Skip to main content
NCBI home page
NeuroImage: Clinical logoLink to NeuroImage: Clinical
. 2025 Jul 10;48:103840. doi: 10.1016/j.nicl.2025.103840

Multi-voxel pattern analysis for characterizing functional connectivity and neurocognitive function in major depression: A CAN-BIND-1 report

Alice Rueda a,1, Ilya Demchenko a,b,1, Vanessa K Tassone a,b, Fatemeh Gholamali Nezhad a, Vanessa Peters c, Nathan W Churchill d,e, Benicio N Frey f,g, Stefanie Hassel h, Raymond W Lam i, Roumen V Milev j,k, Daniel J Müller l,m,n,o, Tom A Schweizer d,e, Stephen C Strother p,q, Valerie H Taylor h, Sidney H Kennedy b,e,m,r, Sheeba Arnold Anteraper c,2, Venkat Bhat a,b,e,r,2,⁎; CAN-BIND Investigator Teama
PMCID: PMC12302774  PMID: 40684717

Graphical abstract

graphic file with name ga1.jpg

Open in a new tab

Keywords: Mood disorders, Functional neuroimaging, Magnetic resonance imaging, Brain networks, Data-driven analysis, Multi-voxel pattern analysis

Highlights

  • •

    Six whole-brain clusters revealed distinct connectivity shifts in depression.

  • •

    The left cerebellar crus I emerged as the most robust and extensive cluster.

  • •

    These clusters uncovered 24 abnormal connectivity patterns in depression.

  • •

    Five altered connections were tied to impairments in five neurocognitive domains.

  • •

    Brain-cognition associations seen in healthy people were disrupted in depression.

Abstract

Background

Major depressive disorder (MDD) affects not only mood but also neurocognitive function. In this study, we used whole-brain functional connectivity multi-voxel pattern analysis (fc-MVPA) to examine the relationship between resting-state functional connectivity (rsFC) and neurocognitive function in individuals with MDD compared to healthy controls (HC).

Methods

Baseline functional magnetic resonance imaging (fMRI) scans from the CAN-BIND-1 dataset were analyzed using a data-driven whole-brain fc-MVPA approach in 147 individuals with MDD and 98 HC. All participants completed the Computerized Neurocognitive Assessment Vital Signs (CNS-VS) battery outside the scanner, and correlations between rsFC differences and CNS-VS domain scores were explored.

Results

The fc-MVPA reduced the dimensionality of fMRI data at both individual and group levels, identifying six clusters with altered rsFC in MDD relative to HC: left cerebellar crus I, right precuneus, left superior lateral occipital cortex, right ventral caudate, left superior parietal lobule, and left dorsal anterior cingulate cortex. Using these clusters as seeds, post-hoc analyses identified 24 patterns of altered rsFC in MDD involving the default mode, central executive, visual recognition, salience, and sensorimotor networks. Five of these patterns showed significant correlations with CNS-VS domain scores for composite memory, neurocognition index, processing speed, executive function, and simple attention in HC, but these associations were absent in individuals with MDD.

Conclusions

Our findings highlight that MDD is associated with disrupted rsFC across networks relevant to neurocognitive function. The data-driven nature of the fc-MVPA identified the left cerebellar crus I as the most significant region of aberrant rsFC.

1. Introduction

MDD is the leading cause of disability worldwide (Santomauro et al., 2021), and a refined understanding of its pathophysiology is critical for identifying prospective biomarkers (Drysdale et al., 2017, Zhang et al., 2018). Functional magnetic resonance imaging (fMRI) has played a pivotal role in delineating intrinsic connectivity networks (ICNs) that display altered resting-state functional connectivity (rsFC) in MDD cohorts (Mulders et al., 2015, Zeng et al., 2012). According to fMRI studies, MDD patients tend to exhibit hyperconnectivity within the default mode network (DMN) (Manoliu et al., 2014), associated with maladaptive rumination and negative self-focused (Zhou et al., 2020) or self-referential thought (Guo et al., 2015), but hypoconnectivity within the central executive (CEN) (Liu et al., 2020), salience (SN) (Dai et al., 2019), and sensorimotor (SMN) (Buyukdura et al., 2011, Hou et al., 2022) networks, linked with failure of effective cognitive control, poor incentive salience, and psychomotor retardation, respectively (Demchenko et al., 2022). MDD is also marked by the presence of cognitive deficits that can significantly impact one’s daily functioning and quality of life, characterized by memory impairment, executive dysfunction (Snyder, 2013), attention and concentration difficulties, slower processing speed, cognitive flexibility issues, and physical manifestations such as psychomotor retardation (Steffens et al., 2006, Hammar et al., 2022, Kriesche et al., 2023). Indeed, as confirmed by a meta-analysis by McDermott and Ebmeier (2009), severity of MDD is correlated with performance in the domains of episodic memory, executive function, and processing speed.

Several studies have compared rsFC between participants with MDD and healthy controls (HC), but few have tried to bridge the gap between altered rsFC and cognitive deficits. Nowadays, the treatment of cognitive deficits in MDD has started to gain attention in the literature (Pan et al., 2019, Hammar et al., 2022). A smaller study by Pan et al. (2020) with 28 MDD patients and 24 matched HC conducted a seed-based rsFC analysis with seeding at the dorsolateral prefrontal cortex (dlPFC), revealing that rsFC of the CEN was associated with symptom severity. Similarly, with a larger group of participants (100 MDD and 100 HC), Liu et al. (2020) conducted another seed-based connectivity analysis using 27 seeds in the DMN, CEN, SN, and limbic system (LS). Altered rsFC was associated with performance on neuropsychological tests, including the Stroop Test, verbal fluency test, and Symbol Digit Modalities Test. In particular, the Stroop interference effect and rsFC between the right anterior prefrontal cortex and left cerebellum IV/V were disrupted in individuals with MDD.

Most fMRI studies (Mulders et al., 2015, Drysdale et al., 2017), however, continue to employ seed-based methods to examine rsFC and their respective neurocognitive indices. While the results of such studies are promising, this analytical approach requires a priori knowledge to select regions of interest (ROI). In fact, hypothesis-driven analyses continue to predominate in the fMRI literature to this day (Cole et al., 2010). The requirement for an a priori hypothesis restricts the capacity to prospectively identify novel signatures of MDD, and as such, present-day research fosters the application of data-driven approaches to explore rsFC (Huys et al., 2016). One such approach is whole-brain functional connectivity multi-voxel pattern analysis (fc-MVPA), which enables model-free analysis of fMRI data (Mahmoudi et al., 2012). For example, fc-MVPA can be directly combined with machine learning tools to perform discriminative analysis (Zeng et al., 2012). However, whole-brain fc-MVPA can be computationally intensive due to voxel-level correlation calculations, although some studies have mitigated this challenge by applying gray matter masks (Murrough et al., 2016, Scheinost et al., 2018, Wang et al., 2014a).

In this study, we aimed to analyze a large dataset from the initial trial of the Canadian Biomarker Integration Network in Depression (CAN-BIND-1) (MacQueen et al., 2019, Kennedy et al., 2019, Lam et al., 2016, Dunlop et al., 2020, Alders et al., 2020, van der Wijk et al., 2022, Sajjadian et al., 2023, Ayyash et al., 2024), which enrolled participants with MDD and HC, to investigate group differences in whole-brain rsFC in an unbiased fashion. Specifically, our primary objective was to perform the whole-brain fc-MVPA (Nieto-Castanon, 2022) of the fMRI data to reveal ROIs with altered rsFC in MDD. Our secondary objective was to study the relationship between the identified rsFC patterns and neurocognitive indices in MDD. The approach employed in this study marks a departure from previous studies using the CAN-BIND-1 dataset (van der Wijk et al., 2022, Anteraper et al., 2022), as it enables a more scalable analysis of a large number of fMRI scans due to dimensionality reduction.

As a data-driven voxel-based method, fc-MVPA examines the correlation of the fMRI signal between each voxel and every other voxel in the brain through principal component analysis (PCA), thereby reducing the dimensionality of the data. Investigating rsFC using fc-MVPA can help identify novel and useful network signatures of MDD and facilitate consensus in network topology, connectivity strength, and signal quality, thus mitigating the impact of hypothesis- or atlas-based techniques on the fMRI data. Furthermore, significant clusters rendered by fc-MVPA can simply serve as seed-ROIs for subsequent spatial characterization of rsFC in the post-hoc analysis, and the resulting rsFC patterns can be examined in the context of their relationship with clinical or neurocognitive measures.

2. Materials and methods

2.1. Study Population

As part of the CAN-BIND-1 trial (ClinicalTrials.gov Identifier: NCT01655706; Registration date: August 2nd, 2012) (MacQueen et al., 2019, Lam et al., 2016), 211 MDD patients and 112 HC (age range: 18–60 years) were recruited from seven Canadian clinical centres: Toronto General and Toronto Western Hospitals of the University Health Network (Toronto, Ontario), Centre for Addiction and Mental Health (Toronto, Ontario), University of Calgary (Calgary, Alberta), St. Joseph’s Healthcare Hamilton (Hamilton, Ontario), Providence Care Mental Health Services (Kingston, Ontario), and Djavad Mowafaghian Centre for Brain Health (Vancouver, British Columbia). All participants provided written informed consent, and standard participation and data transfer agreements were regulated by the Ontario Brain Institute and local ethics and legislative bodies of each respective centre. Participants with MDD were outpatients with a primary diagnosis of MDD as per the Diagnostic and Statistical Manual of Mental Disorders, fourth edition, text revision (DSM-IV-TR) and Mini International Diagnostic Interview (MINI). They had a Montgomery-Åsberg Depression Rating Scale (MADRS) (Montgomery and Åsberg, 1979) total score of ≥ 24 at screening (Lam et al., 2016), indicating moderate-to-severe symptoms. Detailed eligibility criteria for MDD and HC groups are provided in the Supplementary Material.

