slightly from the categories for the task scans since the former was only used to define the regions of interest (ROIs). ROIs related to the false font and word stimuli will not be considered in this paper.

Stimuli were presented using the Psychophysics Toolbox package (Brainard 1997) in MATLAB (The MathWorks). Stimulus images were projected onto a screen and were viewed through a mirror mounted on the head coil. All stimuli were presented centrally on a gray background.

Scanning Procedure

The study consisted of two sessions [mean separation of 16.6 (16.8 SD) days]. In session 1, subjects received three resting-state runs, two localizer runs, and eight task runs, with the procedure for each run schematically shown in Fig. 2. In session 2, subjects received two resting-state runs, two localizer runs, eight task runs, and two posttask resting-state runs. One subject had a total of 13 task runs over the two sessions instead of 16.

Fig. 2. — Experimental design, with separate resting scans, blocked-design localizer scans, and event-related task scans. The photographs shown in this figure are different than the photographs used in the experiments, which were obtained from Downing et al. (2006). TR, repetition time.

Resting-state runs.

Participants received a total of seven resting-state scans, each lasting 5 min [300 repetition times (TRs)]. During a scan the participant was asked to maintain fixation on a cross that was displayed at the center of the screen during the entire run. Five resting scans (3 for the first session and 2 for the second session) were conducted before any localizer or task scans to collect stimulus-free intrinsic activities. For the second session only, two additional 5-min resting-state scans were conducted after the task scans to investigate potential post-stimuli-driven effects on intrinsic activity. The results from the posttask resting scans will not be discussed here.

Localizer runs.

Each session included two localizer runs (4 in total), each lasting 5 min and 40 s (340 TRs), and each localizer scan was presented in a blocked fMRI design. Each block of a localizer run contained 20 images of a single category, and those images were different from the images used in the task scans. A fully randomized sequence of eight blocks, consisting of the seven stimulus categories and a fixation block, was repeated twice within each run. At the beginning and the end of each run, an additional fixation block was presented for 4 s and 16 s. Within each category block, images were presented for 300 ms with an interstimulus interval (ISI) of 700 ms. A fixation cross was continuously present at the center of the screen during the ISI and during fixation blocks. During category blocks, participants performed a minimally cognitively engaging task by pressing a button if the initially presented image was changed in size or position during the 300-ms presentation.

Task runs.

Each session included 8 task runs (16 in total), each lasting 5 min and 15 s (315 TRs). For each subject and for each run, stimulus presentation order and interstimulus intervals were fully randomized using Optseq2 (Dale 1999). Each stimulus presentation lasted for 300 ms and the interval between stimuli was jittered between 3.7 s and 8.7 s. A fixation cross was continuously present at the center of the screen during the ISI. In each whole-object category, there were 24 separate exemplars (e.g., 24 different faces) and each exemplar was repeated four times. In each scrambled category, there were 96 exemplars, each presented once. Participants performed a minimally cognitively engaging task by pressing a button if the presented image changed its size or position during a 300-ms presentation, the same task as that performed during the localizer scans. The stimulus exemplars for each category presented in task scans were all different from the stimulus exemplars presented in localizer scans.

Imaging Parameters and fMRI Preprocessing

Structural and fMRI images were obtained from a Siemens 3T Prisma MRI scanner (https://www.siemens-healthineers.com/magnetic-resonance-imaging/3t-mri-scanner/magnetom-prisma). Structural images for atlas transformation and lesion segmentation were acquired using T1-weighted magnetization prepared-rapid gradient echo (MP-RAGE) [1 × 1 × 1 mm voxels; echo time (TE) = 2.36 ms, repetition time (TR) = 1,700 ms, inversion time (TI) = 1,000 ms, flip angle = 8°] and T2-weighted fast spin echo sequences (1 × 1 × 1 mm voxels; TE = 564 ms, TR = 3,200 ms). FMRI scans were collected using a gradient echo-planar sequence sensitive to blood-oxygenation-level-dependent (BOLD) contrast (TE = 26.6 ms, flip angle = 58°, 2.4 × 2.4 × 2.4 mm voxels, 48 contiguous slices, TR = 1.0 s, and multiband factor of 4).

fMRI Preprocessing

fMRI data underwent preprocessing as previously described (Siegel et al. 2016). Each participant’s anatomical images were linearly registered using 12-parameter affine transformation to the Montreal Neurological Institute (MNI) MNI152 atlas brain using the Washington University Neuroimaging Laboratories 4dfp toolset (https://github.com/robbisg/4dfp_tools). First the T1 was registered to the target atlas brain. The T2 was then registered to the unregistered T1 image. To increase accuracy in registration to the target atlas brain, we used FNIRT from the FSL toolbox to perform a nonlinear registration from the linearly atlas registered T1 to the atlas target. All transforms and the nonlinear warp were combined into a single warpfield for each modality. The volume fMRI data underwent the following steps: slice timing correction using sinc interpolation, elimination of odd/even slice intensity differences due to interleaved acquisition, computation of mode 1,000 whole brain intensity normalization, removal of distortion using measured gradient echo field mapping, and spatial realignment within and across fMRI runs. For BOLD atlas registration, a BOLD reference image was created by computing the mean across time of the cross realigned BOLD images. The BOLD reference image was then linearly registered to the T2 anatomical image. A one-step resampling to 2.4-mm cubic voxels in atlas space including realignment was computed for each BOLD volume by multiplying the transformation matrices and using FSL to combine the linear transforms, distortion correction, and nonlinear warp (Griffis et al. 2020). The first four timepoints of each fMRI scan were eliminated.

Surface generation and processing of functional data followed procedures similar to Glasser et al. (2013). First, anatomical surfaces were generated for each subject’s T1 MRI using FreeSurfer automated segmentation (Fischl et al. 1999). This step included brain extraction, segmentation, generation of white matter and pial surface, inflation of the surfaces to a sphere, and surface shape-based spherical registration to the subjects’ “native” surface to the fs_average surface. Segmentations were manually checked for accuracy. The left and right hemispheres were then resampled to 164,000 vertices and registered to each other (Van Essen et al. 2001) for projection of functional data.

BOLD data were sampled to each subject’s individual surface (between white matter and pial surface) using a ribbon-constrained sampling available in Connectome Workbench (Marcus et al. 2013). Voxels with a high coefficient of variation (1 standard deviations above the mean coefficient of variation of all voxels in a 5-mm sigma Gaussian neighborhood) were excluded from volume to surface mapping (Glasser et al. 2013).

After the fMRI data were projected to the surface, it passed through several additional preprocessing steps to reduce noise. White matter and ventricle regressors were computed in volume using the average timeseries of the tissue types based on FreeSurfer segmentation (Fischl et al. 1999). Movement regressors were obtained by using the six parameters obtained by rigid body correction of head motion. The whole brain signal was obtained by averaging voxels in volume over the whole brain. All regressors were then regressed from the surface projected fMRI data. To minimize motion artifact, frames that exceeded a frame-to-frame displacement of 0.5 mm were eliminated (“scrubbing”) and the timepoint following each scrubbed frame was also eliminated (Power et al. 2014). Finally, temporal bandpass filtering, retaining frequencies in the 0.009- to 0.08-Hz band with linear interpolation over eliminated frames, was performed on the regressed surface fMRI timeseries. All brain surface visualizations were generated using Connectome Workbench (Marcus et al. 2013).

To account for magnitude variability between different task and resting-state runs, the BOLD timeseries for each vertex were z-normalized across time within the task and the resting-state runs. This z-normalization was not applied to the localizer scans. Also, it was not applied to the task scans for a separate analysis described below in which task-evoked activation magnitudes were determined (see Task Scans: Estimation of Multivertex Activation Patterns).

Defining ROIs from Localizer Activation Contrasts

The next step was to define ROIs for each subject that showed a preference for specific categories or features. For this purpose, subject-specific ROIs were defined from univariate vertexwise statistical contrasts on the localizer activation magnitudes for different categories. For example, face-selective areas were defined from the vertices for the contrast of faces minus objects, where objects consisted of chairs and tools. First, for each participant a general linear model (GLM) was applied to their functional localizer scans. The GLM consisted of separate regressors for each stimulus category (e.g., faces) using an assumed hemodynamic response function from the Statistical Parametric Mapping (SPM12), a baseline term, and a linear trend term. Condition contrasts were formed to identify vertices showing a preference for each category, using a scheme similar to that of Bracci and Op de Beeck (2016): body preference (body > objects, i.e., chairs and tools), face preference (face > objects), scene preference (scene > objects), whole objects preference [face + body + scene + object (chair+tool) > phase-scrambled], and phase-scrambled-objects preference [phase-scrambled > face + body + scene + object (chair+tool)].

A group random-effect statistical z-map for each contrast was then computed from the single-subject localizer GLMs (see Fig. 3A for the group z-statistic maps for body, face, and scene preferences). The z-values obtained were sorted in magnitude. From the highest z-values from the map, the group peak with the next highest z-value was generated until the z-value was ≤2.0. Group peaks had to be separated by at least 38.4 mm (9.6 mm × 4) in the sphere mesh to prevent a vertex being assigned to multiple ROIs in a subject. ROIs were then defined separately for each participant based on the individual’s univariate statistical maps (Oosterhof et al. 2012; Wurm et al. 2016). From each group peak defined above, the corresponding peak for an individual subject peak was defined as the vertex with the highest z-value within a sphere of 9.6-mm radius centered around the group peak in each subject’s sphere mesh. The single-subject ROI was formed from the vertices exceeding z = 2.0 in a sphere of 9.6-mm radius centered around the peak in the subject’s mesh. All ROIs used in the following analysis contained at least 175 vertices in at least 14 subjects. ROIs in individual subjects with less than 175 vertices were discarded.

To remove differences in BOLD magnitude across magnetic resonance frames, for each ROI a z-normalization was applied across the vertices of each frame of the resting and task scans. This within-frame z-normalization was not applied to the localizer scans. Also, it was not applied to the task scans for a separate analysis described below in which task-evoked activation magnitudes were determined (see below, Task Scans: Estimation of Multivertex Activation Patterns).

Two sets of ROIs were created for use in different analyses. The first set was created from the localizer-defined ROIs that preferred a specific category (face, body, or scene) as compared with the object category (chairs + tools) and was used to compare the similarity of stimulus-evoked and resting multivertex patterns in high-level, category-preferring regions of visual cortex. Vertices from all “constituent” ROIs that preferred a category (e.g., bodies) were grouped into a single “joint-ROI,” excluding all vertices located in early visual areas (V1 to V3) (Strappini et al. 2019), as estimated from surface topology using the template created by Wang et al. (Wang et al. 2015). For instance, the body joint-ROI included constituent regions such as left and right EBA and left and right fusiform body area (FBA), and the scene joint-ROI included constituent regions such as PPA, the transverse occipital sulcus (TOS), and retrosplenial cortex (RSC).

The rationale for combining all the vertices that prefer a category into a single “joint” ROI was as follows. First, we had no a priori reason in our analysis of spontaneous activity patterns to expect different results for regions preferring the same category. Second, the use of joint-ROIs reduced the total number of statistical comparisons in our subsequent analyses of spontaneous activity patterns, simplifying the analysis. Third, the use of joint-ROIs increased the number of vertices over which spontaneous activity patterns were assessed, increasing the reliability of the analysis.

A second set of contrasts identified ROIs that preferred whole objects relative to phase-scrambled objects (face + body + scene + object > phase-scrambled) or the reverse (phase-scrambled > face + body + scene + object). This set of ROIs was used to compare resting and task-evoked patterns for whole objects versus low level features. whole-object and phase-scrambled-object constituent ROIs were grouped, respectively, into a whole-object joint-ROI and a phase-scrambled-object joint-ROI. Table 1 summarizes the mean MNI coordinate, mean z-score for the obtained group peak, and mean number of vertices for all constituent ROIs in each joint-ROI. Figure 3B schematically indicates the position of all constituent ROIs in the face, the body, and the scene joint-ROIs based on their group peak locations. Figure 4B shows the location of all constituent ROIs in the whole-object joint-ROI and the phase-scrambled-object joint-ROI superimposed on a surface map of V1–V3 using the template from Wang et al. (2015).

Table 1.

ROI summary

		Mean MNI Across Subjects					No. of Vertices
Contrast	Hemisphere	x	y	z	Group Peak Z-value	No. of Subjects	Mean	SD	ROI Name
Face > objects	RH	41	−50	−16	3.76	14	262	70	R FFA
		54	−49	12	2.71	15	232	74	R STG
Body > objects	LH	−41	−73	12	6.61	16	289	45	L EBA
		−37	−44	−17	4.42	14	253	73
	RH	42	−65	15	6.28	15	312	52	R EBA
		41	−44	−13	4.62	14	280	69
		56	−52	4	3.91	14	276	73
Scene > objects	LH	−21	−47	−4	8.14	16	327	37	L PPA
		−14	−51	7	5.88	16	294	56	L RSC
		−17	−59	21	3.95	14	236	42
		−30	−80	23	3.19	14	211	57	L TOS
	RH	20	−41	−10	8.35	16	279	45	R PPA
		27	−68	−10	6.79	16	254	43
		21	−54	10	6.02	14	285	76	R RSC
		35	−77	29	4.55	15	273	64	R TOS
		13	−36	41	4.01	14	237	65
Whole objects > Phase-scrambled objects	LH	−41	−73	−1	8.46	16	374	31	L LO
		−35	−66	−12	7.50	16	334	36
		−20	−72	30	5.92	15	252	42
		−30	−83	11	5.71	16	280	63
		−26	−39	−14	5.57	15	243	48
		−22	−56	45	4.12	14	245	70
		−51	−59	8	3.10	14	231	49
	RH	44	−72	−7	8.18	16	322	38	R LO
		36	−44	−20	6.47	16	295	43
		26	−74	30	5.03	15	237	60
		40	−64	16	4.74	14	263	82
Phase-scrambled objects > whole objects	LH	−5	−89	3	8.25	16	276	34
		−8	−80	−11	5.84	16	267	40
		−19	−89	16	3.71	15	282	54
	RH	10	−93	4	7.41	16	354	37
		9	−77	−4	7.4	16	247	35

Open in a new tab

EBA, extrasiate body area; FFA, fusiform face area; L, left; LH, left hemisphere; LO, lateral occipital; MNI, Montreal Neurological Institute; PPA, parahippocampal place area; RH, right hemisphere; ROI, region of interest; RSC, retrosplenial cortex; STG, superior temporal gyrus; TOS, transverse occipital sulcus.

