
==== Front
Cereb Cortex
Cereb Cortex
cercor
Cerebral Cortex (New York, NY)
1047-3211
1460-2199
Oxford University Press

10.1093/cercor/bhae381
bhae381
Original Article
AcademicSubjects/MED00310
AcademicSubjects/MED00385
AcademicSubjects/SCI01870
Transdiagnostic depression severity and its relationship to global and prefrontal-amygdala structural properties in people with major depression and post-traumatic stress disorder
Li Lei Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Guoxuexiang 37#, Wuhou, Chengdu, 640041, China
Research Unit of Psychoradiology, Chinese Academy of Medical Sciences, Guoxuexiang 37#, Wuhou, Chengdu, 640041, China

Jiang Jing Research Unit of Psychoradiology, Chinese Academy of Medical Sciences, Guoxuexiang 37#, Wuhou, Chengdu, 640041, China
The Third People’s Hospital, Yangshijie 19#, Qingyang, Chengdu, 610031, China

Zhong Shitong West China School of Medicine, Sichuan University, Renminnanlu 16#, Wuhou, Chengdu, 640041, China

Lin Jinping Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Guoxuexiang 37#, Wuhou, Chengdu, 640041, China

Yao Yuhao West China School of Medicine, Sichuan University, Renminnanlu 16#, Wuhou, Chengdu, 640041, China

Kemp Graham J Liverpool Magnetic Resonance Imaging Centre and Institute of Life Course and Medical Sciences, University of Liverpool, Foundation Building, Brownlow Hill, Liverpool, L69 7ZX, United Kingdom

https://orcid.org/0000-0002-4376-8744
Chen Ying Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Guoxuexiang 37#, Wuhou, Chengdu, 640041, China

Gong Qiyong Huaxi MR Research Center, Department of Radiology, West China Hospital of Sichuan University, Guoxuexiang 37#, Wuhou, Chengdu, 640041, China
Department of Radiology, West China Xiamen Hospital of Sichuan University, Jinyuanxilu 699#, Jimei, Xiamen, 361022, China

Corresponding author: Huaxi MR Research Center (HMRRC), Department of Radiology, West China Hospital of Sichuan University, Chengdu, Sichuan 610041, China. Email: qiyonggong@hmrrc.org.cn; HMRRC, Department of Radiology, West China Hospital of Sichuan University, Chengdu, Sichuan 610041, China. Email: chenying85285@163.com
9 2024
24 9 2024
24 9 2024
34 9 bhae38129 5 2024
27 8 2024
12 9 2024
© The Author(s) 2024. Published by Oxford University Press.
2024
https://creativecommons.org/licenses/by-nc/4.0/ This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (https://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com

Abstract

While some studies have used a transdiagnostic approach to relate depression to metabolic or functional brain alterations, the structural substrate of depression across clinical diagnostic categories is underexplored. In a cross-sectional study of 52 patients with major depressive disorder and 51 with post-traumatic stress disorder, drug-naïve, and spanning mild to severe depression severity, we examined transdiagnostic depressive correlates with regional gray matter volume and the topological properties of gray matter-based networks. Locally, transdiagnostic depression severity correlated positively with gray matter volume in the right middle frontal gyrus and negatively with nodal topological properties of gray matter-based networks in the right amygdala. Globally, transdiagnostic depression severity correlated positively with normalized characteristic path length, a measure implying brain integration ability. Compared with 62 healthy control participants, both major depressive disorder and post-traumatic stress disorder patients showed altered nodal properties in regions of the fronto–limbic–striatal circuit, and global topological organization in major depressive disorder in particular was characterized by decreased integration and segregation. These findings provide evidence for a gray matter-based structural substrate underpinning depression, with the prefrontal–amygdala circuit a potential predictive marker for depressive symptoms across clinical diagnostic categories.

depression
post-traumatic stress disorder
amygdala
magnetic resonance imaging
prefrontal cortex
National Natural Science Foundation 82302159 82027808 81820108018 81621003
==== Body
pmcIntroduction

Depressive symptoms often follow trauma and are prevalent in post-traumatic stress disorder (PTSD); indeed, PTSD can predict subsequent major depressive disorder (MDD) (Stander et al. 2014). In the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV), PTSD and MDD share features such as anhedonia, sleep disturbance, and impaired concentration (Flory and Yehuda 2015). They also show common behavioral abnormalities, with high rates of comorbidity (Elhai et al. 2008). Given the social dysfunction and high suicide incidence in both conditions, exploring their shared neuropathological dysfunction may help us understand transdiagnostic depressive vulnerabilities, assess depressive symptom severity, and develop interventions.

Research traditionally distinguishes MDD and PTSD, but this risks overlooking common neural abnormalities. Comparing either MDD or PTSD patients with healthy controls (HCs), neuroimaging studies have found similar abnormalities in both, particularly in the prefrontal cortex (PFC; Kroes et al. 2011; Janiri et al. 2020) and amygdala (AMYG; Karl et al. 2006; Daftary et al. 2019), regions strongly influenced by stress and critical for regulating emotion in depression (McEwen 2017). However, mapping transdiagnostic symptoms to their neural substrates may be more useful than focusing solely on categorical disorders (Cuthbert 2014). Some studies have used such a transdiagnostic approach to explore metabolic or functional alterations: for example, across individuals with MDD and PTSD, depression severity was inversely correlated with synaptic vesicle glycoprotein 2A density in the PFC and hippocampus (Holmes et al. 2019); and in a connectome-wide functional analysis in women with MDD and PTSD, depression severity was associated with frontal-AMYG hypoconnectivity (Satterthwaite et al. 2016).