2.2. Neuroimaging data collection

At all seven sites, baseline T1-weighted anatomical and resting-state fMRI (rs-fMRI) scans were acquired at baseline using 3 T magnetic resonance imaging (MRI) scanners. T1-weighted anatomical images were acquired using site-specific sequences, but all conformed to a resolution of 1 mm isotropic voxels. Rs-fMRI images were acquired over ∼ 10 min (300 volumes) using TR = 2000 ms, TE = 30 ms, FA = 75°, and 4.0 mm isotropic voxels, and sequences were standardized across sites. Sequence parameters for T1-weighted and rs-fMRI scans are provided in Supplementary Tables S1 and S2 (adapted from MacQueen et al., 2019). All scans were pre-harmonized, consistent with previous CAN-BIND-1 publications (Ayyash et al., 2024, Harris et al., 2022, van der Wijk et al., 2022, Alders et al., 2020).

2.3. Data analysis

2.3.1. Analysis pipeline summary

The analysis pipeline aimed to identify altered rsFC patterns in MDD and examine their association with neurocognitive function. As illustrated in Fig. 1, fMRI scans were first preprocessed and assessed for quality. Whole-brain fc-MVPA was then applied to extract clusters with altered rsFC in MDD compared to HC. These clusters served as seeds in a post-hoc seed-based connectivity (SBC) analysis to map whole-brain rsFC differences. Neurocognitive performance was assessed using the Computerized Neurocognitive Assessment Vital Signs (CNS-VS) battery (Iverson et al., 2009), and between-group differences were tested using a two-sided t-test. Pearson correlation and linear models were then used to examine the relationship between altered rsFC patterns and neurocognitive domains that differed significantly between groups. Appropriate corrections for multiple comparisons were applied at each stage of the analysis to control for Type I error.

Fig. 1.

Fig. 1

Open in a new tab

Process flow diagram illustrating the relationship between significantly altered resting-state functional connectivity (rsFC) in MDD vs. HC and neurocognitive performance as assessed by the Computerized Neurocognitive Assessment Vital Signs (CNS-VS) battery. In the fMRI analysis pathway, functional connectivity multi-voxel pattern analysis (fc-MVPA) was first used to identify regions (clusters) with altered rsFC, which were then used as seeds for seed-based connectivity (SBC) analysis. In parallel, statistical comparisons were conducted on CNS-VS domain scores to identify differences between MDD and HC. Finally, correlations were examined between these scores and the connectivity strength of the rsFC patterns identified through SBC analysis.

2.3.2. Preprocessing

Prior to distribution, the CAN-BIND-1 dataset had been denoised using B0 field maps and time-slice corrected by the CAN-BIND Investigator Team. Anatomical and functional MRI data were preprocessed using the default preprocessing pipeline in MATLAB-based CONN Toolbox v.22a (RRID: SCR_009550, v.22a) (Nieto-Castanon and Whitfield-Gabrieli, 2021). Steps included realignment to the first volume, unwarping to correct geometric distortions, outlier volume detection for quality assurance (QA), normalization to the Montreal Neurological Institute (MNI) template, segmentation, resampling to 2 mm isotropic voxels, and smoothing using the 6 mm full-width half-maximum (FWHM) Gaussian kernel. Potential outlier volumes were identified using the Artifact Detection Tools (ART) toolbox v.7(19)11, defined as volumes with movement greater than 0.5 mm or global blood-oxygen-level-dependent (BOLD) signal intensity changes greater than 3 standard deviations. fMRI scans were denoised using a component-based noise correction method (CompCor) (Behzadi et al., 2007). Nuisance regressors included five principal components from white matter and cerebrospinal fluid (CSF) masks, six motion parameters and their first-order derivatives, outlier volumes (ART), and the first two Discrete Cosine Transform (DCT) terms (constant + linear drift). Residuals were then band-pass filtered (0.008 Hz and 0.09 Hz).

2.3.3. Functional connectivity multi-voxel pattern analysis

We performed whole-brain data-driven fc-MVPA as described in our previous work (Arnold Anteraper et al., 2019, Anteraper et al., 2020, Guell et al., 2021, Anteraper et al., 2021, Westfall et al., 2020, Morris et al., 2021). fc-MVPA employs a voxel-to-voxel approach that generates ROIs for standard seed-to-voxel analysis (Fig. 2). First, a principal components analysis (PCA) is used to reduce the dimensionality of resultant data by estimating a multivariate representation of the connectivity pattern, which involves computing pairwise connectivity patterns between each voxel and the rest of the brain. Then, a second PCA is run across all participants to retain the first N components. An omnibus F-test is then performed simultaneously on all N fc-MVPA components to identify voxels that show significant differences in connectivity patterns between groups/contrast of interest. In this study, 64 indicates the number of subject-specific PCA components retained when characterizing each subject’s voxel-to-voxel correlation structure. This is a form of subject-level dimensionality reduction typically used (e.g., in independent component analysis, as well as across most of CONN's voxel-to-voxel analyses) as the first step to reduce the complexity of subsequent analyses.

Fig. 2.

Fig. 2

Open in a new tab

For each voxel, the rsFC pattern between this voxel and the entire brain (top left) is computed separately for each subject. Each subject’s rsFC pattern is then characterized using a lower-dimensional representation (dots in the top-right graph).

Then, in fc-MVPA, “retaining N MVPA components” means that, for each seed voxel, CONN will compute the variability in seed-to-voxel connectivity maps across subjects, characterize that variability using PCA, and retain the first or largest N PCA factors/components. Instead of using training/test datasets, the component scores of these N factors for each subject (N values per subject) are used as a proxy for the entire connectivity pattern between this voxel and the rest of the brain in multivariate second-level analyses (e.g., to determine whether those patterns vary between MDD and HC groups). This entire process is repeated for every seed-voxel to obtain voxel-level maps of the second-level statistics. To maintain a conservative 20:1 ratio of subjects-to-components for principal component estimation, we chose N to be 5 in this study. To minimize the risk of false positives in parametric analyses (Eklund et al., 2016) and statistical inference of cluster-extent-based thresholding (Woo et al., 2014), we applied a more stringent Threshold-Free Cluster Enhancement (TFCE) approach using 1000 residual-randomization iterations at a familywise error-corrected 5 % false positive level (p-FWE < 0.05) and voxel thresholding of p < 0.001.

2.3.4. Post-hoc seed-based connectivity analysis

In the post-hoc SBC analysis, clusters identified by whole-brain fc-MVPA were used as seed-ROIs. The main purpose of this analysis was to identify fc-MVPA clusters showing significant between-group differences in rsFC. Correlation coefficients were Fisher’s z-transformed to improve normality prior to statistical testing. Group comparisons of rsFC between individuals with MDD and HC were conducted using two-sample t-tests, with TFCE applied as in the original fc-MVPA. Scanner site was included as a covariate at the group level (Fig. 1).

2.3.5. Computerized Neurocognitive Assessment Vital Signs (CNS-VS)

Outside the scanner, all participants completed the CNS-VS battery (Iverson et al., 2009). This 30-minute test is designed for quick clinical assessments. Unlike many computerized neurocognitive tests, it includes well-known and established tests such as verbal and visual memory, finger tapping, symbol digit coding, the Stroop test, a test of shifting attention, and the continuous performance test (see Supplementary Tables S3 and S4). CNS-VS scores are presented in a summary domain dashboard, where either one or multiple tests contribute to a particular domain score.

2.3.6. Correlation between functional connectivity and neurocognitive function

Out of 16 CNS-VS domain scores, the student’s t-test was first used to identify neurocognitive domains that were significantly different between MDD and HC. F-test was used to calculate the ratio of variance of the HC:MDD, and Cohen’s d was used to estimate the effect size. Pearson correlation was then run to examine the relationship between Fisher Z-transformed (z-scores) rsFC (relative rsFC strength) and statistically different CNS-VS scores (rsFC-to-score correlation). The α-level was set at 0.05, and the |r|>0.2 thresholding was applied to exclude very weak rsFC-to-score correlations. The false discovery rate (FDR) correction was applied to control for multiple comparisons. Since Pearson correlation represents a linear relationship normalized by standard deviations, linear models were used to visualize the correlation between rsFC and CNS-VS scores.

3. Results

Out of 211 participants with MDD and 112 HC enrolled in the CAN-BIND-1 trial, 151 MDD and 99 HC had baseline fMRI scans available. After QA and outlier detection, scans of 147 MDD participants and 98 HC were included in the analysis. Four scans were excluded from the MDD group (3 did not pass the QA, and 1 had excessive head motion), and 1 scan was excluded from the HC group (due to excessive head motion). Fig. S1 in the Supplementary Materials presents QA results. Table 1 provides a summary of the sociodemographic, clinical, and neurocognitive characteristics of our sample. Participants with MDD had significantly lower scores than HC along 8 CNS-VS domains, with the largest difference seen for psychomotor speed (Table 1 and Fig. 3a).

Table 1.

Sociodemographic, clinical, and neurocognitive characteristics of participants with MDD and HC at baseline.

