GINNA, a 33 resting-state networks atlas with meta-analytic decoding-based cognitive characterization

doi:10.21203/rs.3.rs-4803512/v1

Download PDF

Article

GINNA, a 33 resting-state networks atlas with meta-analytic decoding-based cognitive characterization

https://doi.org/10.21203/rs.3.rs-4803512/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Since resting-state networks were first observed using magnetic resonance imaging (MRI), their cognitive relevance has been widely suggested. These networks have often been labeled based on their visual resemblance to task activation networks, suggesting possible functional equivalence. However, to date, the empirical cognitive characterization of these networks has been limited. The present study introduces the Groupe d’Imagerie Neurofonctionnelle Network Atlas, a comprehensive brain atlas featuring 33 resting-state networks. Based on the resting-state data of 1812 participants, the atlas was developed by classifying independent components extracted individually, ensuring that the GINNA networks are consistently detected across subjects. We further explored the cognitive relevance of each GINNA network using meta-analytic decoding and generative null hypothesis testing, linking each network with cognitive terms derived from Neurosynth meta-analytic maps. Six independent authors then assigned one or two cognitive processes to each network based on significant terms. The GINNA atlas showcases a diverse range of topological profiles, including cortical, subcortical, and cerebellar gray matter, reflecting a broad spectrum of the known human cognitive repertoire. The processes associated with each network are named according to the standard Cognitive Atlas ontology, informed by two decades of task-related functional magnetic resonance imaging, thus providing opportunities for empirical validation.

Biological sciences/Neuroscience/Cognitive neuroscience

Biological sciences/Neuroscience/Cognitive neuroscience/Attention

Biological sciences/Neuroscience/Cognitive neuroscience/Cognitive control

Biological sciences/Neuroscience/Cognitive neuroscience/Language

Biological sciences/Neuroscience/Cognitive neuroscience/Perception

fMRI

resting state networks

functional connectivity

functional decoding

meta-analytic decoding

atlasing

The brain is intrinsically organized into sets of tightly coupled brain regions or networks. Using resting-state functional magnetic resonance imaging (rs-fMRI), it has been observed that distant brain regions display synchrony in their low-frequency spontaneous fluctuations of the blood oxygen level-dependent (BOLD) signal (Biswal et al., 1995). Understanding the functional role of these so-called resting-state networks (RSNs) remains a central goal of cognitive neuroscience. It has long been posited that the coordinated activity of distributed brain regions supports cognition (Goldman-Rakic, 1988; McIntosh, 2000; Mesulam, 1990). Since their first observations, it has been noted that RSNs follow an organization reflecting the boundaries of the main cognitive systems observed during tasks - e.g., somatomotor (Biswal et al., 1995), vision (Hampson et al., 2004), episodic memory (Vincent et al., 2006), language (Cordes et al., 2000) - suggesting that RSNs may represent functionally relevant systems (Damoiseaux et al., 2006; De Luca et al., 2006; Fox & Raichle, 2007). However, while it has been suggested that networks should be named according to an anatomically grounded taxonomy (Uddin et al., 2019), researchers tend to name networks according to their putative cognitive functions (Uddin et al., 2023).

Initially, the brain at rest has been proposed to be intrinsically segregated into two anticorrelated extrinsic, “task-positive”, and intrinsic, “task-negative” systems (Fox et al., 2005). Later, it was shown that these two systems could be hierarchically decomposed into 5 modules, thought to subserve distinct functions: switching/control, sensory/motor/attentional, visual, manipulation/maintenance of information, and the emergence of spontaneous thoughts (Doucet et al., 2011). Finally, the currently most widely used atlas proposes a segregation into 7 functional networks: visual, somatomotor, dorsal attention, ventral attention (salience), limbic, frontoparietal (control), and default network (Yeo et al., 2011). Yet, the functional relevance of RSNs remains indirectly established: such cognitive functions of networks have been inferred because of a visual similarity with task-based networks. This method of deducing a functional role for a given network based on visual similarity alone has been shown to be poorly reliable, even when performed by neuroimaging specialists (Uddin et al., 2023). This can be particularly problematic given that these inferred cognitive functions are frequently used to interpret results. Further complicating the association of cognitive functions to resting-state networks, it has been suggested that there may not be a clear one-to-one mapping between large-scale networks and cognitive processes with the low granularity (i.e., 7 networks) typically observed in RSN atlases (Thompson & Fransson, 2017).

Nonetheless, several lines of empirical evidence have supported the cognitive relevance of RSNs. The first direct link between the organization of the resting brain and the brain undergoing tasks was made by Smith et al. (2009) showing that networks extracted using independent component analysis (ICA) at rest closely spatially matched networks extracted using the same methodology from the BrainMap activation maps database (Fox et al., 2005). Since this seminal observation, numerous studies have reinforced the link between resting-state and task-based network architectures, providing evidence that the network architecture observed during tasks is shaped by resting-state networks’ architecture (Cole et al., 2014, 2016), or that resting-state networks implement cognition modularly (Bertolero et al., 2015; Yeo et al., 2015). All in all, this supports the claim that resting-state networks may reflect critical cognitive units.

Laird et al. (2011) were the first to deliver an extensive description of the cognitive interpretation of intrinsic connectivity networks. Using the same methodology as in Smith et al. (2009) (i.e., ICA) to derive intrinsic connectivity networks from the BrainMap database (Fox et al., 2005), these interpretations were made possible thanks to the richness of the manual annotations provided in the database relative to cognitive processes. In a similar attempt, another study used the BrainMap taxonomy (Fox et al., 2005) to extract the "functional fingerprint" of networks across 20 cognitive domains (e.g., "Audition-Perception", "Memory-Cognition"), allowing to reveal each network’s unique cognitive profile (Anderson et al., 2013). However, despite the highly valuable insights provided by these studies, some limitations appear. For instance, while links to cognitive-behavioral terminology were made possible by analyzing network maps derived from task studies ( Smith et al., 2009, Laird et al., 2011), the interpretation of RSN was still inferred from their apparent resemblance to the BrainMap-derived networks. This approach carries the risk that minor differences in topology or regional activation levels could lead to significant differences in their respective cognitive profile. Alternatively, other studies defined each network’s cognitive profile according to 20 broad behavioral domains (Anderson et al., 2013), therefore dampening the fine description of the involved cognitive processes. Finally, despite the high metadata quality of the activation maps found in the BrainMap database, these metadata come from manual annotations of a subset of the available studies, therefore preventing it from incorporating a large part of the fMRI literature. By contrast, relying on natural language processing, Neurosynth (Yarkoni et al., 2011) is a database that relies on the automatic generation of term-related meta-analytic maps from thousands of fMRI activation studies. Therefore, compared to BrainMap, it scales much more extensively and encompasses more specific terms than BrainMap’s broad behavioral domains.

More recently, building on the development of Neurosynth (Yarkoni et al., 2011), a new class of decoding methods has emerged to probe the potential involvement of cognitive functions given some brain activity ( Poldrack et al., 2009; Poldrack, 2011; Rubin et al., 2017). Such decoding methods can be used to infer the presence of putative cognitive processes by spatial comparison to the meta-analytic database. For instance, meta-analytic decoding has been successfully used to infer the cognitive content of short task-fMRI blocks (Wegrzyn et al., 2018) or to reveal the cognitive profiles of the human intraparietal sulcus (Boeken & Markett, 2023) or the cognitive states of participants watching a video (Pacella et al., 2024). These methods can potentially assess the cognitive relevance of RSNs by allowing the direct comparison of their topology to meta-analytic maps. However, to the best of our knowledge, meta-analytic decoding has never been applied for the empirical characterization of the potential cognitive relevance of RSNs.

The present study introduces the Groupe d’Imagerie Neurofonctionnelle Network Atlas (GINNA), a new fine-grained 33 networks atlas with an empirical characterization of each network’s cognitive relevance. Contrasting with existing atlases, GINNA creation relied on the classification of independent components obtained at the individual level, providing finer granularity and ensuring that components are reliably present across individuals. We developed a strategy for cognitive labeling of the networks relying on decoding cognitive terms from a manually curated subset of the Neurosynth database to provide an exhaustive characterization of their putative cognitive functions in light of the current neuroimaging literature. Such an atlas grounded in a careful characterization of its putative cognitive functions could be a useful guide when combined with newly emerging methods, allowing the reveal of mechanistic accounts in future direct/causal assessments of large-scale networks’ functions.

Data acquisition

MRi-Share study protocol

All participants (n = 1812; 1304 females (71.96%), 508 males (28.04%); mean age ± s.d.: 22.10 ± 2.29) were part of MRi-Share (Tsuchida et al., 2021), the Magnetic Resonance Imaging (MRI)-based substudy of the larger internet-based Students Health Research Enterprise (i-Share) cohort, launched in 2013. Detailed study protocol, demographic information and methodological description are available in Tsuchida et al. (2021). All procedures were performed in compliance with the declaration of Helsinki, and were approved by the local ethical committee (CPP2015-A00850-49).