However, the structural substrate associated with depression across clinical diagnostic categories remains underexplored. Voxel-based morphometry (VBM) is a traditional method for assessing regional gray matter volume (GMV), while a newer approach involves graph-based analysis of the GM structural connectome, characterizing the topological properties of the individual’s GM covariance pattern at global and nodal levels (Tijms et al. 2012; Chen et al. 2022; Chen et al. 2023). Here, we applied both analyses to structural imaging data from MDD, PTSD, and HC, examining the associations between morphological properties and depression severity measured by the Hamilton Depression Scale 17 (HAMD-17). The clinical subjects represented a broad spectrum of depression severity, from mild depressive symptoms in PTSD to severe MDD. Importantly, all participants were drug-naive, revealing depression-related brain changes without confounding effects from medication. Our hypothesis was that both regional structural abnormalities (expected in the PFC and AMYG) and global GM-based network topology would be transdiagnostically associated with depressive symptoms.

Methods

Participants

We studied 165 participants: 52 with MDD, 51 with PTSD, and 62 age- and gender-matched HCs. Demographic and clinical details are shown in Table 1. All clinical participants were medication-naive. MDD participants were recruited from the Mental Health Centre of West China Hospital, Chengdu, China; PTSD participants were survivors from the Wenchuan 8.0-magnitude earthquake in China; HCs were recruited from the same sociodemographic environment via poster advertisements.

Table 1 Demographic and clinical characteristics of study groups.

	MDD (n = 52)	PTSD (n = 51)	HC (n = 62)	P	
ANOVA	MDD vs PTSD	
Age (y)	40.9 ± 8.1	44.1 ± 6.2	43.0 ± 13.9	0.277	0.116	
Sex (m:f)	19:33	12:39	26:36	0.115a	0.150a	
Education (y)	10.5 ± 4.7	8.7 ± 3.9	10.7 ± 5.5	0.065	0.059	
HAMD-17 total score	26.4 ± 6.6	12.6 ± 5.7	–	–	<0.001	
Mood/emotional symptoms	8.35 ± 2.72	2.44 ± 1.95	–		< 0.001	
Sleep disturbances	4.7 ± 1.89	3.51 ± 1.78			0.020	
Cognitive symptoms	3.45 ± 1.39	1.24 ± 1.02			< 0.001	
Psychomotor symptoms	3.55 ± 1.15	0.95 ± 1.05			< 0.001	
Anxiety symptoms	4.55 ± 1.23	1.98 ± 1.42			< 0.001	
Somatic symptoms	3.85 ± 1.39	1.39 ± 1.38			< 0.001	
The comparisons of demographic data among 3 groups were performed by general linear model analysis, then post hoc comparisons of each pair used LSD. The HAMD-17 scores of the MDD and PTSD groups were compared using independent-sample t-test. The threshold was set at P < 0.05. Values are presented as mean ± standard deviation (SD) unless otherwise indicated; numbers in parentheses represent range. aP calculated by Chi-squared test. Age and years of education are as defined at the time of MRI scanning.

Diagnosis of MDD and PTSD was made by 2 experienced psychiatrists according to the DSM-IV criteria using the Structured Clinical Interview for Diagnosis. Depression symptom severity was assessed using HAMD-17 (Hamilton 1967). All individuals with MDD were during the first episode of depression at screening. The Clinician-Administered Posttraumatic Stress Disorder Scale (CAPS) was used to confirm PTSD diagnosis and assess symptom severity (Blake et al. 1995). Exclusion criteria were current or history of affective or psychotic comorbidity other than MDD or PTSD; history of > 5 min loss of consciousness; physical injury or serious head trauma during earthquake; alcohol or drug abuse; age < 18 or > 60 years; standard contraindications to magnetic resonance imaging (MRI); and left-handedness. Exclusion criteria were the same for HCs, plus current or history of any DSM-IV diagnosis. The local research ethics committee of West China Hospital, Sichuan University, approved the study. All participants gave written fully-informed consent.

MRI acquisition

T1-weighted MRI scans were acquired on a 3.0 tesla MRI system (Siemens 3 T Trio; Erlangen, Germany) with a 12-channel phased-array head coil. Participants were instructed to keep their eyes closed without directed thought. Foam padding stabilized the head and ear plugs reduced noise. High-resolution 3D T1-weighted images were acquired using a spoiled gradient recalled sequence with these parameters: repetition time/echo time (TR/TE), 1900/2.26 ms; flip angle, 9°; matrix, 256 × 256; field of view, 240 × 240 mm2; slice thickness, 1 mm without gap; 176 axial slices; and in-plane resolution, 0.94 × 0.94 mm2. A radiologist with 9 years’ experience evaluated image quality, excluding any subjects with gross brain abnormalities.