MDD N = 147 HC N = 98
mean/N SD/% mean/N SD/% p-value
Sociodemographic characteristics
Sex, N (male | %) 56 38.10 35 35.71 0.71
Age in years (mean | SD) 34.25 12.20 33.08 10.91 0.43
Ethnicity (N white | %) 113 76.87 70 71.43 0.34
Marital status (N non-marrieda | %) 105 71.43 64 65.31 0.31
Employment (N working/student | %) 95 64.63 91 92.86 <0.001
Handedness (N right-handed | %) 129 87.76 84 85.71 0.64
Clinical characteristics mean SD mean SD p-value RoV Cohen’s d
MADRS Total at baseline (mean | SD) 29.76 5.65 0.90 1.76 < 0.001 0.097 6.386
CNS-VS domains
Psychomotor Speed 104.43 18.8 111.41 18.2 0.004 0.940 0.376
Neurocognition Index 99.72 12.7 104.04 10.8 0.006 0.731 0.360
Processing Speed 108.69 19.3 115.15 20.4 0.014 1.111 0.327
Motor Speed 100.12 15.7 104.48 14.0 0.026 0.804 0.290
Cognitive Flexibility 99.12 18.4 104.01 16.1 0.032 0.760 0.279
Executive Function 100.44 18.1 105.13 15.5 0.033 0.739 0.274
Composite Memory 100.67 17.1 105.18 15.9 0.037 0.864 0.271
Simple Attention 96.60 17.9 100.71 12.7 0.041 0.503 0.257
Visual Memory 103.05 16.3 106.59 15.6 0.090 0.913 0.220
Complex Attention 94.05 17.0 97.87 17.4 0.101 1.053 0.223
Verbal Memory 98.76 17.9 101.92 17.1 0.167 0.918 0.180
Non-Verbal Reasoning 104.71 12.4 102.49 13.1 0.199 1.119 0.174
Sustained Attention 103.30 10.5 104.99 10.2 0.247 0.955 0.163
Reaction Time 98.81 16.7 100.54 14.6 0.396 0.769 0.109
Working Memory 103.66 12.1 104.98 12.6 0.439 1.085 0.106
Social Acuity 99.88 15.2 101.40 15.3 0.452 1.023 0.099
Open in a new tab

p-values correspond to the results of the Student’s t-test (α-level < 0.05) or the Chi-Square test. The ratio of variance (RoV) represents the results of the F-test.

Cohen’s d displays the standardized difference between the two means.

Abbreviations: CNS-VS = Computerized Neurocognitive Assessment Vital Signs; MADRS = Montgomery-Åsberg Depression Rating Scale; RoV = Ratio of Variance (HC:MDD); SD = Standard Deviation.

a

Includes single, divorced, widowed, and separated.

Fig. 3.

Fig. 3

Open in a new tab

Distributions of (A) significantly different CNS-VS domain scores between MDD and HC, and (B) significant fc-MVPA clusters using the first five principal components, with variance normalized via Fisher Z-transformation.

3.1. Resting-state functional connectivity in MDD

3.1.1. Seed-clusters extracted from fc-MVPA

At baseline, six significant clusters with altered rsFC were extracted from fc-MVPA comparing MDD and HC. Fig. 3b illustrates the variance distribution (z-scores) of the six fc-MVPA clusters normalized using Fisher Z-transformation. The largest and most significant cluster was localized at the border of the left cerebellar crus I and lobule VI (MNI: [-36––74 –22]; cluster size: 117 voxels; referred to as cerebellar crus I). With scanner site included in the fc-MVPA as a covariate, the ranked clusters, in the order of cluster size and statistical significance level, were 1) left cerebellar crus I, 2) right precuneus, 3) left superior lateral occipital cortex (sLOC), 4) right ventral caudate, 5) left superior parietal lobule (SPL), and 6) left dorsal anterior cingulate cortex (dACC). Detailed descriptions of each fc-MVPA cluster are provided in Table 2. Fig. 4 illustrates the spatial representation of the size and anatomy of each fc-MVPA cluster.

Table 2.

Six significant clusters identified by group fc-MVPA, reflecting significant resting-state functional connectivity differences between participants with MDD and HC.

fc-MVPA Cluster # Peak MNI coordinates [x, y, z] Cluster Size p-FWE Peak Brodmann Area Brain Region
1 [-36, −74, –22] 117 0.00003 – (Left) Cerebellar Crus I
2 [+18, −66, +34] 98 0.0002 19, (Right) Visual AC (Right) Precuneus
3 [-24, −80, +32] 75 0.001 19, (Left) Visual AC (Left) sLOC
4 [+08, +06, +02] 75 0.001 – (Right) Ventral Caudate
5 [-20, −60, +56] 63 0.005 7, (Left) Visual Motor (Left) SPL
6 [-02, +38, +22] 47 0.03 32, (Left) dACC (Left) dACC
Open in a new tab

“–” indicates brain regions outside of defined Brodmann Areas.

Site was included as a covariate in the fc-MVPA.

Abbreviations: dACC = dorsal anterior cingulate cortex; fc-MVPA = functional connectivity multi-voxel pattern analysis; MNI = Montreal Neurological Institute; sLOC = superior lateral occipital cortex; SPL = superior parietal lobule; Visual AC = visual association cortex.

Fig. 4.

Fig. 4

Open in a new tab

Spatial representation of the six clusters identified by fc-MVPA. These clusters reflect brain regions showing significant rsFC differences between MDD and HC, identified using a voxel-wise threshold of p < 0.001 and a cluster-level FWE-corrected threshold of p < 0.05. Scanner site was included as a covariate. The six clusters are localized in the: (A) left cerebellar crus I, (B) right precuneus, (C) left superior lateral occipital cortex (sLOC), (D) right ventral caudate, (E) left superior parietal lobule (SPL), and (F) left dorsal anterior cingulate cortex (dACC). Abbreviations: dACC = dorsal anterior cingulate cortex; fc-MVPA = functional connectivity multi-voxel pattern analysis; FWE = family-wise error; HC = healthy controls; MDD = major depressive disorder; SPL = superior parietal lobule; sLOC = superior lateral occipital cortex.

3.1.2. Post-hoc seed-to-voxel connectivity analysis

The six clusters identified by fc-MVPA were then used as seed-ROIs for the post-hoc whole-brain seed-to-voxel rsFC analyses (i.e., SBC analysis), which revealed a total of 24 significantly different rsFC patterns between MDD and HC (Fig. 5 and Table 3).

Fig. 5.

Fig. 5

Open in a new tab

Post-hoc seed-to-voxel analyses of the six fc-MVPA clusters (A-F) comparing rsFC between MDD and HC. Scanner site was included as a covariate. Warm colors (yellow/red) indicate greater rsFC in MDD (MDD > HC); cool colors (blue/purple) indicate reduced rsFC in MDD (MDD < HC). Clusters R01-R24 represent 24 significantly different seed-based rsFC patterns. Boxplots show Fisher Z-transformed rsFC strength values for each cluster in HC (green) and MDD (orange). Abbreviations: dACC = dorsal anterior cingulate cortex; L = left; R = right; sLOC = superior lateral occipital cortex; SPL = superior parietal lobule.

Table 3.

Post-hoc seed-to-voxel analyses with fc-MVPA clusters as seed-ROIs, reflecting significant resting-state functional connectivity differences between participants with MDD and HC.

fc-MVPA cluster SBC cluster Peak MNI coordinates [x, y, z] Cluster size p-FWE Peak MNI Brodmann Area Brain Region
(Left) Cerebellar Crus I R01 [+06, +26, +54] 264 0.002 8, (Right) Frontal Eye Fields (−) (Right) Superior Frontal Gyrus/dmPFC
R02 [+00, +50, +12] 203 0.009 32, (Right) dACC (−) (Right) dACC
R03 [-08, −84, +24] 173 0.02 19, (Left) Visual AC (+) (Left) Visual AC



(Right) Precuneus R04 [+00, +36, +24] 2913 <0.00001 32, (Right) dACC (−) (Right) dACC/Pregenual area
R05 [+26, −88, +26] 2627 <0.00001 19, (Right) Visual AC (+) (Right) Middle Occipital Gyrus
R06 [-56, −42, +42] 761 <0.00001 39, (Left) Angular Gyrus (−) (Left) IPL
R07 [-40, −66, −16] 711 <0.00001 37, (Left) Fusiform (+) (Left) Fusiform Gyrus
R08 [–22, −80, +32] 424 0.00005 19, (Left) Visual AC (+) (Left) sLOC
R09 [+64, −42, +36] 353 0.0002 40 (Right) Supramarginal Gyrus (−) (Right) IPL
R10 [-38, +26, +36] 345 0.0003 9, (Left) dlPFC (−) (Left) Middle Frontal Gyrus/dlPFC
R11 [-34, +18, +00] 254 0.002 13, (Left) Insula (−) (Left) Insula
R12 [+20, +46, +30] 168 0.02 9, (Right) dlPFC (−) (Right) Middle Frontal Gyrus/dlPFC
R13 [+28, –32, +08] 149 0.03 – (+) (Right) Tapetum fibers



(Left) sLOC R14 [+02, −58, +08] 1593 <0.00001 – (+) (Right) Precuneus
R15 [+40, −56, +20] 263 0.002 39, (Right) Angular Gyrus (+) (Right) Angular Gyrus/IPL
R16 [-24, −42, −12] 136 <0.05 37, (Left) Fusiform (+) (Left) Fusiform Gyrus



(Right) Ventral Caudate R17 [+18, −12, +00] 342 0.0002 – (−) (Right) Posterior limb of the Internal Capsule
R18 [-02, −74, –32] 310 0.0004 – (−) (Left) Cerebellar Crus II/Vermis
R19 [-44, −56, –32] 186 0.009 – (−) (Left) Cerebellar Crus I
R20 [-10, +02, +02] 153 0.02 – (−) (Left) Genu of the Internal Capsule



(Left) SPL R21 [+36, −78, +16] 344 0.0004 19, (Right) Visual AC (+) (Right) IPL/Middle Occipital Gyrus
R22 [+02, +46, +44] 330 0.0005 – (−) (Right) Superior Frontal Gyrus/dmPFC



(Left) dACC R23 [+22, −64, +32] 1406 <0.00001 19, (Right) Visual AC (−) (Right) sLOC
R24 [-12, −70, +56] 244 0.002 7, (Left) Visual Motor (−) (Left) SPL
Open in a new tab

“–“ indicates brain regions outside of defined Brodmann Areas.