Resting-state fMRI acquisition

All neuroimaging data were acquired between November 2015 and November 2017. Participants underwent a single run of resting-state echo-planar imaging (EPI) functional MRI (fMRI) acquisition (Siemens 3T Prisma, 64-channels head coil; voxel size: 2.4x2.4x2.4 mm³; Time of Repetition (TR) = 850ms; Time of Echo (TE) = 35.0ms; flip angle = 56°, multi-band factor = 6), for a duration of approximately 15 minutes. This resulted in 1054 brain volumes acquired for each participant. Prior to the resting-state fMRI (rs-fMRI) acquisition, participants were instructed to “keep their eyes closed, to relax, to refrain from moving, to stay awake, and to let their thoughts come and go”. In addition to the resting-state acquisition, other imaging modalities were acquired, including T1-weighted (T1w) structural images (one volume; three-dimensional Magnetization Prepared Rapid Gradient Echo (3D MPRAGE) sequence; Siemens 3T Prisma, 64-channels head coil; voxel size: 1.0x1.0x1.0 mm³; Time of Repetition (TR) = 2000ms; Time of Echo (TE) = 2.0 ms; Time of Inversion = 880 ms). In total, a whole acquisition session lasted for approximately 45 minutes.

Resting-state fMRI preprocessing

The preprocessing pipeline is described in detail in the supplementary materials of Tsuchida et al. (2021). Briefly, the distortion-corrected rs-fMRI data were registered to anatomical space, spatially filtered with a gaussian full width at half maximum of 5mm in each orthogonal dimensions, and time band-pass filtered to a frequency window of 0.01–0.1 Hz. Images were subsequently registered to Montreal Neurological Institute (MNI) standard space (sampling of 2x2x2 mm3). The temporal signal was corrected from the movement parameters and their derivatives, as well as the average of the signal recorded in the white matter, cerebrospinal fluid and the gray matter. Preprocessing primarily involved tools from FSL v5.0.10 (Smith et al., 2004) and AFNI v10.0.05 (Cox, 1996) encapsulated into a singularity container.

GINNA Resting state Atlas

Briefly, the three steps of the GINNA atlas creation are as follows: 1) Subject individual independent component (IC) analysis (ICA): individual fMRI data were individually processed using the ICA program MELODIC (multivariate exploratory linear optimized decomposition into independent components, version 3.14) available in the FMRIB Software Library (FSL; Smith et al., 2004). For each subject, the number of ICs was estimated using the Laplace approximation (Minka, 2000). 2) ICs classification into 41 classes: this was done using the MICCA clustering algorithm (Naveau et al., 2012), Icasso (Himberg & Hyvarinen, 2003), and supervised deep learning-based classification of ICs (Nozais et al., 2021). 3) Atlas creation: for each class, the individual ICs were averaged to obtain a group-level IC map. This class map was thresholded by selecting voxels that belong to at least 50% of the individual ICs (individual z-maps thresholded using a mixture model of a Gaussian plus two gamma distributions at p = 0.5).

Data analysis

Decoding the cognitive terms associated to resting-state networks

The meta-analytic decoding procedure (Margulies et al., 2016; Peraza et al., 2024) is illustrated in Fig. 1. In order to decode the cognitive terms associated with each GINNA resting-state network (RSN), the extracted group-level IC maps of RSNs were spatially compared with Meta-analytic Activation Maps (MAMs) extracted from the Neurosynth database (Yarkoni et al., 2011). Each meta-analytic map corresponds to a term used in the fMRI literature (e.g., fear, face) and aggregates results based on the coordinates of activation reported in all studies using the term at a frequency of at least 1/1000 words. As Neurosynth proceeds by an automatic scraping of the literature, the database includes terms not limited to cognitive behavioral denomination, such as anatomical (e.g., accumbens) or vague/broad terms (e.g., accurately). Therefore, we employed a subset of the Neurosynth database manually curated to be specifically representative of cognition (Karolis et al., 2019; Pacella et al., 2024). This subset comprises 506 MAMs (association test maps), encompassing 11406 studies of the fMRI literature, published between 1999 and 2017.

Both RSN and thresholded MAMs (z = 3.4) were first parcellated using a combination of a first atlas for cortical and subcortical volume that accounts for homotopy, a major aspect of the human brain organization (AICHA v2) (Joliot et al., 2015), and AAL3 (Rolls et al., 2020) for cerebellum. For each brain map, the average value of all voxels within each parcel was extracted using the Nilearn object NiftiLabelMasker, resulting in 410 values. Spatial comparison of each RSN map to the Neurosynth database was performed by computing Pearson product-moment correlations between each resting-state network/meta-analytic maps pairs, resulting in 506 correlations for each of the considered RSNs.

In order to determine the statistical significance of the correlation between an input RSN map and the meta-analytic maps of the database, we implemented a procedure of generative null-hypothesis testing. Namely, for each RSN, we computed a null distribution of 10,000 autocorrelation-preserving surrogate maps using the BrainSmash package (Burt et al., 2020) as implemented in Python (https://brainsmash.readthedocs.io). Information about the distance between brain regions was calculated using the Euclidean distance (Burt et al., 2020) and provided to the software as a 410x410 distance matrix. Spatial autocorrelation has been identified as a key component of brain organization. Taking it into account in the null hypothesis modeling effectively leads to consequent false positive rate reduction as compared to spatially-unconstrained surrogates (Markello et al., 2022).

The Pearson product-moment correlations between each surrogate map and the database were computed, leading to a distribution of null correlations for each term, further allowing to compare the observed correlation value for the considered term to those that would be observed by chance (506 x 10,000). To account for multiple comparisons, the maximal correlation across all terms was retained for each exemplar of the null distribution, resulting in a null distribution (n = 10,000) accounting for family-wise error rate (fwer). This null distribution was used to determine the MAMs that were significantly correlated with the considered RSN (p_fwer < 0.05). The terms corresponding to the MAMs significantly correlated with the RSNs were further extracted, effectively resulting in the association of RSNs to one or several (if any) cognitive terms.

We performed principal component analysis (PCA) at the single RSN level as a complementary analysis. This procedure aimed at characterizing whether the network formed a homogeneous functional unit or was associated with multiple cognitive processes, by revealing if the terms associated with a given network could be described by one or multiple principal “cognitive” components. We first created a binary mask of the parcellated versions of the RSNs, by considering a parcel as belonging to a RSN if it exceeded a liberal z-score threshold of 1. We subsequently extracted the parcellated MAMs of each term significantly associated with the considered network. To consider only RSN-relevant regions, MAMs were further masked with their respective RSN mask. PCA was computed using the R FactoMineR package (Lê et al., 2008), by treating individual brain regions composing the network as samples, parametrized by their activation values across the set of significantly associated term meta-analytic maps. From the PCA, we retrieved the percentage of explained variance of each component. To assess the contribution of both cognitive terms and brain regions to each component, their coordinates were extracted. Coordinates for cognitive terms were ordered to reveal the terms that were the most representative of the component. PCA was not computed for RSNs with less than 3 significantly associated MAMs.

Anatomical nomenclature

Following guidelines of the Organization for Human Brain Mapping (OHBM) Workgroup for HArmonized Taxonomy of NETworks (WHATNET) consortium (Uddin et al., 2023), we did not restrict the labeling of GINNA RSNs to cognitive terminology but grounded it into an anatomical taxonomy. This is to prevent any confusion regarding the identification of GINNA RSNs, which are mainly referred to by their anatomical description, then supplemented with their suggested cognitive relevance. RSNs anatomical nomenclature comprised the cerebral lobes (Frontal (F), Temporal (T), Parietal (P), Occipital (O), Cingular (Cing), Insular (Ins)), as well as the pericentral sulcus (Pericentral (Pc)) and subcortical nuclei (basal ganglia, BG). The cingulum was specified with a, m, and p (for anterior, mid, and posterior). A lobe was included in the nomenclature of a given RSN if it contained voxels among the 50% highest z-values of the RSN maps. In addition, a left or right lateralization (L/R, respectively), Dorsal (D), and medial (med) indicator was added as a prefix to the anatomical nomenclature when relevant. In cases where multiple RSNs were attributed the same anatomical nomenclature, a numbering following the principal gradient of cortical connectivity extending from low-level, unimodal sensory cortices, up to higher-level, heteromodal association cortices (Margulies et al., 2016) was added as a suffix.