Fig. 4. — Whole-object and phase-scrambled regions of interest (ROIs). A: group z-statistic Localizer maps of visual regions that prefer whole objects or phase-scrambled objects. B: schematic rendering of whole-object and phase-scrambled ROIs. ROIs were separately defined for each individual from their localizer maps using the group foci as a constraint. Surface renderings of V1, V2, and V3 from the Wang et al. (2015) template are superimposed. LH, left hemisphere; RH, right hemisphere.

Task Scans: Estimation of Multivertex Activation Patterns

For each ROI from each subject, we separately estimated the multivertex activity pattern evoked by each stimulus and by each category in the task scans using two general linear models (GLM). One model used stimulus-specific β weights to estimate the multivertex activity pattern evoked by each individual stimulus, and the other used category-specific β weights to estimate the stimulus-evoked activity pattern associated with each category (e.g., the pattern outlined by the red outline in Fig. 3A). Each GLM also included a separate target regressor for trials in which a stimulus was perturbed in size or position, and baseline and linear trend regressors for each scan. To determine the task-evoked magnitude for each stimulus category, a β weight matrix was separately computed using spatially nonnormalized BOLD timeseries from the task scans. As with the localizer scans, all GLMs for the task scans were constructed using an assumed hemodynamic response.

Task Scans: Representational Similarity Analysis of Stimulus-Evoked Patterns

We conducted representational similarity analyses (RSA) (Bracci et al. 2015; Devereux et al. 2013; Kriegeskorte et al. 2008b) to determine the within-category and between-category similarities of exemplars from a joint-ROI’s preferred category and exemplars from each of the other categories. A representational similarity matrix (RSM) for each subject was computed by correlating across vertices the obtained β weights for each stimulus exemplar with the weights for each other stimulus exemplar. The individual RSMs were then averaged across all 16 subjects, with Fisher z-transformations and reverse Fisher z-transformations, to produce a group-averaged RSM. RSM matrices were computed for three classic category-preferring regions (Fig. 5A; L EBA, R FFA, R PPA) and for each joint-ROI (Fig. 5B). We then conducted two statistical analyses using the group-averaged RSM matrices for each joint-ROI. First, unpaired t tests were used to determine for each joint-ROI whether the similarity of multivertex BOLD patterns for pairs of exemplars from the preferred category of the joint-ROI was larger than the similarity of BOLD patterns for pairs of exemplars from each of the other whole-object categories (Fig. 5C). Second, unpaired t tests were used to determine for each joint-ROI whether the within-category similarity of BOLD patterns for pairs of exemplars from the preferred category of a joint-ROI was larger than the between-category similarity of BOLD patterns between each exemplar of the preferred category and each exemplar from a different whole-object category (Fig. 5D).

Fig. 5. — Representational similarity analyses (RSA) of multivertex blood-oxygenation-level-dependent (BOLD) patterns for individual stimuli and categories. A: RSA of exemplars for each “whole object” category in 3 classical category-preferential areas. B: RSA of individual exemplars for each whole-object category for each joint-ROI. C: for each joint-ROI (e.g., faces), the graphs show the average within-category similarity for all pairs of exemplars from the joint-ROI’s preferred category (e.g., faces, green symbol) and the average within-category similarity for all pairs of exemplars from each of the other whole-object categories (e.g., body, red symbols). The similarity of two patterns was determined by spatial correlation over vertices. Black asterisks indicate the significance in each joint-ROI of an unpaired t test that compared the magnitudes of the mean pairwise within-category similarities for the preferred category versus each of the other categories (* and ** indicate P < 0.01 and P < 0.0001; error bars indicate ± SE). D: for each joint-ROI (e.g., faces), the graphs show the average within-category similarity for all pairs of exemplars from the joint-ROI’s preferred category (e.g., faces, green symbol), and the average between-category similarity of all pairs of exemplars in which one exemplar was from the preferred category and the other exemplar was from a different category (e.g., the mean pairwise similarities of faces and bodies, blue symbol). Asterisks indicate the significance in each joint-ROI of an unpaired t test that compared the magnitudes of the mean pairwise within-category similarity for the preferred category to the mean pairwise between-category similarity for the preferred category with each other category (* and ** indicate P < 0.01, and P < 0.0001; error bars indicate ± SE). E: between-category RSA in each joint-ROI based on the multivertex pattern evoked for each category. EBA, extrastriate body area; FFA, fusiform face area; L, left; PPA, parahippocampal place area; R, right; ROI, region of interest; Scr, scrambled.

We also conducted a second representational similarity analysis to determine whether the multivertex patterns for each category showed the expected cross-category relationships, such as greater similarity between the patterns for face and body categories than for face and scene categories. For each individual an RSM was computed by correlating across vertices the obtained categorical β weights (e.g., the average scene-evoked multivertex pattern outlined by the red outline in Fig. 3A) across categories, i.e., we correlated the multivertex pattern for scenes with the multivertex patterns for phase- or grid-scrambled tools, chairs, mammals, bodies, and faces. A group-averaged RSM was then computed (Fig. 5E).

Determining Similarity of Resting Multivertex Patterns and Stimulus-Evoked Patterns

The next step in the analysis determined the similarity of stimulus-evoked multivertex patterns to the patterns observed on each frame of resting scans. For each participant’s individual joint-ROI and the associated constituent ROIs, we determined the degree to which the stimulus-evoked multivertex pattern for a category matched the multivertex pattern on each resting frame. The procedure is illustrated in Fig. 6A for a single subject using real data. In the first step, as described above, the multivertex pattern evoked by a category in a region was determined (e.g., the “Scene” activity pattern outlined by the red outline in Fig. 6A, Task BOLD). Then, the stimulus-evoked pattern for a category was correlated across vertices with the resting activity pattern on a single frame of the resting-state scans (Fig. 6A, Resting-state BOLD). A high positive correlation coefficient indicates that the resting multivertex pattern on a given frame was very similar to the pattern evoked by the category (e.g., the resting frame with a scene-like resting activity pattern outlined by the magenta outline in Fig. 6A). A near-zero correlation coefficient indicates that the resting multivertex pattern on a given frame was not similar to the pattern evoked by the category (e.g., the resting frame with a not-scene-like resting activity pattern outlined by the green outline in Fig. 6A). Finally, a high negative correlation coefficient indicates that the resting multivertex pattern on a given frame was very similar to the inverse of the pattern evoked by the category (e.g., the resting frame with a scene-inverted resting activity pattern outlined by the cyan outline in Fig. 6A). This procedure was repeated across all resting frames, resulting in a “stimulus-pattern-to-rest” correlation timeseries (one correlation coefficient per resting frame) for each category in each ROI, as shown by the timeseries in Fig. 6A.

From each timeseries, we constructed a corresponding distribution of correlation coefficients (Fig. 6A, distribution shown in blue). Our initial expectation was that the mean of the distribution of correlation coefficients for the preferred category of a joint-ROI would be shifted toward correlations (i.e., higher similarity values) relative to the mean similarity value for nonpreferred categories. However, as shown in the results section, the mean similarity value for all categories was essentially zero. We therefore analyzed the second-order statistics of the distributions and found systematic differences between preferred and nonpreferred categories that reflected a greater spread of positive and negative similarity values for preferred than nonpreferred categories (see results). We indexed the spread of the distribution of correlation coefficients, which reflects a spread across resting timepoints of the spatial similarity of resting activity patterns to the stimulus-evoked pattern for a category, by the upper 90% value of the distribution, hereafter termed the U90 value (Fig. 6C). The U90 value computed for a category and ROI served as a measure of the relationship between resting activity patterns and the patterns evoked by a category mean. The U90 value was used as an alternative measure of the variance of a distribution since the U90 value refers to the value of a correlation coefficient, indicating the degree of pattern similarity between the task-evoked and resting state activity pattern, and therefore is a more transparent measure of spatial similarity. However, similar results were found using the variance of the distribution as a summary measure instead of the U90 value. For analyses that involved the whole-object joint-ROI rather than joint-ROIs that preferred a particular object category such as faces, U90 values for the six whole-object categories (face, body, mammal, chair, tool, scene) were averaged together to form a whole-object U90 value (Fig. 7B).

Fig. 7. — Stimulus-evoked-to-rest pattern similarity analysis in regions preferring phase-scrambled objects or whole objects. A: representational similarity analyses for whole-object and phase-scrambled-object joint-ROIs based on the multivertex pattern for each category. B: group-averaged upper 90% of the distribution (U90) values. Black symbols indicate significant paired t test between the whole-object category and scrambled category (++Bonferroni–Holm corrected P-value ≤ 0.005). Effect size (Hedge’s G) for black symbols are 0.90 for whole object ROI and are lower than −0.63 for phase-scrambled-object joint-ROI. Obj, object; Scr, scrambled. Error bars indicate ± SE.

Finally, to observe a potential relationship in a joint-ROI between the magnitude of the category-evoked response (Fig. 8, left column) and the strength of the relationship between category-evoked multivertex patterns and spontaneous activity patterns (U90 values), the correlation across category between task activation magnitudes and U90 values of stimulus-evoked-to-rest pattern similarity was measured for each joint-ROI (Fig. 8, middle and right columns). This correlation was computed for each joint-ROI in two ways. First, for each stimulus category we averaged the U90 values and task activation magnitudes across participants and then computed the correlation across categories (Fig. 8, middle column). Second, within each participant we calculated the correlation between the U90 value and the task-activation magnitude across categories (Fig. 8, right column). We then conducted a one-sample t test on these single-subject correlation coefficients (one coefficient from each participant) to determine whether the correlation was significantly different than zero. The latter technique determines whether a nonzero value generalizes over the population.

Statistical Analysis of U90 Values

To statistically analyze the similarity between stimulus-evoked and resting multivertex patterns, U90 values were analyzed via repeated-measures ANOVAs and post hoc paired t tests. For example, the statistical significance of an overall dependence of U90 values for a joint-ROI on the stimulus category was determined by conducting repeated-measures ANOVAs with Category Type as factors. Paired t tests were then conducted to test specific contrasts, with a Bonferroni–Holm correction for multiple comparisons. We also separately analyzed U90 values from sessions 1 and 2 and conducted paired t tests to determine whether there were significant differences between sessions.

Pattern-Based Resting Functional Connectivity

The preceding analyses examined the similarity between stimulus-evoked patterns and the patterns measured on single frames of resting scans. The analyses described next determined the temporal correlation over resting frames of the amplitude of stimulus-evoked patterns in two different regions (pattern-based FC). For the pattern-based FC analysis, we used the constituent ROIs from the joint-ROI that preferred a category (face, body, or scene) relative to the object category (chairs + tools). Since only two face constituent ROIs were found and one of those ROIs largely overlapped with a body constituent ROI in ventral temporal cortex, face constituent ROIs were not included in the FC analysis. Therefore, pattern-based FC was computed over 14 ROIs: 5 from the body joint-ROI and 9 from the scene joint-ROI.

The correlation over vertices of a stimulus-evoked pattern with the pattern on a single resting frame results in a single correlation coefficient. Repeating this process for each successive resting frame results in a timeseries of correlation values, which we call a stimulus-pattern-to-rest correlation timeseries (see Determining Similarity of Resting Multivertex Patterns and Stimulus-Evoked Patterns and Fig. 6A). Three pattern-based FC matrices were computed using stimulus-pattern-to-rest correlation timeseries. Figure 9A illustrates the procedure for computing the cells of an FC matrix using the stimulus-pattern-to-rest correlation timeseries from two scene regions (TOS, PPA) and two body regions (EBA, FBA). Figure 9B shows the resulting matrices. First, for each participant, stimulus-pattern-to-rest correlation timeseries for each of the 14 ROIs were generated using only the body-evoked pattern for each ROI (i.e., the multivertex activity pattern evoked by bodies in that ROI during the task scans). The correlation between these body pattern-to-rest timeseries for all pairings of the 14 ROIs was then computed (i.e., body-ROI-to-body-ROI, scene-ROI-to-scene-ROI, and body-ROI-to-scene-ROI pairings). For example, the leftmost graphs in Fig. 9A show the correlation between TOS and PPA (top graph) and the correlation between EBA and FBA (bottom graph) using the body-pattern-to-rest correlation timeseries generated in each ROI from the body-evoked pattern. The correlation coefficients were then entered into the corresponding cells of the pattern-based FC matrix, “Using Body pattern-to-rest correlation timeseries only,” shown in Fig. 9B.

A similar procedure was used to generate a pattern-based FC matrix using only the scene-evoked pattern for each ROI. For example, the middle graphs in Fig. 9A show the correlation between TOS and PPA, and between EBA and FBA using the stimulus-pattern-to-rest correlation timeseries generated in each ROI using the scene-evoked pattern. Finally, the correlation coefficients were entered into the corresponding cells of the pattern-based FC matrix, “Using Scene pattern-to-rest correlation timeseries only,” shown in Fig. 9B.

To generate the third pattern-based FC matrix in Fig. 9B (“Using Preferred pattern-to-rest correlation timeseries”), the stimulus-pattern-to-rest correlation timeseries in a body ROI was generated using the body-evoked pattern for that ROI and the stimulus-pattern-to-rest correlation timeseries in a scene ROI was generated using the scene-evoked pattern for that ROI. Then the correlation between the correlation timeseries for all pairings of the 14 ROIs was computed and entered into the appropriate cells of the pattern-based FC matrix.

Finally, a vertex-averaged FC matrix was computed by first averaging the resting BOLD timeseries across all vertices of an ROI to generate a vertex-averaged timeseries and then temporally correlating these averaged timeseries for all pairs of ROIs (Fig. 10). Vertex-averaged FC matrices, which correspond to the standard regional FC matrices found in the literature, eliminate any information carried by the multivertex pattern of BOLD activity within ROIs. Vertex-averaged FC matrices were computed to allow an initial evaluation of whether the sets of regions linked by pattern-based FC differed from the sets of regions that have been linked by previous studies of resting-state organization using standard regional FC matrices.

Fig. 10. — Vertex-averaged functional connectivity (FC). A standard non-pattern-based FC matrix was computed by first averaging the vertexwise timeseries of the blood-oxygenation-level-dependent (BOLD) signal across the vertices within each region, and then correlating the resulting vertex-averaged timeseries across all pairs of regions (*left*). All cells involving FC between body regions, between scene regions, and between body and scene regions are averaged and plotted (*right*). Black symbols indicate significant group paired t test comparing correlation (ρ) values (*P ≤ 0.05, effect size = 2.11 and 2.83). Error bars indicate ± SE. ROI, region of interest.