(+) indicates increased resting-state functional connectivity in participants with MDD (MDD > HC).

(−) indicates decreased resting-state functional connectivity in participants with MDD (MDD < HC).

Abbreviations:dACC = dorsal anterior cingulate cortex; dlPFC = dorsolateral prefrontal cortex; dmPFC = dorsomedial prefrontal cortex; fc-MVPA = functional connectivity multi-voxel pattern analysis; HC = healthy controls; IPL = inferior parietal lobule; MDD = major depressive disorder; MNI = Montreal Neurological Institute; SBC = seed-based connectivity; sLOC = superior lateral occipital cortex; SPL = superior parietal lobule; Visual AC = visual association cortex.

Fig. 4 illustrates altered rsFC for each fc-MVPA seed, and boxplots display the differences in normalized rsFC strength between MDD and HC. Compared to HC, individuals with MDD had both increased and reduced rsFC patterns across the brain. Left cerebellar crus I (fc-MVPA cluster 1) had reduced rsFC with the right dorsomedial prefrontal cortex (dmPFC) and right dACC, but increased rsFC with the left visual association cortex (AC).

Right precuneus (fc-MVPA cluster 2) also had reduced rsFC with the right dACC, bilateral inferior parietal lobule (IPL), bilateral dlPFC, and left insula, but increased rsFC with the right visual association cortex (AC), left fusiform gyrus, and left sLOC. Left sLOC (fc-MVPA cluster 3) had increased rsFC with the right precuneus, right IPL, and left fusiform gyrus. Right ventral caudate (fc-MVPA cluster 4) had reduced rsFC with the cerebellar crus II/vermis and crus I. Left SPL (fc-MVPA cluster 5) had reduced rsFC with the dmPFC and increased rsFC with the right IPL. Left dACC (fc-MVPA cluster 6) had reduced rsFC with the right sLOC and left SPL.

3.1.3. Correlation between functional connectivity and neurocognitive function

Out of 24 significantly different rsFC patterns identified in the post-hoc SBC analysis, five unique rsFC patterns were found to be significantly correlated with CNS-VS domain scores. On the other hand, out of the eight statistically different neurocognitive domains between MDD and HC (Table 1), only five exhibited association with rsFC: composite memory, neurocognition index, processing speed, executive function, and simple attention. The associations between the five rsFC patterns and five neurocognitive domains are summarized in Table 4. The strength of the associations is illustrated by slopes of the linear models in Fig. 6, along with the corresponding rsFC. The steepness of the slope provides a visual indication of the strength of the correlation between rsFC and neurocognitive functions.

Table 4.

Pearson correlation coefficients (r) between the strength of resting-state functional connectivity in fc-MVPA clusters and CNS-VS domain scores in MDD and HC.

rsFC CNS-VS Domain MDD HC
r p-FDR r p-FDR
(A) (L) Cerebellar Crus 1—(R) dACC Composite Memory 0.011 0.898 −0.332 0.007
Neurocognition Index −0.025 0.875 −0.277 0.033
Processing Speed 0.030 0.875 −0.243 0.045
Executive Function −0.027 0.875 −0.224 < 0.05
Simple Attention 0.034 0.875 −0.221 < 0.05
(B) (R) Precuneus—(L) IPL Processing Speed −0.037 0.686 −0.293 0.032
Simple Attention 0.011 0.898 −0.332 0.007
(C) (R) Precuneus—(R) IPL Simple Attention −0.027 0.875 −0.224 < 0.05
(D) (R) Ventral Caudate—(L) Cerebellar Crus I Simple Attention −0.016 0.846 0.285 0.019
(E) (L) dACC—(R) sLOC Executive Function 0.001 0.995 0.310 0.018
Open in a new tab

Abbreviations:CNS-VS = Computerized Neurocognitive Assessment Vital Signs; dACC = dorsal anterior cingulate cortex; dlPFC = dorsolateral prefrontal cortex; dmPFC = dorsomedial prefrontal cortex; fc-MVPA = functional connectivity multi-voxel pattern analysis; HC = healthy controls; IPL = inferior parietal lobule; (L) = left; MDD = major depressive disorder; p-FDR = p-value adjusted for multiple comparisons using the Benjamini-Hochberg false discovery rate procedure; (R) = Right; rsFC = resting-state functional connectivity; sLOC = superior lateral occipital cortex; SPL = superior parietal lobule.

Fig. 6.

Fig. 6

Open in a new tab

Linear models illustrating significant correlations (p < 0.05) between CNS-VS scores in 5 out of 16 neurocognitive domains and 5 out of 24 seed-based rsFC patterns from the post-hoc analysis. Blue lines represent linear models for MDD; red lines represent HC. Shaded areas denote 95 % confidence intervals. The steepness of the slope provides a visual indication of the strength of the correlation between rsFC and neurocognitive functions. Significant correlations between neurocognitive performance and rsFC were observed in HC but were absent in participants with MDD. Abbreviations: CRCrI = cerebellar crus I; dACC = dorsal anterior cingulate cortex; IPL = inferior parietal lobule; PCu = precuneus; sLOC = superior lateral occipital cortex; VC = ventral caudate.

In general, a weak-to-moderate correlation was present between rsFC strength and CNS-VS scores in the HC group, whereas this correlation was significantly diminished or absent in MDD. After correction for multiple comparisons, the rsFC between the left cerebellar crus I (fc-MVPA cluster 1) and right dACC showed a weakened negative association with CNS-VS scores in MDD across several domains, including composite memory, neurocognition index, processing speed, executive function, and simple attention (Fig. 7). The rsFC between the right precuneus (fc-MVPA cluster 2) and IPL showed a weakened association with processing speed (left IPL only) and simple attention (bilateral IPL). The rsFC between the right ventral caudate (fc-MVPA cluster 4) and left cerebellar crus I had a weakened association with simple attention. The rsFC between the left dACC (fc-MVPA cluster 6) and right sLOC had a weakened association with executive function. The rsFC of the left sLOC (fc-MVPA cluster 3) and left SPL (fc-MVPA cluster 5) had no significant association with CNS-VS domain scores in MDD or HC.

Fig. 7.

Fig. 7

Open in a new tab

Spatial mapping of the five rsFC patterns that were significantly correlated with CNS-VS domain scores across four fc-MVPA clusters (blue nodes): (A) left cerebellar crus I, (B) right precuneus, (C) right ventral caudate, and (D) left dACC. No significant correlations were observed for the remaining two fc-MVPA clusters (left sLOC and left SPL). Golden nodes represent functionally connected brain regions identified in the post-hoc SBC analysis. Thick golden lines indicate five rsFC patterns with significant correlation to CNS-VS scores; thin black lines represent 14 rsFC patterns with no significant correlation.

4. Discussion

The current study used rs-fMRI scans from participants enrolled in the CAN-BIND-1 trial to compare rsFC between MDD and HC. In lieu of using atlas-based parcellation or fixed-radius spherical ROIs, we applied data-driven voxel-to-voxel fc-MVPA. Six clusters were identified by fc-MVPA, indicating whole-brain rsFC differences in MDD compared to HC. The identified fc-MVPA clusters were then used as seed-ROIs for further post-hoc characterization, and the association between rsFC strength and neurocognitive indices was explored.

The largest cluster was localized at the border of the left cerebellar crus I and lobule VI, which overlaps with cerebellar regions involved in attentional processes (King et al., 2019). Consistent with these results, previous fMRI studies have demonstrated a cerebello-cerebral coupling of lobule VI with prefrontal, posterior parietal, and limbic regions in patients with MDD (Depping et al., 2018, Habas, 2018). Krienen and Buckner (2009) demonstrated the correlation of crus I to the dlPFC and medial prefrontal cortices, and of lobule VI to the anterior prefrontal cortex, which may confirm the involvement of crus I/lobule VI in cognitive control and executive function. This is supported by research mapping crus I to association networks, including those related to executive and cognitive control, sensory-motor integration, and the DMN (Krienen and Buckner, 2009, Buckner et al., 2011).

From the functional connectivity perspective, participants with MDD in our study demonstrated reduced rsFC from the left cerebellar crus I/lobule VI to the right dmPFC/superior frontal gyrus/frontal eye fields and right dACC, but increased rsFC to the left visual AC relative to HC. The dmPFC, situated in the superior frontal gyrus, overlaps with the area known in the literature as the “dorsal nexus,” which serves as an intersection point of multiple ICNs, notably the CEN, affective network, and DMN (Sheline et al., 2010). Previous studies have suggested that the dorsal nexus may be implicated in attention and cognitive control (Brefczynski-Lewis et al., 2007, Le et al., 2017), and it has been speculated that increased DMN-dorsal nexus rsFC might contribute to shifts in attention and increased self-focus, both of which may impair task performance (Sheline et al., 2010).