Attributing cognitive labels to GINNA resting-state networks

Because the terms present in the Neurosynth database are uncontextualized, having meta-analytic terms significantly associated with resting-state networks is not sufficient to assign to them one or multiple cognitive processes. A term such as ‘face’, for instance, could indeed be reliably reported in studies about the somatomotor homonculus of the face, their visual perception, or because they are the support of an emotion recognition task. For this reason, we decided to resort to a qualitative attribution of cognitive processes to RSNs by neuroimaging experts. Six authors were supplemented with the results of the procedures described above (set of decoded terms, their strength of association with the RSN and their repartitions in PCA cognitive components) and asked to independently attribute one or several cognitive processes to the networks. All answers were collected via an online questionnaire. A final step involved harmonizing the resulting labels so that they would correspond to those found in the Cognitive Atlas (R. Poldrack et al., 2011) (https://cognitiveatlas.org/), without meaningfully altering the labels obtained from the experts consensus.

GINNA resting-state networks

The GINNA atlas is presented in Fig. 2 and Fig. 3, and is made publicly available at https://github.com/Achillegillig/ginna. The procedure of resting-state network (RSN) decomposition resulted in 41 networks. Out of the 41 RSNs identified by the procedure, 7 that did not overlap the brain (mainly venous artifacts), as well as one limited to the brainstem, were excluded from further analyses, resulting in a total of 33 RSNs further analyzed. The RSNs numbering (RSN01 to RSN33) reflects the reliability of their identification: RSNs with the lowest numbering are detected in the most subjects (highest between-subject reliability).

Meta-analytic cognitive terms decoding

Our procedure allowed to identify an average of 12 terms per network, with associations ranging from 38 (n = 1; RSN32) to 0 terms (n = 3; RSNs 01, 02, 26) (Fig. 4, Table 1). The MAMs correlating significantly with their respective RSNs resulted in an average correlation of r = 0.61, with a maximum of r = 0.91 (RSN14) and a minimum of r = 0.32 (RSN31) (Fig. 4). Detailed results for each network are available in the supplementary materials.

Table 1

Meta-analytic decoding. Number of significantly decoded terms and the 3 most correlated terms for each network. Bold typeface indicates significance after correction for multiple comparisons (p_fwer < 0.05). RSNs are ordered by their decreasing order of detection, RSN01 being the network most reliably identified at the individual level.
RSN	Sig. terms (n)	3 closest terms (Pearson r, p)
01	0	mnemonic (r = 0.40, p = 0.07); retrieval (r = 0.38, p = 0.1); recollection (r = 0.37, p = 0.1)
02	0	monitoring (r = 0.32, p = 0.1); signal_task (r = 0.29, p = 0.2); cognitive_control (r = 0.28, p = 0.4)
03	20	visual (r = 0.76, p < 0.001); visual_field (r = 0.54, p < 0.001); attended (r = 0.52, p < 0.001)
04	7	speech_production (r = 0.69, p < 0.001); oral (r = 0.67, p < 0.001); naming (r = 0.39, p = 0.01)
05	17	autobiographical_memory (r = 0.71, p < 0.001); episodic (r = 0.66, p < 0.001); autobiographical (r = 0.64, p < 0.001)
06	7	foot (r = 0.88, p < 0.001); limb (r = 0.72, p < 0.001); arm (r = 0.59, p < 0.001)
07	17	default_mode (r = 0.84, p < 0.001); default_network (r = 0.72, p < 0.001); self_referential (r = 0.64, p < 0.001)
08	13	spatial (r = 0.70, p < 0.001); orienting (r = 0.64, p < 0.001); visuospatial (r = 0.53, p < 0.001)
09	3	early_visual (r = 0.74, p < 0.001); primary_visual (r = 0.72, p < 0.001); visual_stimulus (r = 0.66, p < 0.001)
10	22	hands (r = 0.70, p < 0.001); action_observation (r = 0.69, p < 0.001); action (r = 0.68, p < 0.001)
11	15	secondary_somatosensory (r = 0.81, p < 0.001); painful (r = 0.79, p < 0.001); noxious (r = 0.74, p < 0.001)
12	3	reasoning (r = 0.43, p = 0.02); judgments (r = 0.42, p = 0.03); solving (r = 0.40, p = 0.05)
13	6	money (r = 0.56, p < 0.001); preferences (r = 0.50, p = 0.001); decision_making (r = 0.47, p = 0.004)
14	27	pitch (r = 0.91, p < 0.001); musical (r = 0.88, p < 0.001); auditory (r = 0.87, p < 0.001)
15	9	verb (r = 0.61, p < 0.001); verbs (r = 0.50, p < 0.001); syntactic (r = 0.45, p = 0.002)
16	5	memory_load (r = 0.49, p = 0.002); wm (r = 0.40, p = 0.03); memory_wm (r = 0.40, p = 0.03)
17	12	nogo (r = 0.51, p < 0.001); response_inhibition (r = 0.51, p < 0.001); intentions (r = 0.47, p = 0.002)
18	1	expectancy (r = 0.49, p = 0.002); response_inhibition (r = 0.28, p = 0.4); tools (r = 0.26, p = 0.5)
19	21	read (r = 0.60, p < 0.001); sentence (r = 0.58, p < 0.001); comprehension (r = 0.56, p < 0.001)
20	6	calculation (r = 0.58, p < 0.001); subtraction (r = 0.54, p < 0.001); arithmetic (r = 0.49, p = 0.003)
21	27	index_finger (r = 0.78, p < 0.001); hand_movements (r = 0.62, p < 0.001); hand (r = 0.60, p < 0.001)
22	12	conflict (r = 0.69, p < 0.001); stop_signal (r = 0.50, p < 0.001); error (r = 0.49, p < 0.001)
23	7	demands (r = 0.62, p < 0.001); verbal (r = 0.55, p = 0.005); judgment (r = 0.54, p = 0.007)
24	17	gain (r = 0.56, p < 0.001); monetary (r = 0.56, p < 0.001); anticipation (r = 0.55, p < 0.001)
25	6	hand (r = 0.51, p = 0.003); hands (r = 0.51, p = 0.004); motor_task (r = 0.51, p = 0.004)
26	0	integrate (r = 0.35, p = 0.2); early_visual (r = 0.29, p = 0.5); visual_stimulus (r = 0.27, p = 0.6)
27	23	arithmetic (r = 0.59, p < 0.001); orthographic (r = 0.49, p = 0.002); judgment (r = 0.48, p = 0.003)
28	13	inferences (r = 0.57, p < 0.001); judgments (r = 0.49, p < 0.001); mentalizing (r = 0.48, p < 0.001)
29	2	matching (r = 0.52, p < 0.001); matching_task (r = 0.47, p = 0.004); face (r = 0.33, p = 0.1)
30	6	pointing (r = 0.41, p = 0.003); movements (r = 0.37, p = 0.01); imitation (r = 0.37, p = 0.01)
31	4	cognitive_control (r = 0.34, p = 0.02); interference (r = 0.34, p = 0.02); conflicting (r = 0.34, p = 0.02)
32	38	motion (r = 0.83, p < 0.001); visual_motion (r = 0.76, p < 0.001); viewing (r = 0.75, p < 0.001)
33	27	comprehension (r = 0.83, p < 0.001); sentences (r = 0.79, p < 0.001); language_comprehension (r = 0.76, p < 0.001)

Resting-state networks cognitive labeling

The cognitive processes that emerged as consensual for each network, corresponding to Cognitive Atlas concepts (Poldrack et al., 2011), are presented in Table 2, Fig. 2 and Fig. 3. The attributed processes covered a large span of cognitive functions. This included sensory and motor functions, with auditory (n = 1; TN-01 (RSN14) - auditory perception), somatomotor (n = 5; PcN-01 (RSN06) - movement (limb); PcN-02 (RSN04) - articulation; PcN-03 (RSN11) - somatosensation; R-PcN (RSN25) - movement (left hand); L-PcN (RSN21) - movement (right hand) ), and visual processes (n = 4; ON-01 (RSN09) - visual perception; ON-02 (RSN29) - visual form discrimination ON-04 (RSN03) - motion detection, visual object recognition; OTN (RSN32) - object perception). We also uncovered several resting-state networks related to more integrated visuomotor and visuospatial processes (n = 3; D-FPN-01 (RSN10) - motor planning; D-FPN-02 (RSN30) - motor imagery; D-FPN-03 (RSN08) - spatial selective attention).

We also evidenced RSNs related to other high-level cognition (n = 5; R-FTPN-02 (RSN20) - mental arithmetic; L-InsFPN (RSN23) - phonological working memory; L-FTPN-02 (RSN12) - reasoning; mCingFPN (RSN16) - working memory; FTPN-02 (RSN27) – reading, mental arithmetic). Other high-level processes included cognitive control (n = 3; R-FInsN (RSN31) - cognitive control; mCingInsN (RSN22) - performance monitoring; FTPN-01 (RSN18) - expectancy), decision-making (n = 2; BGN (RSN24) - reward anticipation; aCingN (RSN13) - decision making).