Pattern-based FC values were analyzed via repeated-measures ANOVAs and paired t tests. For example, we statistically evaluated whether the magnitude of pattern-based FC depended on both the category of the stimulus-evoked multivertex activity pattern and the preferred category of the ROIs by conducting a repeated-measures ANOVA with the Category-evoked pattern (body, scene) and ROI-type (body, scene) as factors. Paired t tests were conducted to test differences between specific evoked patterns/ROI combinations. For example, pattern-based FC values for body ROIs versus scene ROIs were compared for correlation timeseries generated using body-evoked multivertex patterns.

RESULTS

The first goal of the experiment was to compare multivertex activity patterns measured in the resting state with fMRI to stimulus-evoked patterns for different stimulus categories, including ecological stimuli (e.g., photographs of faces, tools, and scenes) and stimuli that emphasized low-level features (e.g., phase-scrambled or grid-scrambled images of those stimuli). This comparison was conducted in regions of higher-order visual cortex that activated more strongly to specific stimulus categories (e.g., bodies) relative to other categories (e.g., chairs and tools). In addition, resting and stimulus-evoked patterns for phase-scrambled, grid-scrambled, and whole objects were compared in regions of visual cortex that responded more strongly to phase-scrambled objects than to whole objects or showed the reverse relationship. To measure spontaneous activity, we ran a set of resting-state scans in which human observers fixated a central point on a blank screen. This activity was measured first to prevent possible learning effects from the other conditions (Fig. 2).

Localization of Regions with Visual Category Preferences

To identify category-specific visual regions, we ran a set of localizer scans in which multiple stimuli belonging to one of five stimulus categories [faces, bodies, scenes, man-made objects (chairs and tools), and phase-scrambled versions of these stimuli] were presented in a blocked design (Figs. 1 and 2). We used standard contrasts (as in Bracci and Op de Beeck 2016) to identify category-preferring regions. For instance, activity evoked by body stimuli was subtracted from activity evoked by man-made objects (tools and chairs) to localize body-preferring regions such as EBA (see Fig. 3, A and B, and Table 1 for all category-preferring regions). Separate contrasts identified regions more active for whole objects [face+scene+bodies+(tools+chairs)] than for low-level visual features (phase-scrambled objects). Phase-scrambled objects activated more strongly in regions of early visual cortex [V1–V3 based on the maps of Wang et al. (2015)], while whole objects activated more strongly in lateral and ventral occipital cortex, including some category-preferring regions (Fig. 4, A and B and Table 1).

Representational Similarity Analysis of Task-Evoked Patterns

During task scans (Fig. 2), we randomly presented individual stimuli belonging to each category to extract the stimulus-evoked multivertex pattern in a particular ROI for each stimulus and corresponding category. Two general linear models (GLMs) were conducted to estimate the stimulus-evoked patterns. One model used stimulus-specific β weights to estimate the multivertex activity pattern evoked by each individual stimulus, and the other used category-specific β weights to estimate the stimulus-evoked multivertex activity pattern associated with each category.

To show that our stimuli and procedure generated multivertex patterns that were consistent with the literature, we conducted a representational similarity analysis (RSA) (Bracci et al. 2015; Devereux et al. 2013; Kriegeskorte et al. 2008b). The representational similarity analysis was conducted using the task scans, which were completely independent of the localizer scans used to determine category-preferring ROIs. Figure 5A shows the similarity of multivertex patterns evoked by individual stimuli within several classical category-preferring ROIs. In left EBA the highest representational similarity was found between human bodies, and the next highest between pictures of mammals, which included their bodies. In the right fusiform face area [FFA (Kanwisher et al. 1997)], faces and other animate stimuli (bodies, mammals) generated more similar patterns than stimuli from the inanimate categories chair, tool, and scene, with the most consistent representational similarity found between face exemplars (Grill-Spector and Weiner 2014; Kriegeskorte et al. 2008a). In the scene-preferring region right parahippocampal place area [PPA (Epstein and Kanwisher 1998)], the activity patterns evoked by different scenes were well correlated, with low correlations between and within all other categories. RSA analyses of each joint-ROI (Fig. 5B) yielded similar results.

To evaluate the statistical significance of these results, we used unpaired t tests to compare 1) the within-category similarity of pairs of exemplars from the preferred category for a joint-ROI versus the within-category similarity of exemplar pairs from each of the other categories (Fig. 5C) and 2) the within-category similarity of pairs of exemplars from the preferred category for a joint-ROI versus the between-category similarity of exemplars pairs consisting of a preferred exemplar and a nonpreferred exemplar (Fig. 5D) (see methods, Task Scans: Representational Similarity Analysis of Stimulus-Evoked Patterns). These analyses showed the expected results for each joint-ROI, with significantly higher similarity values for pairs of preferred exemplars (green symbols in Fig. 5C) than for pairs of exemplars from each of the other whole-object categories (red symbols in Fig. 5C). The only exceptions were that within the body joint-ROI, within-category similarity for faces (another animate category) was highest, and no difference was found between the body and mammal categories, which was not surprising since the mammal photographs included their bodies. The comparison of within- and between-category similarity also yielded the expected results, with significantly higher similarity values for exemplar pairs from the joint-ROI’s preferred category (green symbols in Fig. 5D) than for exemplar pairs involving the preferred category and a different category (blue symbols in Fig. 5D).

We conducted a second representational similarity analysis using the pattern evoked by a stimulus category, as estimated by a category regressor in a separate GLM, rather than using the patterns evoked by individual stimuli. In addition, we grouped each set of category-preferential ROIs for an individual into a single joint-ROI instead of conducting the analysis separately within each localizer-defined ROI. For instance, the body joint-ROI included left and right EBA, left and right fusiform body area (FBA), and so forth, and the scene joint-ROI included constituent regions such as PPA, the transverse occipital sulcus (TOS), and retrosplenial cortex (RSC) (see Table 1 for a complete listing). Therefore, a joint-ROI included all the vertices from visual regions that preferred a certain visual category. This procedure simplified the analysis of the similarity between stimulus-evoked and resting multivertex patterns (see methods, Defining ROIs from Localizer Activation Contrasts for the full rationale behind the use of joint-ROIs).

The results of this second RSA procedure (Fig. 5E) were also consistent with the literature. Both in body and face joint-ROIs, the highest representational similarity was found between animate categories (face, bodies, mammals) as compared with other categories (Grill-Spector and Weiner 2014; Kriegeskorte et al. 2008a). For instance, the similarity of body- and face-evoked multivertex activity patterns in the face joint-ROI was ρ = 0.73, while the similarity of face- and scene-evoked patterns in the face joint-ROI was ρ = 0.41. Table 2 indicates the representational similarity between the task-evoked multivertex patterns corresponding to the face, body, and scene categories within each joint-ROI.

Table 2.

Correlation over vertices between face-evoked, body-evoked, and scene-evoked multivertex patterns in three joint-ROIs

	Similarity (ρ) Between Face-Template and Body-Template	Similarity (ρ) Between Face-Template and Scene-Template	Similarity (ρ) Between Body-Template and Scene-Template
Face Joint-ROI	0.729	0.415	0.550
Body Joint-ROI	0.618	0.421	0.423
Scene Joint-ROI	0.506	0.356	0.379

Open in a new tab

ROI, region of interest.

Stimulus-Evoked-to-Rest Pattern Similarity Analysis in Category-Preferring Regions

We next tested the first prediction of the representation hypothesis, namely that multivertex patterns of spontaneous activity in category-preferring regions should be more related to the stimulus-evoked multivertex pattern for preferred than nonpreferred categories. For each joint-ROI and for each category, the category-evoked pattern for the joint-ROI was correlated over the vertices of the joint-ROI with the resting pattern in the joint-ROI on each resting frame to determine a resting timeseries of correlation coefficients (termed a stimulus-pattern-to-rest correlation timeseries) and a corresponding frequency distribution of coefficient values. The upper 90% value (U90 value) of the distribution was used as a summary measure of the relationship between the stimulus-evoked and resting multivertex activity patterns (see methods, Determining Similarity of Resting Multivertex Patterns and Stimulus-Evoked Patterns for the rationale behind using U90 values as opposed to measures of variance).

Figure 6A illustrates this procedure in a single subject using a single ROI that prefers scenes (PPA) and the category-evoked pattern for scenes, which is a multivertex set of normalized activation values (Fig. 6A, left). This evoked pattern is correlated over the vertices of PPA (ρ) with the spontaneous patterns of activity on each frame of the resting-state scans. A timeseries of ρ-values is generated (Fig. 6A, middle), as well as a corresponding frequency distribution of ρ-values (Fig. 6A, the histogram in blue). A U90 value for the ROI and category is then determined from the distribution. The insets in the middle panel show resting frames in which the spontaneous activity (real data) was not correlated (ρ = 0.003, outlined in green), positively correlated (ρ = 0.81, outlined in magenta), or negatively correlated (ρ = −0.74, outlined in cyan) with the scene-evoked multivertex pattern. The same procedure was used to generate a U90 value for each of eight categories (faces, bodies, mammals, chairs, tools, scenes, grid-scrambled, phase-scrambled) in each joint-ROI (body, face, scene).

Figure 6B shows the distributions of correlation coefficients across all subjects for each preferred category within its corresponding joint-ROI (green hue). For example, the leftmost graph shows the distribution using the body-evoked multivertex pattern within the body joint-ROI. A second distribution, generated using the pattern evoked by phase-scrambled objects within the same body joint-ROI, has been superimposed (black hue). Theoretically, the two distributions might differ in the mean, variance, skewness, or some other parameter. For each joint-ROI, the distribution of correlation coefficients for both the preferred stimulus category and the phase-scrambled category were symmetric and centered on zero. However, the spread of the distribution was higher for the preferred category-evoked multivertex pattern, meaning that larger correlation coefficients, both positive and negative, were observed for the preferred than phase-scrambled category (red arrows in Fig. 6B). Therefore, a larger U90 value indicates the presence of larger positive matches and larger negative matches between the resting multivertex pattern and the category-evoked pattern. Similar findings were obtained for face (middle panel) and scene multivertex patterns (rightmost panel) in the corresponding joint-ROIs, as compared with phase-scrambled patterns.

The categorical specificity of spontaneous multivertex patterns in each joint-ROI was tested by comparing U90 values for different categories. Figure 6C shows mean U90 values for a joint-ROI’s preferred category, defined from its localizer contrast (green symbol; e.g., body in the body joint-ROI), “nonpreferred” categories (red symbols), grid-scrambled category (blue symbol) and phase-scrambled category (black symbol) averaged across subjects. We first conducted an overall repeated-measures ANOVA on U90 values with joint-ROI (body, face, and scene) and Category (8 levels) as factors. The main effects of joint-ROI [F(2, 30) = 50.4, P < 0.0001] and Category [F(7, 105) = 3.46, P = 0.002], and the interaction of joint-ROI by Category [F(14, 210) = 4.37, P < 0.0001] were all significant (η² = 0.047). Separate repeated-measures ANOVAs for each joint-ROI with Category (8 levels) as a factor indicated a highly significant main effect of Category in each joint-ROI [body: F(7,105) = 7.25, P < 0.0001, η² = 0.205; face: F(7,105) = 3.25, P = 0.004, η² = 0.103; scene: F(7,105) = 3.93, P = 0.0008, η² = 0.117]. Therefore, for each joint-ROI, the spread of stimulus-pattern-to-rest similarity values significantly depended on the category of the stimulus-evoked pattern.

The significant main effect of Category for each joint-ROI indicated that resting multivertex patterns in each joint-ROI consistently showed larger U90 values for some category-evoked patterns than for other patterns. The significant interaction of joint-ROI by Category indicated that these modulations of U90 values across category reliably differed across joint-ROIs. These results are consistent with representation hypothesis.

To compare the U90 value for the joint-ROI’s preferred category versus each other category, we conducted paired t tests with a Bonferroni–Holm correction for multiple comparisons. Significant, multiple-corrected comparisons are indicated in Fig. 6C by plus signs. In the body joint-ROI, the U90 value for bodies was significantly larger than for chairs, scenes, and phase-scrambled stimuli. In the face joint-ROI, the U90 value for faces was significantly larger than for scenes. Conversely, in the scene joint-ROI, the U90 value for scenes was significantly larger than for all other categories.

Therefore, the “animate” body and face joint-ROIs and the “inanimate” scene joint-ROI showed in part the opposite category selectivity, with U90 values in the face and body joint-ROIs significantly greater for the face and body categories, respectively, than for scene categories, and the U90 value in the scene joint-ROI significantly greater for scenes than for either faces or bodies. The U90 values for face and body categories within each joint ROI were similar, reflecting the fact that both are animate categories and have greater cross-category representational similarity with each other than with scenes (Table 2). Finally, paired t tests did not yield any significant differences, uncorrected, between the U90 values for a stimulus category/joint-ROI on sessions 1 versus 2, indicating no consistent effect of one session on the other.

These results provide some support for the first prediction of the representation hypothesis, namely that spontaneous multivertex patterns in category-preferring regions are more related to the patterns for some categories than for others. However, “more related” means a greater spread of extreme similarity values, both positive and negative, rather than a shift in the mean to more positive similarity values.

Stimulus-Evoked-to-Rest Pattern Similarity Analysis in Regions Preferring Whole Versus Phase-Scrambled Objects

The above results showed that the multivertex pattern of spontaneous activity in regions of high-level visual cortex that respond preferentially to ecological visual categories was more related to the pattern evoked by one category than another (e.g., bodies versus scenes). We next asked whether a similar result would be found in regions that show stimulus preferences for low-level features as compared with more ecological categories such as face or body. This result would support a general conclusion that the stimulus preferences of a region largely drive the multivertex pattern of spontaneous activity. We used the localizer scans to identify ROIs in which stimulus-evoked responses were stronger or weaker for phase-scrambled objects than for the union of the whole-object categories (face, body, mammal, chair, tool, and scene). The resulting “phase-scrambled-object” joint-ROI comprised medial posterior visual regions in early visual cortex (V1–V3 according to the Wang template (Wang et al. 2015) while the “whole-object” joint-ROI comprised regions in lateral and ventral visual cortex (Fig. 4B).

A representational similarity analysis in the phase-scrambled-object joint-ROI showed high similarity between phase-scrambled, grid-scrambled, and scene stimuli, while the whole-object joint-ROI showed low similarity between those categories (Fig. 7A). Figure 7B shows the results of a stimulus-evoked-to-rest pattern similarity analysis based on U90 values in each joint-ROI, which support the general conclusion that stimulus-evoked-to-rest pattern similarities are not necessarily stronger for more ecological stimuli. Instead, stimulus-evoked-to-rest correspondences reflect a region’s stimulus preferences, which are different in high- and low-level visual cortical regions (Fig. 7B). An ANOVA with ROI-type (whole objects, phase-scrambled objects) and Stimulus-type (whole objects, grid-scrambled, phase-scrambled objects) as factors indicated that the critical interaction of ROI-type by Category [F(2,30) = 14.2, P < 0.0001, η² = 0.069] was significant. A significant interaction was also found for a 2 × 2 sub-ANOVA restricted to the categories whole objects and phase-scrambled objects [F(1,15) = 23.5, P < 0.0001, η² = 0.097]. These results are again consistent with the representation hypothesis.