GMV analysis

Structural images were preprocessed using Statistical Parametric Mapping 12 (SPM12) software (http://www.fil.ion.ucl.ac.uk/spm/software/spm12/). Individual images were segmented into GM, white matter, cerebrospinal fluid, bone, soft tissue, and air/background using the default tissue probability maps as priors (Ashburner and Friston 2005); automatic segmentations were visually confirmed. Gray and white matter images were spatially normalized to Montreal Neurological Institute (MNI) coordinate space using diffeomorphic anatomical registration through exponentiated lie algebra tools in SPM12: GM data were resampled to 2 × 2 × 2 mm3 voxels and spatially smoothed (Gaussian kernel with 6 mm full width at half maximum). Then, a whole-brain multiple regression analysis in SPM12 was performed to identify brain GM volume correlates underlying depression symptoms. The GM volume of each voxel was taken as the dependent variable and the HAMD-17 score as the independent variable; age, sex, and total intracranial brain volume (TIV) were covariates and an absolute threshold masking of 0.2 was used; the significance threshold was taken as P < 0.05 for family-wise error (FWE) correction. Next, the regional GM volume values of the identified regions were extracted using the MarsBaR toolbox in SPM12.

Extraction of GM-based networks

Single-subject GM-based networks were extracted by an automated data-driven method (Tijms et al. 2012). In these large-scale morphological networks, the nodes are defined as small regions corresponding to 3 × 3 × 3 cubic voxels, and edges connect regions that have a statistically similar structure. These cubes preserve the 3D structure of the cortex, so they include spatial information including local thickness, cortex curvature, as well as voxel values. The structural similarity between every pair of cubes was quantified by correlation coefficients, after rotation to identify their maximum correlation value with seed cubes. Next, unweighted and undirected graphs were constructed by binarizing the similarity matrices, using a threshold for each individual network with a permutation-based method to ensure < 5% spurious correlations; only the positive similarity values survived this threshold.

The resulting single-subject GM-based networks have different sizes, and as network properties vary with network size, it is critical to normalize them all to the same node number and node sizes. We did this using the unified Automated Anatomical Labeling (AAL) parcellation template (Batalle et al. 2013). Each cube was assigned to an AAL region/node based on the region to which most of its voxels belong. Each pair of AAL regions/nodes was considered to be connected with a weight reflecting the strength of connection, corresponding to the ratio of actual significant cube-to-cube correlations to the total possible connections among cubes in pairs of nodes; the weight obtained is bounded by 0 and 1, and self-connections were excluded. This resulted in a 90 × 90 weighted normalized network for each subject, from which the network measures were calculated.

Network properties and threshold selection

The network properties were calculated using GRETNA (www.nitrc.org/projects/gretna/) software. Both global and nodal properties were calculated at each sparsity threshold. Global graph metrics included: clustering coefficient Cp, characteristic path length Lp, normalized clustering coefficient γ, normalized characteristic path length λ, small-worldness σ, local efficiency Eloc, and global efficiency Eglob (Watts and Strogatz 1998). The nodal graph properties included nodal degree, nodal efficiency and nodal betweenness examined in each ALL region.

A range of network sparsity (S) thresholds was applied to the correlation matrices, the upper and lower limits of S being chosen to ensure that the thresholded networks were estimable for σ scalar with sparse properties, and that the small-world index was larger than 1.0. With these limits, the area under the curve (AUC) was calculated over the threshold range of 0.10 < S < 0.34 with interval 0.01. The AUC provides a summarizing scalar for the topological characterization of brain networks, free of the potential bias of using an arbitrary single threshold. Both clinical and HC groups exhibited small-world properties in GM-based network architecture with γ > 1 and λ ≈ 1 (Supplementary Fig. S1).

Statistical analysis

For the regions showing significant correlation with transdiagnostic depression severity identified in VBM analysis of GM volume, partial correlation analysis was conducted between the extracted regional GM volume values and transdiagnostic depression severity measured by HAMD-17 scores, in SPSS software (version 16.0), using age, sex, and total TIV as covariates.

To investigate effects of depression on structural properties of GM-based networks, 2 analyses were conducted. First, to identify altered global and nodal GM-based network topological characteristics in MDD/PTSD compared with HC, we conducted nonparametric permutation tests with a design model of 1-way univariate 1-way analysis of variance (ANOVA) to test for group differences among MDD, PTSD, and HC groups followed by post hoc pairwise permutation tests, using a permutation analysis of linear models (PALM) (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/PALM). The AUC values of all the topological metrics among the 3 groups were compared, with Benjamini Hochberg false discovery rate (FDR) correction for the multiple comparisons of each of the 90 nodal metrics. The randomization was repeated 10,000 times. Second, to examine the relationship between the AUC values of each topological property (regional and global) and transdiagnostic depression severity, a partial correlation analysis was conducted in SPSS software (version 16.0) using age and gender as covariates. All statistical significance was set as P < 0.05.

Statistical analysis for demographic and clinical variables among 3 groups were performed by 1-way ANOVA, followed by pairwise post hoc Fisher’s least significance difference (LSD) tests. The continuous variables of clinical groups were compared using the independent-sample t-test, the categorical variables of 3 groups using the Chi-square test. We also conducted exploratory partial correlation analysis to assess associations of depression symptoms measured by HAMD-17 and PTSD symptoms measured by CAPS-total scores and sub-scores in individuals with PTSD, using age and gender as covariates. All statistical significance was set as P < 0.05.