Three networks were associated with language-related terms, two of which were strongly left-lateralized (L-FTN (RSN15) - syntactic processing; L-FTPN-01 (RSN19) - sentence comprehension), and another encompassing the bilateral temporal gyrus (TN-02 (RSN33) - speech perception). In addition, two networks were associated with social cognition (n = 2: med-FN (RSN28) - theory of mind; R-FTPN-01 (RSN17) - self-monitoring, theory of mind).

Two networks were related to the default mode, as evidenced by a significant decoding of the terms “default mode” and “default network”, and were associated with memory (n = 1; med-TN (RSN05) - memory retrieval); and thoughts about the self (n = 1; med-FPN (RSN07) - self-referential processing).

RSNs for which the meta-analytic decoding procedure did not result in any significant association with meta-analytic maps (n = 3; pCing-medPN (RSN01), R-FTPN-03 (RSN02), ON-03 (RSN26)), were labeled as non-significant (n.s.).

Table 2

Anatomical and cognitive labels of GINNA networks. The cognitive labels were attributed by six independent authors based on the results of the meta-analytic cognitive decoding. The anatomical nomenclature is described in the methods section. Abbreviations: BG, basal ganglia; a/m/pCing, anterior/middle/posterior cingulate ; D, dorsal; F, frontal; Ins, insular; L, left; med, median; N, network; n.s., non-significant; O, occipital; P, parietal; Pc, pericentral; R, right; T, temporal
RSN	Anatomical label	Cognitive Atlas label	RSN	Anatomical label	Cognitive Atlas label
01	pCing-medPN	n.s.	18	FTPN-01	expectancy
02	R-FTPN-03	n.s.	19	L-FTPN-01	sentence comprehension
03	ON-04	motion detection, visual object recognition	20	R-FTPN-02	mental arithmetic
04	PcN-02	articulation	21	L-PcN	movement (right hand)
05	med-TN	memory retrieval	22	mCingInsN	performance monitoring
06	PcN-01	movement (limb)	23	L-InsFPN	phonological working memory
07	med-FPN	self-referential processing	24	BGN	reward anticipation
08	D-FPN-03	spatial selective attention	25	R-PcN	movement (left hand)
09	ON-01	visual perception	26	ON-03	n.s.
10	D-FPN-01	motor planning	27	FTPN-02	reading, mental arithmetic
11	PcN-03	somatosensation	28	med-FN	theory of mind
12	L-FTPN-02	reasoning	29	ON-02	visual form discrimination
13	aCingN	decision making	30	D-FPN-02	motor imagery
14	TN-01	auditory perception	31	R-FInsN	interference resolution
15	L-FTN	syntactic processing	32	OTN	object perception
16	mCingFPN	working memory	33	TN-02	speech perception
17	R-FTPN-01	self monitoring, theory of mind

Detailed results for default mode and language networks

In this section, we focus on a few networks that illustrate well how our results relate to the state of the art and how the results from the principal component analyses were used as a basis for the consensus among authors to choose the cognitive labels. Exhaustive results for each RSN are available in the supplementary materials.

Default mode networks: RSN05, 07 – med-TN, med-FPN

The two networks were associated with terms related to the default mode (“default network”, “default mode”): RSN05 (med-TN) and RSN07 (med-FPN) (Fig. 5).

RSN05 (med-TN) encompassed the hippocampal gyri, posterior cingulum, precuneus and bilateral angular gyri. In addition to terms related to the default mode, RSN05 was associated with terms related to memory, with the 3 most correlated terms referring to autobiographical aspects of declarative memory (autobiographical memory: r = 0.71, p < 0.001; episodic: r = 0.66, p < 0.001; autobiographical: r = 0.64, p < 0.001; Table 1, Fig. 5). The cognitive label resulting from the consensus for this RSN was “memory retrieval” (Table 2).

RSN07 (med-FPN) comprised the medial prefrontal cortex (mPFC), posterior cingulate cortex (PCC), precuneus, and the bilateral angular gyri, therefore corresponding to the canonical anatomical definition of the default mode network (DMN). Seventeen terms were decoded for this network. Unsurprisingly, the two most correlated terms referred to the default mode (default mode: r = 0.84, p < 0.001; default network: r = 0.72, p < 0.001) (Table 1, Fig. 5). The principal component analysis computed on the network regions’ activations indicated that the terms significantly associated with the network could be summarized in two principal components (PCs) that explained 41.6% and 20.7% of the variance, respectively (Supplementary Fig. 7). PC1 loaded positively on all terms that generally referred to internal thoughts (with theory of mind and mentalizing at the top of the loadings), while PC2 represented an opposition between thoughts oriented to one owns’ experience (autobiographical memory, self-referential, personal) and thoughts associated to a more social context (beliefs, moral, social). The consensual cognitive process that emerged for this network was “self-referential processing”.

RSN15, 19, 33: L-FTN, L-FTPN-01, TN-02 - language networks

The three networks associated with language processes were L-FTN (RSN15) and L-FTPN-01 (RSN19), lateralized to the left hemisphere, and the TN-02 (RSN33), encompassing the superior temporal gyri and extending ventrally in the middle temporal gyri of both hemispheres (Fig. 6). Anatomically, all three networks overlapped in a region centered on the posterior part of the left superior temporal sulcus (STS), extending in the superior and middle temporal gyri. In addition, L-FTN and L-FTPN-01 displayed a partial overlap in the inferior frontal gyrus, the precentral sulcus, and the superior frontal gyrus, with L-FTPN-01 always more anterior than L-FTN. L-FTPN-01 extended also more posteriorly in the left angular gyrus, while the L-FTN stopped in the supramarginal gyrus. The cognitive terms significantly associated to TN-02 grouped into two cognitive principal components (Supplementary Fig. 33), with a first component (56.5% explained variance) regrouping terms related to the auditory modality (spoken, listening, speech, auditory, acoustic). The second component (24.3% explained variance) opposed the comprehension of language (sentence comprehension, language network) to its less linguistic dimension (communication, sounds, acoustic). High positive loadings were most present in the left hemisphere STS and middle temporal regions and strong negative loadings were observed in the bilateral superior temporal gyri and the right STS and right middle temporal gyrus (Supplementary Fig. 33). TN-02 was labeled as “speech perception” with respect to the cognitive atlas terminology.

Though L-FTN and L-FTPN-01 at least partially overlapped, their respective set of associated terms differed. L-FTN, attributed to “syntactic processing”, was explained by a single unitary component (90.7% explained variance) most represented by the term’s verbs, syntactic, and sentence comprehension (Supplementary Fig. 15). L-FTPN-01 which was attributed to “sentence comprehension” comprised a main component (56.4% explained variance) related to linguistic aspects of language comprehension (language comprehension, sentence comprehension, syntactic, semantic) and a minor one (12.3% explained variance) related to the representation of meaning (mentalizing, theory of mind, inference). The first component was best represented in the left STS and anterior part of the inferior frontal gyrus, while the second was best represented in the left angular gyrus, temporal pole, and the anterior part of the medial frontal gyrus (Supplementary Fig. 19).

The cognitive relevance of resting-state networks (RSNs) is poorly understood. To resolve this issue, we propose the Groupe d’Imagerie Neurofonctionnelle Network Atlas (GINNA), a comprehensive RSN atlas derived from the resting-state data of 1,812 participants, providing an exhaustive cognitive characterization of the human brain into 33 distinct networks reliably detected at the individual level.

We systematically analyzed the topographical similarity between GINNA networks and meta-analytic maps extracted from the Neurosynth database (Yarkoni et al., 2011). Although the method we propose relies on a simple measure of spatial similarity using Pearson correlation, here, we demonstrate its usefulness for investigating the cognitive processes potentially linked to RSNs. By relying on an approach of quantitative meta-analytic decoding of cognitive terms related to GINNA RSNs, we provide, to the best of our knowledge, the first empirical cognitive characterization of RSNs. Comparing task-derived meta-analytic maps from the literature and RSNs is particularly relevant if we consider that RSNs represent the prospective exploration of an available repertoire of cognitive functions (Deco et al., 2013). In addition, in the context of multivariate pattern analysis, decoding from Neurosynth-derived maps has been shown to perform similarly to more complex, multivariate decoders (Jabakhanji et al., 2022), and to allow to decode from short blocks of task fMRI the cognitive domains that a single participant was engaged in (Wegrzyn et al., 2018). This suggests that despite its simplicity and low computational cost, decoding based on topographical similarity with activation maps is theoretically justified.

To date, existing brain network parcellations that provide cognitive labeling (e.g., Yeo et al., 2011) have performed an association of function from visual similarity with networks obtained using task paradigms, a method that has proven to result in poor identifiability of RSNs (Uddin et al., 2023). Alternatively, it has been suggested that networks should be labeled according to their anatomical profile (Uddin et al., 2019). GINNA RSNs are provided with an anatomically grounded taxonomy accompanied by suggested cognitive process(es) to reconcile both views. Previous attempts of empirical cognitive characterization of RSNs have relied on topographically similar task-based networks as a proxy (Laird et al., 2011), or have done so with respect to broad cognitive domains extracted from BrainMap (Anderson et al., 2013). By contrast, our approach allows the direct assessment of networks obtained at rest, and benefits from the single cognitive term precision enabled by Neurosynth.