Within each joint-ROI, we compared the U90 value for the preferred category versus the two nonpreferred categories using paired t tests with a Bonferroni–Holm correction for the four comparisons over the two joint-ROIs. In the phase-scrambled joint-ROI, U90 values were significantly higher for both scrambled stimulus categories than for the whole-object category, and in the whole-object joint-ROI, the U90 value for the whole-object category was significantly greater than for the phase-scrambled object category. The grid-scrambled pattern, which contains both high-level and low-level features (e.g., a high density of contour terminators), showed U90 values both in early visual and higher-order visual cortex that were not distinguishable from the region’s preferred stimulus category.

These results comprise a second demonstration that the category selectivity of resting U90 values varied with a joint-ROI’s category preferences: we observed significantly larger U90 values for category A than B in a joint-ROI preferring category A (e.g., whole objects) simultaneously with significantly larger U90 values for category B than A in a different joint-ROI preferring category B (e.g., phase-scrambled objects). Therefore, the sign of significant U90 differences between two categories depended on the joint-ROI in which the multivertex similarity of the evoked pattern to spontaneous patterns was evaluated, and was governed by the joint-ROIs’ stimulus preferences. These results are consistent with the interpretation that spontaneous activity patterns in visual cortex are affected by the stimulus preferences of the region, irrespective of whether those preferences favor more or less ecological categories.

U90 Values Correlate with Activation Strength

The representation hypothesis maintains that task-evoked patterns entrain spontaneous activity patterns during development and through experience. Therefore, one might expect a positive relationship between the magnitude of the category-evoked response and the strength of the relationship between category-evoked multivertex patterns and spontaneous activity patterns (i.e., the U90 value). Figure 8, left column shows the mean activation strengths for different categories during the task scans. The magnitude of the stimulus-evoked response in a joint-ROI was generally strongest for the preferred category. Since joint-ROIs were defined from localizer scans that were independent of the task scans, this result indicates the stability of the ROI assignments.

Figure 8, middle column shows for each joint-ROI the correlation across categories between group-averaged task activation magnitudes and U90 values of stimulus-evoked-to-rest pattern similarity. The correlation was positive and significant in each joint-ROI. The greater the activation magnitude of a category, the greater the U90 value. This relationship also was significant when the correlation coefficient between activation magnitudes and U90 values was computed separately for each participant (Fig. 8, right column), and a group one-sample t test was conducted on the correlation coefficients across participants (body, P < 0.001, face, P-value = 0.005, scene, P-value = 0.001; whole-object P-value < 0.001; phase-scrambled, P-value < 0.001).

Pattern-Based Functional Connectivity at Rest

FC analyses typically evaluate the correlation between the timeseries of activity for single voxels or between voxel-averaged timeseries. However, recent studies have also measured the interregional temporal correlation of multivertex patterns during tasks (Anzellotti and Coutanche 2018; Chen et al. 2018; Coutanche and Thompson-Schill 2013). We used a similar approach to determine whether the signed magnitude of the resting multivertex pattern for a category in each constituent ROI of a joint-ROI, as determined on a resting frame by correlating over vertices the category-evoked and resting activity pattern, fluctuated synchronously or in an uncorrelated fashion over frames between pairs of constituent ROIs. For instance, we determined whether the amplitude of the multivertex patterns for scenes in regions such as PPA, TOS, and RSC, which were previously combined to form the scene joint-ROI (Table 1), fluctuated synchronously at rest. Synchronous fluctuations (i.e., temporally correlated fluctuations) would indicate temporal variations of the amplitude of an interregional brain state specific for a particular category. We conducted this analysis across the constituent ROIs of the body and scene joint-ROIs. The face joint-ROI was not included in these analyses, since that ROI only included two regions and one of them overlapped with a constituent body ROI. In contrast, the scene and body joint-ROIs contained multiple ROIs that were all disjoint.

For each body- and scene-preferring constituent ROI, separate body and scene “stimulus-pattern-to-rest” correlation timeseries were computed based, respectively, on the correlation over vertices of the multivertex pattern on each resting frame with the body-evoked pattern and the scene-evoked pattern. This procedure is illustrated in Fig. 9A using data from segments of resting scans in one subject. Stimulus-pattern-to-rest-correlation timeseries are shown for two scene-preferring regions (right PPA and right TOS) and two body-preferring regions (right EBA and right FBA). Each stimulus-pattern-to-rest-correlation timeseries shows the similarity values over time of resting multivertex patterns to a category-evoked pattern in a single ROI. For example, the left lower dark blue timeseries in Fig. 9A shows the similarity of the resting pattern on each frame to the body-evoked pattern in the body constituent region EBA. The left lower graph shows that the body-pattern-to-rest correlation timeseries for body-preferring regions (right EBA and right FBA), which were computed using body-evoked multivertex patterns, are positively correlated. In contrast, the left upper graph shows that the body-pattern-to-rest correlation timeseries for scene-preferring regions (right PPA and right TOS), which again were computed using body-evoked multivertex patterns, are uncorrelated. Conversely, when stimulus-pattern-to-rest correlation timeseries were computed using the scene-evoked multivertex pattern, the opposite results are found. The scene-pattern-to-rest correlation timeseries for scene-preferring regions show positively correlated fluctuations (right upper graph), while the scene-pattern-to-rest correlation timeseries for body-preferring regions are weakly correlated (right lower graph).

The data from all resting scans of all subjects were analyzed and the results are summarized in Fig. 9, B and C. Figure 9B shows three resting “pattern-based FC” matrices. A pattern-based “body” FC matrix (leftmost matrix) was constructed by computing all pairwise interregional correlations between the body-pattern-to-rest correlation timeseries, which were computed using body-evoked multivertex patterns. Similarly, a pattern-based “scene” FC matrix (middle matrix) was computed using scene-pattern-to-rest correlation timeseries (i.e., timeseries computed using scene-evoked patterns). Qualitatively, body-preferring ROIs showed stronger positively correlated spontaneous fluctuations for body-evoked than scene-evoked patterns, and scene-preferring ROIs showed stronger positively correlated spontaneous fluctuations for scene-evoked than body-evoked patterns.

The graphs in Fig. 9C summarize the pattern-based body and scene FC matrices by averaging the interregional correlations for body-preferring regions (the lower left block of each matrix in Fig. 9B outlined in blue) and scene-preferring regions (the upper right block of each matrix outlined in red). The left and middle graphs show respectively the results when stimulus-pattern-to-rest timeseries were computed using body-evoked and scene-evoked multivertex patterns. A two-factor ANOVA on the mean pairwise pattern-based FC values with ROI-type (body, scene) and Category-evoked-pattern (body, scene) as factors yielded a main effect of ROI-type [F(1,15) = 5.41, P = 0.035], reflecting the larger FC values in body-preferring regions and, critically, a significant interaction of ROI-type by Category-evoked pattern [F(1,15) = 8.96, P = 0.009] with effect size (η²) of 0.097. This effect was further supported by paired t tests of specific contrasts. When stimulus-pattern-to-rest timeseries were computed using a body-evoked multivertex pattern, interregional correlations were significantly higher in body- than scene-preferring ROIs. Conversely, when stimulus-pattern-to-rest timeseries were computed using a scene-evoked multivertex pattern, interregional correlations were significantly higher in scene- than body-preferring ROIs.

Therefore, spontaneous fluctuations of multivertex patterns of activity were more strongly correlated for multivertex patterns corresponding to the regions’ preferred category. This result suggests that resting pattern-based FC is modulated by the putative representational content of spontaneous activity. In addition, a paired t test indicated that in the pattern-based body FC matrix, the average FC was less in the scene-body block than in the body-body block of the matrix (P < 0.001, effect size = 1.89; Fig. 9B, leftmost matrix, orange versus blue outlined blocks). Similarly, in the pattern-based scene FC matrix, the average FC was less in the scene-body block than in the scene-scene block of the matrix (P = 0.035, effect size = 0.75; Fig. 9B, middle matrix, gray vs red outlined blocks). Therefore, pattern-based FC was greater between regions preferring the same category than between regions preferring different categories.

A related question was whether fluctuations of putative body and scene representations were uncorrelated at rest. The rightmost matrix in Fig. 9B shows a Preferred-Category pattern-based FC matrix in which body-pattern-to-rest correlation timeseries were computed in body-preferring regions (i.e., timeseries were computed using body-evoked multivertex patterns) and scene-pattern-to-rest correlation timeseries were computed in scene-preferring regions (i.e., timeseries were computed using scene-evoked multivertex patterns). Accordingly, the lower left and upper right blocks of the Preferred-Category matrix match, respectively, the lower left block of the body pattern-based FC matrix and the upper right block of the scene pattern-based FC matrix.

The “scene-body” block of the Preferred-Category matrix outlined in green is of primary interest. The correlation between scene and body regions was uniformly low under conditions in which the interregional correlation involved timeseries from scene and body regions that respectively indicated the fluctuations of scene- and body-evoked multivertex patterns (see rightmost graph, Fig. 9C, for average correlation values for scene-body blocks). Therefore, periods in which a body-evoked pattern was maximally present in body-preferring ROIs were largely uncorrelated with periods in which a scene-evoked pattern was maximally present in scene-preferring ROIs. Paired t tests indicated that correlations in scene-body blocks from the Preferred-Category matrix were significantly lower than the correlations from scene-scene (P = 0.009, effect size = 1.25) and body-body region blocks (P < 0.0001, effect size = 1.80).

Finally, a standard FC matrix (Fig. 10, left) was constructed by computing vertex-averaged resting timeseries for each region, followed by pairwise correlation of the regional timeseries. As in previous work (Hutchison et al. 2014; Konkle and Caramazza 2017; Wang et al. 2016; Zhu et al. 2011), vertex-averaged FC was category specific, with stronger FC between body-preferring regions and between scene-preferring regions, than between body- and scene-preferring regions. Pattern-based FC matrices were moderately-to-strongly correlated with the vertex-averaged FC matrix. The largest correlation was with the preferred-category matrix rather than the matrices generated using a single category (body-evoked pattern, r = 0.57; scene-evoked pattern, r = 0.48; preferred-category, r = 0.65). Qualitatively, we did not find that pattern-based FC linked regions that are normally uncorrelated in seed-based studies of resting FC.

The Relation of Pattern-Based FC in Constituent ROIs to U90 Values in Joint-ROIs

The first part of the results section analyzed the similarity on single resting frames of stimulus-evoked and resting activity patterns within a joint-ROI based on the distribution of single-frame stimulus-pattern-to-rest correlation coefficients (Figs. 6 and 7). We found a larger spread of positive and negative similarity values within a joint-ROI for preferred versus nonpreferred categories (e.g., a larger U90 value for bodies than scenes within the body joint-ROI). The second part of the results section on pattern-based FC showed that the timeseries of single-frame stimulus-pattern-to-rest correlation coefficients for the constituent ROIs of a joint-ROI fluctuated more synchronously over frames for the constituent ROIs’ preferred versus nonpreferred categories (Fig. 9). In this section, we present evidence supporting the hypothesis that the synchronous fluctuations of representational content between constituent ROIs (pattern-based FC) is partly responsible for the larger spread of positive and negative similarity values for a joint-ROI’s preferred category. This section therefore unifies the two main sets of results in the paper.

The basic argument is that, within a joint-ROI, large positive or negative similarity values on a resting frame will occur when the similarity values on that frame within the constituent ROIs of the joint-ROI are simultaneously large and have the same sign. If the signs are different or if the correlation coefficients in some constituent ROIs are small, the similarity values for different constituent ROIs will tend to cancel or average to a lower value when similarity is computed over the entire joint-ROI. Therefore, more extreme similarity values in a joint-ROI, resulting in larger U90 values, are more likely to be observed if the similarity values in the constituent ROIs fluctuate in a temporally correlated or synchronous fashion.

This mechanism explains the pattern of U90 values across the different joint-ROIs shown in Fig. 6C. U90 values averaged across categories were highest in the face joint-ROI, intermediate for the body joint-ROI, and lowest for the scene joint-ROI, i.e., were inversely related to the number of constituent regions in each joint-ROI (2 for face, 5 for body, and 9 for scenes). The greater the number of constituent ROIs in a joint-ROI, the more the overall U90 value for the joint-ROI was likely to be decreased by suboptimal synchronicity. This conclusion was supported statistically. As noted earlier, in a two-factor ANOVA on U90 values with joint-ROI (body, face, scene) and Category (8 levels) as factors, the main effect of joint-ROI was significant. We also computed a single U90 value for each joint-ROI for each participant by averaging over categories. Paired t tests on these averaged U90 values, with a Bonferroni–Holm correction for multiple comparisons (3 tests), indicated significantly larger U90 values within the face than body joint-ROIs (P < 0.0001), face than scene joint-ROIs (P < 0.0001), and body than scene joint-ROIs (P = 0.01).

The postulated mechanism linking pattern-based FC in constituent ROIs and U90 values in joint-ROIs makes a second prediction. Since the multivertex activity patterns for a nonpreferred category do not fluctuate as synchronously across constituent ROIs as those for a preferred category (Fig. 9), the resulting similarity values across frames in the joint-ROI for that nonpreferred category should be more likely to show less variation from zero, resulting in smaller U90 values. Therefore, within a joint-ROI, differences in U90 values between categories should be more reliable for joint-ROIs comprised of more constituent ROIs. This prediction is also consistent with the results in Fig. 6C, with the fewest significant differences found for the face joint-ROI and the most for the scene joint-ROI. Although other factors besides the number of constituent ROIs are clearly important in determining the category selectivity of U90 values in a joint-ROI, the larger point is that the greater pattern-based FC between constituent ROIs for their preferred category increases the incidence of extreme similarity values between stimulus-evoked and spontaneous patterns in a category-specific fashion.

DISCUSSION

The goal of the experiment was to test representational theories of spontaneous activity by determining whether in regions of human visual cortex there is a link between multivertex patterns of spontaneous activity, measured in the resting state, and the multivertex patterns evoked by ecological visual stimuli such as bodies or stimuli that emphasized low-level features such as phase-scrambled bodies.