Results

Demographic characteristics and depressive symptoms

Table 1 summarizes the descriptive statistics. There were no significant differences in age, sex, and education between the 3 groups (Table 1). HAMD-17 scores were significantly lower in MDD than PTSD (P < 0.001).

Regional structural alterations and correlation with transdiagnostic depression severity

Figure 1A shows that transdiagnostic depression severity correlated with regional GM volume in right middle frontal gyrus (MNI coordinates: x = 24, y = 57, z = 27, t = 5.09, cluster size 345 voxels, FWE-corrected after controlling for age, sex, and TIV). Figure 1B shows the significant positive correlation of GM volume with transdiagnostic depression severity (r = 0.352, P < 0.001 after controlling for age, sex, and TIV). Transdiagnostic depression severity also correlated with the nodal topological property of GM-based networks in the right AMYG (r = −0.345, P < 0.001, Fig. 1C and D), from all regions that showed significant nodal differences among MDD, PTSD, and HC; these are 6 in total, including right middle frontal gyrus (orbital part), right insula, right AMYG, right pallidum (PAL), left caudate (CAU), and left thalamus (THA; P < 0.05, FDR corrected; Fig. 2, Table 2). Of these, lower nodal efficiency in the right PAL, left CAU and left THA, and higher nodal efficiency in right middle frontal gyrus were found in both MDD and PTSD compared with HC; nodal efficiency in the right AMYG was lower in MDD but higher in PTSD relative to HC, and differed significantly in direct comparison between MDD and PTSD; and nodal efficiency in the right insula was higher in PTSD compared with both MDD and HC. No significant differences were found in nodal degree or betweenness centrality.

Fig. 1 Relationships between transdiagnostic depression severity and regional morphometric/topological brain properties in the whole clinical group (both MDD and PSTD). A) The location (in the right middle frontal gyrus) of voxels whose GMV shows significant correlation with HAMD-17 scores. B) A scatter plot of GMV in the right middle frontal gyrus against HAMD-17 scores; the line represents a whole-group regression line, showing the significant positive correlation; each point represents 1 clinical individual (with PTSD in orange and MDD in blue). C) Scatter plot of nodal efficiency in the right AMYG against HAMD-17 scores; the line represents a whole group regression, showing the significant negative correlation; each point represents 1 clinical individual (with PTSD in orange and MDD in blue). D) The right AMYG region in automated ALL template.

Fig. 2 Brain regions which show abnormalities of nodal efficiency in brain gray matter-based topological networks among MDD, PTSD, and HC groups using 1-way ANOVA. Abbreviations: INS, insular; ORBmid, middle frontal gyrus, orbital part.

Table 2 Topological properties of GMV-based brain networks showing differences among MDD, PTSD, and HC groups.

Measurements	MDD
Mean ± SD	PTSD
Mean ± SD	HC
Mean ± SD	ANOVA
P (F)	Post hoc P (t)	
MDD vs PTSD	MDD vs HC	PTSD vs HC	
Global								
Eglob	0.0664 ± 0.0014	0.0667 ± 0.0020	0.0671 ± 0.0015	0.0481* (3.089)	0.2017	0.0084$ (−2.435)	0.0638	
Eloc	0.0980 ± 0.0025	0.0982 ± 0.0036	0.0994 ± 0.0023	0.0110* (4.535)	0.3510	0.0039$ (−2.744)	0.0094$ (−2.351)	
Cp	0.0766 ± 0.0038	0.0772 ± 0.0053	0.0794 ± 0.0035	0.0008* (7.131)	0.2417	0.0004$ (−3.541)	0.0029$ (−2.777)	
λ	0.2388 ± 0.0006	0.2386 ± 0.0006	0.2389 ± 0.0006	0.0129* (4.453)	0.0090(2.359)	0.3697	0.0023$ (−2.796)	
Lp	0.8773 ± 0.0188	0.873 ± 0.0275	0.8675 ± 0.0198	0.0596 (2.853)	0.2097	0.0103$ (2.3424)	0.0699	
Nodal efficiency								
R orbital frontal gyrus	0.0669 ± 0.0040	0.0663 ± 0.0043	0.0640 ± 0.0049	0.0022# (6.464)	0.2502	0.0005$ (3.364)	0.0045$ (2.657)	
R insula	0.0778 ± 0.0053	0.0807 ± 0.0051	0.0778 ± 0.0043	0.0023# (6.329)	0.0012$ (−3.032)	0.4870	0.0006$ (3.178)	
R AMYG	0.0443 ± 0.0181	0.0572 ± 0.00646	0.0511 ± 0.0128	0.0001# (11.989)	0.0001$ (−0.489)	0.0042$ (−2.723)	0.0086$ (2.391)	
L CAU	0.1008 ± 0.0024	0.1009 ± 0.0035	0.1024 ± 0.0025	0.0028# (6.046)	0.4118	0.0015$ (−3.077)	0.0022$ (−2.840)	
R PAL	0.0971 ± 0.0035	0.0967 ± 0.0043	0.0990 ± 0.0028	0.0021# (6.689)	0.3012	0.0030$ (−2.819)	0.0005$ (−3.356)	
L THA	0.0708 ± 0.0024	0.0714 ± 0.0024	0.0729 ± 0.0023	0.0002# (10.015)	0.1491	0.0002$ (−4.260)	0.0013$ (−3.156)	
Abbreviations: L, left; R, right, GMV. *Significant between-group difference at P < 0.05. #Significant between-group difference at P < 0.05 after FDR correction. $Significant between-group difference at P < 0.05/3 for post hoc analysis. P is before FDR correction.