Our results also demonstrate that participants with MDD exhibit decreased cerebellar rsFC with the dACC, a node of the SN involved in cognitive control and response inhibition (Buckner et al., 2011). Thus, through altered rsFC with the dACC, the cerebellum might contribute to salience allocation and active attention in MDD. In fact, it has been suggested that crus I/lobule VI is potentially involved in determining the valence of salient emotional cues and behavioral responses to such cues (Habas, 2018). This decrease in rsFC between the cerebellum and more critical anterior nodes involved in cognitive processes and executive function may be compensated by an increase in rsFC with the posterior nodes of the occipital cortex. For instance, we observed increased rsFC between the left cerebellar crus I/lobule VI and left visual AC. The primary role of the visual AC is to integrate visual information with other sensory inputs and cognitive processes, and the visual AC has been shown to be associated with dysregulation of visual working memory updating and retention of unrelated negative information in MDD (Le et al., 2017). This information may represent negative self-referential thoughts and rumination, also reflective of the role of the cerebellum in the activity of the DMN (Guo et al., 2015).

The second, third, and fourth largest fc-MVPA clusters were localized in the precuneus, sLOC, and ventral caudate, respectively, with the cluster size (98 voxels) for the precuneus being close to the size of the cerebellar crus I/lobule VI cluster (fc-MVPA cluster 1). Precuneus is a node of the posterior DMN (Fransson, 2005) and is involved in a variety of higher-order functions, including a sense of self, mental imagery strategies, cue reactivity, integrative processing of information related to the perception of the environment, and episodic memory recall (Zhang and Li, 2012). It is also a key area involved in spatial information processing, memory, and navigation (Cavanna and Trimble, 2006). Previously, precuneal rsFC has been reported in the pathophysiology of MDD (Demchenko et al., 2022), most often associated with the hyperconnectivity of the DMN and linked with a tendency of individuals with MDD to ruminate. sLOC, on the other hand, is one of the major areas involved in visual object processing and face perception (Nagy et al., 2012, Grill-Spector et al., 2001). Together, precuneus and sLOC form visual recognition circuits, which have been implicated in MDD and shown to be associated with antidepressant response (Wang et al., 2014b). Although their precise functional role is yet to be determined, the precuneus and sLOC can be conceptualized as providing visual information from the occipital cortex and spatial information from the parietal cortex, with access to the hippocampus and its episodic memory system. Such information connects to and from the cornu ammonis 3 (CA3) region of the hippocampus, forming a basis for episodic memory with spatial, object, and face components (Kesner and Rolls, 2015). Due to the connectivity between the hippocampus and limbic regions involved in negative emotions, this system may contribute to increased ruminating thoughts in depression and referring these thoughts to the sense of self, which behaviorally may be associated with low self-esteem in MDD.

Post-hoc SBC analyses demonstrated vast patterns of altered rsFC of the precuneus and sLOC with other nodes of the DMN, CEN, SN, and related visual information processing network. More specifically, the precuneus had reduced rsFC with the bilateral dlPFC in MDD, consistent with theories supporting CEN hypoconnectivity contributing to diminished cognitive control and top-down regulation over emotional processing (Demchenko et al., 2022). The rsFC with bilateral IPL was also disrupted: right precuneus-bilateral IPL rsFC was reduced, whereas left sLOC-right IPL rsFC was elevated. Anatomically, IPL includes the angular gyrus (AG) – one of the core nodes of the DMN, which serves as a “connector hub” integrating multimodal sensory information and serving as a top-down regulator of self-referential processing (Kim, 2010). In MDD, AG dysconnectivity contributes to impairments in integrating mental representations and contextual information with the social environment (Kim, 2010). In fact, our previous examination of the CAN-BIND-1 dataset (Anteraper et al., 2022), where we evaluated changes in the cerebello-cerebral rsFC using the functional gradient analysis method, revealed increased rsFC between cerebellar crus I and bilateral AG, which, together with the results of the current study, support the general contribution of AG and association circuits to network dysconnectivity profile in MDD.

Further, the ventral caudate was another significant fc-MVPA cluster with disrupted rsFC. Caudate, as part of the striatum, feeds into the affective division of the cortico-striato-thalamic loop that mediates reward-seeking behaviors and reward valuation (Gold, 2003). Reduced caudate volumes have been found in MDD participants (Koolschijn et al., 2009), which may map onto our fc-MVPA findings of altered rsFC. At the same time, the caudate is vastly interconnected with the basal ganglia and cerebellum, forming the SMN – a set of interconnected brain regions coordinating motor planning and initiation, error detection, and executing response in relation to incoming inputs (Demchenko et al., 2022). A recent meta-analysis (Yan et al., 2019) supports reduced within-network connectivity of the SMN in MDD, which is likely associated with symptoms of psychomotor retardation, as well as the contribution of an imbalance between SMN and DMN connectivity to perseverative cognition in psychiatric disorders (Makovac et al., 2020, Stern et al., 2022). The results of our study further corroborate this theory, as we observed decreased rsFC between the ventral caudate and regions of the left cerebellum, including crura I and II, as well as the white matter of the internal capsule, which contains sensory and motor projection fibers from the basal ganglia.

Other clusters identified by fc-MVPA were localized in the left SPL and dACC. SPL comprises a node of the CEN involved in spatial orientation, visual input processing, and general parietal lobe functions (Koenigs et al., 2009). Together with the precuneus and sLOC, SPL also feeds into the visual recognition circuits (Wang et al., 2014b), and its general involvement in the identified rsFC patterns indicates that spaces, objects, and faces may potentially contribute to the formation of memories associated with negative thoughts seen in MDD and may even trigger their retrieval (Kesner and Rolls, 2015). The left SPL had increased rsFC with the right IPL, likely feeding into the DMN hyperactivity profile (Demchenko et al., 2022). Reduced rsFC with the right dmPFC (the “dorsal nexus”) was also seen, which is consistent with theories supporting CEN within-network hypoconnectivity contributing to diminished cognitive control and top-down regulation over emotional processing (Demchenko et al., 2022) and the presence of compensatory mechanisms whereby nodes of the CEN, such as SPL, become more functionally coherent with non-cognitive nodes of the neocortex, as well as extracerebral regions (Sheng et al., 2021, Hartwigsen, 2018).

On the other hand, the dACC is a part of the SN and is one of the most commonly reported regions affected in MDD (Demchenko et al., 2022). Decreased SN connectivity contributes to attentional difficulties, anhedonia, and poor incentive salience in MDD (Demchenko et al., 2022). Our study demonstrates that these changes are also detectable at rest. Specifically, we found decreased rsFC of the left dACC with the right sLOC and left SPL when dACC was used as a seed (fc-MVPA cluster 6), as well as reduced left cerebellar crus I/lobule VI-right dACC (from fc-MVPA cluster 1) and right precuneus-right dACC (from fc-MVPA cluster 2) rsFC, further supporting the theory of reduced connectivity of SN in MDD. These disruptions are likely associated with impaired allocation of attention to the stimuli coming from the environment via visual association networks, as well as undermining the integration of this information with the social context and the sense of self. Whereas the dACC has been traditionally conceptualized as interconnected with the anterior insula, another prominent node of the SN, future studies may perform a more in-depth investigation of dACC rsFC with nodes of the cerebellum and posterior cortex to elucidate their functional significance for MDD pathophysiology and symptoms.

Our CNS-VS results are consistent with the literature, suggesting that participants with MDD are subject to multiple neurocognitive deficits (Hammar et al., 2022). Namely, deficits in multiple test domains of composite memory, psychomotor speed, cognitive flexibility, and neurocognition index were prominent in our sample. Further, deficits in certain single test domains, such as executive function from the shifting attention test, simple attention from the continuous performance test, motor speed from the finger tapping test, and processing speed from the symbol digit coding test, may be more important than others. These processes tend to be mediated by neocortical regions of the CEN, SN, DMN, and SMN, although the cerebellum, as an extracerebral region and also part of the DMN and SMN, appears to contribute to the “fine-tuning” of cortical neural outputs involved in these higher-order neurocognitive processes (Koziol et al., 2014, Demchenko et al., 2022, Anteraper et al., 2022).

A weak-to-moderate correlation between certain CNS-VS domains and rsFC was present in the HC group but not in the MDD group, likely reflecting a pathological shift in the rsFC of brain regions that function as nodes of the DMN, SN, and SMN. Based on the results of our study, the domain of composite memory appears to be most associated with the rsFC of the left cerebellar crus I/lobule VI (fc-MVPA cluster 1) and right dACC, reflecting the involvement of SN (Demchenko et al., 2022, Li et al., 2018). The domain of processing speed displays an association with the rsFC of the left cerebellar crus I/lobule VI (fc-MVPA cluster 1) and right precuneus (fc-MVPA cluster 2) in healthy brains with nodes of the SN (right dACC) and DMN (left IPL), respectively, whereas the single test domains of executive function and simple attention are associated with the rsFC of the left cerebellar crus I/lobule VI (fc-MVPA cluster 1), right precuneus (fc-MVPA cluster 2), right ventral caudate (fc-MVPA cluster 4), and left dACC (fc-MVPA cluster 6). Notably, the rsFC between the left cerebellar crus I (fc-MVPA cluster 1) and right dACC was the only identified rsFC pattern that exhibited an association with five domains of the CNS-VS, highlighting the prominent role of these two areas across different neurocognitive functions in healthy brains that may be compromised in MDD. While this diverse association may be expected for the dACC (Shenhav et al., 2013), a node of the SN, our study is one of the first to showcase the simultaneous involvement of the left cerebellum in composite memory, neurocognition index, processing speed, executive function, and simple attention, where the cerebellum may be involved in the fine-tuning and adjustment of these processes. Interestingly, this simultaneous contribution to many domains of neurocognition is qualitatively proportional to the cluster size of the cerebellar region, which was the largest and most significant across the entire brain in our whole-brain fc-MVPA.