Positioning RSNs cognitive characterization with respect to specific cognitive processes is a much-needed endeavor for the field of cognitive neuroscience. For one, this referencing to well-defined psychological constructs allows to empirically test the predictions we make for each RSN, contrasting with broad cognitive domains (“visual”, “control”) that are not always informative. Second, as the rationale behind the inference of functions rests upon comparison with the neuroimaging literature, the present cognitive characterization is effectively an accurate summary of the current knowledge in both the conceptualization of cognitive concepts (as reflected by the terms present in studies and extracted by Neurosynth), as well as their brain underpinnings (as reflected by the topography of the meta-analytic maps). As such, if the goal is to understand the brain organization of cognition, it may prove more useful to describe RSN putative processes in terms related to those used in cognitive theories (e.g., theory of mind, visual perception) rather than using broad terms that do not necessarily relate to any psychological reality (e.g., limbic, visual). Moreover, because almost all attributed processes are referenced in the Cognitive Atlas Ontology (with the exception of “self-referential processing”, for which we found no equivalent), users can refer to the definitions provided in order to disambiguate the meaning of the concept and provide a common ground to all researchers (Poldrack et al., 2011). Cognitive atlas definitions are available at https://www.cognitiveatlas.org/concepts/categories/all.

The proposed atlas contrasts with existing atlases in its granularity; though rarely considered, this finer granularity might prove beneficial. This is supported by evidence showing that when grouping together a large span of networks constructs taken from psychology (e.g., “fear network”, “working memory network”) into higher-order, large-scale networks, markedly dissimilar cognitive processes become regrouped together (Thompson & Fransson, 2017). Additionally, while functional lateralization of brain circuits is a crucial organizational principle of the human brain, most existing atlases propose networks that are organized bilaterally. Bilateral RSNs are still observed, but the finer granularity in GINNA also resulted in the fragmention of some bilateral networks into homotopical counterparts. For instance, the hand somatomotor system is fragmented into two homotopical systems that correspond to the somatomotor homunculus of left-hand and right-hand motricity. The same can be observed for other networks, such as the FrontoTemporoParietal networks, that fragment into left and right counterparts, associated with distinct processes (e.g., L-FTPN01: sentence comprehension and R-FTPN01: self-monitoring-theory of mind). The identification of lateralized RSNs in GINNA supports the idea that they represent relevant functional units.

The cognitive processes attributed to GINNA RSNs range from low-order sensorimotor (visual, auditory, sensorimotor), up to more integrated, higher-order domains (decision-making, control, memory, social cognition, language, executive). This high diversity suggests that the GINNA atlas covers an extensive share of the known human cognitive repertoire. Of note, only 3 RSNs could not be significantly associated with any Neurosynth term. More importantly, though different in many aspects from existing atlases, some RSNs of the presently proposed atlas align with some of the main large scale networks described in the literature (Uddin et al., 2019). The closest resemblance is observed for RSNs related to the visual (ON-01 to ON-04, OTN), somatomotor (PcN-01 to PcN-03, L-PcN, R-PcN), and default mode systems (med-TN, med-FPN, pCing-medPN). D-FPN-03 - RSN08 associated with selective spatial attention corresponds to the dorsal frontoparietal network linked to attention. The mCingInsN – RSN22, which is associated with performance monitoring, resembles the MidCingulo-Insular network (Uddin et al., 2019), commonly referred to as the salience network. Performance monitoring implies detecting errors and conflicts during tasks and signaling the need for cognitive control adjustments, a role that seems in accordance with the functional definition of the salience network (Seeley, 2019; Seeley et al., 2007).

For a set of regions to be significantly associated with a given cognitive process, it is important that this association exhibits some specificity for this process: any set of regions that would systematically engage in many other tasks could lose its specificity and fail to be significantly more associated to a term than the others. Interestingly, this was the case for three networks (pCing-medPN- RSN01, R-FTPN-03 - RSN02, and ON-3- RSN26). The fact that two of these networks were the most consistently detected RSNs across all individuals may indicate their prime importance in diverse cognitive activities, although their exact contribution remains to be determined. pCingmedPN is a subpart of the classically defined default-mode network (Menon, 2023), and R-FTPN-03 shows a substantial overlap with the 'Multiple Demand (MD) system' (Duncan, 2013), although the latter is usually reported as bilateral rather than right-lateralized as in our case.

Detailed investigation of the terms decoded for well-studied networks, namely, default mode and language networks, highlights the precision of the method and its accordance with the literature. The cognitive labels that we associated with the DMN in its canonical definition (here, med-FPN - RSN07), almost exactly match the cognitive functions reported to be associated with increased activity within its nodes, namely autobiographical memory, self-referential cognition, and theory of mind, as recently reviewed (Menon, 2023).

The three networks that we uncover as related to language processes (L-FTN - RSN15, L-FTPN-01 - RSN19, TN-02 - RSN33) summarize well the current state of knowledge of the brain supports of language processes, and reveal the superiority of the meta-analytic decoding over visual attribution of function. Indeed, though they share a consequent amount of overlap, each RSN’s unique topographical pattern allows the segregation of their associated processes. Despite the overlap in the left superior temporal gyrus (STG), only the TN-02 – RSN33 is bilateral and, therefore, is associated with more perceptual aspects of speech, in line with the highest phonological specificity for the bilateral STG (Turker et al., 2023). L-FTPN-01 – RSN19, that we associate to sentence comprehension, encompasses regions that situated along the inferior bank of the superior temporal sulcus, the temporal pole, the angular gyrus, the left frontal pole, and the left superior frontal gyrus, all reported to be associated with semantic processing (Turker et al., 2023). This network is very similar to the core network of the SENtence Supramodal Areas AtlaS (SENSAAS) describing the essential areas for sentence reading, listening and production (Labache et al., 2019).

The present work is not without limitations. Our approach inherits all shortcomings from performing decoding from a meta-analytic database of task studies. Namely, our study is anchored in the risks associated with reverse inference: it cannot be concluded that because a cognitive process P engages a given brain region R, the activity in R implies the presence of the cognitive process P (R. Poldrack, 2006). In other terms, inferring cognitive processes to RSNs by analyzing their spatial similarity with task activations does not provide evidence of an explanatory relationship, but rather, of a coarse associative one (Mill et al., 2017).

Another limitation is related to the nature of the maps in the Neurosynth database: all task-fMRI studies proceed by contrasting some condition of interest to a control condition. By relying on this assumption of pure insertion (R. A. Poldrack & Yarkoni, 2016; Sternberg, 1969), any process that would be shared by the task of interest and the control condition would be masked out. Moreover, most task-based studies report results at the level of a restricted set of active brain regions or regions of interest. As our observed correlations rarely indicate a near-perfect match between RSNs and meta-analytic maps, our analysis does not allow to firmly determine that the decoded processes are implemented at the whole-network level, as opposed to a (subset of) region-level. The neural context hypothesis proposed that the functional relevance of a brain region relies on its co-activation with other brain regions (McIntosh, 2000). This means that a given region, reported to be implicated in, e.g., working memory, may perform markedly different computations when inscribed in a larger network comprising regions related to, e.g., language. Similarly, there is no guarantee that discrete cognitive processes map onto discrete brain representations. Therefore, it remains to be determined whether RSNs correspond to the prospective exploration of specific cognitive functions or, alternatively, to lower-level “cognitive building blocks” that cannot be isolated from the contrast logic,

We decided to rely on the qualitative attribution of networks’cognitive labels based on an expert consensus procedure. Establishing the relationship between the terms would necessitate a cognitive ontology (Francken et al., 2022; R. A. Poldrack & Yarkoni, 2016) that has yet to emerge despite significant efforts pushed in that direction (the most developed one being the Cognitive Atlas; Poldrack et al., 2011). As a consequence, and because there is to date no clear understanding of how distinct cognitive processes relate to one another (R. A. Poldrack & Yarkoni, 2016), a data-driven clustering of cognitive terms into broader cognitive domains (see, for instance, Wegrzyn et al., 2018), though it may provide an easily interpretable solution, is likely to be imperfect. The reader is invited to confront his/her own interpretation of the inferred processes attributed on the basis of the available results to the one proposed here. The fact that RSNs are labeled with respect to the Cognitive Atlas processes makes the present propositions amenable to further validation through empirical testing.

From the statistical standpoint, although we employed a method of spatial autocorrelation-preserving null hypothesis modeling that effectively reduces false positive rates as compared with spatial naive models (e.g., random shuffling of voxels), slightly inflated false positive rates may remain (Markello & Misic, 2021), leaving room for further methodological developments.