Global structural topological alterations and correlation with transdiagnostic depression severity

There were significant differences among MDD, PTSD, and HC in global topological properties (P < 0.05, FDR-corrected, Table 2); Eloc, Eglob, and Cp were higher in MDD compared to PTSD (Fig. 3A), while λ was lower in PTSD compared with both MDD and HC (Fig. 3B) and positively correlated with transdiagnostic depression severity (r = 0.254, P = 0.010; Fig. 3C). No significant group differences were found in Lp, γ, and σ.

Fig. 3 Global topological properties of the gray matter-based network. A) The AUC of 3 global topological properties (labeled on x-axis) in the 3 groups (see key); AUC was significantly lower for Eloc, Eglob, and Cp in both MDD and PTSD vs HC. Data points represent all participants; error bars represent SD; and P-values are shown for pair-wise comparisons. Note: The y-axis does not start from zero. B) Column graph showing increased λ in MDD and HC groups vs PTSD. Note: The y-axis does not start from zero. C) A scatter plot of λ against HAMD-17 scores in both MDD and PTSD; the line represents a whole group regression line, showing the significant positive correlation; each point represents 1 clinical individual (PTSD in orange and MDD in blue). Abbreviation: AUC, area under the curve of the network property across the range of network sparsity thresholds.

Relationships between PTSD symptoms and depression severity

PTSD patients showed significant positive correlation between HAMD-17 scores and CAPS-total scores (r = 0.365, P = 0.010), CAPS-avoidance scores (r = 0.366, P = 0.010), and CAPS-hyperarousal scores (r = 0.348, P = 0.014; Fig. S2).

Discussion

This study explored the regional GM volume and topological properties of GM-based network, and their association with transdiagnostic depression, in a large sample of medication-naïve individuals with MDD or PTSD. A key finding is the significant association between prefrontal-AMYG structural properties and global topological organization, represented by the global topological index λ, with transdiagnostic depressive symptoms. Additionally, regions within the fronto-limbic-striatal circuit showed altered nodal properties in individuals with MDD and PTSD compared with HC. Categorical analyses showed greater global abnormalities including lower Eglob (implying decreased integration ability) and lower Eloc and Cp (implying decreased segregation ability) in MDD than PTSD. This may suggest progressive effects of depression on GM-based network properties in individuals with symptoms ranging from mild (PTSD) to severe (MDD). We next discuss the possible significance of these findings.

Notable findings were the positive correlation between transdiagnostic depression severity and GM volume of right middle frontal gyrus, and the higher nodal efficiency in the orbital part of right middle frontal gyrus in MDD and PTSD compared with HC. The PFC, including middle frontal gyrus, plays regulatory role in top–down control of behavior, thought, and mood, as well as in stress-induced remodeling (Miller and Cohen 2001). Stress-induced alterations in synaptic connections and resultant network dysfunction in the PFC have been demonstrated in MDD and PTSD (Duman and Aghajanian 2012; Kang et al. 2012). A meta-analysis of resting-state functional neuroimaging studies has reported hyperactivity in the orbital part of middle frontal gyrus in MDD (Ma et al. 2019). However, structural studies have reported inconsistent cortical volume or thickness alterations in MDD. We speculate that PFC volume may increase during episodes of illness or close to illness onset in first-episode individuals but decrease over multiple episodes in chronic recurrent patients (Frodl et al. 2008). Longitudinal studies are needed to address this.

Transdiagnostic depression severity was significantly correlated with nodal efficiency of right AMYG, and nodal efficiency in AMYG was altered in both in MDD and PTSD, but in opposite directions: lower in MDD, higher in PTSD. Lower nodal efficiency in AMYG has been consistently reported in MDD alongside lower AMYG volume and impaired AMYG–PFC functional connectivity (Amidfar et al. 2020; Gray et al. 2020). The prefrontal-AMYG circuit plays important roles in emotion regulation and stress response (Kaul et al. 2021). As nodal efficiency evaluates the capacity of a given node for information communication (Achard and Bullmore 2007), the decreased nodal efficiency in right AMYG in MDD may result in dysfunctional transmission of emotional information. The contrasting higher nodal efficiency in AMYG in PTSD may accord with the meta-analysis finding of AMYG hyperactivity in PTSD (Etkin and Wager 2007), perhaps due to loss of prefrontal top–down inhibitory control. The AMYG contains several nuclei connecting with different cortical areas and has functions contributing to fear conditioning, specifically focusing attention on threat (Chiba et al. 2021).

PTSD showed higher nodal efficiency compared to HC in the right insula. The AMYG and anterior insula are closely related in the salience network (Seeley et al. 2007), which is involved detecting threatening information, guiding behavior, and promoting survival in PTSD (Seeley et al. 2007; Patel et al. 2012). Together, AMYG and insula arousal may underlie vivid trauma recollections, causing re-experiencing symptoms such as flashback. This hypothesis should be tested explicitly in future research.