5. Conclusions

This study suggests that aberrant rsFC involving the cerebellar crus I, precuneus, sLOC, ventral caudate, SPL, and dACC, and its association with specific neurocognitive functions in healthy brains but lack thereof in MDD, may represent a potential neurobiological signature of MDD pathophysiology. Among these, the left cerebellar crus I emerged as the largest and most significant contributor of the altered rsFC profile in MDD, likely playing a critical role in the “fine-tuning” of multiple neurocognitive domains. Further research is needed to determine whether the rsFC patterns identified via fc-MVPA, using a voxel-to-voxel correlation coefficient matrix, can be replicated across independent samples and study settings. Future work should also explore the laterality of these rsFC features. As a data- driven method, fc-MVPA captures individual variability in brain connectivity without relying on a priori assumptions, offering a valuable framework for supporting existing findings and enabling more detailed analyses of functional dysconnectivity.

CRediT authorship contribution statement

Alice Rueda: Writing – review & editing, Writing – original draft, Visualization, Methodology, Formal analysis, Conceptualization. Ilya Demchenko: Writing – review & editing, Writing – original draft, Investigation, Conceptualization. Vanessa K. Tassone: Writing – review & editing, Writing – original draft. Fatemeh Gholamali Nezhad: Writing – review & editing, Writing – original draft. Vanessa Peters: Writing – review & editing, Writing – original draft. Nathan W. Churchill: Writing – review & editing. Benicio N. Frey: Writing – review & editing. Stefanie Hassel: Writing – review & editing. Raymond W. Lam: Writing – review & editing. Roumen V. Milev: Writing – review & editing. Daniel J. Müller: Writing – review & editing. Tom A. Schweizer: Writing – review & editing. Stephen C. Strother: Writing – review & editing. Valerie H. Taylor: Writing – review & editing. Sidney H. Kennedy: Writing – review & editing. Sheeba Arnold Anteraper: Writing – review & editing, Conceptualization. Venkat Bhat: Writing – review & editing, Supervision, Conceptualization.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors would like to acknowledge the contributions of Mojdeh Zamyadi and Jacqueline Harris for data quality control, and of Stephen R. Arnott, Andrew D. Davis, and Geoffrey B. Hall for sequence assessment and standardization. CAN-BIND Investigator Team: Benicio N. Frey, Pierre Blier, Raymond Lam, Claudio Soares, Roumen Milev, Sagar Parikh, Susan Rotzinger, Jane Foster, Faranak Farzan, Gustavo Turecki, Francesco Leri, Daniel J. Mueller, Valerie H. Taylor, Rudolf Uher, Nicholas Bock, Frank Rudzicz, Joshua Rosenblat, Nathan Churchill, Lena Quilty, Kate Harkness, Andrew KComt, Kathryn Schade, Keith Ho, Franca Placenza, Stefanie Hassel, Sidney H. Kennedy, Glenda M. MacQueen. ClinicalTrials.gov Registration: NCT01655706 (https://clinicaltrials.gov/study/NCT01655706).

Funding

This project did not receive any specific grants from funding agencies in the public, commercial, or not-for-profit sectors.

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.nicl.2025.103840.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Supplementary Data 1
mmc1.docx (145.5KB, docx)

Data availability

Data will be made available on request.

References

  1. Alders G.L., Davis A.D., MacQueen G., Strother S.C., Hassel S., Zamyadi M., Sharma G.B., Arnott S.R., Downar J., Harris J.K., et al. Escitalopram ameliorates differences in neural activity between healthy comparison and major depressive disorder groups on an fMRI emotional conflict task: a can-bind-1 study. J. Affect. Disord. 2020;264:414–424. doi: 10.1016/j.jad.2019.11.068. [DOI] [PubMed] [Google Scholar]
  2. Anteraper S.A., Collin G., Guell X., Schneidert T., Molokotos E., Henriksen M.T., Mesholam-Gately R., Thermenos H.W., Seidman L.J., Keshavan M.S., Gabrieli J.D.E., Whitfield-Gabrieli S. Altered resting-state functional connectivity in young children at familial high risk for psychotic illness: a preliminary study. Schizophr. Res. 2020 doi: 10.1016/j.schres.2020.01.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Anteraper S.A., Gopinath K., Hoch M.J., Jiang H., Cheng H., Sammet C.L., He T., Pula S., Taylor M.J., Tassiopoulos K., Dewhurst S., Gongvatana A., Cohen R.A. A comprehensive data-driven analysis framework for detecting impairments in brain function networks with resting state fMRI in HIV-infected individuals on cART. J. Neurovirol. 2021;27(2):239–248. doi: 10.1007/s13365-020-00906-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Anteraper S.A., Guell X., Lee Y.J., Raya J., Demchenko I., Churchill N.W., Frey B.N., Hassel S., Lam R.W., MacQueen G.M., Milev R., Schweizer T.A., Strother S.C., Whitfield-Gabrieli S., Kennedy S.H., Bhat V. Cerebello-cerebral Functional Connectivity Networks in Major Depressive Disorder: a CAN-BIND-1 Study Report. The Cerebellum URL. 2022 doi: 10.1007/s12311-021-01353-5. https://link.springer.com/10.1007/ [DOI] [PubMed] [Google Scholar]
  5. Arnold Anteraper S., Guell X., D'Mello A.M., Joshi N., Whitfield-Gabrieli S., Joshi G. Disrupted cerebrocerebellar intrinsic functional connectivity in young adults with high-functioning autism spectrum disorder: a data-driven, whole-brain, high-temporal resolution functional magnetic resonance imaging study. Brain Connect. 2019;9(1):48–59. doi: 10.1089/brain.2018.0581. [DOI] [PubMed] [Google Scholar]
  6. Ayyash S., Davis A.D., Alders G.L., MacQueen G., Strother S.C., Hassel S., Zamyadi M., Arnott S.R., Harris J.K., Lam R.W., et al. Assessing remission in major depressive disorder using a functional-structural data fusion pipeline: a can-bind-1 study. IBRO Neuroscience Reports. 2024;16:135–146. doi: 10.1016/j.ibneur.2023.12.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Behzadi Y., Restom K., Liau J., Liu T.T. A component based noise correction method (CompCor) for BOLD and perfusion based fMRI. Neuroimage. 2007;37:90–101. doi: 10.1016/j.neuroimage.2007.04.042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Brefczynski-Lewis J.A., Lutz A., Schaefer H.S., Levinson D.B., Davidson R.J. Neural correlates of attentional expertise in long-term meditation practitioners. PNAS. 2007;104:11483–11488. doi: 10.1073/pnas.0606552104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Buckner R.L., Krienen F.M., Castellanos A., Diaz J.C., Thomas Yeo B.T. The organization of the human cerebellum estimated by intrinsic functional connectivity. J. Neurophysiol. 2011;106:2322–2345. doi: 10.1152/jn.00339.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Buyukdura J.S., McClintock S.M., Croarkin P.E. Psychomotor retardation in depression: biological underpinnings, measurement, and treatment. Prog. Neuropsychopharmacol. Biol. Psychiatry. 2011;35:395–409. doi: 10.1016/j.pnpbp.2010.10.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cavanna A.E., Trimble M.R. The precuneus: a review of its functional anatomy and behavioural correlates. Brain. 2006;129:564–583. doi: 10.1093/brain/awl004. [DOI] [PubMed] [Google Scholar]
  12. Cole D.M., Smith S.M., Beckmann C.F. Advances and pitfalls in the analysis and interpretation of resting-state fmri data. Front. Syst. Neurosci. 2010;4:1459. doi: 10.3389/fnsys.2010.00008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dai L., Zhou H., Xu X., Zuo Z. Brain structural and functional changes in patients with major depressive disorder: a literature review. PeerJ. 2019;7:e8170. doi: 10.7717/peerj.8170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Demchenko I., Tassone V.K., Kennedy S.H., Dunlop K., Bhat V. Intrinsic connectivity networks of glutamate-mediated antidepressant response: a neuroimaging review. Front. Psych. 2022;13 doi: 10.3389/fpsyt.2022.864902. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Depping, M.S., Schmitgen, M.M., Kubera, K.M., Wolf, R.C., 2018. Cerebellar Contributions to Major Depression. Frontiers in Psychiatry 9, Article 634. URL: https://www.frontiersin.org/article/10.3389/fpsyt.2018.00634/full, doi:10.3389/fpsyt.2018.00634. [DOI] [PMC free article] [PubMed]
  16. Drysdale A.T., Grosenick L., Downar J., Dunlop K., Mansouri F., Meng Y., Fetcho R.N., Zebley B., Oathes D.J., Etkin A., et al. Resting-state connectivity biomarkers define neurophysiological subtypes of depression. Nat. Med. 2017;23:28–38. doi: 10.1038/nm.4246. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Dunlop K., Rizvi S.J., Kennedy S.H., Hassel S., Strother S.C., Harris J.K., Zamyadi M., Arnott S.R., Davis A.D., Mansouri F., et al. Clinical, behavioral, and neural measures of reward processing correlate with escitalopram response in depression: a canadian biomarker integration network in depression (can-bind-1) report. Neuropsychopharmacology. 2020;45:1390–1397. doi: 10.1038/s41386-020-0688-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Eklund A., Nichols T.E., Knutsson H. Cluster failure: why fMRI inferences for spatial extent have inflated false-positive rates. PNAS. 2016;113:7900–7905. doi: 10.1073/pnas.1602413113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Fransson P. Spontaneous low-frequency BOLD signal fluctuations: an fMRI investigation of the resting-state default mode of brain function hypothesis. Hum. Brain Mapp. 2005;26:15–29. doi: 10.1002/hbm.20113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Gold J.I. Linking reward expectation to behavior in the basal ganglia. Trends Neurosci. 2003;26:12–14. doi: 10.1016/s0166-2236(02)00002-4. [DOI] [PubMed] [Google Scholar]
  21. Grill-Spector K., Kourtzi Z., Kanwisher N. The lateral occipital complex and its role in object recognition. Vision Res. 2001;41:1409–1422. doi: 10.1016/S0042-6989(01)00073-6. [DOI] [PubMed] [Google Scholar]
  22. Guell X., Anteraper S.A., Gardner A., Whitfield-Gabrieli S., Kay-Lambkin F., Gabrieli J.D.E., Stanwell P. Functional connectivity changes in retired rugby league players: a data-driven, whole-brain, functional magnetic resonance imaging study. J. Neurotrauma. 2021;38(2):220–227. doi: 10.1089/neu.2020.7091. [DOI] [PubMed] [Google Scholar]
  23. Guo, W., Liu, F., Liu, J., Yu, M., Zhang, Z., Liu, G., Xiao, C., Zhao, J., 2015. Increased Cerebellar-Default-Mode-Network Connectivity in Drug-Naive Major Depressive Disorder at Rest. Medicine 94, e560. URL: https://journals.lww.com/00005792-201503010-00003, doi:10.1097/MD.0000000000000560. [DOI] [PMC free article] [PubMed]
  24. Habas C. Research note: a resting-state, cerebello-amygdaloid intrinsically connected network. Cerebellum & Ataxias. 2018;5:1–4. doi: 10.1186/s40673-018-0083-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Hammar Å., Ronold E.H., Rekkedal G.Å. Cognitive impairment and neurocognitive profiles in major depression—a clinical perspective. Front. Psych. 2022;13 doi: 10.3389/fpsyt.2022.764374. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Harris J.K., Hassel S., Davis A.D., Zamyadi M., Arnott S.R., Milev R., Lam R.W., Frey B.N., Hall G.B., Müller D.J., et al. Predicting escitalopram treatment response from pre-treatment and early response resting state fmri in a multi-site sample: a can-bind-1 report. NeuroImage: Clinical. 2022;35 doi: 10.1016/j.nicl.2022.103120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hartwigsen, G., 2018. Flexible Redistribution in Cognitive Networks. Trends in Cognitive Sciences 22, 687–698. URL: http://dx. doi.org/10.1016/j.tics.2018.05.008https://linkinghub.elsevier.com/retrieve/pii/S1364661318301220, doi:10.1016/ j.tics.2018.05.008. [DOI] [PubMed]
  28. Hou X., Mei B., Wang F., Guo H., Li S., Wu G., Zang C., Cao B. Neural activity in adults with major depressive disorder differs from that in healthy individuals: a resting-state functional magnetic resonance imaging study. Front. Psych. 2022;13 doi: 10.3389/fpsyt.2022.1028518. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Huys Q.J., Maia T.V., Frank M.J. Computational psychiatry as a bridge from neuroscience to clinical applications. Nat. Neurosci. 2016;19:404–413. doi: 10.1038/nn.4238. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Iverson G.L., Brooks B.L., Young A.H. Identifying neurocognitive impairment in depression using computerized testing. Appl. Neuropsychol. 2009;16:254–261. doi: 10.1080/09084280903297594. [DOI] [PubMed] [Google Scholar]
  31. Kennedy S.H., Lam R.W., Rotzinger S., Milev R.V., Blier P., Downar J., Evans K.R., Farzan F., Foster J.A., Frey B.N., et al. Symptomatic and functional outcomes and early prediction of response to escitalopram monotherapy and sequential adjunctive aripiprazole therapy in patients with major depressive disorder: a can-bind-1 report. J. Clin. Psychiatry. 2019;80:2236. doi: 10.4088/JCP.18m12202. [DOI] [PubMed] [Google Scholar]
  32. Kesner, R.P., Rolls, E.T., 2015. A computational theory of hippocampal function, and tests of the theory: New developments. Neuroscience and Biobehavioral Reviews 48, 92–147. URL: http://dx.doi.org/10.1016/j.neubiorev.2014.11.009, doi:10.1016/j.neubiorev. 2014.11.009. [DOI] [PubMed]
  33. Kim H. Dissociating the roles of the default-mode, dorsal, and ventral networks in episodic memory retrieval. Neuroimage. 2010;50:1648–1657. doi: 10.1016/j.neuroimage.2010.01.051. [DOI] [PubMed] [Google Scholar]
  34. King M., Hernandez-Castillo C.R., Poldrack R.A., Ivry R.B., Diedrichsen J. Functional boundaries in the human cerebellum revealed by a multi-domain task battery. Nat. Neurosci. 2019;22:1371–1378. doi: 10.1038/s41593-019-0436-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Koenigs M., Barbey A.K., Postle B.R., Grafman J. Superior parietal cortex is critical for the manipulation of information in working memory. J. Neurosci. 2009;29:14980–14986. doi: 10.1523/JNEUROSCI.3706-09.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Koolschijn P.C.M., Van Haren N.E., Lensvelt-Mulders G.J., Hulshoff Pol H.E., Kahn R.S. Brain volume abnormalities in major depressive disorder: a meta-analysis of magnetic resonance imaging studies. Hum. Brain Mapp. 2009;30:3719–3735. doi: 10.1002/hbm.20801. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Koziol L.F., Budding D., Andreasen N., D’Arrigo S., Bulgheroni S., Imamizu H., Ito M., Manto M., Marvel C., Parker K., Pezzulo G., Ramnani N., Riva D., Schmahmann J., Vandervert L., Yamazaki T. Consensus paper: the cerebellum’s role in movement and cognition. Cerebellum. 2014;13:151–177. doi: 10.1007/s12311-013-0511-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Krienen F.M., Buckner R.L. Segregated fronto-cerebellar circuits revealed by intrinsic functional connectivity. Cereb. Cortex. 2009;19:2485–2497. doi: 10.1093/cercor/bhp135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Kriesche D., Woll C.F.J., Tschentscher N., Engel R.R., Karch S. Neurocognitive deficits in depression: a systematic review of cognitive impairment in the acute and remitted state. Eur. Arch. Psychiatry Clin. Neurosci. 2023;273:1105–1128. doi: 10.1007/s00406-022-01479-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Lam R.W., Milev R., Rotzinger S., Andreazza A.C., Blier P., Brenner C., Daskalakis Z.J., Dharsee M., Downar J., Evans K.R., et al. Discovering biomarkers for antidepressant response: protocol from the canadian biomarker integration network in depression (can-bind) and clinical characteristics of the first patient cohort. BMC Psychiatry. 2016;16:1–13. doi: 10.1186/s12888-016-0785-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Le, T.M., Borghi, J.A., Kujawa, A.J., Klein, D.N., Leung, H.C., 2017. Alterations in visual cortical activation and connectivity with prefrontal cortex during working memory updating in major depressive disorder. NeuroImage: Clinical 14, 43–53. URL: http://dx. doi.org/10.1016/j.nicl.2017.01.004 https://linkinghub.elsevier.com/retrieve/pii/S2213158217300049, doi:10.1016/ j.nicl.2017.01.004. [DOI] [PMC free article] [PubMed]
  42. Li B., Friston K., Mody M., Wang H., Lu H., Hu D. A brain network model for depression: from symptom understanding to disease intervention. CNS Neurosci. Ther. 2018;24:1004–1019. doi: 10.1111/cns.12998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Liu Y., Chen Y., Liang X., Li D., Zheng Y., Zhang H., Cui Y., Chen J., Liu J., Qiu S. Altered resting-state functional connectivity of multiple networks and disrupted correlation with executive function in major depressive disorder. Front. Neurol. 2020;11:272. doi: 10.3389/fneur.2020.00272. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. MacQueen G.M., Hassel S., Arnott S.R., Addington J., Bowie C.R., Bray S.L., Davis A.D., Downar J., Foster J.A., Frey B.N., et al. The canadian biomarker integration network in depression (can-bind): magnetic resonance imaging protocols. J. Psychiatry Neurosci. 2019;44:223–236. doi: 10.1503/jpn.180036. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Mahmoudi A., Takerkart S., Regragui F., Boussaoud D., Brovelli A. Multivoxel pattern analysis for fmri data: a review. Comput. Math. Methods Med. 2012;2012 doi: 10.1155/2012/961257. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Makovac E., Fagioli S., Rae C.L., Critchley H.D., Ottaviani C. Can’t get it off my brain: Meta-analysis of neuroimaging studies on perseverative cognition. Psychiatry Res. Neuroimaging. 2020;295 doi: 10.1016/j.pscychresns.2019.111020. [DOI] [PubMed] [Google Scholar]
  47. Manoliu A., Meng C., Brandl F., Doll A., Tahmasian M., Scherr M., Schwerthöffer D., Zimmer C., Förstl H., Bäuml J., et al. Insular dysfunction within the salience network is associated with severity of symptoms and aberrant inter-network connectivity in major depressive disorder. Front. Hum. Neurosci. 2014;7:930. doi: 10.3389/fnhum.2013.00930. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. McDermott L.M., Ebmeier K.P. A meta-analysis of depression severity and cognitive function. J. Affect. Disord. 2009;119:1–8. doi: 10.1016/j.jad.2009.04.022. [DOI] [PubMed] [Google Scholar]
  49. Montgomery S.A., Åsberg M. A new depression scale designed to be sensitive to change. Br. J. Psychiatry. 1979;134:382–389. doi: 10.1192/bjp.134.4.382. [DOI] [PubMed] [Google Scholar]
  50. Morris T.P., Chaddock-Heyman L., Ai M., Anteraper S.A., Kramer A.F., Hillman C.H., Cohen N.J. Enriching activities during childhood are associated with variations in functional connectivity patterns later in life. Neurobiol. Aging. 2021;106:40–48. doi: 10.1016/j.neurobiolaging.2021.05.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Mulders P.C., van Eijndhoven P.F., Schene A.H., Beckmann C.F., Tendolkar I. Resting-state functional connectivity in major depressive disorder: a review. Neurosci. Biobehav. Rev. 2015;56:330–344. doi: 10.1016/j.neubiorev.2015.07.014. [DOI] [PubMed] [Google Scholar]
  52. Murrough J.W., Abdallah C.G., Anticevic A., Collins K.A., Geha P., Averill L.A., Schwartz J., DeWilde K.E., Averill C., Jia-Wei Yang G., et al. Reduced global functional connectivity of the medial prefrontal cortex in major depressive disorder. Hum. Brain Mapp. 2016;37:3214–3223. doi: 10.1002/hbm.23235. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Nagy K., Greenlee M.W., Kovács G. The lateral occipital cortex in the face perception network: an effective connectivity study. Front. Psychol. 2012;3:1–12. doi: 10.3389/fpsyg.2012.00141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Nieto-Castanon A. Brain-wide connectome inferences using functional connectivity multivariate pattern analyses (fc-mvpa) PLoS Comput. Biol. 2022;18 doi: 10.1371/journal.pcbi.1010634. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Nieto-Castanon A., Whitfield-Gabrieli S. Technical Report. Hilbert Press; 2021. CONN functional connectivity toolbox (RRID : SCR_009550), Version 21. [Google Scholar]
  56. Pan F., Xu Y., Zhou W., Chen J., Wei N., Lu S., Shang D., Wang J., Huang M. Disrupted intrinsic functional connectivity of the cognitive control network underlies disease severity and executive dysfunction in first-episode, treatment-naive adolescent depression. J. Affect. Disord. 2020;264:455–463. doi: 10.1016/j.jad.2019.11.076. [DOI] [PubMed] [Google Scholar]
  57. Pan Z., Park C., Brietzke E., Zuckerman H., Rong C., Mansur R.B., Fus D., Subramaniapillai M., Lee Y., McIntyre R.S. Cognitive impairment in major depressive disorder. CNS Spectr. 2019;24:22–29. doi: 10.1017/S1092852918001207. [DOI] [PubMed] [Google Scholar]
  58. Sajjadian M., Uher R., Ho K., Hassel S., Milev R., Frey B.N., Farzan F., Blier P., Foster J.A., Parikh S.V., et al. Prediction of depression treatment outcome from multimodal data: a can-bind-1 report. Psychol. Med. 2023;53:5374–5384. doi: 10.1017/S0033291722002124. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Santomauro D.F., Herrera A.M.M., Shadid J., Zheng P., Ashbaugh C., Pigott D.M., Abbafati C., Adolph C., Amlag J.O., Aravkin A.Y., et al. Global prevalence and burden of depressive and anxiety disorders in 204 countries and territories in 2020 due to the covid-19 pandemic. Lancet. 2021;398:1700–1712. doi: 10.1016/S0140-6736(21)02143-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Scheinost D., Holmes S.E., DellaGioia N., Schleifer C., Matuskey D., Abdallah C.G., Hampson M., Krystal J.H., Anticevic A., Esterlis I. Multimodal investigation of network level effects using intrinsic functional connectivity, anatomical covariance, and structure-to-function correlations in unmedicated major depressive disorder. Neuropsychopharmacology. 2018;43:1119–1127. doi: 10.1038/npp.2017.229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Sheline Y.I., Price J.L., Yan Z., Mintun M.A. Resting-state functional MRI in depression unmasks increased connectivity between networks via the dorsal nexus. PNAS. 2010;107:11020–11025. doi: 10.1073/pnas.1000446107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Sheng X., Chen H., Shao P., Qin R., Zhao H., Xu Y., Bai F. Brain structural network compensation is associated with cognitive impairment and alzheimer’s disease pathology. Front. Neurosci. 2021;15:1–13. doi: 10.3389/fnins.2021.630278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Shenhav A., Botvinick M.M., Cohen J.D. The expected value of control: an integrative theory of anterior cingulate cortex function. Neuron. 2013;79:217–240. doi: 10.1016/j.neuron.2013.07.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Snyder H.R. Major depressive disorder is associated with broad impairments on neuropsychological measures of executive function: a meta-analysis and review. Psychol. Bull. 2013;139:81. doi: 10.1037/a0028727. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Steffens D.C., Otey E., Alexopoulos G.S., Butters M.A., Cuthbert B., Ganguli M., Geda Y.E., Hendrie H.C., Krishnan R.R., Kumar A., et al. Perspectives on depression, mild cognitive impairment, and cognitive decline. Arch. Gen. Psychiatry. 2006;63:130–138. doi: 10.1001/archpsyc.63.2.130. [DOI] [PubMed] [Google Scholar]
  66. Stern, E.R., Eng, G.K., De Nadai, A.S., Iosifescu, D.V., Tobe, R.H., Collins, K.A., 2022. Imbalance between default mode and sensorimotor connectivity is associated with perseverative thinking in obsessive-compulsive disorder. Translational Psychiatry 12, 19. URL: https://www.nature.com/articles/s41398-022-01780-w, doi:10.1038/s41398-022-01780-w. [DOI] [PMC free article] [PubMed]
  67. van der Wijk G., Harris J.K., Hassel S., Davis A.D., Zamyadi M., Arnott S.R., Milev R., Lam R.W., Frey B.N., Hall G.B., et al. Baseline functional connectivity in resting state networks associated with depression and remission status after 16 weeks of pharmacotherapy: a can-bind report. Cereb. Cortex. 2022;32:1223–1243. doi: 10.1093/cercor/bhab286. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Wang L., Dai Z., Peng H., Tan L., Ding Y., He Z., Zhang Y., Xia M., Li Z., Li W., et al. Overlapping and segregated resting-state functional connectivity in patients with major depressive disorder with and without childhood neglect. Hum. Brain Mapp. 2014;35:1154–1166. doi: 10.1002/hbm.22241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Wang L.J., Kuang W.H., Xu J.J., Lei D., Yang Y.C. Resting-state brain activation correlates with short-time antidepressant treatment outcome in drug-naïve patients with major depressive disorder. J. Int. Med. Res. 2014;42:966–975. doi: 10.1177/0300060514533524. [DOI] [PubMed] [Google Scholar]
  70. Westfall D.R., Anteraper S.A., Chaddock-Heyman L., Drollette E.S., Raine L.B., Whitfield-Gabrieli S., Kramer A.F., Hillman C.H. Resting-state functional connectivity and scholastic performance in preadolescent children: a data-driven multivoxel pattern analysis (MVPA) J. Clin. Med. 2020;9(10):3198. doi: 10.3390/jcm9103198. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Woo C.W., Krishnan A., Wager T.D. Cluster-extent based thresholding in fMRI analyses: pitfalls and recommendations. Neuroimage. 2014;91:412–419. doi: 10.1016/j.neuroimage.2013.12.058. [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Yan C.G., Chen X., Li L., Castellanos F.X., Bai T.J., Bo Q.J., Cao J., Chen G.M., Chen N.X., Chen W., Cheng C., Cheng Y.Q., Cui X.L., Duan J., Fang Y.R., Gong Q.Y., Guo W.B., Hou Z.H., Hu L., Kuang L., Li F., Li K.M., Li T., Liu Y.S., Liu Z.N., Long Y.C., Luo Q.H., Meng H.Q., Peng D.H., Qiu H.T., Qiu J., Shen Y.D., Shi Y.S., Wang C.Y., Wang F., Wang K., Wang L., Wang X., Wang Y., Wu X.P., Wu X.R., Xie C.M., Xie G.R., Xie H.Y., Xie P., Xu X.F., Yang H., Yang J., Yao J.S., Yao S.Q., Yin Y.Y., Yuan Y.G., Zhang A.X., Zhang H., Zhang K.R., Zhang L., Zhang Z.J., Zhou R.B., Zhou Y.T., Zhu J.J., Zou C.J., Si T.M., Zuo X.N., Zhao J.P., Zang Y.F. Reduced default mode network functional connectivity in patients with recurrent major depressive disorder. PNAS. 2019;116:9078–9083. doi: 10.1073/pnas.1900390116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Zeng L.L., Shen H., Liu L., Wang L., Li B., Fang P., Zhou Z., Li Y., Hu D. Identifying major depression using whole-brain functional connectivity: a multivariate pattern analysis. Brain. 2012;135:1498–1507. doi: 10.1093/brain/aws059. [DOI] [PubMed] [Google Scholar]
  74. Zhang F.F., Peng W., Sweeney J.A., Jia Z.Y., Gong Q.Y. Brain structure alterations in depression: Psychoradiological evidence. CNS Neurosci. Ther. 2018;24:994–1003. doi: 10.1111/cns.12835. [DOI] [PMC free article] [PubMed] [Google Scholar]
  75. Zhang S., Li C.S. Functional connectivity mapping of the human precuneus by resting state fMRI. NeuroImage. 2012;59:3548–3562. doi: 10.1016/j.neuroimage.2011.11.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Zhou H.X., Chen X., Shen Y.Q., Li L., Chen N.X., Zhu Z.C., Castellanos F.X., Yan C.G. Rumination and the default mode network: Meta-analysis of brain imaging studies and implications for depression. Neuroimage. 2020;206 doi: 10.1016/j.neuroimage.2019.116287. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplementary Data 1
mmc1.docx (145.5KB, docx)

Data Availability Statement

Data will be made available on request.

Articles from NeuroImage : Clinical are provided here courtesy of Elsevier

RESOURCES