Finally, it is worth noting that the employed methodology does not account for the relevance of the dynamics in the expression of resting-state networks. Several studies have demonstrated that RSNs that appear over the course of relatively long resting-state acquisitions are, in fact, superordinate approximations of underlying dynamic states (Ciric et al., 2017; Sporns et al., 2021; Tagliazucchi et al., 2012). Therefore, the exact cognitive relevance of RSNs, seen as a prospective exploration of cognitive states (Deco et al., 2013), might be better understood in light of their instantaneous interactions with the rest of the brain, as observed at any given time.

Overall, we provide the Groupe d’Imagerie Fonctionnelle Network Atlas (GINNA), a 33 resting-state networks atlas of the human brain grounded in a meta-analytic decoding-based characterization of its cognitive relevance. Each resting-state network’s cognitive relevance is provided in terms of well-defined cognitive processes taken from the Cognitive Atlas ontology, and, as such, should represent better guides for future investigations. The atlas covers a broad spectrum of the human cognitive repertoire, with processes that align with brain laterality and display high associative precision. Potential use cases for GINNA include the selection of a priori regions of interest belonging to a specific network for neuroimaging analyses in the absence of task-derived functional data acquisition, as well as using the maps to analyze a posteriori whether significant regions/edges are distributed within specific cognitive systems.

Competing interests

The authors have no relevant financial or non-financial interests to disclose.

Ethics approval

The study protocol was approved by the Comité de Protection des Personnes Sud-Ouest et Outre-Mer (local ethics committee CPP SOOMIII) with agreement nr 2015-A00850-49.

Consent to participate
All participants signed an informed written consent form.

Funding

The i-Share cohort has been funded by a grant ANR-10COHO-05-01 (P.I. C Tzourio) as part of the Programme pour les Investissements d’Avenir. Supplementary funding was received from the Conseil Régional of Nouvelle-Aquitaine, Reference 4370420 (P.I. C Tzourio). The MRi-Share cohort has been supported by grants ANR-10-LABX-57 (P.I. B Mazoyer) and ANR-16-LCV2-0006 (GINESISLAB for the software, P.I. M Joliot). The bio-Share cohort and some regulatory and ethical aspects of MRiShare have been supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme under Grant Agreement No 640643 (P.I. S Debette) and the FHU SMART. Achille Gillig has benefited from state support managed by the Agence Nationale de la Recherche (French National Research Agency) under reference 17-EURE-0028.

Acknowledgements

Computer time for this study was provided by the computing facilities of the MCIA (Mésocentre de Calcul Intensif Aquitain, Bordeaux, France)

Data availability

The GINNA atlas is available at https://github.com/Achillegillig/ginna. Due to French regulations regarding sharing of the medical imaging data, individual raw data used for this study cannot be shared through a public repository. Rather, for MRi-Share de-identified data, a request can be submitted to the i-Share Scientific Collaborations Coordinator ([email protected]), the procedure is described on the i-share web site (https://research.i-share.fr/).

Code availability

The code is available on request to AG.