Both MDD and PTSD show lower nodal efficiency compared to HC in left THA and the striatum including left CAU and right PAL. This is consistent with reports of lower GM volume in the THA and striatum in first-episode medication-naïve MDD (Zhang et al. 2020). The THA has been suggested as a key region for treatment response in MDD; transient significant volume increase in THA was observed after electroconvulsive therapy (Jehna et al. 2021), and thalamic hyperactivity has been reported in individuals with selective serotonin reuptake inhibitor treatment-resistant depression (Yamamura et al. 2016). The THA sends sensory information about the fear-eliciting stimulus to the lateral nucleus of AMYG (Rodrigues et al. 2009). The striatum is proposed to participate in discriminating intermediate reward magnitudes, and a study has linked inflammation to reduced striatal activation in MDD (Burrows et al. 2021). Altogether, the altered nodal properties in PFC, AMYG, THA, and striatum provide evidence for impairment of fronto–limbic–striatal circuit involving in the depressed brain (Zhang et al. 2020).

Transdiagnostic depression severity was positively correlated with the global structural property λ, which reflects the average distance between all possible pairs of nodes in the network, and is an index of the network’s integration ability (Watts and Strogatz 1998). Thus, across both clinical groups, the more severe the depressive symptoms, the lower the integration ability; with higher levels of depression than PTSD, MDD showed significantly higher λ than PTSD. Another integration index is Eglob (lower Eglob implying decreased integration ability), and this is lower in MDD. In addition, MDD compared to HC show lower Eloc and Cp, implying decreased segregation ability. Thus, though it retains overall small-world architecture; the brain structural network in MDD is less efficient at both local and global scales, implying weaker “small-worldness”. Studies have reported inconsistent structural topological organization alterations in MDD, for which possible explanations include differences in MRI modalities (structural MRI or diffusion tensor imaging), structural properties (GM-based or cortical thickness-based), node definition atlas (AAL, Desikan–Killiany atlas or freesurfer), and network types (binary or weighted). Our result is consistent with reports of decreased segregation and decreased integration in GM-based networks, and decreased Eloc, Eglob, and Cp in MDD (Chen et al. 2016).

The CAPS-avoidance, CAPS-hyperarousal, and CAPS-total scores were positively correlated with depressive symptom severity measured by HAMD-17. The likeliest reason is the overlap of symptoms between PTSD and MDD, including diminished interest (in the avoidance or numbing cluster), sleep disturbance, and poor concentration (in the hyperarousal cluster) in the PTSD diagnosis in DSM-IV (Flory and Yehuda 2015). PTSD showed lower depression than MDD and can be thought of as high-risk for subsequent MDD development. Our correlation results support PTSD as a proposed predictor for developing MDD.

Limitations of the study. First, all individuals with MDD were in the depressed episode, thus our results may represent the effects of mood state rather than trait effects. Participants in remission could be included in future studies to examine this. Second, traumatized individuals without PTSD were not studied, so we were not able to disentangle alterations induced by PTSD disease from the stress exposure per se. Future study including a group of trauma-exposed controls would help to elucidate this. Third, the anatomical AAL template was used to facilitate network comparisons across individuals. Though there is no gold standard for node parcellation, the use of an anatomic atlas is known to introduce potential bias due to its inhomogeneity. Finally, though PTSD participants are perhaps high-risk for subsequent MDD development, this cross-sectional study cannot confirm this assumption. Longitudinal studies that track participants for mood evaluation will be important to help to elucidate the neurology of depression in future.

In conclusion, our study provides evidence for a GM-based property underpinning depression psychopathology, with the prefrontal-AMYG circuit identified as a candidate predictor of depressive symptoms across clinical diagnostic categories. Our results underscore the utility of the transdiagnostic approach in informing etiologic models and in the development of more targeted, transdiagnostic symptom-based treatments for this debilitating symptom.

Supplementary Material

Supplementary_Materials_bhae381

Acknowledgments

We acknowledge and appreciate the efforts of all the authors of the included studies who responded to our requests for further information not included in published manuscripts.

Author contributions

Lei Li (Data curation, Formal analysis, Funding acquisition, Methodology, Writing—original draft), Jing Jiang (Data curation), Shitong Zhong (Data curation), Jinping Lin (Data curation), Yuhao Yao (Data curation), Graham J. Kemp (Writing—review & editing), Ying Chen (Formal analysis, Methodology, Writing—review & editing), and Qiyong Gong (Funding acquisition, Resources, Supervision).

Funding

This study was supported by the National Natural Science Foundation (grant numbers 81621003, 81820108018, 82027808, and 82302159).

Conflict of interest statement: None declared.