We obtained two main results. First, resting multivertex activity patterns in regions of visual cortex were more closely related to the patterns evoked by the regions’ preferred stimulus categories. This relationship did not reflect a greater average similarity of resting patterns to the patterns for preferred categories, but instead a greater spread of resting similarity values for preferred categories. As resting multivertex patterns in a region fluctuated over time, we observed both larger positive and larger negative similarity values for the patterns evoked by stimulus categories preferred by the region. This result was demonstrated statistically by showing that significant differences between U90 values for two categories changed sign depending on the category preference of the joint-ROI in which the similarity of resting and task-evoked multivertex patterns were evaluated. Body- and face-preferring regions showed larger U90 values for faces and bodies respectively than for scenes, indexing a larger spread of similarity values for those categories in those regions, while scene-preferring regions showed larger U90 values for scenes than for faces and bodies (Fig. 6). Regions preferring whole objects versus phase-scrambled objects showed a similar result (Fig. 7), with regions preferring whole objects showing significantly greater U90 values for whole objects than for phase-scrambled objects and regions preferring phase-scrambled objects showing significantly greater U90 values for phase-scrambled objects than whole objects. This result was further strengthened by the positive correlation between U90 values and stimulus-specific activation magnitudes. The more strongly a stimulus activated a region, the higher the spread of similarity values over resting frames between the stimulus-evoked multivertex pattern and the spontaneous pattern (Fig. 8). The latter result is consistent with the notion that task-evoked patterns entrain spontaneous activity patterns during development and through experience (Fig. 1).

The second main result was that multivertex activity patterns evoked by a category fluctuated more synchronously at rest between cortical regions preferring that category. For example, in the resting state the pattern evoked by bodies was more positively correlated between body-preferring ROIs than between scene-preferring ROIs. The pattern evoked by scenes showed the opposite result (Fig. 9). Finally, resting fluctuations of the multivertex patterns evoked by scenes and bodies were largely uncorrelated within the respective preferred regions for those categories, showing that multiple representational states fluctuated in a largely uncorrelated fashion.

Therefore, the current results show that multivertex patterns of spontaneous activity within regions of human cortex, and fluctuations of those patterns between regions, code for stimulus- and category-specific information. In this respect, our work is more related to seminal work conducted in cats (Kenet et al. 2003) and monkeys (Fukushima et al. 2012; Omer et al. 2019) than to previous work on category-selective visual regions in humans, which has focused on measurements of voxelwise functional connectivity rather than on measurements of the resting multivertex pattern of activity at a timepoint (Hutchison et al. 2014; Stevens et al. 2017; Strappini et al. 2019; Turk-Browne et al. 2010; Wilf et al. 2017; Zhang et al. 2009; Zhu et al. 2011). The most closely related previous work in humans was reported by Chen and colleagues, who compared voxelwise functional connectivity measurements to evoked multivertex patterns (Chen et al. 2017).

Spontaneous Activity Patterns for Objects and Features in Visual Cortex

Spontaneous activity patterns were not more similar on average to preferred stimulus evoked-patterns but showed the greatest variation with respect to those patterns (Fig. 6). Interestingly, animal studies thus far have found the same result with some caveats. For instance, Kenet et al. (2003) recorded voltage-sensitive dye imaging in anesthetized cat visual cortex and found a significant correlation (r = 0.6) between spontaneous activity patterns and orientation selective stimulus-evoked patterns. Positive and negative values of the correlation distribution were higher (as in our experiment), rather than the mean, as compared with a control distribution obtained by flipping the orientation selective map. This finding was replicated recently in anesthetized monkey visual cortex (Omer et al. 2019). In auditory monkey cortex, spontaneous spatial covariations of gamma activity recorded with cortical grids from auditory cortex resemble tonotopic maps derived with auditory stimuli. Also in this case, the correlation involves both positive and negative high correlation values as compared with a control distribution obtained by shuffling the tonotopic map (Fukushima et al. 2012). In our experiment, the control distribution was not spatially shuffled because this control might not preserve the local structure of the vascular architecture that is the anatomical basis of the measured BOLD signal. Therefore, we instead compared stimulus-evoked-to-rest correlation distributions for two different stimuli (e.g., bodies versus scenes). Although the animal and human experiments differed in many ways, in all experiments the reported match between spontaneous and task-evoked activity was not a shift in the mean, but rather a higher frequency of more extreme matches/mismatches of the spatial patterns (e.g., compare our Fig. 6B to Omer et al. 2019 Figs. 1 and 2).

One rationale for postulating a representational function of resting activity patterns is that they serve as a prior for task processing, mitigating limits on the information processing capacity of the brain by incorporating useful prior information. For example, ongoing resting activity patterns may facilitate the analysis of sensory input during object recognition with respect to biologically important or frequently occurring categories. Recent work in monkey inferotemporal cortex indicates that individual faces can be coded by face cell assemblies whose firing rate is distributed along a small number of orthogonal dimensions (Chang and Tsao 2017). Correspondingly, the resting activity patterns found in the present work may reflect fluctuations along canonical low-dimensional stimulus configurations.

Some aspects of an appropriate prior will depend on context, such as walking in a forest versus eating a family meal at the dinner table, and will change dynamically, perhaps reflecting generative models of the expected input via top-down pathways (Mumford 1992). Because resting scans are usually conducted under conditions in which subjects lie in a dark tube while fixating a cross in an otherwise blank display, however, the activity patterns measured at rest presumably reflect canonical biases or predispositions rather than these dynamic changes (Teufel and Fletcher 2020).

Synchronous Fluctuations of Representational Content

The second important result is that multivertex patterns that putatively code a stimulus category fluctuate more synchronously at rest between visual cortical regions that prefer the same category (what we call pattern-based FC). Pattern-based FC between body-preferring regions was significantly larger when computed using a body- than scene-evoked pattern, while pattern-based FC between scene-preferring regions was significantly larger when computed using a scene-evoked than body-evoked pattern (Fig. 9).

An interesting result from the pattern-based FC analysis was that at rest different putative representational states, as indexed by multivertex patterns, fluctuated in a largely uncorrelated fashion. Synchronous fluctuations associated with body-evoked patterns in body-preferring ROIs were largely uncorrelated with synchronous fluctuations associated with scene-evoked patterns in scene-preferring ROIs, as shown by the very low correlations between scene and body regions in the analysis of the Preferred-Category FC matrix (Fig. 9). Therefore, resting activity across category-selective regions of visual cortex cannot be described in terms of a single representational state.

An interesting approach for identifying the representational content of resting FC has been reported in studies of early visual cortex, which have shown that resting FC respects the tuning of single voxels for polar angle, eccentricity, and low-level stimulus features (Arcaro et al. 2015; Heinzle et al. 2011; Raemaekers et al. 2014; Ryu and Lee 2018). Most task-based studies of representation in higher-order visual and associative regions, however, have not involved measurements of voxelwise tuning functions but instead have identified task-evoked representations through measurements of regional patterns. Therefore, pattern-based FC (Anzellotti and Coutanche 2018; Chen et al. 2018; Coutanche and Thompson-Schill 2013) can provide insights into the putative representational FC of spontaneous activity in high-level brain regions that are complementary to those provided by approaches based on the tuning properties of single voxels.

We suggest two additional ways in which pattern-based FC might inform studies of resting-state organization. First, pattern-based FC may help fractionate existing resting-state networks and identify the functional factors associated with that fractionation. For example, pattern-based resting FC between regions that prefer a particular category might depend on selectivity for features within the category, such as gender for face-preferring regions.

Second, pattern-based FC might uncover resting FC organizations that differ substantially from the normative whole-brain structure that has been described over the past decade (Cole et al. 2016; Gordon et al. 2016; Power et al. 2011; Yeo et al. 2011) although this structure does vary over individuals (Gordon et al. 2017; Gratton et al. 2018; Laumann et al. 2015). In the current work, pattern-based FC was measured within category-preferring regions. Because regions that coactivate tend to show greater resting FC (Smith et al. 2009) and, correspondingly, regions preferring the same category show preferential resting FC (Hutchison et al. 2014; Stevens et al. 2017; Turk-Browne et al. 2010; Zhang et al. 2009; Zhu et al. 2011), we expected that in the current study the regions linked by pattern-based FC would be similar to those linked using standard interregional FC (i.e., FC in which timeseries are averaged over all the voxels/vertices within a region). Accordingly, pattern-based FC matrices were moderately-to-strongly correlated with interregional vertex-averaged FC matrices.

However, if task-evoked activity is measured in paradigms that combine frequently occurring processes from different domains, pattern-based FC that measures interregional fluctuations of the corresponding task-evoked patterns in the resting state may link regions that are normally uncorrelated using standard resting-state methodology. For example, activity patterns based on integration of voice and face information during person-to-person interactions may link visual and auditory regions at rest, activity patterns measured during object manipulation may link sensory and motor regions at rest, and biologically significant stimulus-reward or response-reward contingencies may link sensory and reward-related regions at rest.

We suggested above that relationships between task-evoked and spontaneous activity could help mitigate limitations on the brain’s capacity to process information by incorporating useful information. A related possibility is that resting patterns incorporate synergies that reduce the dimensionality of possible brain states. For example, studies of the motor system indicate that the brain partly deals with the large number of possible hand movements through synergies (Leo et al. 2016; Santello et al. 2013; Schieber and Santello 2004). Studies of human hand movements under experimental and naturalistic conditions (Ingram et al. 2008; Santello et al. 1998) have isolated a small number of principal components that code for large amounts of variance and distinguish a variety of movements. A recent fMRI study reported that these components map onto multivoxel patterns of movement-evoked BOLD activity in regions of motor and premotor cortex, SMA, SPL, and anterior IPS (Leo et al. 2016). Therefore, natural movements can be described by a small number of correlated components or muscle synergies, that involve both proximal and distal movements and control movements concurrently across multiple joints, and are represented in patterns of neural activity at multiple levels in the motor system (Cheung et al. 2009, 2012; Howard et al. 2009; Ingram et al. 2008). These synergies reduce the dimensionality of the space in which movements are programmed, lessening the computational difficulties engendered by the high degrees of freedom associated with arbitrary limb movements. The existence of these modal axes or synergies increases the plausibility that resting activity might encode biologically important representations. More speculatively, cognitive systems might also be described by synergies that link frequently occurring processes to reduce dimensionality and these synergies might be reflected in resting-state patterns.

Low- and High-Level Visual Correspondences at Rest

The spread of similarity values between resting activity patterns and stimulus-evoked patterns was determined by how well a stimulus activated the region, irrespective of whether the stimulus was more or less ecological. In many higher-level visual ROIs, stimulus preferences favored a particular whole-stimulus category (e.g., bodies) over another whole-stimulus category (e.g., scenes) or over the phase-scrambled category. Conversely, in early visual cortex, preferences favored stimuli that weighted low-level features, resulting in larger U90 values for scrambled than whole-stimulus categories. The larger U90 values for scrambled stimuli in early visual cortex do not contradict an overall framework in which resting activity patterns reflect the statistical distribution of features in the environment. Rather, this result suggests that resting activity patterns in regions that primarily extract low-level visual features are relatively uncorrelated with the patterns associated with higher-order features/statistics that define categories of more ecological stimuli.

Grid-scrambled objects showed greater U90 values than whole objects in the phase-scrambled joint-ROI and equivalent U90 values to whole objects in the whole object joint-ROI (Fig. 7). This latter equivalence may have reflected the fact that the union of different category-preferential regions in the whole object joint-ROI eliminated or reduced the importance of features selective for a specific ecological category. Instead, the U90 value reflected features common to different ecological categories that were also present in grid-scrambled objects.

Therefore, the relationship between resting patterns and stimulus-evoked patterns can be driven by a variety of stimulus features that reflect local (e.g., contour-related features) or global (e.g., faces) stimulus characteristics depending on the tested regions.

Limitations

Stimuli were not controlled for low-level variables that might have differentially activated visual regions. As noted, grid-scrambled stimuli may have included contour terminators to a larger degree than many whole object stimuli, increasing the activation of early visual cortex. However, this factor was not explicitly controlled or manipulated. Also, stimuli were presented in a nonnaturalistic context. Wilf et al. (2017) have shown that in early visual cortex, resting FC patterns are better accounted for by movies than by standard retinotopic stimuli, while Strappini et al. (2019) have shown that in higher-level visual cortex, resting FC patterns are better accounted for by movies than by static pictures of stimuli similar to those used here. Therefore, the present results may have underestimated correspondences between resting and evoked multivertex patterns.

GRANTS

This work was supported by the National Institutes of Health RO1 MH096482 to M. Corbetta.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

D.K., T.L., M.C., and G.S. conceived and designed research; D.K. and N.M. performed experiments; D.K. and N.M. analyzed data; D.K., M.C., and G.S. interpreted results of experiments; D.K. prepared figures; D.K. and G.S. drafted manuscript; D.K., M.C., and G.S. edited and revised manuscript; D.K., M.C., and G.S. approved final version of manuscript.