Anderson, M.L., Kinnison, J., Pessoa, L.: Describing functional diversity of brain regions and brain networks. NeuroImage. 73, 50–58 (2013). https://doi.org/10.1016/j.neuroimage.2013.01.071
Bertolero, M.A., Yeo, B.T.T., D’Esposito, M.: The modular and integrative functional architecture of the human brain. Proceedings of the National Academy of Sciences, 112(49), E6798–E6807. (2015). https://doi.org/10.1073/pnas.1510619112
Biswal, B., Zerrin Yetkin, F., Haughton, V.M., Hyde, J.S.: Functional connectivity in the motor cortex of resting human brain using echo-planar mri. Magn. Reson. Med. 34(4), 537–541 (1995). https://doi.org/10.1002/mrm.1910340409
Boeken, O.J., Markett, S.: Systems-level decoding reveals the cognitive and behavioral profile of the human intraparietal sulcus. Front. Neuroimaging. 1, 1074674 (2023). https://doi.org/10.3389/fnimg.2022.1074674
Burt, J.B., Helmer, M., Shinn, M., Anticevic, A., Murray, J.D.: Generative modeling of brain maps with spatial autocorrelation. NeuroImage, 220. (2020). https://doi.org/10.1016/j.neuroimage.2020.117038
Ciric, R., Nomi, J.S., Uddin, L.Q., Satpute, A.B.: Contextual connectivity: A framework for understanding the intrinsic dynamic architecture of large-scale functional brain networks. Sci. Rep. 7(1) (2017). Article 1. https://doi.org/10.1038/s41598-017-06866-w
Cole, M.W., Bassett, D.S., Power, J.D., Braver, T.S., Petersen, S.E.: Intrinsic and task-evoked network architectures of the human brain. Neuron. 83(1), 238–251 (2014). https://doi.org/10.1016/j.neuron.2014.05.014
Cole, M.W., Ito, T., Bassett, D.S., Schultz, D.H.: Activity flow over resting-state networks shapes cognitive task activations. Nat. Neurosci. 19(12), 1718–1726 (2016). https://doi.org/10.1038/nn.4406
Corbetta, M., Shulman, G.L.: Control of goal-directed and stimulus-driven attention in the brain. Nat. Rev. Neurosci. 3(3), 201–215 (2002). https://doi.org/10.1038/nrn755
Cordes, D., Haughton, V.M., Arfanakis, K., Wendt, G.J., Turski, P.A., Moritz, C.H., Quigley, M.A., Meyerand, M.E.: Mapping Functionally Related Regions of Brain with Functional Connectivity MR Imaging. Am. J. Neuroradiol. 21(9), 1636–1644 (2000)
Cox, R.W.: AFNI: Software for analysis and visualization of functional magnetic resonance neuroimages. Comput. Biomed. Res. Int. J. 29(3), 162–173 (1996). https://doi.org/10.1006/cbmr.1996.0014
Damoiseaux, J.S., Rombouts, S.A.R.B., Barkhof, F., Scheltens, P., Stam, C.J., Smith, S.M., Beckmann, C.F.: Consistent resting-state networks across healthy subjects. Proceedings of the National Academy of Sciences, 103(37), 13848–13853. (2006). https://doi.org/10.1073/pnas.0601417103
De Luca, M., Beckmann, C.F., De Stefano, N., Matthews, P.M., Smith, S.M.: fMRI resting state networks define distinct modes of long-distance interactions in the human brain. NeuroImage. 29(4), 1359–1367 (2006). https://doi.org/10.1016/j.neuroimage.2005.08.035
Deco, G., Jirsa, V.K., McIntosh, A.R.: Resting brains never rest: Computational insights into potential cognitive architectures. Trends Neurosci. 36(5), 268–274 (2013). https://doi.org/10.1016/j.tins.2013.03.001
Dosenbach, N.U.F., Fair, D.A., Miezin, F.M., Cohen, A.L., Wenger, K.K., Dosenbach, R.A.T., Fox, M.D., Snyder, A.Z., Vincent, J.L., Raichle, M.E., Schlaggar, B.L., Petersen, S.E.: Distinct brain networks for adaptive and stable task control in humans. Proceedings of the National Academy of Sciences, 104(26), 11073–11078. (2007). https://doi.org/10.1073/pnas.0704320104
Dosenbach, N.U.F., Visscher, K.M., Palmer, E.D., Miezin, F.M., Wenger, K.K., Kang, H.C., Burgund, E.D., Grimes, A.L., Schlaggar, B.L., Petersen, S.E.: A Core System for the Implementation of Task Sets. Neuron. 50(5), 799–812 (2006). https://doi.org/10.1016/j.neuron.2006.04.031
Doucet, G., Naveau, M., Petit, L., Delcroix, N., Zago, L., Crivello, F., Jobard, G., Tzourio-Mazoyer, N., Mazoyer, B., Mellet, E., Joliot, M.: Brain activity at rest: A multiscale hierarchical functional organization. J. Neurophysiol. 105(6), 2753–2763 (2011). https://doi.org/10.1152/jn.00895.2010
Duncan, J.: The structure of cognition: Attentional episodes in mind and brain. Neuron. 80(1), 35–50 (2013). https://doi.org/10.1016/j.neuron.2013.09.015
Fox, M.D., Raichle, M.E.: Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging. Nat. Rev. Neurosci. 8(9), 700–711 (2007). https://doi.org/10.1038/nrn2201
Fox, M.D., Snyder, A.Z., Vincent, J.L., Corbetta, M., Van Essen, D.C., Raichle, M.E.: The human brain is intrinsically organized into dynamic, anticorrelated functional networks. Proceedings of the National Academy of Sciences, 102(27), 9673–9678. (2005). https://doi.org/10.1073/pnas.0504136102
Fox, P.T., Laird, A.R., Fox, S.P., Fox, P.M., Uecker, A.M., Crank, M., Koenig, S.F., Lancaster, J.L.: Brainmap taxonomy of experimental design: Description and evaluation. Hum. Brain. Mapp. 25(1), 185–198 (2005). https://doi.org/10.1002/hbm.20141
Francken, J.C., Slors, M., Craver, C.F.: Cognitive ontology and the search for neural mechanisms: Three foundational problems. Synthese. 200(5), 378 (2022). https://doi.org/10.1007/s11229-022-03701-2
Goldman-Rakic, P.S.: Topography of Cognition: Parallel Distributed Networks in Primate Association Cortex. Annual Review of Neuroscience, 11(Volume 11, 1988), 137–156. (1988). https://doi.org/10.1146/annurev.ne.11.030188.001033
Hampson, M., Olson, I.R., Leung, H.-C., Skudlarski, P., Gore, J.C.: Changes in functional connectivity of human MT/V5 with visual motion input. NeuroReport. 15(8), 1315 (2004). https://doi.org/10.1097/01.wnr.0000129997.95055.15
Himberg, J., Hyvarinen, A.: Icasso: Software for investigating the reliability of ICA estimates by clustering and visualization. 2003 IEEE XIII Workshop on Neural Networks for Signal Processing (IEEE Cat. No.03TH8718), 259–268. (2003). https://doi.org/10.1109/NNSP.2003.1318025
Jabakhanji, R., Vigotsky, A.D., Bielefeld, J., Huang, L., Baliki, M.N., Iannetti, G., Apkarian, A.V.: Limits of decoding mental states with fMRI. Cortex. 149, 101–122 (2022). https://doi.org/10.1016/j.cortex.2021.12.015
Joliot, M., Jobard, G., Naveau, M., Delcroix, N., Petit, L., Zago, L., Crivello, F., Mellet, E., Mazoyer, B., Tzourio-Mazoyer, N.: AICHA: An atlas of intrinsic connectivity of homotopic areas. J. Neurosci. Methods. 254, 46–59 (2015). https://doi.org/10.1016/j.jneumeth.2015.07.013
Karolis, V.R., Corbetta, M., De Thiebaut, M.: The architecture of functional lateralisation and its relationship to callosal connectivity in the human brain. Nat. Commun. 10(1), 1417 (2019). https://doi.org/10.1038/s41467-019-09344-1
Labache, L., Joliot, M., Saracco, J., Jobard, G., Hesling, I., Zago, L., Mellet, E., Petit, L., Crivello, F., Mazoyer, B., Tzourio-Mazoyer, N.: A SENtence Supramodal Areas AtlaS (SENSAAS) based on multiple task-induced activation mapping and graph analysis of intrinsic connectivity in 144 healthy right-handers. Brain Struct. Function. 224(2), 859–882 (2019). https://doi.org/10.1007/s00429-018-1810-2
Laird, A.R., Fox, P.M., Eickhoff, S.B., Turner, J.A., Ray, K.L., McKay, D.R., Glahn, D.C., Beckmann, C.F., Smith, S.M., Fox, P.T.: Behavioral Interpretations of Intrinsic Connectivity Networks. J. Cogn. Neurosci. 23(12), 4022–4037 (2011). https://doi.org/10.1162/jocn_a_00077
Lê, S., Josse, J., Husson, F.: FactoMineR: An R Package for Multivariate Analysis. J. Stat. Softw. 25(1) (2008). https://doi.org/10.18637/jss.v025.i01
Margulies, D.S., Ghosh, S.S., Goulas, A., Falkiewicz, M., Huntenburg, J.M., Langs, G., Bezgin, G., Eickhoff, S.B., Castellanos, F.X., Petrides, M., Jefferies, E., Smallwood, J.: Situating the default-mode network along a principal gradient of macroscale cortical organization. Proc. Natl. Acad. Sci. U.S.A. 113(44), 12574–12579 (2016). https://doi.org/10.1073/pnas.1608282113
Markello, R.D., Hansen, J.Y., Liu, Z.-Q., Bazinet, V., Shafiei, G., Suárez, L.E., Blostein, N., Seidlitz, J., Baillet, S., Satterthwaite, T.D., Chakravarty, M.M., Raznahan, A., Misic, B.: neuromaps: Structural and functional interpretation of brain maps. Nat. Methods. 19(11) (2022). Article 11 https://doi.org/10.1038/s41592-022-01625-w
Markello, R.D., Misic, B.: Comparing spatial null models for brain maps. NeuroImage. 236, 118052 (2021). https://doi.org/10.1016/j.neuroimage.2021.118052
McIntosh, A.R.: Towards a network theory of cognition. Neural Netw. 13(8), 861–870 (2000). https://doi.org/10.1016/S0893-6080(00)00059-9
Menon, V.: 20 years of the default mode network: A review and synthesis. Neuron. 111(16), 2469–2487 (2023). https://doi.org/10.1016/j.neuron.2023.04.023
Mesulam, M.-M.: Large-scale neurocognitive networks and distributed processing for attention, language, and memory. Ann. Neurol. 28(5), 597–613 (1990). https://doi.org/10.1002/ana.410280502
Mill, R.D., Ito, T., Cole, M.W.: From connectome to cognition: The search for mechanism in human functional brain networks. NeuroImage. 160, 124–139 (2017). https://doi.org/10.1016/j.neuroimage.2017.01.060
Minka, T.: Automatic choice of dimensionality for PCA. In T. Leen, T. Dietterich, & V. Tresp (Eds.), Advances in neural information processing systems (Vol. 13). MIT Press. (2000). https://proceedings.neurips.cc/paper_files/paper/2000/file/7503cfacd12053d309b6bed5c89de212-Paper.pdf
Naveau, M., Doucet, G., Delcroix, N., Petit, L., Zago, L., Crivello, F., Jobard, G., Mellet, E., Tzourio-Mazoyer, N., Mazoyer, B., Joliot, M.: A Novel Group ICA Approach Based on Multi-scale Individual Component Clustering. Application to a Large Sample of fMRI Data. Neuroinformatics. 10(3), 269–285 (2012). https://doi.org/10.1007/s12021-012-9145-2
Nozais, V., Boutinaud, P., Verrecchia, V., Gueye, M.-F., Hervé, P.-Y., Tzourio, C., Mazoyer, B., Joliot, M.: Deep Learning-based Classification of Resting‐state fMRI Independent‐component Analysis. Neuroinformatics. 19(4), 619–637 (2021). https://doi.org/10.1007/s12021-021-09514-x
Pacella, V., Nozais, V., Talozzi, L., Abdallah, M., Wassermann, D., Forkel, S.J., De Schotten, T.: M. The morphospace of the brain-cognition organisation. Nature Communications. In press. (2024)
Peraza, J.A., Salo, T., Riedel, M.C., Bottenhorn, K.L., Poline, J.-B., Dockès, J., Kent, J.D., Bartley, J.E., Flannery, J.S., Hill-Bowen, L.D., Lobo, R.P., Poudel, R., Ray, K.L., Robinson, J.L., Laird, R.W., Sutherland, M.T., de la Vega, A., Laird, A.R.: Methods for decoding cortical gradients of functional connectivity. Imaging Neurosci. 2, 1–32 (2024). https://doi.org/10.1162/imag_a_00081
Poldrack, R.: Can cognitive processes be inferred from neuroimaging data? Trends Cogn. Sci. 10(2), 59–63 (2006). https://doi.org/10.1016/j.tics.2005.12.004
Poldrack, R.A.: Neuron. 72(5), 692–697 (2011). https://doi.org/10.1016/j.neuron.2011.11.001 Inferring Mental States from Neuroimaging Data: From Reverse Inference to Large-Scale Decoding
Poldrack, R.A., Halchenko, Y.O., Hanson, S.J.: Decoding the Large-Scale Structure of Brain Function by Classifying Mental States Across Individuals. Psychol. Sci. 20(11), 1364–1372 (2009). https://doi.org/10.1111/j.1467-9280.2009.02460.x
Poldrack, R.A., Yarkoni, T.: From Brain Maps to Cognitive Ontologies: Informatics and the Search for Mental Structure. Ann. Rev. Psychol. 67(1), 587–612 (2016). https://doi.org/10.1146/annurev-psych-122414-033729
Poldrack, R., Kittur, A., Kalar, D., Miller, E., Seppa, C., Gil, Y., Parker, D., Sabb, F., Bilder, R.: The Cognitive Atlas: Toward a Knowledge Foundation for Cognitive Neuroscience. Frontiers in Neuroinformatics, 5. https://www.frontiersin.org/articles/ (2011). 10.3389/fninf.2011.00017
Rolls, E.T., Huang, C.-C., Lin, C.-P., Feng, J., Joliot, M.: Automated anatomical labelling atlas 3. NeuroImage, 206, 116189. (2020). https://doi.org/10.1016/j.neuroimage.2019.116189
Rubin, T.N., Koyejo, O., Gorgolewski, K.J., Jones, M.N., Poldrack, R.A., Yarkoni, T.: Decoding brain activity using a large-scale probabilistic functional-anatomical atlas of human cognition. PLoS Comput. Biol. 13(10), e1005649 (2017). https://doi.org/10.1371/journal.pcbi.1005649
Seeley, W.W.: The Salience Network: A Neural System for Perceiving and Responding to Homeostatic Demands. J. Neurosci. 39(50), 9878–9882 (2019). https://doi.org/10.1523/JNEUROSCI.1138-17.2019
Seeley, W.W., Menon, V., Schatzberg, A.F., Keller, J., Glover, G.H., Kenna, H., Reiss, A.L., Greicius, M.D.: Dissociable Intrinsic Connectivity Networks for Salience Processing and Executive Control. J. Neurosci. 27(9), 2349–2356 (2007). https://doi.org/10.1523/JNEUROSCI.5587-06.2007
Smith, S.M., Fox, P.T., Miller, K.L., Glahn, D.C., Fox, P.M., Mackay, C.E., Filippini, N., Watkins, K.E., Toro, R., Laird, A.R., Beckmann, C.F.: Correspondence of the brain’s functional architecture during activation and rest. Proceedings of the National Academy of Sciences, 106(31), 13040–13045. (2009). https://doi.org/10.1073/pnas.0905267106
Smith, S.M., Jenkinson, M., Woolrich, M.W., Beckmann, C.F., Behrens, T.E.J., Johansen-Berg, H., Bannister, P.R., De Luca, M., Drobnjak, I., Flitney, D.E., Niazy, R.K., Saunders, J., Vickers, J., Zhang, Y., De Stefano, N., Brady, J.M., Matthews, P.M.: Advances in functional and structural MR image analysis and implementation as FSL. NeuroImage. 23, S208–S219 (2004). https://doi.org/10.1016/j.neuroimage.2004.07.051
Sporns, O., Faskowitz, J., Teixeira, A.S., Cutts, S.A., Betzel, R.F.: Dynamic expression of brain functional systems disclosed by fine-scale analysis of edge time series. Netw. Neurosci. 5(2), 405–433 (2021). https://doi.org/10.1162/netn_a_00182
Sternberg, S.: Memory-Scanning: Mental Processes Revealed by Reaction-Time Experiments. Am. Sci. 57(4), 421–457 (1969)
Tagliazucchi, E., Balenzuela, P., Fraiman, D., Chialvo, D.R.: Criticality in Large-Scale Brain fMRI Dynamics Unveiled by a Novel Point Process Analysis. Frontiers in Physiology, 3. (2012). https://doi.org/10.3389/fphys.2012.00015
Thompson, W.H., Fransson, P.: Spatial confluence of psychological and anatomical network constructs in the human brain revealed by a mass meta-analysis of fMRI activation. Sci. Rep. 7(1) (2017). Article 1. https://doi.org/10.1038/srep44259
Tsuchida, A., Laurent, A., Crivello, F., Petit, L., Joliot, M., Pepe, A., Beguedou, N., Gueye, M.-F., Verrecchia, V., Nozais, V., Zago, L., Mellet, E., Debette, S., Tzourio, C., Mazoyer, B.: The MRi-Share database: Brain imaging in a cross-sectional cohort of 1870 university students. Brain Struct. Function. 226(7), 2057–2085 (2021). https://doi.org/10.1007/s00429-021-02334-4
Turker, S., Kuhnke, P., Eickhoff, S.B., Caspers, S., Hartwigsen, G.: Cortical, subcortical, and cerebellar contributions to language processing: A meta-analytic review of 403 neuroimaging experiments. Psychol. Bull. 149(11–12), 699–723 (2023). https://doi.org/10.1037/bul0000403
Uddin, L.Q., Betzel, R.F., Cohen, J.R., Damoiseaux, J.S., De Brigard, F., Eickhoff, S.B., Fornito, A., Gratton, C., Gordon, E.M., Laird, A.R., Larson-Prior, L., McIntosh, A.R., Nickerson, L.D., Pessoa, L., Pinho, A.L., Poldrack, R.A., Razi, A., Sadaghiani, S., Shine, J.M., Spreng, R.N.: Controversies and progress on standardization of large-scale brain network nomenclature. Netw. Neurosci. 7(3), 864–905 (2023). https://doi.org/10.1162/netn_a_00323
Uddin, L.Q., Yeo, B.T.T., Spreng, R.N.: Towards a Universal Taxonomy of Macro-scale Functional Human Brain Networks. Brain Topogr. 32(6), 926–942 (2019). https://doi.org/10.1007/s10548-019-00744-6
Vincent, J.L., Snyder, A.Z., Fox, M.D., Shannon, B.J., Andrews, J.R., Raichle, M.E., Buckner, R.L.: Coherent Spontaneous Activity Identifies a Hippocampal-Parietal Memory Network. J. Neurophysiol. 96(6), 3517–3531 (2006). https://doi.org/10.1152/jn.00048.2006
Wegrzyn, M., Aust, J., Barnstorf, L., Gippert, M., Harms, M., Hautum, A., Heidel, S., Herold, F., Hommel, S.M., Knigge, A.-K., Neu, D., Peters, D., Schaefer, M., Schneider, J., Vormbrock, R., Zimmer, S.M., Woermann, F.G., Labudda, K.: Thought experiment: Decoding cognitive processes from the fMRI data of one individual (p. 341594). bioRxiv. (2018). https://doi.org/10.1101/341594
Yarkoni, T., Poldrack, R.A., Nichols, T.E., Van Essen, D.C., Wager, T.D.: Large-scale automated synthesis of human functional neuroimaging data. Nat. Methods. 8(8), 665–670 (2011). https://doi.org/10.1038/nmeth.1635
Yeo, B.T.T., Krienen, F.M., Eickhoff, S.B., Yaakub, S.N., Fox, P.T., Buckner, R.L., Asplund, C.L., Chee, M.W.L.: Functional Specialization and Flexibility in Human Association Cortex. Cereb. Cortex. 25(10), 3654–3672 (2015). https://doi.org/10.1093/cercor/bhu217
Yeo, B.T.T., Krienen, F.M., Sepulcre, J., Sabuncu, M.R., Lashkari, D., Hollinshead, M., Roffman, J.L., Smoller, J.W., Zöllei, L., Polimeni, J.R., Fischl, B., Liu, H., Buckner, R.L.: The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J. Neurophysiol. 106(3), 1125–1165 (2011). https://doi.org/10.1152/jn.00338.2011