Data availability

All the data that support the findings of the present study are available from the corresponding author through request.
==== Refs
References

Achard S , BullmoreE. Efficiency and cost of economical brain functional networks. PLoS Comput Biol. 2007:3 :e17. 10.1371/journal.pcbi.0030017.17274684
Amidfar M , QuevedoJ, Z RéusG, KimYK. Grey matter volume abnormalities in the first depressive episode of medication-naive adult individuals: a systematic review of voxel based morphometric studies. Int J Psychiatry Clin Pract. 2020:25 :407–420. 10.1080/13651501.2020.1861632.33351672
Ashburner J , FristonKJ. Unified segmentation. NeuroImage. 2005:26 :839–851. 10.1016/j.neuroimage.2005.02.018.15955494
Batalle D , Munoz-MorenoE, FiguerasF, BargalloN, EixarchE, GratacosE. Normalization of similarity-based individual brain networks from gray matter MRI and its association with neurodevelopment in infants with intrauterine growth restriction. NeuroImage. 2013:83 :901–911. 10.1016/j.neuroimage.2013.07.045.23886985
Blake DD , WeathersFW, NagyLM, KaloupekDG, GusmanFD, CharneyDS, KeaneTM. The development of a clinician-administered PTSD scale. J Trauma Stress. 1995:8 :75–90.7712061
Burrows K , StewartJL, KuplickiR, Figueroa-HallL, SpechlerPA, ZhengH, GuinjoanSM, TulsaI, SavitzJB, Kent TeagueT, et al. Elevated peripheral inflammation is associated with attenuated striatal reward anticipation in major depressive disorder. Brain Behav Immun. 2021:93 :214–225. 10.1016/j.bbi.2021.01.016.33508469
Chen JH , YaoZJ, QinJL, YanR, HuaLL, LuQ. Aberrant global and regional topological organization of the fractional anisotropy-weighted brain structural networks in major depressive disorder. Chin Med J. 2016:129 :679–689. 10.4103/0366-6999.178002.26960371
Chen Y , LeiD, CaoH, NiuR, ChenF, ChenL, ZhouJ, HuX, HuangX, GuoL, et al. Altered single-subject gray matter structural networks in drug-naïve attention deficit hyperactivity disorder children. Hum Brain Mapp. 2022:43 :1256–1264. 10.1002/hbm.25718.34797010
Chen Y , CaoH, PanN, LongY, ZhouJ, ZhouC, WangY, GongQ. The alpha-2A-adrenergic receptor gene polymorphism modulates gray matter structural networks, visual memory, and inhibitory cognitive control in children with attention deficit/hyperactivity disorder. Cereb Cortex (New York, NY : 1991). 2023:33 :9088–9094.37310179
Chiba T , IdeK, TaylorJE, BokuS, TodaH, KanazawaT, KatoS, HoriuchiY, HishimotoA, MaruyamaT, et al. A reciprocal inhibition model of alternations between under−/overemotional modulatory states in patients with PTSD. Mol Psychiatry. 2021:26 (9):5023–5039.
Cuthbert BN . The RDoC framework: facilitating transition from ICD/DSM to dimensional approaches that integrate neuroscience and psychopathology. World Psychiatry. 2014:13 :28–35. 10.1002/wps.20087.24497240
Daftary S , Van EnkevortE, KulikovaA, LegacyM, BrownES. Relationship between depressive symptom severity and amygdala volume in a large community-based sample. Psychiatry Res Neuroimaging. 2019:283 :77–82. 10.1016/j.pscychresns.2018.12.005.30554129
Duman RS , AghajanianGK. Synaptic dysfunction in depression: potential therapeutic targets. Science. 2012:338 :68–72. 10.1126/science.1222939.23042884
Elhai JD , GrubaughAL, KashdanTB, FruehBC. Empirical examination of a proposed refinement to DSM-IV posttraumatic stress disorder symptom criteria using the National Comorbidity Survey Replication data. J Clin Psychiatry. 2008:69 :597–602. 10.4088/JCP.v69n0411.18294026
Etkin A , WagerTD. Functional neuroimaging of anxiety: a meta-analysis of emotional processing in PTSD, social anxiety disorder, and specific phobia. Am J Psychiatry. 2007:164 :1476–1488. 10.1176/appi.ajp.2007.07030504.17898336
Flory JD , YehudaR. Comorbidity between post-traumatic stress disorder and major depressive disorder: alternative explanations and treatment considerations. Dialogues Clin Neurosci. 2015:17 :141–150. 10.31887/DCNS.2015.17.2/jflory.26246789
Frodl TS , KoutsoulerisN, BottlenderR, BornC, JagerM, ScupinI, ReiserM, MollerHJ, MeisenzahlEM. Depression-related variation in brain morphology over 3 years: effects of stress? Arch Gen Psychiatry. 2008:65 :1156–1165. 10.1001/archpsyc.65.10.1156.18838632
Gray JP , MullerVI, EickhoffSB, FoxPT. Multimodal abnormalities of brain structure and function in major depressive disorder: a meta-analysis of neuroimaging studies. Am J Psychiatry. 2020:177 :422–434. 10.1176/appi.ajp.2019.19050560.32098488
Hamilton M . Development of a rating scale for primary depressive illness. Br J Soc Clin Psychol. 1967:6 :278–296. 10.1111/j.2044-8260.1967.tb00530.x.6080235