REFERENCES

Anzellotti S, Coutanche MN. Beyond functional connectivity: investigating networks of multivariate representations. Trends Cogn Sci 22: 258–269, 2018. doi: 10.1016/j.tics.2017.12.002. [DOI] [PubMed] [Google Scholar]
Arcaro MJ, Honey CJ, Mruczek RE, Kastner S, Hasson U. Widespread correlation patterns of fMRI signal across visual cortex reflect eccentricity organization. eLife 4: e03952, 2015. doi: 10.7554/eLife.03952. [DOI] [PMC free article] [PubMed] [Google Scholar]
Berkes P, Orbán G, Lengyel M, Fiser J. Spontaneous cortical activity reveals hallmarks of an optimal internal model of the environment. Science 331: 83–87, 2011. doi: 10.1126/science.1195870. [DOI] [PMC free article] [PubMed] [Google Scholar]
Biswal B, Yetkin FZ, Haughton VM, Hyde JS. Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magn Reson Med 34: 537–541, 1995. doi: 10.1002/mrm.1910340409. [DOI] [PubMed] [Google Scholar]
Bracci S, Caramazza A, Peelen MV. Representational similarity of body parts in human occipitotemporal cortex. J Neurosci 35: 12977–12985, 2015. doi: 10.1523/JNEUROSCI.4698-14.2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bracci S, Op de Beeck H. Dissociations and associations between shape and category representations in the two visual pathways. J Neurosci 36: 432–444, 2016. doi: 10.1523/JNEUROSCI.2314-15.2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
Brainard DH. The psychophysics toolbox. Spat Vis 10: 433–436, 1997. doi: 10.1163/156856897X00357. [DOI] [PubMed] [Google Scholar]
Chang L, Tsao DY. The code for facial identity in the primate brain. Cell 169: 1013–1028.e14, 2017. doi: 10.1016/j.cell.2017.05.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chen Q, Garcea FE, Almeida J, Mahon BZ. Connectivity-based constraints on category-specificity in the ventral object processing pathway. Neuropsychologia 105: 184–196, 2017. doi: 10.1016/j.neuropsychologia.2016.11.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chen RH, Ito T, Kulkarni KR, Cole MW. The human brain traverses a common activation-pattern state space across task and rest. Brain Connect 8: 429–443, 2018. doi: 10.1089/brain.2018.0586. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cheung VCK, Piron L, Agostini M, Silvoni S,Turolla A, Bizzi E. Stability of muscle synergies for voluntary actions after cortical stroke in humans. Proc Natl Acad Sci USA 106: 19563–10568, 2009. doi: 10.1073/pnas.0910114106. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cheung VCK, Turolla A, Agostini M, Silvoni S,Bennis C, Kasi P, Paganoni S, Bonato PBizzi E. Muscle synergy patterns as physiological markers of motor cortical damage. Proc Natl Acad Sci USA 109: 14652–14656, 2012. doi: 10.1073/pnas.1212056109. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cole MW, Ito T, Bassett DS, Schultz DH. Activity flow over resting-state networks shapes cognitive task activations. Nat Neurosci 19: 1718–1726, 2016. doi: 10.1038/nn.4406. [DOI] [PMC free article] [PubMed] [Google Scholar]
Coutanche MN, Thompson-Schill SL. Informational connectivity: identifying synchronized discriminability of multi-voxel patterns across the brain. Front Hum Neurosci 7: 15, 2013. doi: 10.3389/fnhum.2013.00015. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dale AM. Optimal experimental design for event-related fMRI. Hum Brain Mapp 8: 109–114, 1999. doi:. [DOI] [PMC free article] [PubMed] [Google Scholar]
de Pasquale F, Della Penna S, Snyder AZ, Lewis C, Mantini D, Marzetti L, Belardinelli P, Ciancetta L, Pizzella V, Romani GL, Corbetta M. Temporal dynamics of spontaneous MEG activity in brain networks. Proc Natl Acad Sci USA 107: 6040–6045, 2010. doi: 10.1073/pnas.0913863107. [DOI] [PMC free article] [PubMed] [Google Scholar]
Deco G, Ponce-Alvarez A, Mantini D, Romani GL, Hagmann P, Corbetta M. Resting-state functional connectivity emerges from structurally and dynamically shaped slow linear fluctuations. J Neurosci 33: 11239–11252, 2013. doi: 10.1523/JNEUROSCI.1091-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
Devereux BJ, Clarke A, Marouchos A, Tyler LK. Representational similarity analysis reveals commonalities and differences in the semantic processing of words and objects. J Neurosci 33: 18906–18916, 2013. doi: 10.1523/JNEUROSCI.3809-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
Downing PE, Chan AW, Peelen MV, Dodds CM, Kanwisher N. Domain specificity in visual cortex. Cereb Cortex 16: 1453–1461, 2006. doi: 10.1093/cercor/bhj086. [DOI] [PubMed] [Google Scholar]
Downing PE, Jiang Y, Shuman M, Kanwisher N. A cortical area selective for visual processing of the human body. Science 293: 2470–2473, 2001. doi: 10.1126/science.1063414. [DOI] [PubMed] [Google Scholar]
Epstein R, Kanwisher N. A cortical representation of the local visual environment. Nature 392: 598–601, 1998. doi: 10.1038/33402. [DOI] [PubMed] [Google Scholar]
Fischl B, Sereno MI, Tootell RB, Dale AM. High-resolution intersubject averaging and a coordinate system for the cortical surface. Hum Brain Mapp 8: 272–284, 1999. doi:. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fiser J, Berkes P, Orbán G, Lengyel M. Statistically optimal perception and learning: from behavior to neural representations. Trends Cogn Sci 14: 119–130, 2010. doi: 10.1016/j.tics.2010.01.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fiser J, Chiu C, Weliky M. Small modulation of ongoing cortical dynamics by sensory input during natural vision. Nature 431: 573–578, 2004. doi: 10.1038/nature02907. [DOI] [PubMed] [Google Scholar]
Fox MD, Snyder AZ, Vincent JL, Raichle ME. Intrinsic fluctuations within cortical systems account for intertrial variability in human behavior. Neuron 56: 171–184, 2007. doi: 10.1016/j.neuron.2007.08.023. [DOI] [PubMed] [Google Scholar]
Fukushima M, Saunders RC, Leopold DA, Mishkin M, Averbeck BB. Spontaneous high-gamma band activity reflects functional organization of auditory cortex in the awake macaque. Neuron 74: 899–910, 2012. doi: 10.1016/j.neuron.2012.04.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
Glasser MF, Sotiropoulos SN, Wilson JA, Coalson TS, Fischl B, Andersson JL, Xu J, Jbabdi S, Webster M, Polimeni JR, Van Essen DC, Jenkinson M; WU-Minn HCP Consortium . The minimal preprocessing pipelines for the Human Connectome Project. Neuroimage 80: 105–124, 2013. doi: 10.1016/j.neuroimage.2013.04.127. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gordon EM, Laumann TO, Adeyemo B, Huckins JF, Kelley WM, Petersen SE. Generation and evaluation of a cortical area parcellation from resting-state correlations. Cereb Cortex 26: 288–303, 2016. doi: 10.1093/cercor/bhu239. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gordon EM, Laumann TO, Gilmore AW, Newbold DJ, Greene DJ, Berg JJ, Ortega M, Hoyt-Drazen C, Gratton C, Sun H, Hampton JM, Coalson RS, Nguyen AL, McDermott KB, Shimony JS, Snyder AZ, Schlaggar BL, Petersen SE, Nelson SM, Dosenbach NUF. Precision functional mapping of individual human brains. Neuron 95: 791–807.e7, 2017. doi: 10.1016/j.neuron.2017.07.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gratton C, Laumann TO, Nielsen AN, Greene DJ, Gordon EM, Gilmore AW, Nelson SM, Coalson RS, Snyder AZ, Schlaggar BL, Dosenbach NUF, Petersen SE. Functional brain networks are dominated by stable group and individual factors, not cognitive or daily variation. Neuron 98: 439–452.e5, 2018. doi: 10.1016/j.neuron.2018.03.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
Grefkes C, Fink GR. Connectivity-based approaches in stroke and recovery of function. Lancet Neurol 13: 206–216, 2014. doi: 10.1016/S1474-4422(13)70264-3. [DOI] [PubMed] [Google Scholar]
Greicius MD, Krasnow B, Reiss AL, Menon V. Functional connectivity in the resting brain: a network analysis of the default mode hypothesis. Proc Natl Acad Sci USA 100: 253–258, 2003. doi: 10.1073/pnas.0135058100. [DOI] [PMC free article] [PubMed] [Google Scholar]
Griffis JC, Metcalf NV, Corbetta M, Shulman GL. Damage to the shortest structural paths between brain regions is associated with disruptions of resting-state functional connectivity after stroke. Neuroimage 210: 116589, 2020. doi: 10.1016/j.neuroimage.2020.116589. [DOI] [PMC free article] [PubMed] [Google Scholar]
Grill-Spector K, Weiner KS. The functional architecture of the ventral temporal cortex and its role in categorization. Nat Rev Neurosci 15: 536–548, 2014. doi: 10.1038/nrn3747. [DOI] [PMC free article] [PubMed] [Google Scholar]
Harmelech T, Malach R. Neurocognitive biases and the patterns of spontaneous correlations in the human cortex. Trends Cogn Sci 17: 606–615, 2013. doi: 10.1016/j.tics.2013.09.014. [DOI] [PubMed] [Google Scholar]
He BJ, Snyder AZ, Zempel JM, Smyth MD, Raichle ME. Electrophysiological correlates of the brain’s intrinsic large-scale functional architecture. Proc Natl Acad Sci USA 105: 16039–16044, 2008. doi: 10.1073/pnas.0807010105. [DOI] [PMC free article] [PubMed] [Google Scholar]
Heinzle J, Kahnt T, Haynes JD. Topographically specific functional connectivity between visual field maps in the human brain. Neuroimage 56: 1426–1436, 2011. doi: 10.1016/j.neuroimage.2011.02.077. [DOI] [PubMed] [Google Scholar]
Howard IS, Ingram JN, Körding KP, Wolpert DM. Statistics of natural movements are reflected in motor errors. J Neurophysiol 102: 1902–1910, 2009. doi: 10.1152/jn.00013.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hutchison RM, Culham JC, Everling S, Flanagan JR, Gallivan JP. Distinct and distributed functional connectivity patterns across cortex reflect the domain-specific constraints of object, face, scene, body, and tool category-selective modules in the ventral visual pathway. Neuroimage 96: 216–236, 2014. doi: 10.1016/j.neuroimage.2014.03.068. [DOI] [PubMed] [Google Scholar]
Ingram JN, Körding KP, Howard IS, Wolpert DM. The statistics of natural hand movements. Exp Brain Res 188: 223–236, 2008. doi: 10.1007/s00221-008-1355-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kanwisher N, McDermott J, Chun MM. The fusiform face area: a module in human extrastriate cortex specialized for face perception. J Neurosci 17: 4302–4311, 1997. doi: 10.1523/JNEUROSCI.17-11-04302.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kenet T, Bibitchkov D, Tsodyks M, Grinvald A, Arieli A. Spontaneously emerging cortical representations of visual attributes. Nature 425: 954–956, 2003. doi: 10.1038/nature02078. [DOI] [PubMed] [Google Scholar]
Konkle T, Caramazza A. The large-scale organization of object-responsive cortex is reflected in resting-state network architecture. Cereb Cortex 27: 4933–4945, 2017. doi: 10.1093/cercor/bhw287. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kriegeskorte N, Mur M, Bandettini P. Representational similarity analysis - connecting the branches of systems neuroscience. Front Syst Neurosci 2: 4, 2008b. doi: 10.3389/neuro.06.004.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
Kriegeskorte N, Mur M, Ruff DA, Kiani R, Bodurka J, Esteky H, Tanaka K, Bandettini PA. Matching categorical object representations in inferior temporal cortex of man and monkey. Neuron 60: 1126–1141, 2008a. doi: 10.1016/j.neuron.2008.10.043. [DOI] [PMC free article] [PubMed] [Google Scholar]
Laumann TO, Gordon EM, Adeyemo B, Snyder AZ, Joo SJ, Chen MY, Gilmore AW, McDermott KB, Nelson SM, Dosenbach NU, Schlaggar BL, Mumford JA, Poldrack RA, Petersen SE. Functional system and areal organization of a highly sampled individual human brain. Neuron 87: 657–670, 2015. doi: 10.1016/j.neuron.2015.06.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
Leo A, Handjaras G, Bianchi M, Marino H, Gabiccini M, Guidi A, Scilingo EP, Pietrini P, Bicchi A, Santello M, Ricciardi E. A synergy-based hand control is encoded in human motor cortical areas. eLife 5: e13420, 2016. doi: 10.7554/eLife.13420. [DOI] [PMC free article] [PubMed] [Google Scholar]
Marcus DS, Harms MP, Snyder AZ, Jenkinson M, Wilson JA, Glasser MF, Barch DM, Archie KA, Burgess GC, Ramaratnam M, Hodge M, Horton W, Herrick R, Olsen T, McKay M, House M, Hileman M, Reid E, Harwell J, Coalson T, Schindler J, Elam JS, Curtiss SW, Van Essen DC; WU-Minn HCP Consortium . Human Connectome Project informatics: quality control, database services, and data visualization. Neuroimage 80: 202–219, 2013. doi: 10.1016/j.neuroimage.2013.05.077. [DOI] [PMC free article] [PubMed] [Google Scholar]
Mumford D. On the computational architecture of the neocortex. II. The role of cortico-cortical loops. Biol Cybern 66: 241–251, 1992. doi: 10.1007/BF00198477. [DOI] [PubMed] [Google Scholar]
Northoff G, Duncan NW. How do abnormalities in the brain’s spontaneous activity translate into symptoms in schizophrenia? From an overview of resting state activity findings to a proposed spatiotemporal psychopathology. Prog Neurobiol 145-146: 26–45, 2016. doi: 10.1016/j.pneurobio.2016.08.003. [DOI] [PubMed] [Google Scholar]
Omer DB, Fekete T, Ulchin Y, Hildesheim R, Grinvald A. Dynamic patterns of spontaneous ongoing activity in the visual cortex of anesthetized and awake monkeys are different. Cereb Cortex 29: 1291–1304, 2019. doi: 10.1093/cercor/bhy099. [DOI] [PubMed] [Google Scholar]
Oosterhof NN, Tipper SP, Downing PE. Viewpoint (in)dependence of action representations: an MVPA study. J Cogn Neurosci 24: 975–989, 2012. doi: 10.1162/jocn_a_00195. [DOI] [PubMed] [Google Scholar]
Osher DE, Brissenden JA, Somers DC. Predicting an individual’s dorsal attention network activity from functional connectivity fingerprints. J Neurophysiol 122: 232–240, 2019. doi: 10.1152/jn.00174.2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
Petersen SE, Sporns O. Brain networks and cognitive architectures. Neuron 88: 207–219, 2015. doi: 10.1016/j.neuron.2015.09.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
Power JD, Cohen AL, Nelson SM, Wig GS, Barnes KA, Church JA, Vogel AC, Laumann TO, Miezin FM, Schlaggar BL, Petersen SE. Functional network organization of the human brain. Neuron 72: 665–678, 2011. doi: 10.1016/j.neuron.2011.09.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
Power JD, Mitra A, Laumann TO, Snyder AZ, Schlaggar BL, Petersen SE. Methods to detect, characterize, and remove motion artifact in resting state fMRI. Neuroimage 84: 320–341, 2014. doi: 10.1016/j.neuroimage.2013.08.048. [DOI] [PMC free article] [PubMed] [Google Scholar]
Raemaekers M, Schellekens W, van Wezel RJ, Petridou N, Kristo G, Ramsey NF. Patterns of resting state connectivity in human primary visual cortical areas: a 7T fMRI study. Neuroimage 84: 911–921, 2014. doi: 10.1016/j.neuroimage.2013.09.060. [DOI] [PubMed] [Google Scholar]
Raichle ME. The restless brain. Brain Connect 1: 3–12, 2011. doi: 10.1089/brain.2011.0019. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ryu J, Lee SH. Stimulus-tuned structure of correlated fMRI activity in human visual cortex. Cereb Cortex 28: 693–712, 2018. doi: 10.1093/cercor/bhw411. [DOI] [PubMed] [Google Scholar]
Santello M, Baud-Bovy G, Jörntell H. Neural bases of hand synergies. Front Comput Neurosci 7: 23, 2013. doi: 10.3389/fncom.2013.00023. [DOI] [PMC free article] [PubMed] [Google Scholar]
Santello M, Flanders M, Soechting JF. Postural hand synergies for tool use. J Neurosci 18: 10105–10115, 1998. doi: 10.1523/JNEUROSCI.18-23-10105.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
Schieber MH, Santello M. Hand function: peripheral and central constraints on performance. J Appl Physiol (1985) 96: 2293–2300, 2004. doi: 10.1152/japplphysiol.01063.2003. [DOI] [PubMed] [Google Scholar]
Siegel JS, Ramsey LE, Snyder AZ, Metcalf NV, Chacko RV, Weinberger K, Baldassarre A, Hacker CD, Shulman GL, Corbetta M. Disruptions of network connectivity predict impairment in multiple behavioral domains after stroke. Proc Natl Acad Sci USA 113: E4367–E4376, 2016. doi: 10.1073/pnas.1521083113. [DOI] [PMC free article] [PubMed] [Google Scholar]
Smith SM, Fox PT, Miller KL, Glahn DC, Fox PM, Mackay CE, Filippini N, Watkins KE, Toro R, Laird AR, Beckmann CF. Correspondence of the brain’s functional architecture during activation and rest. Proc Natl Acad Sci USA 106: 13040–13045, 2009. doi: 10.1073/pnas.0905267106. [DOI] [PMC free article] [PubMed] [Google Scholar]
Stevens WD, Kravitz DJ, Peng CS, Tessler MH, Martin A. Privileged functional connectivity between the visual word form area and the language system. J Neurosci 37: 5288–5297, 2017. doi: 10.1523/JNEUROSCI.0138-17.2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
Strappini F, Wilf M, Karp O, Goldberg H, Harel M, Furman-Haran E, Golan T, Malach R. Resting-state activity in high-order visual areas as a window into natural human brain activations. Cereb Cortex 29: 3618–3635, 2019. doi: 10.1093/cercor/bhy242. [DOI] [PubMed] [Google Scholar]
Tavor I, Parker Jones O, Mars RB, Smith SM, Behrens TE, Jbabdi S. Task-free MRI predicts individual differences in brain activity during task performance. Science 352: 216–220, 2016. doi: 10.1126/science.aad8127. [DOI] [PMC free article] [PubMed] [Google Scholar]
Teufel C, Fletcher PC. Forms of prediction in the nervous system. Nat Rev Neurosci 21: 231–242, 2020. [Erratum in Nat Rev Neurosci 21: 297, 2020]. doi: 10.1038/s41583-020-0275-5. [DOI] [PubMed] [Google Scholar]
Tsodyks M, Kenet T, Grinvald A, Arieli A. Linking spontaneous activity of single cortical neurons and the underlying functional architecture. Science 286: 1943–1946, 1999. doi: 10.1126/science.286.5446.1943. [DOI] [PubMed] [Google Scholar]
Turk-Browne NB, Norman-Haignere SV, McCarthy G. Face-specific resting functional connectivity between the fusiform gyrus and posterior superior temporal sulcus. Front Hum Neurosci 4: 176, 2010. doi: 10.3389/fnhum.2010.00176. [DOI] [PMC free article] [PubMed] [Google Scholar]
Van Essen DC, Drury HA, Dickson J, Harwell J, Hanlon D, Anderson CH. An integrated software suite for surface-based analyses of cerebral cortex. J Am Med Inform Assoc 8: 443–459, 2001. doi: 10.1136/jamia.2001.0080443. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wang L, Mruczek RE, Arcaro MJ, Kastner S. Probabilistic maps of visual topography in human cortex. Cereb Cortex 25: 3911–3931, 2015. doi: 10.1093/cercor/bhu277. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wang X, Zhen Z, Song Y, Huang L, Kong X, Liu J. The hierarchical structure of the face network revealed by its functional connectivity pattern. J Neurosci 36: 890–900, 2016. doi: 10.1523/JNEUROSCI.2789-15.2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
Watson DM, Hymers M, Hartley T, Andrews TJ. Patterns of neural response in scene-selective regions of the human brain are affected by low-level manipulations of spatial frequency. Neuroimage 124, Pt A: 107–117, 2016. doi: 10.1016/j.neuroimage.2015.08.058. [DOI] [PubMed] [Google Scholar]
Wilf M, Strappini F, Golan T, Hahamy A, Harel M, Malach R. Spontaneously emerging patterns in human visual cortex reflect responses to naturalistic sensory stimuli. Cereb Cortex 27: 750–763, 2017. doi: 10.1093/cercor/bhv275. [DOI] [PubMed] [Google Scholar]
Wurm MF, Ariani G, Greenlee MW, Lingnau A. Decoding concrete and abstract action representations during explicit and implicit conceptual processing. Cereb Cortex 26: 3390–3401, 2016. doi: 10.1093/cercor/bhv169. [DOI] [PubMed] [Google Scholar]
Yeo BT, Krienen FM, Sepulcre J, Sabuncu MR, Lashkari D, Hollinshead M, Roffman JL, Smoller JW, Zöllei L, Polimeni JR, Fischl B, Liu H, Buckner RL. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J Neurophysiol 106: 1125–1165, 2011. doi: 10.1152/jn.00338.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
Zhang H, Tian J, Liu J, Li J, Lee K. Intrinsically organized network for face perception during the resting state. Neurosci Lett 454: 1–5, 2009. doi: 10.1016/j.neulet.2009.02.054. [DOI] [PMC free article] [PubMed] [Google Scholar]
Zhu Q, Zhang J, Luo YL, Dilks DD, Liu J. Resting-state neural activity across face-selective cortical regions is behaviorally relevant. J Neurosci 31: 10323–10330, 2011. doi: 10.1523/JNEUROSCI.0873-11.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B1] Anzellotti S, Coutanche MN. Beyond functional connectivity: investigating networks of multivariate representations. Trends Cogn Sci 22: 258–269, 2018. doi: 10.1016/j.tics.2017.12.002. [DOI] [PubMed] [Google Scholar]