There is NO Competing Interest.

Supplementary.pdf

Download PDF

Version 1

posted

You are reading this latest preprint version

GINNA, a 33 resting-state networks atlas with meta-analytic decoding-based cognitive characterization

Status:

Version 1

Abstract

Figures

Introduction

Material and methods

Data acquisition

MRi-Share study protocol

Resting-state fMRI acquisition

Resting-state fMRI preprocessing

GINNA Resting state Atlas

Data analysis

Decoding the cognitive terms associated to resting-state networks

Anatomical nomenclature

Attributing cognitive labels to GINNA resting-state networks

Results

GINNA resting-state networks

Meta-analytic cognitive terms decoding

Resting-state networks cognitive labeling

Detailed results for default mode and language networks

Default mode networks: RSN05, 07 – med-TN, med-FPN

RSN15, 19, 33: L-FTN, L-FTPN-01, TN-02 - language networks

Discussion

Declarations

Competing interests

Ethics approval

Consent to participate
All participants signed an informed written consent form.

Funding

Acknowledgements

Data availability

Code availability

References

Additional Declarations

Supplementary Files

Status:

Version 1

GINNA, a 33 resting-state networks atlas with meta-analytic decoding-based cognitive characterization

Status:

Version 1

Abstract

Figures

Introduction

Material and methods

Data acquisition

MRi-Share study protocol

Resting-state fMRI acquisition

Resting-state fMRI preprocessing

GINNA Resting state Atlas

Data analysis

Decoding the cognitive terms associated to resting-state networks

Anatomical nomenclature

Attributing cognitive labels to GINNA resting-state networks

Results

GINNA resting-state networks

Meta-analytic cognitive terms decoding

Resting-state networks cognitive labeling

Detailed results for default mode and language networks

Default mode networks: RSN05, 07 – med-TN, med-FPN

RSN15, 19, 33: L-FTN, L-FTPN-01, TN-02 - language networks

Discussion

Declarations

Competing interests

Ethics approval

Consent to participate All participants signed an informed written consent form.

Funding

Acknowledgements

Data availability

Code availability

References

Additional Declarations

Supplementary Files

Status:

Version 1

Consent to participate
All participants signed an informed written consent form.