Holmes SE , ScheinostD, FinnemaSJ, NaganawaM, DavisMT, DellaGioiaN, NabulsiN, MatuskeyD, AngaritaGA, PietrzakRH, et al. Lower synaptic density is associated with depression severity and network alterations. Nat Commun. 2019:10 :1529. 10.1038/s41467-019-09562-7.30948709
Janiri D , MoserDA, DoucetGE, LuberMJ, RasgonA, LeeWH, MurroughJW, SaniG, EickhoffSB, FrangouS. Shared neural phenotypes for mood and anxiety disorders: a meta-analysis of 226 task-related functional imaging studies. JAMA Psychiatry. 2020:77 :172–179. 10.1001/jamapsychiatry.2019.3351.31664439
Jehna M , WurmW, PinterD, VogelK, HollA, HofmannP, EbnerC, RopeleS, FuchsG, KapfhammerHP, et al. Do increases in deep grey matter volumes after electroconvulsive therapy persist in patients with major depression? A longitudinal MRI-study. J Affect Disord. 2021:281 :908–917. 10.1016/j.jad.2020.11.035.33279261
Kang HJ , VoletiB, HajszanT, RajkowskaG, StockmeierCA, LicznerskiP, LepackA, MajikMS, JeongLS, BanasrM, et al. Decreased expression of synapse-related genes and loss of synapses in major depressive disorder. Nat Med. 2012:18 :1413–1417. 10.1038/nm.2886.22885997
Karl A , SchaeferM, MaltaLS, DorfelD, RohlederN, WernerA. A meta-analysis of structural brain abnormalities in PTSD. Neurosci Biobehav Rev. 2006:30 :1004–1031. 10.1016/j.neubiorev.2006.03.004.16730374
Kaul D , SchwabSG, MechawarN, MatosinN. How stress physically re-shapes the brain: impact on brain cell shapes, numbers and connections in psychiatric disorders. Neurosci Biobehav Rev. 2021:124 :193–215. 10.1016/j.neubiorev.2021.01.025.33556389
Kroes MC , RuggMD, WhalleyMG, BrewinCR. Structural brain abnormalities common to posttraumatic stress disorder and depression. J Psychiatry Neurosci. 2011:36 :256–265. 10.1503/jpn.100077.21418787
Ma X , LiuJ, LiuT, MaL, WangW, ShiS, WangY, GongQ, WangM. Altered resting-state functional activity in medication-naive patients with first-episode major depression disorder vs. healthy control: a quantitative meta-analysis. Front. Behav Neurosci. 2019:13 :89. 10.3389/fnbeh.2019.00089.
McEwen BS . Neurobiological and systemic effects of chronic stress. Chronic Stress (Thousand Oaks). 2017:1 :247054701769232. 10.1177/2470547017692328.
Miller EK , CohenJD. An integrative theory of prefrontal cortex function. Annu Rev Neurosci. 2001:24 :167–202. 10.1146/annurev.neuro.24.1.167.11283309
Patel R , SprengRN, ShinLM, GirardTA. Neurocircuitry models of posttraumatic stress disorder and beyond: a meta-analysis of functional neuroimaging studies. Neurosci Biobehav Rev. 2012:36 :2130–2142. 10.1016/j.neubiorev.2012.06.003.22766141
Rodrigues SM , LeDouxJE, SapolskyRM. The influence of stress hormones on fear circuitry. Annu Rev Neurosci. 2009:32 :289–313. 10.1146/annurev.neuro.051508.135620.19400714
Satterthwaite T , CookP, BruceS, ConwayC, MikkelsenE, SatchellE, VandekarS, DurbinT, ShinoharaR, ShelineY. Dimensional depression severity in women with major depression and post-traumatic stress disorder correlates with fronto-amygdalar hypoconnectivty. Mol Psychiatry. 2016:21 :894–902. 10.1038/mp.2015.149.26416545
Seeley WW , MenonV, SchatzbergAF, KellerJ, GloverGH, KennaH, ReissAL, GreiciusMD. Dissociable intrinsic connectivity networks for salience processing and executive control. J Neurosci. 2007:27 :2349–2356. 10.1523/JNEUROSCI.5587-06.2007.17329432
Stander VA , ThomsenCJ, Highfill-McRoyRM. Etiology of depression comorbidity in combat-related PTSD: a review of the literature. Clin Psychol Rev. 2014:34 :87–98. 10.1016/j.cpr.2013.12.002.24486520
Tijms BM , SeriesP, WillshawDJ, LawrieSM. Similarity-based extraction of individual networks from gray matter MRI scans. Cereb Cortex. 2012:22 :1530–1541. 10.1093/cercor/bhr221.21878484
Watts DJ , StrogatzSH. Collective dynamics of ‘small-world’ networks. Nature. 1998:393 :440–442. 10.1038/30918.9623998
Yamamura T , OkamotoY, OkadaG, TakaishiY, TakamuraM, MantaniA, KurataA, OtagakiY, YamashitaH, YamawakiS. Association of thalamic hyperactivity with treatment-resistant depression and poor response in early treatment for major depression: a resting-state fMRI study using fractional amplitude of low-frequency fluctuations. Transl Psychiatry. 2016:6 :e754. 10.1038/tp.2016.18.26954981
Zhang Y , YangY, ZhuL, ZhuQ, JiaY, ZhangL, PengQ, WangJ, LiuJ, FanW, et al. Volumetric deficit within the fronto-limbic-striatal circuit in first-episode drug naive patients with major depression disorder. Front Psychiatry. 2020:11 :600583.33551870