[B2] Arcaro MJ, Honey CJ, Mruczek RE, Kastner S, Hasson U. Widespread correlation patterns of fMRI signal across visual cortex reflect eccentricity organization. eLife 4: e03952, 2015. doi: 10.7554/eLife.03952. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B3] Berkes P, Orbán G, Lengyel M, Fiser J. Spontaneous cortical activity reveals hallmarks of an optimal internal model of the environment. Science 331: 83–87, 2011. doi: 10.1126/science.1195870. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B4] Biswal B, Yetkin FZ, Haughton VM, Hyde JS. Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magn Reson Med 34: 537–541, 1995. doi: 10.1002/mrm.1910340409. [DOI] [PubMed] [Google Scholar]

[B5] Bracci S, Caramazza A, Peelen MV. Representational similarity of body parts in human occipitotemporal cortex. J Neurosci 35: 12977–12985, 2015. doi: 10.1523/JNEUROSCI.4698-14.2015. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B6] Bracci S, Op de Beeck H. Dissociations and associations between shape and category representations in the two visual pathways. J Neurosci 36: 432–444, 2016. doi: 10.1523/JNEUROSCI.2314-15.2016. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B7] Brainard DH. The psychophysics toolbox. Spat Vis 10: 433–436, 1997. doi: 10.1163/156856897X00357. [DOI] [PubMed] [Google Scholar]

[B8] Chang L, Tsao DY. The code for facial identity in the primate brain. Cell 169: 1013–1028.e14, 2017. doi: 10.1016/j.cell.2017.05.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B9] Chen Q, Garcea FE, Almeida J, Mahon BZ. Connectivity-based constraints on category-specificity in the ventral object processing pathway. Neuropsychologia 105: 184–196, 2017. doi: 10.1016/j.neuropsychologia.2016.11.014. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B10] Chen RH, Ito T, Kulkarni KR, Cole MW. The human brain traverses a common activation-pattern state space across task and rest. Brain Connect 8: 429–443, 2018. doi: 10.1089/brain.2018.0586. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B10a] Cheung VCK, Piron L, Agostini M, Silvoni S,Turolla A, Bizzi E. Stability of muscle synergies for voluntary actions after cortical stroke in humans. Proc Natl Acad Sci USA 106: 19563–10568, 2009. doi: 10.1073/pnas.0910114106. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B10b] Cheung VCK, Turolla A, Agostini M, Silvoni S,Bennis C, Kasi P, Paganoni S, Bonato PBizzi E. Muscle synergy patterns as physiological markers of motor cortical damage. Proc Natl Acad Sci USA 109: 14652–14656, 2012. doi: 10.1073/pnas.1212056109. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B11] Cole MW, Ito T, Bassett DS, Schultz DH. Activity flow over resting-state networks shapes cognitive task activations. Nat Neurosci 19: 1718–1726, 2016. doi: 10.1038/nn.4406. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B12] Coutanche MN, Thompson-Schill SL. Informational connectivity: identifying synchronized discriminability of multi-voxel patterns across the brain. Front Hum Neurosci 7: 15, 2013. doi: 10.3389/fnhum.2013.00015. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B13] Dale AM. Optimal experimental design for event-related fMRI. Hum Brain Mapp 8: 109–114, 1999. doi:. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B14] de Pasquale F, Della Penna S, Snyder AZ, Lewis C, Mantini D, Marzetti L, Belardinelli P, Ciancetta L, Pizzella V, Romani GL, Corbetta M. Temporal dynamics of spontaneous MEG activity in brain networks. Proc Natl Acad Sci USA 107: 6040–6045, 2010. doi: 10.1073/pnas.0913863107. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B15] Deco G, Ponce-Alvarez A, Mantini D, Romani GL, Hagmann P, Corbetta M. Resting-state functional connectivity emerges from structurally and dynamically shaped slow linear fluctuations. J Neurosci 33: 11239–11252, 2013. doi: 10.1523/JNEUROSCI.1091-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B16] Devereux BJ, Clarke A, Marouchos A, Tyler LK. Representational similarity analysis reveals commonalities and differences in the semantic processing of words and objects. J Neurosci 33: 18906–18916, 2013. doi: 10.1523/JNEUROSCI.3809-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B17] Downing PE, Chan AW, Peelen MV, Dodds CM, Kanwisher N. Domain specificity in visual cortex. Cereb Cortex 16: 1453–1461, 2006. doi: 10.1093/cercor/bhj086. [DOI] [PubMed] [Google Scholar]

[B18] Downing PE, Jiang Y, Shuman M, Kanwisher N. A cortical area selective for visual processing of the human body. Science 293: 2470–2473, 2001. doi: 10.1126/science.1063414. [DOI] [PubMed] [Google Scholar]

[B19] Epstein R, Kanwisher N. A cortical representation of the local visual environment. Nature 392: 598–601, 1998. doi: 10.1038/33402. [DOI] [PubMed] [Google Scholar]

[B20] Fischl B, Sereno MI, Tootell RB, Dale AM. High-resolution intersubject averaging and a coordinate system for the cortical surface. Hum Brain Mapp 8: 272–284, 1999. doi:. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B21] Fiser J, Berkes P, Orbán G, Lengyel M. Statistically optimal perception and learning: from behavior to neural representations. Trends Cogn Sci 14: 119–130, 2010. doi: 10.1016/j.tics.2010.01.003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B22] Fiser J, Chiu C, Weliky M. Small modulation of ongoing cortical dynamics by sensory input during natural vision. Nature 431: 573–578, 2004. doi: 10.1038/nature02907. [DOI] [PubMed] [Google Scholar]

[B23] Fox MD, Snyder AZ, Vincent JL, Raichle ME. Intrinsic fluctuations within cortical systems account for intertrial variability in human behavior. Neuron 56: 171–184, 2007. doi: 10.1016/j.neuron.2007.08.023. [DOI] [PubMed] [Google Scholar]

[B24] Fukushima M, Saunders RC, Leopold DA, Mishkin M, Averbeck BB. Spontaneous high-gamma band activity reflects functional organization of auditory cortex in the awake macaque. Neuron 74: 899–910, 2012. doi: 10.1016/j.neuron.2012.04.014. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B25] Glasser MF, Sotiropoulos SN, Wilson JA, Coalson TS, Fischl B, Andersson JL, Xu J, Jbabdi S, Webster M, Polimeni JR, Van Essen DC, Jenkinson M; WU-Minn HCP Consortium . The minimal preprocessing pipelines for the Human Connectome Project. Neuroimage 80: 105–124, 2013. doi: 10.1016/j.neuroimage.2013.04.127. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B26] Gordon EM, Laumann TO, Adeyemo B, Huckins JF, Kelley WM, Petersen SE. Generation and evaluation of a cortical area parcellation from resting-state correlations. Cereb Cortex 26: 288–303, 2016. doi: 10.1093/cercor/bhu239. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B27] Gordon EM, Laumann TO, Gilmore AW, Newbold DJ, Greene DJ, Berg JJ, Ortega M, Hoyt-Drazen C, Gratton C, Sun H, Hampton JM, Coalson RS, Nguyen AL, McDermott KB, Shimony JS, Snyder AZ, Schlaggar BL, Petersen SE, Nelson SM, Dosenbach NUF. Precision functional mapping of individual human brains. Neuron 95: 791–807.e7, 2017. doi: 10.1016/j.neuron.2017.07.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B28] Gratton C, Laumann TO, Nielsen AN, Greene DJ, Gordon EM, Gilmore AW, Nelson SM, Coalson RS, Snyder AZ, Schlaggar BL, Dosenbach NUF, Petersen SE. Functional brain networks are dominated by stable group and individual factors, not cognitive or daily variation. Neuron 98: 439–452.e5, 2018. doi: 10.1016/j.neuron.2018.03.035. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B29] Grefkes C, Fink GR. Connectivity-based approaches in stroke and recovery of function. Lancet Neurol 13: 206–216, 2014. doi: 10.1016/S1474-4422(13)70264-3. [DOI] [PubMed] [Google Scholar]

[B30] Greicius MD, Krasnow B, Reiss AL, Menon V. Functional connectivity in the resting brain: a network analysis of the default mode hypothesis. Proc Natl Acad Sci USA 100: 253–258, 2003. doi: 10.1073/pnas.0135058100. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B31] Griffis JC, Metcalf NV, Corbetta M, Shulman GL. Damage to the shortest structural paths between brain regions is associated with disruptions of resting-state functional connectivity after stroke. Neuroimage 210: 116589, 2020. doi: 10.1016/j.neuroimage.2020.116589. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B32] Grill-Spector K, Weiner KS. The functional architecture of the ventral temporal cortex and its role in categorization. Nat Rev Neurosci 15: 536–548, 2014. doi: 10.1038/nrn3747. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B33] Harmelech T, Malach R. Neurocognitive biases and the patterns of spontaneous correlations in the human cortex. Trends Cogn Sci 17: 606–615, 2013. doi: 10.1016/j.tics.2013.09.014. [DOI] [PubMed] [Google Scholar]

[B34] He BJ, Snyder AZ, Zempel JM, Smyth MD, Raichle ME. Electrophysiological correlates of the brain’s intrinsic large-scale functional architecture. Proc Natl Acad Sci USA 105: 16039–16044, 2008. doi: 10.1073/pnas.0807010105. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B35] Heinzle J, Kahnt T, Haynes JD. Topographically specific functional connectivity between visual field maps in the human brain. Neuroimage 56: 1426–1436, 2011. doi: 10.1016/j.neuroimage.2011.02.077. [DOI] [PubMed] [Google Scholar]

[B35a] Howard IS, Ingram JN, Körding KP, Wolpert DM. Statistics of natural movements are reflected in motor errors. J Neurophysiol 102: 1902–1910, 2009. doi: 10.1152/jn.00013.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B36] Hutchison RM, Culham JC, Everling S, Flanagan JR, Gallivan JP. Distinct and distributed functional connectivity patterns across cortex reflect the domain-specific constraints of object, face, scene, body, and tool category-selective modules in the ventral visual pathway. Neuroimage 96: 216–236, 2014. doi: 10.1016/j.neuroimage.2014.03.068. [DOI] [PubMed] [Google Scholar]

[B36a] Ingram JN, Körding KP, Howard IS, Wolpert DM. The statistics of natural hand movements. Exp Brain Res 188: 223–236, 2008. doi: 10.1007/s00221-008-1355-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B37] Kanwisher N, McDermott J, Chun MM. The fusiform face area: a module in human extrastriate cortex specialized for face perception. J Neurosci 17: 4302–4311, 1997. doi: 10.1523/JNEUROSCI.17-11-04302.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B38] Kenet T, Bibitchkov D, Tsodyks M, Grinvald A, Arieli A. Spontaneously emerging cortical representations of visual attributes. Nature 425: 954–956, 2003. doi: 10.1038/nature02078. [DOI] [PubMed] [Google Scholar]

[B39] Konkle T, Caramazza A. The large-scale organization of object-responsive cortex is reflected in resting-state network architecture. Cereb Cortex 27: 4933–4945, 2017. doi: 10.1093/cercor/bhw287. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B40] Kriegeskorte N, Mur M, Bandettini P. Representational similarity analysis - connecting the branches of systems neuroscience. Front Syst Neurosci 2: 4, 2008b. doi: 10.3389/neuro.06.004.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B41] Kriegeskorte N, Mur M, Ruff DA, Kiani R, Bodurka J, Esteky H, Tanaka K, Bandettini PA. Matching categorical object representations in inferior temporal cortex of man and monkey. Neuron 60: 1126–1141, 2008a. doi: 10.1016/j.neuron.2008.10.043. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B42] Laumann TO, Gordon EM, Adeyemo B, Snyder AZ, Joo SJ, Chen MY, Gilmore AW, McDermott KB, Nelson SM, Dosenbach NU, Schlaggar BL, Mumford JA, Poldrack RA, Petersen SE. Functional system and areal organization of a highly sampled individual human brain. Neuron 87: 657–670, 2015. doi: 10.1016/j.neuron.2015.06.037. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B43] Leo A, Handjaras G, Bianchi M, Marino H, Gabiccini M, Guidi A, Scilingo EP, Pietrini P, Bicchi A, Santello M, Ricciardi E. A synergy-based hand control is encoded in human motor cortical areas. eLife 5: e13420, 2016. doi: 10.7554/eLife.13420. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B44] Marcus DS, Harms MP, Snyder AZ, Jenkinson M, Wilson JA, Glasser MF, Barch DM, Archie KA, Burgess GC, Ramaratnam M, Hodge M, Horton W, Herrick R, Olsen T, McKay M, House M, Hileman M, Reid E, Harwell J, Coalson T, Schindler J, Elam JS, Curtiss SW, Van Essen DC; WU-Minn HCP Consortium . Human Connectome Project informatics: quality control, database services, and data visualization. Neuroimage 80: 202–219, 2013. doi: 10.1016/j.neuroimage.2013.05.077. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B45] Mumford D. On the computational architecture of the neocortex. II. The role of cortico-cortical loops. Biol Cybern 66: 241–251, 1992. doi: 10.1007/BF00198477. [DOI] [PubMed] [Google Scholar]

[B46] Northoff G, Duncan NW. How do abnormalities in the brain’s spontaneous activity translate into symptoms in schizophrenia? From an overview of resting state activity findings to a proposed spatiotemporal psychopathology. Prog Neurobiol 145-146: 26–45, 2016. doi: 10.1016/j.pneurobio.2016.08.003. [DOI] [PubMed] [Google Scholar]

[B47] Omer DB, Fekete T, Ulchin Y, Hildesheim R, Grinvald A. Dynamic patterns of spontaneous ongoing activity in the visual cortex of anesthetized and awake monkeys are different. Cereb Cortex 29: 1291–1304, 2019. doi: 10.1093/cercor/bhy099. [DOI] [PubMed] [Google Scholar]

[B48] Oosterhof NN, Tipper SP, Downing PE. Viewpoint (in)dependence of action representations: an MVPA study. J Cogn Neurosci 24: 975–989, 2012. doi: 10.1162/jocn_a_00195. [DOI] [PubMed] [Google Scholar]

[B49] Osher DE, Brissenden JA, Somers DC. Predicting an individual’s dorsal attention network activity from functional connectivity fingerprints. J Neurophysiol 122: 232–240, 2019. doi: 10.1152/jn.00174.2019. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B50] Petersen SE, Sporns O. Brain networks and cognitive architectures. Neuron 88: 207–219, 2015. doi: 10.1016/j.neuron.2015.09.027. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B51] Power JD, Cohen AL, Nelson SM, Wig GS, Barnes KA, Church JA, Vogel AC, Laumann TO, Miezin FM, Schlaggar BL, Petersen SE. Functional network organization of the human brain. Neuron 72: 665–678, 2011. doi: 10.1016/j.neuron.2011.09.006. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B52] Power JD, Mitra A, Laumann TO, Snyder AZ, Schlaggar BL, Petersen SE. Methods to detect, characterize, and remove motion artifact in resting state fMRI. Neuroimage 84: 320–341, 2014. doi: 10.1016/j.neuroimage.2013.08.048. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B53] Raemaekers M, Schellekens W, van Wezel RJ, Petridou N, Kristo G, Ramsey NF. Patterns of resting state connectivity in human primary visual cortical areas: a 7T fMRI study. Neuroimage 84: 911–921, 2014. doi: 10.1016/j.neuroimage.2013.09.060. [DOI] [PubMed] [Google Scholar]

[B54] Raichle ME. The restless brain. Brain Connect 1: 3–12, 2011. doi: 10.1089/brain.2011.0019. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B55] Ryu J, Lee SH. Stimulus-tuned structure of correlated fMRI activity in human visual cortex. Cereb Cortex 28: 693–712, 2018. doi: 10.1093/cercor/bhw411. [DOI] [PubMed] [Google Scholar]

[B56] Santello M, Baud-Bovy G, Jörntell H. Neural bases of hand synergies. Front Comput Neurosci 7: 23, 2013. doi: 10.3389/fncom.2013.00023. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B56a] Santello M, Flanders M, Soechting JF. Postural hand synergies for tool use. J Neurosci 18: 10105–10115, 1998. doi: 10.1523/JNEUROSCI.18-23-10105.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B57] Schieber MH, Santello M. Hand function: peripheral and central constraints on performance. J Appl Physiol (1985) 96: 2293–2300, 2004. doi: 10.1152/japplphysiol.01063.2003. [DOI] [PubMed] [Google Scholar]

[B58] Siegel JS, Ramsey LE, Snyder AZ, Metcalf NV, Chacko RV, Weinberger K, Baldassarre A, Hacker CD, Shulman GL, Corbetta M. Disruptions of network connectivity predict impairment in multiple behavioral domains after stroke. Proc Natl Acad Sci USA 113: E4367–E4376, 2016. doi: 10.1073/pnas.1521083113. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B59] Smith SM, Fox PT, Miller KL, Glahn DC, Fox PM, Mackay CE, Filippini N, Watkins KE, Toro R, Laird AR, Beckmann CF. Correspondence of the brain’s functional architecture during activation and rest. Proc Natl Acad Sci USA 106: 13040–13045, 2009. doi: 10.1073/pnas.0905267106. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B60] Stevens WD, Kravitz DJ, Peng CS, Tessler MH, Martin A. Privileged functional connectivity between the visual word form area and the language system. J Neurosci 37: 5288–5297, 2017. doi: 10.1523/JNEUROSCI.0138-17.2017. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B61] Strappini F, Wilf M, Karp O, Goldberg H, Harel M, Furman-Haran E, Golan T, Malach R. Resting-state activity in high-order visual areas as a window into natural human brain activations. Cereb Cortex 29: 3618–3635, 2019. doi: 10.1093/cercor/bhy242. [DOI] [PubMed] [Google Scholar]

[B62] Tavor I, Parker Jones O, Mars RB, Smith SM, Behrens TE, Jbabdi S. Task-free MRI predicts individual differences in brain activity during task performance. Science 352: 216–220, 2016. doi: 10.1126/science.aad8127. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B63] Teufel C, Fletcher PC. Forms of prediction in the nervous system. Nat Rev Neurosci 21: 231–242, 2020. [Erratum in Nat Rev Neurosci 21: 297, 2020]. doi: 10.1038/s41583-020-0275-5. [DOI] [PubMed] [Google Scholar]

[B64] Tsodyks M, Kenet T, Grinvald A, Arieli A. Linking spontaneous activity of single cortical neurons and the underlying functional architecture. Science 286: 1943–1946, 1999. doi: 10.1126/science.286.5446.1943. [DOI] [PubMed] [Google Scholar]

[B65] Turk-Browne NB, Norman-Haignere SV, McCarthy G. Face-specific resting functional connectivity between the fusiform gyrus and posterior superior temporal sulcus. Front Hum Neurosci 4: 176, 2010. doi: 10.3389/fnhum.2010.00176. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B66] Van Essen DC, Drury HA, Dickson J, Harwell J, Hanlon D, Anderson CH. An integrated software suite for surface-based analyses of cerebral cortex. J Am Med Inform Assoc 8: 443–459, 2001. doi: 10.1136/jamia.2001.0080443. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B67] Wang L, Mruczek RE, Arcaro MJ, Kastner S. Probabilistic maps of visual topography in human cortex. Cereb Cortex 25: 3911–3931, 2015. doi: 10.1093/cercor/bhu277. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B68] Wang X, Zhen Z, Song Y, Huang L, Kong X, Liu J. The hierarchical structure of the face network revealed by its functional connectivity pattern. J Neurosci 36: 890–900, 2016. doi: 10.1523/JNEUROSCI.2789-15.2016. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B69] Watson DM, Hymers M, Hartley T, Andrews TJ. Patterns of neural response in scene-selective regions of the human brain are affected by low-level manipulations of spatial frequency. Neuroimage 124, Pt A: 107–117, 2016. doi: 10.1016/j.neuroimage.2015.08.058. [DOI] [PubMed] [Google Scholar]

[B70] Wilf M, Strappini F, Golan T, Hahamy A, Harel M, Malach R. Spontaneously emerging patterns in human visual cortex reflect responses to naturalistic sensory stimuli. Cereb Cortex 27: 750–763, 2017. doi: 10.1093/cercor/bhv275. [DOI] [PubMed] [Google Scholar]

[B71] Wurm MF, Ariani G, Greenlee MW, Lingnau A. Decoding concrete and abstract action representations during explicit and implicit conceptual processing. Cereb Cortex 26: 3390–3401, 2016. doi: 10.1093/cercor/bhv169. [DOI] [PubMed] [Google Scholar]

[B72] Yeo BT, Krienen FM, Sepulcre J, Sabuncu MR, Lashkari D, Hollinshead M, Roffman JL, Smoller JW, Zöllei L, Polimeni JR, Fischl B, Liu H, Buckner RL. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J Neurophysiol 106: 1125–1165, 2011. doi: 10.1152/jn.00338.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B73] Zhang H, Tian J, Liu J, Li J, Lee K. Intrinsically organized network for face perception during the resting state. Neurosci Lett 454: 1–5, 2009. doi: 10.1016/j.neulet.2009.02.054. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B74] Zhu Q, Zhang J, Luo YL, Dilks DD, Liu J. Resting-state neural activity across face-selective cortical regions is behaviorally relevant. J Neurosci 31: 10323–10330, 2011. doi: 10.1523/JNEUROSCI.0873-11.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Spontaneously emerging patterns in human visual cortex and their functional connectivity are linked to the patterns evoked by visual stimuli

DoHyun Kim

Tomer Livne

Nicholas V Metcalf

Maurizio Corbetta

Gordon L Shulman

Series information

Abstract

INTRODUCTION

Fig. 1.

MATERIALS AND METHODS

Participants

Stimuli

Scanning Procedure

Fig. 2.

Resting-state runs.

Localizer runs.

Task runs.

Imaging Parameters and fMRI Preprocessing

fMRI Preprocessing

Defining ROIs from Localizer Activation Contrasts

Fig. 3.

Table 1.

Fig. 4.

Task Scans: Estimation of Multivertex Activation Patterns

Task Scans: Representational Similarity Analysis of Stimulus-Evoked Patterns

Fig. 5.

Determining Similarity of Resting Multivertex Patterns and Stimulus-Evoked Patterns

Fig. 6.

Fig. 7.

Fig. 8.

Statistical Analysis of U90 Values

Pattern-Based Resting Functional Connectivity

Fig. 9.

Fig. 10.

RESULTS

Localization of Regions with Visual Category Preferences

Representational Similarity Analysis of Task-Evoked Patterns

Table 2.

Stimulus-Evoked-to-Rest Pattern Similarity Analysis in Category-Preferring Regions

Stimulus-Evoked-to-Rest Pattern Similarity Analysis in Regions Preferring Whole Versus Phase-Scrambled Objects

U90 Values Correlate with Activation Strength

Pattern-Based Functional Connectivity at Rest

The Relation of Pattern-Based FC in Constituent ROIs to U90 Values in Joint-ROIs

DISCUSSION

Spontaneous Activity Patterns for Objects and Features in Visual Cortex

Synchronous Fluctuations of Representational Content

Low- and High-Level Visual Correspondences at Rest

Limitations

GRANTS

DISCLOSURES

AUTHOR CONTRIBUTIONS

REFERENCES

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases