Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Extended functional connectivity of convergent structural alterations among individuals with PTSD: a neuroimaging meta-analysis

Extended functional connectivity of convergent structural alterations among individuals with... Background: Post‑traumatic stress disorder (PTSD) is a debilitating disorder defined by the onset of intrusive, avoid‑ ant, negative cognitive or affective, and/or hyperarousal symptoms after witnessing or experiencing a traumatic event. Previous voxel‑based morphometry studies have provided insight into structural brain alterations associated with PTSD with notable heterogeneity across these studies. Furthermore, how structural alterations may be associated with brain function, as measured by task‑free and task ‑based functional connectivity, remains to be elucidated. Methods: Using emergent meta‑analytic techniques, we sought to first identify a consensus of structural alterations in PTSD using the anatomical likelihood estimation (ALE) approach. Next, we generated functional profiles of identi‑ fied convergent structural regions utilizing resting‑state functional connectivity (rsFC) and meta‑analytic co ‑activation modeling (MACM) methods. Finally, we performed functional decoding to examine mental functions associated with our ALE, rsFC, and MACM brain characterizations. Results: We observed convergent structural alterations in a single region located in the medial prefrontal cortex. The resultant rsFC and MACM maps identified functional connectivity across a widespread, whole ‑brain network that included frontoparietal and limbic regions. Functional decoding revealed overlapping associations with attention, memory, and emotion processes. Conclusions: Consensus‑based functional connectivity was observed in regions of the default mode, salience, and central executive networks, which play a role in the tripartite model of psychopathology. Taken together, these find‑ ings have important implications for understanding the neurobiological mechanisms associated with PTSD. Keywords: Post‑traumatic stress disorder, Meta‑analysis, Voxel‑based morphometry, Functional connectivity violence, accidents, or combat [99]. Symptoms associ- Background ated with PTSD are categorized into clusters accord- Post-traumatic stress disorder (PTSD) is a psychiatric ing to the DSM 5: (1) intrusion/re-experiencing trauma, disorder in which the onset of symptoms develops after (2) avoidance, (3) negative cognition and mood, and (4) experiencing or witnessing a traumatic event, such as hyperarousal [39, 62]. Approximately 70% of adults expe- rience at least one traumatic event in their lifetime and *Correspondence: bpank001@fiu.edu up to 20% of these people develop PTSD [65]. Individuals Department of Psychology, Florida International University, Miami, FL, USA with PTSD may experience long-term debilitating effects, Full list of author information is available at the end of the article © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 2 of 16 mentally, physically, and cognitively. In the United States, likely reflects the disturbance of distributed, brain-wide roughly 8 million adults suffer from PTSD every year. neural circuitry, we also sought to functionally and Approximately 60% of men experience at least one trau- behaviorally characterize any neuroanatomical altera- matic event in their lives, often associated with combat tions in a task-independent manner. To this end, we first and war, while 50% of women will experience at least one identified convergent regions of gray matter (GM) reduc - traumatic event, typically associated with sexual assault tions in PTSD vs. non-PTSD groups using anatomical and abuse [59]. likelihood estimation (ALE) [21, 22]. Second, we identi- Current theories aim to understand the etiology of fied the task-free resting state functional connectivity PTSD, including behavioral, cognitive, and social mod- (rsFC) patterns, as well as the task-based meta-analytic els. Research suggests that reappraisal of traumatic co-activation modeling (MACM) patterns of conver- events may lead to an overgeneralized threat response gent regions, thus providing multimodal functional con- [20]. Despite progress in understanding the vulnerabil- nectivity profiles for each. Together, the VBM, rsFC, ity, symptomatology, and trajectory of PTSD [1, 39, 64], and MACM meta-analytic approaches have been used the underlying neurobiological determinants of PTSD are in previous clinically related meta-analyses [16, 37, 71], less clear. Substantial prior work has attempted to iden- they provide complementary information, yielding a tify structural brain alterations observed among indi- multimodal functional connectivity profile for a given viduals with PTSD. Voxel-based morphometry (VBM) region of interest. Lastly, we applied meta-analytic func- is a commonly used methodological approach for ana- tional decoding methods to identify the mental processes lyzing structural magnetic resonance imaging (MRI) linked to this functional connectivity profile. Collectively, data, allowing for quantitative statistical comparisons this work utilizes an innovative (meta-) analytic frame- between groups (e.g., differences in gray matter volume; work to quantitatively assess structural alterations associ- GMV) to more clearly understand the structural altera- ated with PTSD and the extended functional profiles of tions associated with neuropsychiatric disorders, such regions implicated in this disorder. A more comprehen- as PTSD. Multiple prior meta-analyses have been con- sive understanding of the neurobiological bases of PTSD ducted to identify convergent gray matter reductions in is needed to delineate future pathways toward improved PTSD patients, although consensus across meta-analyses prevention, diagnosis, and treatment. has not been reached. Each of these meta-analyses was conducted with a different scope, with varied study inclu - Methods sion/exclusion criteria, and subsequently included a wide Analytic overview range of 8 to 20 studies. Varying convergence has been We first conducted a literature search to identify studies observed across these meta-analyses, which have identi- reporting structural alterations comparing the following fied one to five significant clusters in regions that include groups: individuals with PTSD, individuals who experi- medial prefrontal cortex [7, 40, 44, 50, 55], hippocam- enced trauma but were not diagnosed with PTSD, and pus [7, 44], fusiform gyrus [50, 79], and lingual gyrus individuals who did not report experiencing trauma. A [44, 79]. Similarly, from a functional perspective, PTSD coordinate-based meta-analysis was performed using dysfunction has been reported as amygdala and frontal the ALE algorithm to identify convergent brain regions disruptions (e.g., [18] or across alterations of large-scale showing structural alterations associated with PTSD. We functional brain networks (e.g., [41] that are implicated then used multiple connectivity modeling approaches to in the tripartite model of psychopathology [56]. While comprehensively characterize the functional connectiv- some studies have addressed consensus across functional ity of these convergent regions. Specifically, rsFC and neuroimaging studies, it is challenging to assess conver- MACM assessments were applied to identify the func- gence across different psychological states and/or experi - tional profiles of structurally altered regions associated mental paradigms, which has potentially contributed to with PTSD. Lastly, we used functional decoding tech- inconsistent findings in PTSD meta-analyses of resting niques to identify behavioral profiles of the ALE, rsFC, state [3, 92] or task-based [24, 63,  33] studies. Over- and MACM results. An overview of our methodological all, this variability across meta-analytic approaches and approach is provided in Figure 1. results suggests that a consensus neurobiological model of PTSD has not yet been achieved. Literature search and study criteria The objective of the current study was to apply current We conducted a comprehensive literature search to best practices in coordinate-based neuroimaging meth- build a database of peer-reviewed MRI studies reporting ods to investigate the topography of consistently reported structural alterations associated with PTSD from 2002 to structural alterations in PTSD. As PTSD is linked to a 2020. In the first round of identifying studies, we exam - broad spectrum of neuropsychiatric symptoms, which ined previously published voxel-based morphometry P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 3 of 16 Fig. 1 Analysis Pipeline Overview. A We first conducted a literature search to extract structural coordinates and entered them into the ALE algorithm to identify convergent structural alterations among PTSD vs. non‑PTSD groups. B We next created task ‑free and task ‑based functional connectivity profiles for the convergent structural alterations. C Last, we performed functional decoding analyses on these functional profiles to make inferences about which mental functions were associated with our findings meta-analysis papers on PTSD and compiled a list of distributions to address variability within and between included studies [7, 40, 44, 50, 55]. Next, we performed studies. We used the coordinate-based ALE method as a PubMed search to identify additional peer-reviewed, implemented in NiMARE v.0.0.3 (Neuroimaging Meta- structural MRI studies of interest using the search terms Analysis Research Environment; [77], a Python library for “morphometry + PTSD”. The PubMed search aimed to neuroimaging meta-analysis. Reported coordinates were identify any potential studies that were not included in extracted from their original publication,coordinates the previously published meta-analyses. We then con- originally reported in Talairach space converted to were ducted a review of each identified publication to include MNI coordinates [45, 46] so that all coordinates referred the following study criteria: peer-reviewed MRI studies, to MNI space. Once transformed, statistical probability reporting results among adult humans, written in the maps were created for each foci and combined to model English language, focused on gray matter structural dif- the likelihood that a given voxel displayed a between- ferences, and included original data (i.e., not a review). group structural difference for each study. Observed Subsequently, exclusion criteria were as follows: trauma voxel-wise ALE scores characterized the most consist- or stressful life event studies not measuring PTSD, other ently reported foci across the whole brain. Significance non-voxel-based morphometry methods, treatment and testing and correction for multiple comparisons involved longitudinal effects, papers reporting a priori regions of thresholding the voxel-wise ALE map using a cluster- interest (ROIs), within-group effects, null effects, over - forming threshold of P < 0.001. Then, a permutation lapping samples to previous studies, and studies that did procedure was performed in which a null distribution of not report coordinate-based results. maximum cluster sizes was generated from 10,000 itera- tions of replacing reported foci with randomly selected Anatomical likelihood estimation (ALE) gray matter voxels, generating ALE maps from the ran- ALE is a voxel-based meta-analytic technique that iden- domized dataset, and identifying the maximum clus- tifies convergent coordinates (i.e., foci) across a set of ter size after thresholding at P < 0.001. The cluster-level neuroimaging studies. Foci are treated as 3D Gaussian FWE correction threshold was set at P < 0.05, meaning Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 4 of 16 only those clusters from the original, thresholded ALE workflow is described by Glasser and colleagues [27], map were retained if their size was greater than the clus- but consists of typical imaging pre-processing tech- ter size corresponding to the 95th-percentile from the niques that leverage the high-quality data acquired null distribution. We applied the above ALE procedure by the HCP. First, T1- and T2-weighted images were to identify convergent brain regions reflecting structural aligned, bias field corrected, and registered to MNI alterations between individuals with and without PTSD space. Second, the functional fMRI pipeline removed (i.e., PTSD vs. non-PTSD) separately for the contrasts of spatial distortions, realigned volumes to compensate PTSD > non-PTSD and non-PTSD > PTSD. for subject motion, registered the fMRI data to struc- tural volumes (in MNI space), reduced the bias field, normalized each functional acquisition to its corre- sponding global mean, and masked non-brain tissue. Functional profiles of structurally altered regions Noteworthily, care was taken to minimize smooth- associated with PTSD ing induced by interpolation and that no overt volume Next, we sought to characterize the functional connectiv- smoothing was performed. ity patterns associated with regions demonstrating struc- The fMRI signal contains many sources of variabil - tural alterations in PTSD. To this end, we investigated ity, including artifactual and non-neuronal signals, that task-free functional connectivity utilizing a database of make identifying the underlying neuronal activity dif- resting state fMRI data, as well as task-based functional ficult. Using a combination of independent compo - connectivity using a meta-analytic database of co-activa- nent analysis (ICA) and classification techniques, HCP tion results. functional data were automatically denoised using FMRIB’s ICA-based X-noiseifier [75]. Briefly, ICA was performed on each functional dataset independently Task‑free functional connectivity: resting‑state fMRI and characteristics of each component, such as spatial (rs‑fMRI) localization and power in high frequencies, were evalu- Resting-state connectivity analyses typically identify ated by a classifier to determine if a given component brain voxels demonstrating the highest temporal cor- was related to neuronal activity or artifact. The time- relation with the average time series of a seed ROI and series corresponding to artifactual components were provide context about the brain’s underlying functional then regressed out of the data, providing a “cleaned”, architecture. To derive robust rsFC maps for each ROI, denoised dataset for further investigation. we utilized the minimally pre-processed and denoised Using the minimally pre-processed, denoised rest- (or “cleaned”) resting-state fMRI data provided by the ing-state datasets for each participant, the “global sig- Human Connectome Project’s [90] Young Adult Study nal” was removed using FSL’s fsl_glm [36] interface in S1200 Data Release (March 1, 2017). On November 12, NiPype [29]. The “global signal”, although controversial 2019, 150 randomly selected participants (28.7 ± 3.9 in the domain of resting-state analyses, was removed years) were downloaded via the HCP’s Amazon Web under the premise that it performed better than other Services (AWS) Simple Storage Solution (S3) repository. commonly used motion-correction strategies at remov- The randomly chosen participants included 77 females ing motion-related artifacts in the HCP resting-state (30.3 ± 3.5 years) and 73 males (27.1 ± 3.7 years). A dif- data [8]. The resulting data set was then smoothed with ference in age between the two biological sex groups was a FWHM kernel of 6-mm using FSL’s susaan interface significant but is consistent with the 1200 Subjects Data in NyPipe. For each participant, the average time series Release. Detailed acquisition and scanning parameters for each ROI was extracted and a whole-brain correla- for HCP data can be found in consortium manuscripts tion map was calculated and averaged across runs for [82, 89, 91], but relevant scan parameters are briefly sum - a single participant for every ROI. The average corre - marized here. Each participant underwent T1-weighted lation maps for each participant were transformed to and T2-weighted structural acquisitions and four rest- Z-scores using Fisher’s r-to-z transformation. A group- ing-state fMRI acquisitions. Structural images were col- level analysis was then performed to derive a rsFC map lected at 0.7-mm isotropic resolution. Whole-brain EPI for each ROI using FSL’s randomise interface [94] in acquisitions were acquired on the 3T Siemens Connec- NiPype. Images were thresholded non-parametrically tome scanner: 32-channel head coil, TR = 720 msec, using GRF-theory-based maximum height thresholding TE = 33.1 msec, in-plane FOV = 208 × 180 mm, 72 slices, with a (voxel FWE-corrected) significance threshold of 2.0 mm isotropic voxels, and multiband acceleration fac- P < 0.001 [96], such that more spatially specific connec - tor of 8 [25]. tivity maps could be derived when using such a highly The S1200 data release contained minimally pre-pro - powered study [95]. cessed and denoised data. The minimal pre-processing P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 5 of 16 Task‑based functional connectivity: meta‑analytic MACM, and rsFC analyses. To do so, we utilized gener- co‑activation modeling (MACM) alized correspondence latent Dirichlet allocation (GC- Leveraging reported coordinates from task-based fMRI LDA) functional decoding methods in NiMARE applied studies, meta-analytic co-activation is a relatively new to the resulting unthresholded ALE, rsFC, and MACM concept that identifies brain locations that are most likely maps. This type of decoding provides an approach to to be co-activated with a given seed ROI across multiple infer mental processes associated with neuroimaging task states. Differing from rsFC, MACM provides context spatial patterns. GC-LDA utilizes probabilistic Bayesian about neural recruitment during goal-oriented behaviors. statistics that learns latent topics from a large database of We therefore aimed to integrate these two complemen- papers (e.g., NeuroSynth) [74]. From the database, each tary modalities by supplementing the rsFC maps with topic found is treated as a probability distribution and MACM maps for each ROI. To do so, we relied on the creates a spatial distribution in MNI space across voxels Neurosynth database [98], which archives published ste- from the maps entered into the decoding algorithm. The reotactic coordinates from over 14,000 fMRI studies and “topics” encompass terms and associated brain regions 150,000 brain locations. Neurosynth relies on an auto- that co-occur in the literature from a literature database. mated coordinate extraction tool to “scrape” each avail- We set our model to 200 topics. We report 10 terms cor- able fMRI study for reported coordinates. Due to the responding to the highest weights associated with our nature of this automated process, fMRI studies reporting ALE, rsFC, and MACM results. results of multiple experimental contrasts as separate sets of coordinates are amalgamated into a single set of coor- Results dinates; in addition, “activation” and “de-activation” coor- Literature search and study criteria dinates are not distinctly characterized. However, while The literature search yielded a total of 85 articles using this inherent “noise” may limit interpretational abilities, the above-described search terms. Figure  2 provides a the power over manually curated datasets outweighs the PRISMA diagram, which details the review and filtering potential confounds of bi-directional or mixed-contrast of those 85 studies. In the first round of review, records effects. (i.e., titles and abstracts) were screened to exclude 18 To generate a MACM map for each ROI, we utilized studies that corresponded to non-human or non-English NiMARE [77] to search the Neurosynth database for all studies, reviews, or studies reporting white matter dif- studies reporting at least one peak within the defined ferences or differences among children or adolescents. ROI mask. Neurosynth tools implement the multilevel Then, we examined the full-text articles to assess addi - kernel density analysis (MKDA) algorithm for perform- tional study criteria; 44 additional studies were excluded ing meta-analyses based on a subset of studies, such as as being not eligible for the current meta-analysis. that described here. However, we opted to use the ALE The final set of included studies consisted of 23 algorithm as implemented in NiMARE given its optimal publications. Within these publications, gray mat- performance in replicating image-based meta-and mega- ter structural alterations were assessed by compar- analyses [76]. The ALE algorithm requires sample size ing whole-brain VBM results among individuals with information, or the number of subjects, that contributed and without PTSD, reported as 3D coordinates in to a given experimental contrast to generate a smooth- MNI or Talairach space. Control comparison groups ing kernel. However, Neurosynth is not able to capture included individuals who had experienced trauma but sample size (which could also vary across experimental did not develop PTSD and individuals who had not contrasts within a study). u Th s, we utilized a smoothing experienced trauma. Nineteen publications included kernel with a FWHM of 15  mm, which has been shown trauma-exposed controls (TC), while ten publications to yield results with strong correspondence for image- included healthy, non-trauma-exposed controls (HC). based meta- and mega-analyses [76]. The ALE algorithm Altogether, this set of 23 studies collectively examined was applied to the set of studies reporting activation 476 individuals with PTSD and 892 individuals with- within the boundaries of each ROI. Once ALE maps were out PTSD, which included 288 TC and 633 HC. With generated for each ROI, as described above, voxel-FWE respect to the type of structural alterations observed, correction (P < 0.001) was performed to reflect the statis - studies reported multiple different VBM metrics. Sev - tical thresholding approach used for rsFC maps. enteen publications reported group differences in gray matter volume (GMV), seven publications reported Functional decoding: generalized correspondence latent differences in gray matter density (GMD), and one dirichlet allocation (GC‑LDA) reported gray matter concentration (GMC). Collec- We sought to infer what mental processes were most tively, we refer to all of these metrics as gray matter likely linked with brain regions identified in our ALE, (GM) differences among individuals with and without Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 6 of 16 Fig. 2 PRISMA Diagram. PRISMA flow chart detailing the literature search and selection criteria of studies included in the meta‑analysis PTSD. Additional details on the demography of par- non-PTSD for a total of 20 foci, including 3 for PTSD ticipant groups and study design are provided in Addi- (9 foci) vs. TC and 2 contrasts for PTSD vs. HC (9 foci). tional file  1: Table  S1 located in this project’s GitHub repository (https:// github. com/ NBCLab/ meta- analy Anatomical likelihood estimation (ALE) sis_ ptsd). Using NiMARE v.0.0.3 [77], ALE meta-analysis was Within this final set of 23 publications, multiple con - performed to assess convergence for the 25 contrasts trasts of interest were reported. 25 contrasts reported from 22 publications of GM decreases among individu- GM decreases in PTSD vs non-PTSD for a total of 159 als with and without PTSD (i.e., non-PTSD > PTSD); a foci; this included 16 contrasts for PTSD vs. TC (82 complete listing is provided in Table  1. Neuroimaging foci) and 9 contrasts for PTSD vs. HC (77 foci). Con- simulations indicate that a minimum of 20 contrasts versely, 6 contrasts reported GM increases in PTSD vs. are necessary for a well-powered coordinate-based P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 7 of 16 Table 1 Studies Included in ALE Meta‑Analysis Citation Sample size Contrasts 1 [5] Total N = 38; PTSD n = 19 Healthy controls > PTSD 2 [10] Total N = 41; PTSD n = 21 Non‑PTSD > PTSD 3 [11] Total N = 24; PTSD n = 12 Controls > PTSD 4 [12] Total N = 20; PTSD n = 10 Controls > recent onset PTSD 5 [13] Total N = 60; PTSD n = 30 Healthy controls > PTSD 6 [14] Total N = 28; PTSD n = 14 Healthy controls > PTSD 7 [19] Total N = 33; PTSD n = 20 Non‑ Trauma controls > PTSD 8 [26] Total N = 38; PTSD n = 21 Controls > PTSD 9 [31] Total N = 184; PTSD n = 14 Non‑PTSD > PTSD; trauma exposed > PTSD 10 [34] Total N = 28; PTSD n = 13 Trauma exposed > PTSD 11 [38] Total N = 41; PTSD n = 18 Combat‑ exposed Non‑PTSD > PTSD 12 [43] Total N = 53; PTSD n = 24 Controls > PTSD 13 [49] Total N = 24; PTSD n = 12 Controls > PTSD 14 [58] Total N = 43; PTSD n = 21 Non‑PTSD > PTSD 15 [61] Total N = 75; PTSD n = 25 Healthy controls > PTSD; trauma exposed > PTSD 16 [66] Total N = 220; PTSD n = 57 Trauma exposed > PTSD 17 [72] Total N = 32; PTSD n = 16 Trauma exposed controls > PTSD 18 [84] Total N = 31; PTSD n = 11 Healthy controls > PTSD; trauma exposed > PTSD 19 [85] Total N = 50; PTSD n = 25 Healthy controls > PTSD 20 [97] Total N = 25; PTSD n = 9 Non‑PTSD > PTSD 21 [101] Total N = 20; PTSD n = 10 Trauma‑ exposed > PTSD 22 [100] Total N = 39; PTSD n = 14 Non‑PTSD > PTSD 25 contrasts from 22 publications reported GM decreases among individuals with and without PTSD (i.e., non-PTSD > PTSD). Sample sizes are provided for the total number of participants (N) (i.e., PTSD and non-PTSD), as well as the sample sizes for the PTSD groups (n) meta-analysis [23]. Thus, we were unable to assess the vs. TC and PTSD vs. HC contrasts (i.e., GM increases 6 contrasts of GM increases (i.e., PTSD > non-PTSD) and decreases) to determine if the use of different given insufficient power. With respect to GM decreases, comparison groups potentially contributed additional we observed a single cluster of convergence located heterogeneity, limiting assessment of convergence. in the mPFC (x=0, y=46, z=10; BA 32) (Figure  3; P < However, we observed null results for these additional 0.001, FWE-corrected P < 0.05). Given these results, we contrasts as well, likely in part due to the underpow- performed additional ALE meta-analyses for the PTSD ered samples [23]. Fig. 3 ALE Results for non‑PTSD > PTSD. Sagittal brain slices illustrating convergent structural alterations associated with PTSD as determined by an ALE meta‑analysis of GM reductions (P < 0.001, FWE‑ corrected P < 0.05) Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 8 of 16 Functional profiles of structurally altered regions parahippocampus. Next, to further examine function- associated with PTSD ally coupled regions with the mPFC seed, we generated We next investigated the functional connectivity of a MACM map using the Neurosynth database which the mPFC cluster identified above showing convergent demonstrated task-based coactivations with a simi- gray matter reductions among individuals with PTSD. lar pattern as the rsFC map. The locations of rsFC and To this end, we analyzed task-free rsFC and task-based MACM results are provided in Table  2. Figure  4 illus- MACM. First, we generated a rsFC map using the ALE- trates the rsFC (blue) and MACM (red) results, with derived mPFC cluster as a seed region. The resultant overlapping regions, indicating a consensus between rsFC map revealed rsFC with the superior frontal gyrus, rsFC and MACM (pink), revealed in the ACC, medial medial frontal gyrus, inferior frontal gyrus, ACC, thal- prefrontal gyrus, middle temporal gyrus, insula, infe- amus, posterior cingulate (PCC), superior temporal rior parietal lobe, thalamus, precuneus, parahippocam- gyrus, medial temporal gyrus, precuneus, cuneus, and pus, insula, and PCC regions (Table3). Table 2 rsFC and MACM Results rsFC results MACM results Anatomical label x y z Anatomical label x y z Anterior cingulate, BA 32 4 44 10 Medial frontal gyrus, BA 10 − 2 50 6 L Inferior frontal gyrus, BA 47 − 30 14 − 16 Superior frontal gyrus, BA 6 0 14 48 Cingulate gyrus, BA 24 2 − 18 36 Medial frontal gyrus, BA 8 2 26 38 Anterior cingulate, BA 32 0 36 − 6 Posterior cingulate, BA 31 − 4 − 54 26 Posterior cingulate, BA 31 8 − 52 24 L extra‑nuclear, BA 47 − 34 20 − 2 Cingulate gyrus, BA 31 − 8 − 54 26 R extra‑nuclear, BA 47 36 22 − 2 Midbrain 0 − 20 − 20 L angular gyrus, BA 39 − 46 − 68 30 Anterior cingulate, BA 24 4 28 16 L superior parietal lobule, BA 7 − 30 − 62 46 R Inferior frontal gyrus, BA 47 30 16 − 16 L inferior frontal gyrus, BA 9 − 46 10 28 Precuneus, BA 7 0 − 70 34 R superior temporal gyrus, BA 39 52 − 60 26 L caudate − 4 12 − 2 R inferior parietal lobule, BA 40 40 − 52 44 R angular gyrus, BA 39 52 − 64 36 L amygdala − 22 − 8 − 16 L inferior parietal lobule, BA 39 − 50 − 64 40 R amygdala 24 − 6 − 16 Posterior cingulate, BA 30 − 6 − 54 10 R inferior frontal gyrus, BA 9 46 10 28 L parahippocampal gyrus, BA 35 − 22 − 22 − 14 R caudate 12 10 2 L superior frontal gyrus, BA 8 − 22 34 46 L lentiform nucleus − 12 8 − 2 Cingulate gyrus, BA 31 − 4 − 32 38 L thalamus, medial dorsal nucleus − 6 − 14 6 R caudate 10 18 − 4 R thalamus, medial dorsal nucleus 6 − 14 6 L superior frontal gyrus, BA 9 − 20 48 34 L inferior parietal lobule, BA 40 − 42 − 44 44 Cerebellar tonsil 6 − 50 − 36 L inferior temporal gyrus, BA 21 − 56 − 10 − 16 Coordinate locations of the rsFC and MACM results, including the anatomical label and MNI coordinates of local maxima. Negative x values indicate the left (L) hemisphere and positive x values indicate the right (R) hemisphere Fig. 4 rsFC and MACM Results. rsFC (blue) and MACM (red) results; common areas (pink) indicate consensus between connectivity approaches. Images are thresholded at voxel‑ wise FWE P < 0.001 P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 9 of 16 Table 3 Consensus between rsFC and MACM Results separately for the structural ALE, rsFC, and MACM maps. The decoding terms with the top 10 weights from rsFC + MACM consensus the GC-LDA analysis for the structural ALE map were: Anatomical label x y z visual, emotional, memory, novel, reward, motor, self, faces, learning, and face (Table  4a). The decoding terms Medial frontal gyrus, BA 10 − 2 50 6 with the top 10 weights from the GC-LDA analysis for the Medial frontal gyrus, BA 8 2 26 38 rsFC map were: default, default mode network, intrinsic, Posterior cingulate, BA 31 − 4 − 54 26 scale, self, person, reward, bias, judgements, and contexts L angular gyrus, BA 39 − 46 − 68 30 (Table  4b). Topographically speaking, the rsFC results R superior temporal gyrus, BA 39 52 − 60 26 resembled regions of combined default mode [30, 69] and L inferior frontal gyrus, BA 47 − 32 18 − 6 salience networks [57, 78], and the functional decoding R inferior frontal gyrus, BA 47 38 20 − 8 outcomes suggested that the rsFC results were associ- L parahippocampal gyrus, BA 28 − 24 − 16 − 18 ated with self-referential, intrinsic, and reward processes. L lentiform nucleus, putamen − 12 10 − 4 Next, we examined MACM-based decoding results. The R caudate 10 10 − 2 decoding terms with the top 10 weights from the GC- L thalamus, medial dorsal nucleus − 4 − 14 6 LDA analysis for the MACM map were: visual, motor, R thalamus, medial dorsal nucleus 6 − 14 8 emotional, memory, attention, auditory, reward, spatial, L parahippocampal gyrus, BA 34 − 20 2 − 12 schizophrenia, and language (Table  4c). Topographically R hippocampus 26 − 14 − 20 speaking, the MACM results also resembled regions of L inferior temporal gyrus, BA 21 − 56 − 10 − 16 the default mode [30, 69] as well as the frontoparietal L parahippocampal gyrus, BA 28 − 16 − 4 − 14 central executive network [17, 78], and the functional Coordinate locations of the consensus between rsFC and MACM results, decoding outcomes suggested association with execu- including the anatomical label and MNI coordinates of local maxima. Negative x values indicate the left (L) hemisphere and positive x values indicate the right tive emotional and memory processes. A summary of the (R) hemisphere decoding analyses for all three sets of images is shown as a radar plot in Figure 5. Functional decoding: generalized correspondence latent dirichlet allocation (GC‑LDA) Discussion Lastly, we performed functional decoding of the struc- The overall objective of this study was to investigate tural ALE, rsFC, and MACM maps to provide insight convergent alterations in brain structure among indi- into the behavioral functions putatively associated with viduals with PTSD using emergent meta-analytic tech- the observed functional connectivity patterns. Func- niques. Further, we sought to extend the literature and tional decoding was conducted using a GC-LDA analysis assess potential functional consequences associated [74]. Because GC-LDA does not provide correlational or with observed structural alterations in PTSD by applying statistical rankings, the top 10 unique terms computed complementary rsFC and MACM analytic techniques. from the GC-LDA analysis were taken into consideration The current meta-analysis of 23 VBM studies evaluating Table 4 Functional Decoding Results. Functional decoding results for (a) ALE structural meta‑analysis, (b) rsFC, and (c) MACM results as described by Neurosynth terms (a) ALE (b) rsFC (c) MACM Rank Term Weight Rank Term Weight Rank Term Weight 1 Visual 1.886 1 Default 11.234 1 Visual 5900.643 2 Emotional 0.919 2 Default mode network 9.225 2 Motor 3839.578 3 Memory 0.845 3 Intrinsic 7.494 3 Emotional 3665.765 4 Novel 0.616 4 Scale 6.236 4 Memory 3476.688 5 Reward 0.576 5 Self 5.081 5 Attention 2931.357 6 Motor 0.521 6 Person 4.977 6 Auditory 2267.840 7 Self 0.509 7 Reward 4.780 7 Reward 2107.441 8 Faces 0.472 8 Bias 4.568 8 Spatial 2072.742 9 Learning 0.467 9 Judgements 4.279 9 Schizophrenia 2070.157 10 Face 0.450 10 Contexts 4.271 10 Language 2057.731 Rankings display weighted terms listed from highest (1) to lowest (10) Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 10 of 16 brain structure. Previous meta-analyses have identified GM reductions in the mPFC, hippocampus, fusiform gyrus, and lingual gyrus; however, not all of these regions were consistently observed across all meta-analyses [7, 40, 44, 50, 55, 79]. Beyond the mPFC, we did not observe additional convergent GM reductions, indicating that prior findings in these other regions were not replicated. Across the PTSD literature, there is a high degree of variability associated with participant trauma exposure, length of diagnosis of PTSD, medication use, and comor- bidity. Inconsistencies between our findings and previ - ous meta-analytic results could be due to conceptual and methodological differences across the earlier studies, such as the scope of the research question exploring the neurobiology of PTSD, and the subsequent differences in inclusion/exclusion criteria that resulted in different sets Fig. 5 Functional Decoding Results. Functional decoding results for of included studies. Comparison of the included studies the ALE structural meta‑analysis (pink), rsFC (blue), and MACM (red) in this and prior VBM meta-analyses of PTSD indicated results as described by Neurosynth terms. Radar plots display the top varying degrees of overlap, including (from earliest to five terms across all three decoding analyses. The scale of the weights depends on both the GC‑LDA model weights and the input values most recent meta-analyses): 7 of 9 included studies [44], [74], thus, the scale is arbitrary and has been normalized here to 14 of 17 included studies [50], 15 of 20 included studies facilitate visualization [55], 7 of 13 included studies [7], 7 of 8 included studies [40], and 10 out of 12 included studies [79]. Beyond selection of included studies, the meta-ana- GM volume alterations among PTSD versus non-PTSD lytic approach may contribute to the source of variabil- groups identified a single node of convergent gray mat - ity across results. Previous meta-analyses used either the ter loss in the mPFC. GC-LDA-based functional decod- ALE approach [44, 50] or signed differential mapping ing of this cluster was linked to Neurosynth terms of [7, 40, 55, 79]. Consistent with the present results, the visual, emotional, memory, novel, reward, motor, self, meta-analyses by Meng et al. [55] and Klaming et al. [40] faces, learning, and face. Follow-up ALE analyses explor- also yielded a single cluster in mPFC, which used the ing GM reductions in PTSD vs. HC (non-traumatized SDM method while our current results used the ALE controls) and PTSD vs. TC (trauma-exposed controls not approach. However, of all prior meta-analyses, only the diagnosed with PTSD) yielded null findings likely due to study by Meng et al. [55] meets the current threshold of a insufficient power [23]. Subsequent analyses of the ALE- minimum of 20 contrasts for a well-powered coordinate- derived mPFC cluster were conducted to assess task-free based meta-analysis [23]. After reviewing the above prior (rsFC) and task-dependent (MACM) functional connec- meta-analytic work in comparison to our current results, tivity, identifying a consistent and widespread functional we conclude that extensive heterogeneity in the PTSD network implicated in PTSD. These results indicate that literature, combined with varying meta-analytic inclu- structural alterations in the mPFC among individu- sive/exclusion criteria, likely contributed to differences als with PTSD are possibly linked to disruptions across between our results and prior meta-analytic findings. To a larger frontoparietal network that includes the medial, our knowledge, the current meta-analysis of 25 contrasts superior, and inferior frontal gyri, PCC, parahippocam- represents the largest PTSD meta-analysis of structural pal gyri, angular gyri, superior temporal gyrus, thalamus, findings to date, with prior meta-analytic work examin - caudate, and lentiform nucleus. Functional decoding of ing 8-20 included studies. We observed that the mPFC is rsFC and MACM results indicates substantive term over- robustly associated with structural alterations in PTSD; lap with the mPFC ALE results, with additional network- however, it is important to consider how the mPFC is related terms (e.g., default, default mode network, and integrated within existing neurocircuitry models associ- intrinsic). ated with PTSD symptomology. Traditional neurocircuitry models of PTSD utilize a Structural alterations and dysfunction in PTSD fear-conditioning framework, emphasizing hyperreactiv- Our current findings suggest the mPFC appears as the ity of the amygdala in response to fear-related stimuli and most consistently reported brain region across VBM dysfunction between the mPFC and orbitofrontal cortex, neuroimaging studies exploring the impact of PTSD on as well as the hippocampus, in attention and top-down P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 11 of 16 control during threat exposure [70, 81]. However, limit- to a neurobiological theory of psychopathology [28, 56, ing consideration of the psychopathology of PTSD to 57]. The application of the tripartite model to neurobiol - focus on a single brain region (i.e., the amygdala) empha- ogy models of psychiatric disorders define dysfunction sizes fear-related brain activity while minimizing brain within and between connectivity of the DMN, SN, and circuitry implicated in the complex constellation of PTSD CEN networks and relates to a broad range psychiatric symptoms associated with response to trauma exposure, disorders [80], including PTSD [60, 63]. Overall, the cur- such as re-experiencing trauma, avoidance, negative rent meta-analysis identified a functional profile of the mood, and numbing. These additional processes remain mPFC associated with connectivity between the DMN, largely unexplained in original PTSD models. However, SN, and CEN, which broadly supports a network theory more recent neurocircuitry models build from this per- of PTSD [2, 41]. spective, with increased emphasis on altered function of According to the tripartite model of brain function, the the mPFC, its role in contextualization, and how context SN is thought to mediate activity between the DMN and processing is core to the constellation of PTSD symptoms CEN networks in order to orient to external stimuli or [51, 52]. While our results indicated convergent struc- internal salient biological stimuli [57], Sripada et al. [41]. tural alterations in the mPFC, we did not observe similar Altered inter- and intra-network functional connectiv- convergence in the amygdala or other regions that have ity between the DMN, SN, and CEN has previously been been implicated in prior neurocircuitry models of PTSD implicated in PTSD [41]. Specifically, seed-based rest - [32, 42, 70, 81]. However, our results are congruent with ing state studies identified decreased connectivity within the expanded models of PTSD and we provide robust the DMN and SN, yet increased connectivity between evidence in support of the mPFC as a critical node in these two networks among PTSD patients (Sripada et al. PTSD neurocircuitry. Further, our functional decoding [89]). Furthermore, other resting state studies on PTSD results provide additional support for the contextualiza- utilizing graph theory approaches [48] and independ- tion models of PTSD. Taken together, reduced GM in the ent component analysis [102] replicated weakened con- mPFC among individuals diagnosed with PTSD supports nectivity within the DMN, SN, and CEN, yet heightened the premise that these structural alterations may contrib- connectivity between the DMN and SN [35, 89]. Taken ute to deficits in context processing and ultimately play a together, this literature suggests deficits in top-down dominant role in contributing to behaviors related to the control over heightened responses to threatening stim- constellation of symptoms in PTSD [51, 52]. uli and abnormal regulation of orienting attention to threatening stimuli [41, 48, 84, 89, 102]. Patterns from Functional profiles of structural findings in PTSD: support task-based studies reflect previous findings of weakened for the tripartite model of psychopathology connectivity between the SN and DMN and heightened rsFC and MACM analyses characterized mPFC func- connectivity between the SN and CEN [64, 87]. In a study tional connectivity as extending across widespread, among individuals with recent trauma exposure, connec- whole-brain networks engaging frontoparietal and limbic tivity between the DMN, SN, and CEN was reported to regions. These rsFC and MACM results, in conjunction be disrupted among participants who developed PTSD with functional decoding outcomes, identified a func - vs. those who do not [54, 68], providing evidence of dif- tional connectivity profile suggestive of spatial patterns ferential functional connectivity between PTSD patients associated with the default mode network (DMN) [30, and traumatized non-diagnosed individuals. Network 69], salience mode network (SN) [57, 78], and central dysfunction associated with the DMN, SN, and CEN is executive network (CEN) [17, 78]. The DMN is a system also evident in task-based studies, including cues con- of connected brain areas including the mPFC, PCC, infe- taining trauma stimuli [69], eye gaze [87], and a broad rior parietal, and temporal cortices that are often collec- range of behavioral paradigms [64]. Aberrant connectiv- tively observed as displaying anticorrelation with regions ity between and within the DMN, SN, and CEN has also actively engaged during attention-demanding tasks. been associated with PTSD symptoms, such that height- Areas of the DMN are thought to collectively contribute ened connectivity and activity of the DMN was associated to mental processes associated with introspection and with depersonalization/derealization, while weakened self-referential thought [30, 53, 93]. The SN consists of connectivity and activity of the CEN was associated with the dorsolateral ACC and bilateral insula and is involved hyperarousal and hypervigilance [2]. Additionally, weak- in saliency detection and attentional processes [57, 78]. ened inter-network connectivity between the SN and Finally, the CEN consists of the dorsolateral prefrontal DMN has been found to be positively correlated with and posterior parietal cortices and is typically involved in Clinician Administered PTSD Scale (CAPS) scores that attentionally driven cognitive functions, including goal- measure PTSD symptom severity [84, 89]. Moreover, directed behavior [87]. These three networks are central Bluhm et al. [4] found weakened spontaneous activity in Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 12 of 16 regions of the DMN; in addition, posterior cingulate con- the number of participants across each group was some- nectivity was positively correlated with self-reported dis- what unevenly distributed due to small sample sizes in sociated experiences among participants with PTSD. In the original studies. However, the current meta-analysis sum, the literature on abnormal brain function associated met the previously recommended standard of at least with PTSD points to a pattern of results suggesting that 20 experimental contrasts required to conduct a well- symptoms are related to aberrant connectivity within and powered meta-analysis [23]. Second, much heterogeneity between the DMN, SN, and CEN. In a recent review of exists across the studies included in our meta-analysis. the neuroimaging literature on PTSD, Lanius et  al. [47] For example, many of the studies had diagnostic criteria summarized this work to reflect that dysfunction in the for PTSD using different clinical measures and reported DMN is associated with an altered sense of self, dysfunc- different instances of the duration of PTSD (e.g., lifetime tion in the SN is associated with hyperarousal and hyper- vs. first onset). Substantial variability was also present in vigilance, and dysfunction in the CEN is associated with the type of trauma and duration of exposure to trauma cognitive dysfunction, including memory and cognitive within the different groups for this study. Given these control deficits. issues, we were unable to classify PTSD subtypes across The results from the current meta-analysis provide the included studies and thus have reported results that a robust mPFC-centric model of PTSD that is aligned relate to generalized PTSD. Many of the original stud- with the extant literature and compliments the tripartite ies were not able to clearly disentangle comorbidity of model of psychopathology. The mPFC, a core region of PTSD with other psychiatric disorders (e.g., depression, the DMN [30, 69], is often disrupted in individuals with anxiety) or report instances of medication and drug PTSD [15, 68]. The results of the present meta-analysis abuse. Furthermore, studies relied on various neuro- suggest alterations in mPFC structure, and related func- imaging acquisition and analysis methods, which likely tion, may play a crucial role in the underlying neurobi- introduced additional variability associated with meth- ology of PTSD. Dysfunction of the mPFC is thought to odological flexibility [6, 9]. However, the goal of neuro- be associated with poorer regulation of contextualization imaging meta-analysis was to examine consensus despite of PTSD symptoms. Prior literature indicates weakened such variability in the literature. With this in mind, we integration of the DMN and disrupted inter-network are confident that the mPFC is a significant brain region connectivity with the SN and CEN, representing aberrant linked to GM reductions in PTSD, as well as a robust dysfunction of these tripartite networks in the psychopa- node of the DMN that plays an important role in toggling thology of PTSD [73]. Most of the prior functional and between the DMN, SN, and CEN. Future transdiagnostic structural work involved varying analytic approaches, and meta-analytic work is needed to identify similar and examined heterogeneous populations, and utilized region unique neurobiological mechanisms of PTSD in compar- of interest approaches or a priori hypotheses. The cur - ison to other related disorders, including complementary rent application of advanced meta-analytic techniques disease-decoding or structural covariance analysis, which allowed for a whole-brain assessment of structural altera- would further advance clinical insight. tions associated with PTSD and the associated functional profiles of the mPFC. Future work in PTSD should con - Conclusions sider integrating network-based analytic approaches The present study utilized coordinate-based meta-ana - with an mPFC-centric tripartite model to investigate lytic techniques to determine that reduced mPFC GM is differences in neuropathology of PTSD subtypes (e.g., consistently found among individuals with PTSD. Com- trauma experiences, duration of exposures), characteriz- plementary analyses of rsFC and MACM functional ing heterogeneous presentations of PTSD symptoms, and connectivity provided novel insight into how structural potential predispositional developmental effects among alterations may have potential functional consequences. youth, adolescent, and adult populations. Our results indicated that decreases in mPFC GM may be linked to widespread functional systems that are Limitations implicated in behavioral deficits and cluster symptoma - Our study is limited by several considerations. First, the tology of PTSD. Specifically, consensus-based func - present meta-analysis is limited by the small number tional profiles, across task-free and task-based domains, of studies included. The studies that met the standards emphasized brain regions associated with the tripartite of inclusion for this study were considered to reduce model of psychiatric disorders where inter- and intra- instances of variance and consider reliability of study network connectivity involving the DMN, SN, and CEN findings (inclusion and exclusion criteria are shown in are core to PTSD dysfunction. Overall, these results may Fig.  2). By considering the inclusion of trauma-exposed be important in providing a more comprehensive under- controls, healthy controls, and individuals with PTSD, standing of the neurobiological bases of PTSD, which is P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 13 of 16 3. Bao W, Gao Y, Cao L, Li H, Liu J, Liang K, Hu X, Zhang L, Hu X, Gong Q, needed to understand the varying diagnosis, symptoma- Huang X. Alterations in large‑scale functional networks in adult post ‑ tology, and treatment of PTSD, as well as enhanced tar- traumatic stress disorder: a systematic review and meta‑analysis of geting of treatment towards heterogeneous classification resting‑state functional connectivity studies. Neurosci Biobehav Rev. 2021;131:1027–36. and symptom clusters of PTSD. 4. Bluhm RL, Williamson PC, Osuch EA, Frewen PA, Stevens TK, Boksman K, et al. Alterations in default network connectivity in posttraumatic Supplementary Information stress disorder related to early‑life trauma. J Psychiatry Neurosci JPN. 2009;34:187–94. The online version contains supplementary material available at https:// doi. 5. Bossini L, Santarnecchi E, Casolaro I, Koukouna D, Caterini C, Cecchini org/ 10. 1186/ s12993‑ 022‑ 00196‑2. F, et al. Morphovolumetric changes after EMDR treatment in drug‑ naïve PTSD patients. Riv Psichiatr. 2017;52:24–31. https:// doi. org/ 10. Additional file 1: Table S1. Summary of the demographic and clinical 1708/ 2631. 27051. variables for the voxel‑based morphometry studies included in the 6. Botvinik‑Nezer R, Holzmeister F, Camerer CF, et al. Variability in the meta‑analysis analysis of a single neuroimaging dataset by many teams. Nature. 2020. https:// doi. org/ 10. 1038/ s41586‑ 020‑ 2314‑9. 7. Bromis K, Calem M, Reinders AATS, Williams SCR, Kempton MJ. Meta‑ Acknowledgements analysis of 89 structural MRI studies in posttraumatic stress disorder The authors would like to thank the FIU Instructional & Research Comput‑ and comparison with major depressive disorder. Am J Psychiatry. ing Center (IRCC, http:// ircc. fiu. edu) for providing the HPC and computing 2018;175:989–98. https:// doi. org/ 10. 1176/ appi. ajp. 2018. 17111 199. resources that contributed to the research results reported within this paper. 8. Burgess GC, Kandala S, Nolan D, Laumann TO, Power JD, Adeyemo B, et al. Evaluation of denoising strategies to address motion‑ correlated Author contributions artifacts in resting‑state functional magnetic resonance imaging data BP and IC collected and prepared data for meta‑analysis. MCR and JEB from the human connectome Project. Brain Connect. 2016;6:669–80. analyzed data. MCR, JEB, LDHB, and TS contributed scripts and pipelines. BP, https:// doi. org/ 10. 1089/ brain. 2016. 0435. MRC, and ARL wrote the paper. All authors contributed to the revisions and 9. Carp J. On the plurality of (methodological) worlds: estimating the approved the final version. All authors read and approved the final manuscript. analytic flexibility of fMRI experiments. Front Neurosci. 2012. https:// doi. org/ 10. 3389/ fnins. 2012. 00149. Funding 10. Chao LL, Lenoci M, Neylan TC. Eec ff ts of post ‑traumatic stress Funding for this project was provided by NSF 1631325, NIH R01 DA041353, disorder on occipital lobe function and structure. NeuroReport. and NIH U01 DA041156. 2012;23:412–9. https:// doi. org/ 10. 1097/ WNR. 0b013 e3283 52025e. 11. Chen S, Li L, Xu B, Liu J. Insular cortex involvement in declarative Availability of data and materials memory deficits in patients with post ‑traumatic stress disorder. BMC Data and materials are available in a GitHub repository (https:// github. com/ Psychiatry. 2009;9:39. https:// doi. org/ 10. 1186/ 1471‑ 244X‑9‑ 39. NBCLab/ meta‑ analy sis_ ptsd), including the meta‑analytic coordinate files, 12. Chen Y, Fu K, Feng C, Tang L, Zhang J, Huan Y, et al. Different regional data analysis scripts (i.e., code), image‑based results (i.e., ALE, rsFC, and MACM gray matter loss in recent onset PTSD and non PTSD after a single images), and functional decoding results. rsFC analyses used the Human Con‑ prolonged trauma exposure. PLoS ONE. 2012;7:e48298. https:// doi. nectome Project’s [90] Young Adult Study S1200 Data Release (March 1, 2017), org/ 10. 1371/ journ al. pone. 00482 98. which is available at db.humanconnectome.org. 13. Cheng B, Huang X, Li S, Hu X, Luo Y, Wang X, et al. Gray matter altera‑ tions in post‑traumatic stress disorder, obsessive ‑ compulsive disor‑ der, and social anxiety disorder. Front Behav Neurosci. 2015;9:219. Declarations https:// doi. org/ 10. 3389/ fnbeh. 2015. 00219. 14. Corbo V, Clément M‑H, Armony JL, Pruessner JC, Brunet A. Size versus Ethics approval and consent to participate shape differences: contrasting voxel‑based and volumetric analyses This secondary data analysis was approved by the Institutional Review Board of the anterior cingulate cortex in individuals with acute posttrau‑ of Florida International University. matic stress disorder. Biol Psychiatry. 2005;58:119–24. https:// doi. org/ 10. 1016/j. biops ych. 2005. 02. 032. Consent for publication 15. DiGangi JA, Tadayyon A, Fitzgerald DA, Rabinak CA, Kennedy A, Not applicable. Klumpp H, et al. Reduced default mode network connectivity follow‑ ing combat trauma. Neurosci Lett. 2016;615:37–43. https:// doi. org/ Competing interests 10. 1016/j. neulet. 2016. 01. 010. The authors declare no competing interests. 16. Dogan I, Eickhoff CR, Fox PT, Laird AR, Schulz JB, Eickhoff SB, Reetz K. Functional connectivity modeling of consistent cortico‑striatal Author details degeneration in HD. Neuroimage Clin. 2015;7:640–52. Department of Psychology, Florida International University, Miami, FL, USA. 17. Dosenbach NUF, Fair DA, Miezin FM, Cohen AL, Wenger KK, Dosen‑ Department of Physics, Florida International University, Miami, FL, USA. bach RAT, et al. Distinct brain networks for adaptive and stable task Department of Psychology, Old Dominion University, Norfolk, VA, USA. control in humans. Proc Natl Acad Sci. 2007;104:11073–8. https:// doi. org/ 10. 1073/ pnas. 07043 20104. Received: 8 April 2022 Accepted: 27 August 2022 18. Duval E, Liberzon I, Javanbakht A. Neural circuits in anxiety and stress disorders: a focused review. Ther Clin Risk Manag. 2015. https:// doi. org/ 10. 2147/ TCRM. S48528. 19. Eckart C, Stoppel C, Kaufmann J, Tempelmann C, Hinrichs H, Elbert T, et al. Structural alterations in lateral prefrontal, parietal and posterior References midline regions of men with chronic posttraumatic stress disorder. J 1. Agaibi CE, Wilson JP. Trauma, PTSD, and resilience: a review of the litera‑ Psychiatry Neurosci JPN. 2011;36:176–86. https:// doi. org/ 10. 1503/ jpn. ture. Trauma Violence Abuse. 2005;6:195–216. https:// doi. org/ 10. 1177/ 15248 38005 277438. 20. Ehlers A, Clark DM. A cognitive model of posttraumatic stress disor‑ 2. Akiki TJ, Averill CL, Abdallah CG. A Network‑based neurobiological der. Behav Res Ther. 2000;38:319–45. https:// doi. org/ 10. 1016/ S0005‑ model of PTSD: evidence from structural and functional neuroimag‑ 7967(99) 00123‑0. ing studies. Curr Psychiatry Rep. 2017;19:81. https:// doi. org/ 10. 1007/ s11920‑ 017‑ 0840‑4. Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 14 of 16 21. Eickhoff SB, Bzdok D, Laird AR, Kurth F, Fox PT. Activation likelihood esti‑ 40. Klaming R, Harlé KM, Infante MA, Bomyea J, Kim C, Spadoni AD. Shared mation meta‑analysis revisited. Neuroimage. 2012;59:2349–61. https:// gray matter reductions across alcohol use disorder and posttraumatic doi. org/ 10. 1016/j. neuro image. 2011. 09. 017. stress disorder in the anterior cingulate cortex: a dual meta‑analysis. 22. Eickhoff SB, Laird AR, Grefkes C, Wang LE, Zilles K, Fox PT. Coordinate ‑ Neurobiol Stress. 2019;10:100132. https:// doi. org/ 10. 1016/j. ynstr. 2018. based activation likelihood estimation meta‑analysis of neuroimaging 09. 009. data: a random‑ effects approach based on empirical estimates of 41. Koch SBJ, van Zuiden M, Nawijn L, Frijling JL, Veltman DJ, Olff M. spatial uncertainty. Hum Brain Mapp. 2009;30:2907–26. https:// doi. org/ Aberrant resting‑state brain activity in posttraumatic stress disorder: a 10. 1002/ hbm. 20718. meta‑analysis and systematic review: theoretical review—brain activity 23. Eickhoff SB, Nichols TE, Laird AR, Hoffstaedter F, Amunts K, Fox PT, et al. in PTSD during rest. Depress Anxiety. 2016;33:592–605. https:// doi. org/ Behavior, sensitivity, and power of activation likelihood estimation char‑10. 1002/ da. 22478. acterized by massive empirical simulation. Neuroimage. 2016;137:70– 42. Koenigs M, Grafman J. Posttraumatic stress disorder: the role of medial 85. https:// doi. org/ 10. 1016/j. neuro image. 2016. 04. 072. prefrontal cortex and amygdala. Neuroscientist. 2009;15:540–8. https:// 24. Etkin A, Wager TD. Functional neuroimaging of anxiety: a meta‑analysis doi. org/ 10. 1177/ 10738 58409 333072. of emotional processing in PTSD, social anxiety disorder, and specific 43. Kroes MCW, Rugg MD, Whalley MG, Brewin CR. Structural brain abnor‑ phobia. Am J Psychiatry. 2007;164:1476–88. malities common to posttraumatic stress disorder and depression. J 25. Feinberg DA, Moeller S, Smith SM, Auerbach E, Ramanna S, Glasser MF, Psychiatry Neurosci JPN. 2011;36:256–65. https:// doi. org/ 10. 1503/ jpn. et al. Multiplexed echo planar imaging for sub‑second whole brain 100077. FMRI and fast diffusion imaging. PLoS ONE. 2010;5:e15710. https:// doi. 44. Kühn S, Gallinat J. Gray matter correlates of posttraumatic stress disor‑ org/ 10. 1371/ journ al. pone. 00157 10. der: a quantitative meta‑analysis. Biol Psychiatry. 2013;73:70–4. https:// 26. Felmingham K, Williams LM, Whitford TJ, Falconer E, Kemp AH, Peduto doi. org/ 10. 1016/j. biops ych. 2012. 06. 029. A, et al. Duration of posttraumatic stress disorder predicts hippocampal 45. Laird AR, Robinson JL, McMillan KM, Tordesillas‑ Gutiérrez D, Moran ST, grey matter loss. NeuroReport. 2009;20:1402–6. https:// doi. org/ 10. Gonzales SM, et al. Comparison of the disparity between Talairach and 1097/ WNR. 0b013 e3283 300fbc. MNI coordinates in functional neuroimaging data: validation of the 27. Glasser MF, Smith SM, Marcus DS, Andersson JLR, Auerbach EJ, Behrens Lancaster transform. Neuroimage. 2010;51:677–83. https:// doi. org/ 10. TEJ, et al. The Human Connectome Project’s neuroimaging approach. 1016/j. neuro image. 2010. 02. 048. Nat Neurosci. 2016;19:1175–87. https:// doi. org/ 10. 1038/ nn. 4361. 46. Lancaster JL, Tordesillas‑ Gutiérrez D, Martinez M, Salinas F, Evans A, Zilles 28. Goodkind M, Eickhoff SB, Oathes DJ, Jiang Y, Chang A, Jones‑Hagata LB, K, et al. Bias between MNI and Talairach coordinates analyzed using the et al. Identification of a common neurobiological substrate for mental ICBM‑152 brain template. Hum Brain Mapp. 2007;28:1194–205. https:// illness. JAMA Psychiat. 2015;72:305. https:// doi. org/ 10. 1001/ jamap sychi doi. org/ 10. 1002/ hbm. 20345. atry. 2014. 2206. 47. Lanius RA, Frewen PA, Tursich M, Jetly R, McKinnon MC. Restoring 29. Gorgolewski K, Burns CD, Madison C, Clark D, Halchenko YO, Waskom large‑scale brain networks in PTSD and related disorders: a proposal for ML, et al. Nipype: a flexible, lightweight and extensible neuroimaging neuroscientifically‑informed treatment interventions. Eur J Psychotrau‑ data processing framework in python. Front Neuroinformatics. 2011. matology. 2015;6:27313. https:// doi. org/ 10. 3402/ ejpt. v6. 27313. https:// doi. org/ 10. 3389/ fninf. 2011. 00013. 48. Lei D, Li K, Li L, Chen F, Huang X, Lui S, et al. Disrupted functional brain 30. Greicius MD, Krasnow B, Reiss AL, Menon V. Functional connectivity in connectome in patients with posttraumatic stress disorder. Radiology. the resting brain: a network analysis of the default mode hypothesis. 2015;276:818–27. https:// doi. org/ 10. 1148/ radiol. 15141 700. Proc Natl Acad Sci. 2003;100:253–8. https:// doi. org/ 10. 1073/ pnas. 01350 49. Li L, Chen S, Liu J, Zhang J, He Z, Lin X. Magnetic resonance imaging 58100. and magnetic resonance spectroscopy study of deficits in hippocampal 31. Hakamata Y, Matsuoka Y, Inagaki M, Nagamine M, Hara E, Imoto S, et al. structure in fire victims with recent ‑ onset posttraumatic stress disorder. Structure of orbitofrontal cortex and its longitudinal course in cancer‑ Can J Psychiatry. 2006;51:431–7. https:// doi. org/ 10. 1177/ 07067 43706 related post‑traumatic stress disorder. Neurosci Res. 2007;59:383–9. 05100 704. https:// doi. org/ 10. 1016/j. neures. 2007. 08. 012. 50. Li L, Wu M, Liao Y, Ouyang L, Du M, Lei D, et al. Grey matter reduction 32. Hamner MB. Potential role of the anterior cingulate cortex in PTSD: associated with posttraumatic stress disorder and traumatic stress. review and hypothesis. Depress Anxiety. 1999;9:14. Neurosci Biobehav Rev. 2014;43:163–72. https:// doi. org/ 10. 1016/j. neubi 33. Hayes JP, Hayes SM, Mikedis AM. Quantitative meta‑analysis of neural orev. 2014. 04. 003. activity in posttraumatic stress disorder. Biol Mood Anxiety Disord. 51. Liberzon I, Abelson JL. Context processing and the neurobiology of 2012;2:9. post‑traumatic stress disorder. Neuron. 2016;92:14–30. https:// doi. org/ 34. Herringa R, Phillips M, Almeida J, Insana S, Germain A. Post‑traumatic 10. 1016/j. neuron. 2016. 09. 039. stress symptoms correlate with smaller subgenual cingulate, caudate, 52. Liberzon I, Garfinkel SN. Functional neuroimaging in post ‑traumatic and insula volumes in unmedicated combat veterans. Psychiatry Res. stress disorder. In: LeDoux JE, Keane T, Shiromani P, editors. Post‑ 2012;203:139–45. https:// doi. org/ 10. 1016/j. pscyc hresns. 2012. 02. 005. Traumatic stress disorder: basic science and clinical practice. Totowa NJ: 35. Holmes SE, Scheinost D, DellaGioia N, Davis MT, Matuskey D, Pietrzak Humana Press; 2009. p. 219–317 (10.1007/978‑1‑60327‑329‑9). RH, et al. Cerebellar and prefrontal cortical alterations in PTSD: structural 53. Liberzon I, Shulman GL. A default mode of brain function. Proc Natl and functional evidence. Chronic Stress. 2018;2:247054701878639. Acad Sci. 2001;98:676–82. https:// doi. org/ 10. 1073/ pnas. 98.2. 676. https:// doi. org/ 10. 1177/ 24705 47018 786390. 54. Liu Y, Li L, Li B, Feng N, Li L, Zhang X, et al. Decreased triple network 36. Jenkinson M, Beckmann CF, Behrens TEJ, Woolrich MW, Smith SM. FSL. connectivity in patients with recent onset post‑traumatic stress disor ‑ Neuroimage. 2012;62:782–90. https:// doi. org/ 10. 1016/j. neuro image. der after a single prolonged trauma exposure. Sci Rep. 2017;7:12625. 2011. 09. 015.https:// doi. org/ 10. 1038/ s41598‑ 017‑ 12964‑6. 37. Kamalian A, Khodadadifar T, Saberi A, Masoudi M, Camilleri JA, Eickhoff 55. Meng Y, Qiu C, Zhu H, Lama S, Lui S, Gong Q, et al. Anatomical deficits in CR, Zarei M, Pasquini L, Laird AR, Fox PT, Eickhoff SB, Tamasian M. Con‑ adult posttraumatic stress disorder: a meta‑analysis of voxel‑based mor ‑ vergent regional brain abnormalities in behavioral variant frontotempo‑ phometry studies. Behav Brain Res. 2014;270:307–15. https:// doi. org/ 10. ral dementia: a neuroimaging meta‑analysis of 73 studies. Alzheimer’s 1016/j. bbr. 2014. 05. 021. Dement: Diagn, Assess Dis Monit. 2022;14:e12318. 56. Menon V. Large‑scale brain networks and psychopathology: a unifying 38. Kasai K, Yamasue H, Gilbertson MW, Shenton ME, Rauch SL, Pitman RK. triple network model. Trends Cogn Sci. 2011;15:483–506. https:// doi. Evidence for acquired pregenual anterior cingulate gray matter loss org/ 10. 1016/j. tics. 2011. 08. 003. from a twin study of combat‑related posttraumatic stress disorder. Biol 57. Menon V, Uddin LQ. Saliency, switching, attention and control: a Psychiatry. 2008;63:550–6. https:// doi. org/ 10. 1016/j. biops ych. 2007. 06. network model of insula function. Brain Struct Funct. 2010;214:655–67. 022.https:// doi. org/ 10. 1007/ s00429‑ 010‑ 0262‑0. 39. Kirkpatrick HA, Heller GM. Post‑traumatic stress disorder: theory and 58. Nardo D, Högberg G, Looi JCL, Larsson S, Hällström T, Pagani M. Gray treatment update. Int J Psychiatry Med. 2014;47:337–46. https:// doi. matter density in limbic and paralimbic cortices is associated with org/ 10. 2190/ PM. 47.4.h. P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 15 of 16 trauma load and EMDR outcome in PTSD patients. J Psychiatr Res. 77. Salo T, Yarkoni T, Nichols TE, Poline J‑B, Bilgel M, Bottenhorn KL, 2010;44:477–85. https:// doi. org/ 10. 1016/j. jpsyc hires. 2009. 10. 014. Jarecka D, Kent JD, Kimbler A, Nielson DM, Oudyk KM, Peraza JA, 59. National Center for PTSD (2019). PTSD: National Center for PTSD. Com‑ Pérez A, Reeders PC, Yanes JA, Laird AR. NiMARE: neuroimaging meta‑ mon PTSD Adults. Available at: https:// www. ptsd. va. gov/ under stand/ analysis research environment. NeuroLibre. 2022;1:7. https:// doi. org/ common/ common_ adults. asp. Accessed 31 October 2020.10. 55458/ neuro libre. 00007. 60. Nicholson AA, Harricharan S, Densmore M, Neufeld RWJ, Ros T, McKin‑ 78. Seeley WW, Menon V, Schatzberg AF, Keller J, Glover GH, Kenna H, non MC, et al. Classifying heterogeneous presentations of PTSD via the et al. Dissociable intrinsic connectivity networks for salience process‑ default mode, central executive, and salience networks with machine ing and executive control. J Neurosci. 2007;27:2349–56. https:// doi. learning. NeuroImage Clin. 2020;27:102262. https:// doi. org/ 10. 1016/j. org/ 10. 1523/ JNEUR OSCI. 5587‑ 06. 2007. nicl. 2020. 102262. 79. Serra‑Blasco M, Radua J, Soriano ‑Mas C, Gómez‑Benlloch A, Porta‑ 61. O’Doherty DCM, Tickell A, Ryder W, Chan C, Hermens DF, Bennett MR, Casteràs D, Carulla‑Roig M, et al. Structural brain correlates in major et al. Frontal and subcortical grey matter reductions in PTSD. Psychiatry depression, anxiety disorders and post‑traumatic stress disorder: a Res Neuroimaging. 2017;266:1–9. https:// doi. org/ 10. 1016/j. pscyc hresns. voxel‑based morphometry meta‑analysis. Neurosci Biobehav Rev. 2017. 05. 008. 2021;129:269–81. https:// doi. org/ 10. 1016/j. neubi orev. 2021. 07. 002. 62. Pai A, Suris A, North C. Posttraumatic stress disorder in the DSM‑5: con‑ 80. Sha Z, Wager TD, Mechelli A, He Y. Common dysfunction of large‑ troversy, change, and conceptual considerations. Behav Sci. 2017;7:7. scale neurocognitive networks across psychiatric disorders. Biol https:// doi. org/ 10. 3390/ bs701 0007. Psychiatry. 2019;85:379–88. https:// doi. org/ 10. 1016/j. biops ych. 2018. 63. Patel R, Spreng RN, Shin LM, Girard TA. Neurocircuitry models of post‑11. 011. traumatic stress disorder and beyond: a meta‑analysis of functional 81. Shin LM, Rauch SL, Pitman RK. Amygdala, medial prefrontal cortex, neuroimaging studies. Neurosci Biobehav Rev. 2012;36:2130–42. and hippocampal function in PTSD. Ann NY Acad Sci. 2006;1071:67– https:// doi. org/ 10. 1016/j. neubi orev. 2012. 06. 003. 79. https:// doi. org/ 10. 1196/ annals. 1364. 007. 64. Pitman RK, Rasmusson AM, Koenen KC, Shin LM, Orr SP, Gilbertson 82. Smith SM, Beckmann CF, Andersson J, Auerbach EJ, Bijsterbosch MW, et al. Biological studies of post‑traumatic stress disorder. Nat Rev J, Douaud G, et al. Resting‑state fMRI in the Human Connectome Neurosci. 2012;13:769–87. https:// doi. org/ 10. 1038/ nrn33 39. Project. Neuroimage. 2013;80:144–68. https:// doi. org/ 10. 1016/j. neuro 65. PTSD Alliance (2018). Traumatic stress disorder fact sheet. Available at: image. 2013. 05. 039. http:// www. sidran. org/ wp‑ conte nt/ uploa ds/ 2018/ 11/ Post T‑raum atic‑ 83. Sripada RK, King AP, Welsh RC, Garfinkel SN, Wang X, Sripada CS, et al. Stress‑ Disor der‑ Fact‑ Sheet‑. pdf . Accessed 31 October 2020. Neural dysregulation in posttraumatic stress disorder: evidence for 66. Qi R, Luo Y, Zhang L, Weng Y, Surento W, Jahanshad N, et al. Social sup‑ disrupted equilibrium between salience and default mode brain port modulates the association between PTSD diagnosis and medial networks. Psychosom Med. 2012;74:904–11. https:// doi. org/ 10. 1097/ frontal volume in Chinese adults who lost their only child. Neurobiol PSY. 0b013 e3182 73bf33. Stress. 2020;13:100227. https:// doi. org/ 10. 1016/j. ynstr. 2020. 100227. 84. Sui SG, Wu MX, King ME, Zhang Y, Ling L, Xu JM, et al. Abnormal grey 67. Qin L. A preliminary study of alterations in default network connectivity matter in victims of rape with PTSD in Mainland China: a voxel‑based in post‑traumatic stress disorder patients following recent trauma. Brain morphometry study. Acta Neuropsychiatr. 2010;22:118–26. https:// Res. 2012. https:// doi. org/ 10. 1016/j. brain res. 2012. 09. 029.doi. org/ 10. 1111/j. 1601‑ 5215. 2010. 00459.x. 68. Rabellino D, Tursich M, Frewen PA, Daniels JK, Densmore M, Théberge J, 85. Tavanti M, Battaglini M, Borgogni F, Bossini L, Calossi S, Marino D, et al. Intrinsic Connectivity Networks in post‑traumatic stress disorder et al. Evidence of diffuse damage in frontal and occipital cortex in during sub‑ and supraliminal processing of threat ‑related stimuli. Acta the brain of patients with post‑traumatic stress disorder. Neurol Sci. Psychiatr Scand. 2015;132:365–78. https:// doi. org/ 10. 1111/ acps. 12418. 2012;33:59–68. https:// doi. org/ 10. 1007/ s10072‑ 011‑ 0659‑4. 69. Raichle ME. The brain’s default mode network. Annu Rev 86. Thome J, Frewen P, Daniels JK, Densmore M, Lanius RA. Altered con‑ Neurosci. 2015;38:433–47. https:// doi. org/ 10. 1146/ annur nectivity within the salience network during direct eye gaze in PTSD. ev‑ neuro‑ 071013‑ 014030. Borderline Personal Disord Emot Dysregul. 2014;1:17. https:// doi. org/ 70. Rauch SL, Shin LM, Phelps EA. Neurocircuitry models of posttraumatic 10. 1186/ 2051‑ 6673‑1‑ 17. stress disorder and extinction: human neuroimaging research—past, 87. Turner BM, Rodriguez CA, Liu Q, Molloy MF, Hoogendijk M, McClure present, and future. Biol Psychiatry. 2006;60:376–82. https:// doi. org/ 10. SM. On the neural and mechanistic bases of self‑ control. Cereb 1016/j. biops ych. 2006. 06. 004. Cortex. 2019;29:732–50. https:// doi. org/ 10. 1093/ cercor/ bhx355. 71. Reetz K, Dogan I, Rolfs A, Binkofski F, Schulz JB, Laird AR, Fox PT, Eickhoff 88. Tursich M, Ros T, Frewen PA, Kluetsch RC, Calhoun VD, Lanius RA. Dis‑ SB. Investigating function and connectivity of morphometric find‑ tinct intrinsic network connectivity patterns of post‑traumatic stress ings—Exemplified on cerebellar atrophy in spinocerebellar ataxia 17 disorder symptom clusters. Acta Psychiatr Scand. 2015;132:29–38. (SCA17). Neuroimage. 2012;62:1354–66.https:// doi. org/ 10. 1111/ acps. 12387. 72. Rocha‑Rego V, Pereira MG, Oliveira L, Mendlowicz MV, Fiszman A, 89. Uğurbil K, Xu J, Auerbach EJ, Moeller S, Vu AT, Duarte‑ Carvajalino Marques‑Portella C, et al. Decreased premotor cortex volume in JM, et al. Pushing spatial and temporal resolution for functional victims of urban violence with posttraumatic stress disorder. PLoS ONE. and diffusion MRI in the Human Connectome Project. Neuroimage. 2012;7:e42560. https:// doi. org/ 10. 1371/ journ al. pone. 00425 60. 2013;80:80–104. https:// doi. org/ 10. 1016/j. neuro image. 2013. 05. 012. 73. Ross MC, Cisler JM. Altered large‑scale functional brain organization in 90. Van Essen DC, Smith SM, Barch DM, Behrens TEJ, Yacoub E, Ugurbil K. posttraumatic stress disorder: a comprehensive review of univariate The WU‑Minn Human Connectome Project: an overview. Neuroim‑ and network‑level neurocircuitry models of PTSD. NeuroImage Clin. age. 2013;80:62–79. https:// doi. org/ 10. 1016/j. neuro image. 2013. 05. 2020;27:102319. https:// doi. org/ 10. 1016/j. nicl. 2020. 102319. 041. 74. Rubin TN, Koyejo O, Gorgolewski KJ, Jones MN, Poldrack RA, 91. Van Essen DC, Ugurbil K, Auerbach E, Barch D, Behrens TEJ, Bucholz R, Yarkoni T. Decoding brain activity using a large‑scale probabilistic et al. The Human Connectome Project: a data acquisition perspec‑ functional‑anatomical atlas of human cognition. PLOS Comput Biol. tive. Neuroimage. 2012;62:2222–31. https:// doi. org/ 10. 1016/j. neuro 2017;13:e1005649. https:// doi. org/ 10. 1371/ journ al. pcbi. 10056 49.image. 2012. 02. 018. 75. Salimi‑Khorshidi G, Douaud G, Beckmann CF, Glasser MF, Griffanti L, 92. Wang T, Liu J, Zhang J, Zhan W, Li L, Wu M, Huang H, Zhu H, Kemp GJ, Smith SM. Automatic denoising of functional MRI data: combining Gong Q. Altered resting‑state functional activity in posttraumatic stress independent component analysis and hierarchical fusion of classifiers. disorder: a quantitative meta‑analysis. Sci Rep. 2016;6:27131. Neuroimage. 2014;90:449–68. https:// doi. org/ 10. 1016/j. neuro image. 93. Whitfield‑ Gabrieli S, Ford JM. Default mode network activity and con‑ 2013. 11. 046. nectivity in psychopathology. Annu Rev Clin Psychol. 2012;8:49–76. 76. Salimi‑Khorshidi G, Smith SM, Keltner JR, Wager TD, Nichols TE. Meta‑https:// doi. org/ 10. 1146/ annur ev‑ clinp sy‑ 032511‑ 143049. analysis of neuroimaging data: a comparison of image‑based and 94. Winkler AM, Ridgway GR, Webster MA, Smith SM, Nichols TE. Permuta‑ coordinate‑based pooling of studies. Neuroimage. 2009;45:810–23. tion inference for the general linear model. Neuroimage. 2014;92:381– https:// doi. org/ 10. 1016/j. neuro image. 2008. 12. 039. 97. https:// doi. org/ 10. 1016/j. neuro image. 2014. 01. 060. Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 16 of 16 95. Woo C‑ W, Krishnan A, Wager TD. Cluster‑ extent based threshold‑ ing in fMRI analyses: pitfalls and recommendations. Neuroimage. 2014;91:412–9. https:// doi. org/ 10. 1016/j. neuro image. 2013. 12. 058. 96. Worsley KJ. Statistical analysis of activation images. In: Jezzard P, Mat‑ thews PM, Smith SM, editors. Functional magnetic resonance imaging. Oxford: Oxford University Press; 2001. p. 251–70 (10.1093/acprof: oso/9780192630711.003.0014). 97. Yamasue H, Kasai K, Iwanami A, Ohtani T, Yamada H, Abe O, et al. Voxel‑ based analysis of MRI reveals anterior cingulate gray‑matter volume reduction in posttraumatic stress disorder due to terrorism. Proc Natl Acad Sci USA. 2003;100:9039–43. https:// doi. org/ 10. 1073/ pnas. 15304 98. Yarkoni T, Poldrack RA, Nichols TE, Van Essen DC, Wager TD. Large‑scale automated synthesis of human functional neuroimaging data. Nat Methods. 2011;8:665–70. https:// doi. org/ 10. 1038/ nmeth. 1635. 99. Yehuda R, Hoge CW, McFarlane AC, Vermetten E, Lanius RA, Niev‑ ergelt CM, et al. Post‑traumatic stress disorder. Nat Rev Dis Primer. 2015;1:15057. https:// doi. org/ 10. 1038/ nrdp. 2015. 57. 100. Zhang H, Chen X, Chen S, Li Y, Chen C, Long Q, et al. Facial expression enhances emotion perception compared to vocal prosody: behavioral and fMRI studies. Neurosci Bull. 2018;34:801–15. https:// doi. org/ 10. 1007/ s12264‑ 018‑ 0231‑9. 101. Zhang J, Tan Q, Yin H, Zhang X, Huan Y, Tang L, et al. Decreased gray matter volume in the left hippocampus and bilateral calcarine cortex in coal mine flood disaster survivors with recent onset PTSD. Psychiatry Res. 2011;192:84–90. https:// doi. org/ 10. 1016/j. pscyc hresns. 2010. 09. 001. 102. Zhang Y, Liu F, Chen H, Li M, Duan X, Xie B, et al. Intranetwork and internetwork functional connectivity alterations in post‑traumatic stress disorder. J Aec ff t Disord. 2015;187:114–21. https:// doi. org/ 10. 1016/j. jad. 2015. 08. 043. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Behavioral and Brain Functions Springer Journals

Extended functional connectivity of convergent structural alterations among individuals with PTSD: a neuroimaging meta-analysis

Loading next page...
 
/lp/springer-journals/extended-functional-connectivity-of-convergent-structural-alterations-SSbfTPKnw4
Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2022
eISSN
1744-9081
DOI
10.1186/s12993-022-00196-2
Publisher site
See Article on Publisher Site

Abstract

Background: Post‑traumatic stress disorder (PTSD) is a debilitating disorder defined by the onset of intrusive, avoid‑ ant, negative cognitive or affective, and/or hyperarousal symptoms after witnessing or experiencing a traumatic event. Previous voxel‑based morphometry studies have provided insight into structural brain alterations associated with PTSD with notable heterogeneity across these studies. Furthermore, how structural alterations may be associated with brain function, as measured by task‑free and task ‑based functional connectivity, remains to be elucidated. Methods: Using emergent meta‑analytic techniques, we sought to first identify a consensus of structural alterations in PTSD using the anatomical likelihood estimation (ALE) approach. Next, we generated functional profiles of identi‑ fied convergent structural regions utilizing resting‑state functional connectivity (rsFC) and meta‑analytic co ‑activation modeling (MACM) methods. Finally, we performed functional decoding to examine mental functions associated with our ALE, rsFC, and MACM brain characterizations. Results: We observed convergent structural alterations in a single region located in the medial prefrontal cortex. The resultant rsFC and MACM maps identified functional connectivity across a widespread, whole ‑brain network that included frontoparietal and limbic regions. Functional decoding revealed overlapping associations with attention, memory, and emotion processes. Conclusions: Consensus‑based functional connectivity was observed in regions of the default mode, salience, and central executive networks, which play a role in the tripartite model of psychopathology. Taken together, these find‑ ings have important implications for understanding the neurobiological mechanisms associated with PTSD. Keywords: Post‑traumatic stress disorder, Meta‑analysis, Voxel‑based morphometry, Functional connectivity violence, accidents, or combat [99]. Symptoms associ- Background ated with PTSD are categorized into clusters accord- Post-traumatic stress disorder (PTSD) is a psychiatric ing to the DSM 5: (1) intrusion/re-experiencing trauma, disorder in which the onset of symptoms develops after (2) avoidance, (3) negative cognition and mood, and (4) experiencing or witnessing a traumatic event, such as hyperarousal [39, 62]. Approximately 70% of adults expe- rience at least one traumatic event in their lifetime and *Correspondence: bpank001@fiu.edu up to 20% of these people develop PTSD [65]. Individuals Department of Psychology, Florida International University, Miami, FL, USA with PTSD may experience long-term debilitating effects, Full list of author information is available at the end of the article © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 2 of 16 mentally, physically, and cognitively. In the United States, likely reflects the disturbance of distributed, brain-wide roughly 8 million adults suffer from PTSD every year. neural circuitry, we also sought to functionally and Approximately 60% of men experience at least one trau- behaviorally characterize any neuroanatomical altera- matic event in their lives, often associated with combat tions in a task-independent manner. To this end, we first and war, while 50% of women will experience at least one identified convergent regions of gray matter (GM) reduc - traumatic event, typically associated with sexual assault tions in PTSD vs. non-PTSD groups using anatomical and abuse [59]. likelihood estimation (ALE) [21, 22]. Second, we identi- Current theories aim to understand the etiology of fied the task-free resting state functional connectivity PTSD, including behavioral, cognitive, and social mod- (rsFC) patterns, as well as the task-based meta-analytic els. Research suggests that reappraisal of traumatic co-activation modeling (MACM) patterns of conver- events may lead to an overgeneralized threat response gent regions, thus providing multimodal functional con- [20]. Despite progress in understanding the vulnerabil- nectivity profiles for each. Together, the VBM, rsFC, ity, symptomatology, and trajectory of PTSD [1, 39, 64], and MACM meta-analytic approaches have been used the underlying neurobiological determinants of PTSD are in previous clinically related meta-analyses [16, 37, 71], less clear. Substantial prior work has attempted to iden- they provide complementary information, yielding a tify structural brain alterations observed among indi- multimodal functional connectivity profile for a given viduals with PTSD. Voxel-based morphometry (VBM) region of interest. Lastly, we applied meta-analytic func- is a commonly used methodological approach for ana- tional decoding methods to identify the mental processes lyzing structural magnetic resonance imaging (MRI) linked to this functional connectivity profile. Collectively, data, allowing for quantitative statistical comparisons this work utilizes an innovative (meta-) analytic frame- between groups (e.g., differences in gray matter volume; work to quantitatively assess structural alterations associ- GMV) to more clearly understand the structural altera- ated with PTSD and the extended functional profiles of tions associated with neuropsychiatric disorders, such regions implicated in this disorder. A more comprehen- as PTSD. Multiple prior meta-analyses have been con- sive understanding of the neurobiological bases of PTSD ducted to identify convergent gray matter reductions in is needed to delineate future pathways toward improved PTSD patients, although consensus across meta-analyses prevention, diagnosis, and treatment. has not been reached. Each of these meta-analyses was conducted with a different scope, with varied study inclu - Methods sion/exclusion criteria, and subsequently included a wide Analytic overview range of 8 to 20 studies. Varying convergence has been We first conducted a literature search to identify studies observed across these meta-analyses, which have identi- reporting structural alterations comparing the following fied one to five significant clusters in regions that include groups: individuals with PTSD, individuals who experi- medial prefrontal cortex [7, 40, 44, 50, 55], hippocam- enced trauma but were not diagnosed with PTSD, and pus [7, 44], fusiform gyrus [50, 79], and lingual gyrus individuals who did not report experiencing trauma. A [44, 79]. Similarly, from a functional perspective, PTSD coordinate-based meta-analysis was performed using dysfunction has been reported as amygdala and frontal the ALE algorithm to identify convergent brain regions disruptions (e.g., [18] or across alterations of large-scale showing structural alterations associated with PTSD. We functional brain networks (e.g., [41] that are implicated then used multiple connectivity modeling approaches to in the tripartite model of psychopathology [56]. While comprehensively characterize the functional connectiv- some studies have addressed consensus across functional ity of these convergent regions. Specifically, rsFC and neuroimaging studies, it is challenging to assess conver- MACM assessments were applied to identify the func- gence across different psychological states and/or experi - tional profiles of structurally altered regions associated mental paradigms, which has potentially contributed to with PTSD. Lastly, we used functional decoding tech- inconsistent findings in PTSD meta-analyses of resting niques to identify behavioral profiles of the ALE, rsFC, state [3, 92] or task-based [24, 63,  33] studies. Over- and MACM results. An overview of our methodological all, this variability across meta-analytic approaches and approach is provided in Figure 1. results suggests that a consensus neurobiological model of PTSD has not yet been achieved. Literature search and study criteria The objective of the current study was to apply current We conducted a comprehensive literature search to best practices in coordinate-based neuroimaging meth- build a database of peer-reviewed MRI studies reporting ods to investigate the topography of consistently reported structural alterations associated with PTSD from 2002 to structural alterations in PTSD. As PTSD is linked to a 2020. In the first round of identifying studies, we exam - broad spectrum of neuropsychiatric symptoms, which ined previously published voxel-based morphometry P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 3 of 16 Fig. 1 Analysis Pipeline Overview. A We first conducted a literature search to extract structural coordinates and entered them into the ALE algorithm to identify convergent structural alterations among PTSD vs. non‑PTSD groups. B We next created task ‑free and task ‑based functional connectivity profiles for the convergent structural alterations. C Last, we performed functional decoding analyses on these functional profiles to make inferences about which mental functions were associated with our findings meta-analysis papers on PTSD and compiled a list of distributions to address variability within and between included studies [7, 40, 44, 50, 55]. Next, we performed studies. We used the coordinate-based ALE method as a PubMed search to identify additional peer-reviewed, implemented in NiMARE v.0.0.3 (Neuroimaging Meta- structural MRI studies of interest using the search terms Analysis Research Environment; [77], a Python library for “morphometry + PTSD”. The PubMed search aimed to neuroimaging meta-analysis. Reported coordinates were identify any potential studies that were not included in extracted from their original publication,coordinates the previously published meta-analyses. We then con- originally reported in Talairach space converted to were ducted a review of each identified publication to include MNI coordinates [45, 46] so that all coordinates referred the following study criteria: peer-reviewed MRI studies, to MNI space. Once transformed, statistical probability reporting results among adult humans, written in the maps were created for each foci and combined to model English language, focused on gray matter structural dif- the likelihood that a given voxel displayed a between- ferences, and included original data (i.e., not a review). group structural difference for each study. Observed Subsequently, exclusion criteria were as follows: trauma voxel-wise ALE scores characterized the most consist- or stressful life event studies not measuring PTSD, other ently reported foci across the whole brain. Significance non-voxel-based morphometry methods, treatment and testing and correction for multiple comparisons involved longitudinal effects, papers reporting a priori regions of thresholding the voxel-wise ALE map using a cluster- interest (ROIs), within-group effects, null effects, over - forming threshold of P < 0.001. Then, a permutation lapping samples to previous studies, and studies that did procedure was performed in which a null distribution of not report coordinate-based results. maximum cluster sizes was generated from 10,000 itera- tions of replacing reported foci with randomly selected Anatomical likelihood estimation (ALE) gray matter voxels, generating ALE maps from the ran- ALE is a voxel-based meta-analytic technique that iden- domized dataset, and identifying the maximum clus- tifies convergent coordinates (i.e., foci) across a set of ter size after thresholding at P < 0.001. The cluster-level neuroimaging studies. Foci are treated as 3D Gaussian FWE correction threshold was set at P < 0.05, meaning Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 4 of 16 only those clusters from the original, thresholded ALE workflow is described by Glasser and colleagues [27], map were retained if their size was greater than the clus- but consists of typical imaging pre-processing tech- ter size corresponding to the 95th-percentile from the niques that leverage the high-quality data acquired null distribution. We applied the above ALE procedure by the HCP. First, T1- and T2-weighted images were to identify convergent brain regions reflecting structural aligned, bias field corrected, and registered to MNI alterations between individuals with and without PTSD space. Second, the functional fMRI pipeline removed (i.e., PTSD vs. non-PTSD) separately for the contrasts of spatial distortions, realigned volumes to compensate PTSD > non-PTSD and non-PTSD > PTSD. for subject motion, registered the fMRI data to struc- tural volumes (in MNI space), reduced the bias field, normalized each functional acquisition to its corre- sponding global mean, and masked non-brain tissue. Functional profiles of structurally altered regions Noteworthily, care was taken to minimize smooth- associated with PTSD ing induced by interpolation and that no overt volume Next, we sought to characterize the functional connectiv- smoothing was performed. ity patterns associated with regions demonstrating struc- The fMRI signal contains many sources of variabil - tural alterations in PTSD. To this end, we investigated ity, including artifactual and non-neuronal signals, that task-free functional connectivity utilizing a database of make identifying the underlying neuronal activity dif- resting state fMRI data, as well as task-based functional ficult. Using a combination of independent compo - connectivity using a meta-analytic database of co-activa- nent analysis (ICA) and classification techniques, HCP tion results. functional data were automatically denoised using FMRIB’s ICA-based X-noiseifier [75]. Briefly, ICA was performed on each functional dataset independently Task‑free functional connectivity: resting‑state fMRI and characteristics of each component, such as spatial (rs‑fMRI) localization and power in high frequencies, were evalu- Resting-state connectivity analyses typically identify ated by a classifier to determine if a given component brain voxels demonstrating the highest temporal cor- was related to neuronal activity or artifact. The time- relation with the average time series of a seed ROI and series corresponding to artifactual components were provide context about the brain’s underlying functional then regressed out of the data, providing a “cleaned”, architecture. To derive robust rsFC maps for each ROI, denoised dataset for further investigation. we utilized the minimally pre-processed and denoised Using the minimally pre-processed, denoised rest- (or “cleaned”) resting-state fMRI data provided by the ing-state datasets for each participant, the “global sig- Human Connectome Project’s [90] Young Adult Study nal” was removed using FSL’s fsl_glm [36] interface in S1200 Data Release (March 1, 2017). On November 12, NiPype [29]. The “global signal”, although controversial 2019, 150 randomly selected participants (28.7 ± 3.9 in the domain of resting-state analyses, was removed years) were downloaded via the HCP’s Amazon Web under the premise that it performed better than other Services (AWS) Simple Storage Solution (S3) repository. commonly used motion-correction strategies at remov- The randomly chosen participants included 77 females ing motion-related artifacts in the HCP resting-state (30.3 ± 3.5 years) and 73 males (27.1 ± 3.7 years). A dif- data [8]. The resulting data set was then smoothed with ference in age between the two biological sex groups was a FWHM kernel of 6-mm using FSL’s susaan interface significant but is consistent with the 1200 Subjects Data in NyPipe. For each participant, the average time series Release. Detailed acquisition and scanning parameters for each ROI was extracted and a whole-brain correla- for HCP data can be found in consortium manuscripts tion map was calculated and averaged across runs for [82, 89, 91], but relevant scan parameters are briefly sum - a single participant for every ROI. The average corre - marized here. Each participant underwent T1-weighted lation maps for each participant were transformed to and T2-weighted structural acquisitions and four rest- Z-scores using Fisher’s r-to-z transformation. A group- ing-state fMRI acquisitions. Structural images were col- level analysis was then performed to derive a rsFC map lected at 0.7-mm isotropic resolution. Whole-brain EPI for each ROI using FSL’s randomise interface [94] in acquisitions were acquired on the 3T Siemens Connec- NiPype. Images were thresholded non-parametrically tome scanner: 32-channel head coil, TR = 720 msec, using GRF-theory-based maximum height thresholding TE = 33.1 msec, in-plane FOV = 208 × 180 mm, 72 slices, with a (voxel FWE-corrected) significance threshold of 2.0 mm isotropic voxels, and multiband acceleration fac- P < 0.001 [96], such that more spatially specific connec - tor of 8 [25]. tivity maps could be derived when using such a highly The S1200 data release contained minimally pre-pro - powered study [95]. cessed and denoised data. The minimal pre-processing P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 5 of 16 Task‑based functional connectivity: meta‑analytic MACM, and rsFC analyses. To do so, we utilized gener- co‑activation modeling (MACM) alized correspondence latent Dirichlet allocation (GC- Leveraging reported coordinates from task-based fMRI LDA) functional decoding methods in NiMARE applied studies, meta-analytic co-activation is a relatively new to the resulting unthresholded ALE, rsFC, and MACM concept that identifies brain locations that are most likely maps. This type of decoding provides an approach to to be co-activated with a given seed ROI across multiple infer mental processes associated with neuroimaging task states. Differing from rsFC, MACM provides context spatial patterns. GC-LDA utilizes probabilistic Bayesian about neural recruitment during goal-oriented behaviors. statistics that learns latent topics from a large database of We therefore aimed to integrate these two complemen- papers (e.g., NeuroSynth) [74]. From the database, each tary modalities by supplementing the rsFC maps with topic found is treated as a probability distribution and MACM maps for each ROI. To do so, we relied on the creates a spatial distribution in MNI space across voxels Neurosynth database [98], which archives published ste- from the maps entered into the decoding algorithm. The reotactic coordinates from over 14,000 fMRI studies and “topics” encompass terms and associated brain regions 150,000 brain locations. Neurosynth relies on an auto- that co-occur in the literature from a literature database. mated coordinate extraction tool to “scrape” each avail- We set our model to 200 topics. We report 10 terms cor- able fMRI study for reported coordinates. Due to the responding to the highest weights associated with our nature of this automated process, fMRI studies reporting ALE, rsFC, and MACM results. results of multiple experimental contrasts as separate sets of coordinates are amalgamated into a single set of coor- Results dinates; in addition, “activation” and “de-activation” coor- Literature search and study criteria dinates are not distinctly characterized. However, while The literature search yielded a total of 85 articles using this inherent “noise” may limit interpretational abilities, the above-described search terms. Figure  2 provides a the power over manually curated datasets outweighs the PRISMA diagram, which details the review and filtering potential confounds of bi-directional or mixed-contrast of those 85 studies. In the first round of review, records effects. (i.e., titles and abstracts) were screened to exclude 18 To generate a MACM map for each ROI, we utilized studies that corresponded to non-human or non-English NiMARE [77] to search the Neurosynth database for all studies, reviews, or studies reporting white matter dif- studies reporting at least one peak within the defined ferences or differences among children or adolescents. ROI mask. Neurosynth tools implement the multilevel Then, we examined the full-text articles to assess addi - kernel density analysis (MKDA) algorithm for perform- tional study criteria; 44 additional studies were excluded ing meta-analyses based on a subset of studies, such as as being not eligible for the current meta-analysis. that described here. However, we opted to use the ALE The final set of included studies consisted of 23 algorithm as implemented in NiMARE given its optimal publications. Within these publications, gray mat- performance in replicating image-based meta-and mega- ter structural alterations were assessed by compar- analyses [76]. The ALE algorithm requires sample size ing whole-brain VBM results among individuals with information, or the number of subjects, that contributed and without PTSD, reported as 3D coordinates in to a given experimental contrast to generate a smooth- MNI or Talairach space. Control comparison groups ing kernel. However, Neurosynth is not able to capture included individuals who had experienced trauma but sample size (which could also vary across experimental did not develop PTSD and individuals who had not contrasts within a study). u Th s, we utilized a smoothing experienced trauma. Nineteen publications included kernel with a FWHM of 15  mm, which has been shown trauma-exposed controls (TC), while ten publications to yield results with strong correspondence for image- included healthy, non-trauma-exposed controls (HC). based meta- and mega-analyses [76]. The ALE algorithm Altogether, this set of 23 studies collectively examined was applied to the set of studies reporting activation 476 individuals with PTSD and 892 individuals with- within the boundaries of each ROI. Once ALE maps were out PTSD, which included 288 TC and 633 HC. With generated for each ROI, as described above, voxel-FWE respect to the type of structural alterations observed, correction (P < 0.001) was performed to reflect the statis - studies reported multiple different VBM metrics. Sev - tical thresholding approach used for rsFC maps. enteen publications reported group differences in gray matter volume (GMV), seven publications reported Functional decoding: generalized correspondence latent differences in gray matter density (GMD), and one dirichlet allocation (GC‑LDA) reported gray matter concentration (GMC). Collec- We sought to infer what mental processes were most tively, we refer to all of these metrics as gray matter likely linked with brain regions identified in our ALE, (GM) differences among individuals with and without Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 6 of 16 Fig. 2 PRISMA Diagram. PRISMA flow chart detailing the literature search and selection criteria of studies included in the meta‑analysis PTSD. Additional details on the demography of par- non-PTSD for a total of 20 foci, including 3 for PTSD ticipant groups and study design are provided in Addi- (9 foci) vs. TC and 2 contrasts for PTSD vs. HC (9 foci). tional file  1: Table  S1 located in this project’s GitHub repository (https:// github. com/ NBCLab/ meta- analy Anatomical likelihood estimation (ALE) sis_ ptsd). Using NiMARE v.0.0.3 [77], ALE meta-analysis was Within this final set of 23 publications, multiple con - performed to assess convergence for the 25 contrasts trasts of interest were reported. 25 contrasts reported from 22 publications of GM decreases among individu- GM decreases in PTSD vs non-PTSD for a total of 159 als with and without PTSD (i.e., non-PTSD > PTSD); a foci; this included 16 contrasts for PTSD vs. TC (82 complete listing is provided in Table  1. Neuroimaging foci) and 9 contrasts for PTSD vs. HC (77 foci). Con- simulations indicate that a minimum of 20 contrasts versely, 6 contrasts reported GM increases in PTSD vs. are necessary for a well-powered coordinate-based P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 7 of 16 Table 1 Studies Included in ALE Meta‑Analysis Citation Sample size Contrasts 1 [5] Total N = 38; PTSD n = 19 Healthy controls > PTSD 2 [10] Total N = 41; PTSD n = 21 Non‑PTSD > PTSD 3 [11] Total N = 24; PTSD n = 12 Controls > PTSD 4 [12] Total N = 20; PTSD n = 10 Controls > recent onset PTSD 5 [13] Total N = 60; PTSD n = 30 Healthy controls > PTSD 6 [14] Total N = 28; PTSD n = 14 Healthy controls > PTSD 7 [19] Total N = 33; PTSD n = 20 Non‑ Trauma controls > PTSD 8 [26] Total N = 38; PTSD n = 21 Controls > PTSD 9 [31] Total N = 184; PTSD n = 14 Non‑PTSD > PTSD; trauma exposed > PTSD 10 [34] Total N = 28; PTSD n = 13 Trauma exposed > PTSD 11 [38] Total N = 41; PTSD n = 18 Combat‑ exposed Non‑PTSD > PTSD 12 [43] Total N = 53; PTSD n = 24 Controls > PTSD 13 [49] Total N = 24; PTSD n = 12 Controls > PTSD 14 [58] Total N = 43; PTSD n = 21 Non‑PTSD > PTSD 15 [61] Total N = 75; PTSD n = 25 Healthy controls > PTSD; trauma exposed > PTSD 16 [66] Total N = 220; PTSD n = 57 Trauma exposed > PTSD 17 [72] Total N = 32; PTSD n = 16 Trauma exposed controls > PTSD 18 [84] Total N = 31; PTSD n = 11 Healthy controls > PTSD; trauma exposed > PTSD 19 [85] Total N = 50; PTSD n = 25 Healthy controls > PTSD 20 [97] Total N = 25; PTSD n = 9 Non‑PTSD > PTSD 21 [101] Total N = 20; PTSD n = 10 Trauma‑ exposed > PTSD 22 [100] Total N = 39; PTSD n = 14 Non‑PTSD > PTSD 25 contrasts from 22 publications reported GM decreases among individuals with and without PTSD (i.e., non-PTSD > PTSD). Sample sizes are provided for the total number of participants (N) (i.e., PTSD and non-PTSD), as well as the sample sizes for the PTSD groups (n) meta-analysis [23]. Thus, we were unable to assess the vs. TC and PTSD vs. HC contrasts (i.e., GM increases 6 contrasts of GM increases (i.e., PTSD > non-PTSD) and decreases) to determine if the use of different given insufficient power. With respect to GM decreases, comparison groups potentially contributed additional we observed a single cluster of convergence located heterogeneity, limiting assessment of convergence. in the mPFC (x=0, y=46, z=10; BA 32) (Figure  3; P < However, we observed null results for these additional 0.001, FWE-corrected P < 0.05). Given these results, we contrasts as well, likely in part due to the underpow- performed additional ALE meta-analyses for the PTSD ered samples [23]. Fig. 3 ALE Results for non‑PTSD > PTSD. Sagittal brain slices illustrating convergent structural alterations associated with PTSD as determined by an ALE meta‑analysis of GM reductions (P < 0.001, FWE‑ corrected P < 0.05) Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 8 of 16 Functional profiles of structurally altered regions parahippocampus. Next, to further examine function- associated with PTSD ally coupled regions with the mPFC seed, we generated We next investigated the functional connectivity of a MACM map using the Neurosynth database which the mPFC cluster identified above showing convergent demonstrated task-based coactivations with a simi- gray matter reductions among individuals with PTSD. lar pattern as the rsFC map. The locations of rsFC and To this end, we analyzed task-free rsFC and task-based MACM results are provided in Table  2. Figure  4 illus- MACM. First, we generated a rsFC map using the ALE- trates the rsFC (blue) and MACM (red) results, with derived mPFC cluster as a seed region. The resultant overlapping regions, indicating a consensus between rsFC map revealed rsFC with the superior frontal gyrus, rsFC and MACM (pink), revealed in the ACC, medial medial frontal gyrus, inferior frontal gyrus, ACC, thal- prefrontal gyrus, middle temporal gyrus, insula, infe- amus, posterior cingulate (PCC), superior temporal rior parietal lobe, thalamus, precuneus, parahippocam- gyrus, medial temporal gyrus, precuneus, cuneus, and pus, insula, and PCC regions (Table3). Table 2 rsFC and MACM Results rsFC results MACM results Anatomical label x y z Anatomical label x y z Anterior cingulate, BA 32 4 44 10 Medial frontal gyrus, BA 10 − 2 50 6 L Inferior frontal gyrus, BA 47 − 30 14 − 16 Superior frontal gyrus, BA 6 0 14 48 Cingulate gyrus, BA 24 2 − 18 36 Medial frontal gyrus, BA 8 2 26 38 Anterior cingulate, BA 32 0 36 − 6 Posterior cingulate, BA 31 − 4 − 54 26 Posterior cingulate, BA 31 8 − 52 24 L extra‑nuclear, BA 47 − 34 20 − 2 Cingulate gyrus, BA 31 − 8 − 54 26 R extra‑nuclear, BA 47 36 22 − 2 Midbrain 0 − 20 − 20 L angular gyrus, BA 39 − 46 − 68 30 Anterior cingulate, BA 24 4 28 16 L superior parietal lobule, BA 7 − 30 − 62 46 R Inferior frontal gyrus, BA 47 30 16 − 16 L inferior frontal gyrus, BA 9 − 46 10 28 Precuneus, BA 7 0 − 70 34 R superior temporal gyrus, BA 39 52 − 60 26 L caudate − 4 12 − 2 R inferior parietal lobule, BA 40 40 − 52 44 R angular gyrus, BA 39 52 − 64 36 L amygdala − 22 − 8 − 16 L inferior parietal lobule, BA 39 − 50 − 64 40 R amygdala 24 − 6 − 16 Posterior cingulate, BA 30 − 6 − 54 10 R inferior frontal gyrus, BA 9 46 10 28 L parahippocampal gyrus, BA 35 − 22 − 22 − 14 R caudate 12 10 2 L superior frontal gyrus, BA 8 − 22 34 46 L lentiform nucleus − 12 8 − 2 Cingulate gyrus, BA 31 − 4 − 32 38 L thalamus, medial dorsal nucleus − 6 − 14 6 R caudate 10 18 − 4 R thalamus, medial dorsal nucleus 6 − 14 6 L superior frontal gyrus, BA 9 − 20 48 34 L inferior parietal lobule, BA 40 − 42 − 44 44 Cerebellar tonsil 6 − 50 − 36 L inferior temporal gyrus, BA 21 − 56 − 10 − 16 Coordinate locations of the rsFC and MACM results, including the anatomical label and MNI coordinates of local maxima. Negative x values indicate the left (L) hemisphere and positive x values indicate the right (R) hemisphere Fig. 4 rsFC and MACM Results. rsFC (blue) and MACM (red) results; common areas (pink) indicate consensus between connectivity approaches. Images are thresholded at voxel‑ wise FWE P < 0.001 P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 9 of 16 Table 3 Consensus between rsFC and MACM Results separately for the structural ALE, rsFC, and MACM maps. The decoding terms with the top 10 weights from rsFC + MACM consensus the GC-LDA analysis for the structural ALE map were: Anatomical label x y z visual, emotional, memory, novel, reward, motor, self, faces, learning, and face (Table  4a). The decoding terms Medial frontal gyrus, BA 10 − 2 50 6 with the top 10 weights from the GC-LDA analysis for the Medial frontal gyrus, BA 8 2 26 38 rsFC map were: default, default mode network, intrinsic, Posterior cingulate, BA 31 − 4 − 54 26 scale, self, person, reward, bias, judgements, and contexts L angular gyrus, BA 39 − 46 − 68 30 (Table  4b). Topographically speaking, the rsFC results R superior temporal gyrus, BA 39 52 − 60 26 resembled regions of combined default mode [30, 69] and L inferior frontal gyrus, BA 47 − 32 18 − 6 salience networks [57, 78], and the functional decoding R inferior frontal gyrus, BA 47 38 20 − 8 outcomes suggested that the rsFC results were associ- L parahippocampal gyrus, BA 28 − 24 − 16 − 18 ated with self-referential, intrinsic, and reward processes. L lentiform nucleus, putamen − 12 10 − 4 Next, we examined MACM-based decoding results. The R caudate 10 10 − 2 decoding terms with the top 10 weights from the GC- L thalamus, medial dorsal nucleus − 4 − 14 6 LDA analysis for the MACM map were: visual, motor, R thalamus, medial dorsal nucleus 6 − 14 8 emotional, memory, attention, auditory, reward, spatial, L parahippocampal gyrus, BA 34 − 20 2 − 12 schizophrenia, and language (Table  4c). Topographically R hippocampus 26 − 14 − 20 speaking, the MACM results also resembled regions of L inferior temporal gyrus, BA 21 − 56 − 10 − 16 the default mode [30, 69] as well as the frontoparietal L parahippocampal gyrus, BA 28 − 16 − 4 − 14 central executive network [17, 78], and the functional Coordinate locations of the consensus between rsFC and MACM results, decoding outcomes suggested association with execu- including the anatomical label and MNI coordinates of local maxima. Negative x values indicate the left (L) hemisphere and positive x values indicate the right tive emotional and memory processes. A summary of the (R) hemisphere decoding analyses for all three sets of images is shown as a radar plot in Figure 5. Functional decoding: generalized correspondence latent dirichlet allocation (GC‑LDA) Discussion Lastly, we performed functional decoding of the struc- The overall objective of this study was to investigate tural ALE, rsFC, and MACM maps to provide insight convergent alterations in brain structure among indi- into the behavioral functions putatively associated with viduals with PTSD using emergent meta-analytic tech- the observed functional connectivity patterns. Func- niques. Further, we sought to extend the literature and tional decoding was conducted using a GC-LDA analysis assess potential functional consequences associated [74]. Because GC-LDA does not provide correlational or with observed structural alterations in PTSD by applying statistical rankings, the top 10 unique terms computed complementary rsFC and MACM analytic techniques. from the GC-LDA analysis were taken into consideration The current meta-analysis of 23 VBM studies evaluating Table 4 Functional Decoding Results. Functional decoding results for (a) ALE structural meta‑analysis, (b) rsFC, and (c) MACM results as described by Neurosynth terms (a) ALE (b) rsFC (c) MACM Rank Term Weight Rank Term Weight Rank Term Weight 1 Visual 1.886 1 Default 11.234 1 Visual 5900.643 2 Emotional 0.919 2 Default mode network 9.225 2 Motor 3839.578 3 Memory 0.845 3 Intrinsic 7.494 3 Emotional 3665.765 4 Novel 0.616 4 Scale 6.236 4 Memory 3476.688 5 Reward 0.576 5 Self 5.081 5 Attention 2931.357 6 Motor 0.521 6 Person 4.977 6 Auditory 2267.840 7 Self 0.509 7 Reward 4.780 7 Reward 2107.441 8 Faces 0.472 8 Bias 4.568 8 Spatial 2072.742 9 Learning 0.467 9 Judgements 4.279 9 Schizophrenia 2070.157 10 Face 0.450 10 Contexts 4.271 10 Language 2057.731 Rankings display weighted terms listed from highest (1) to lowest (10) Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 10 of 16 brain structure. Previous meta-analyses have identified GM reductions in the mPFC, hippocampus, fusiform gyrus, and lingual gyrus; however, not all of these regions were consistently observed across all meta-analyses [7, 40, 44, 50, 55, 79]. Beyond the mPFC, we did not observe additional convergent GM reductions, indicating that prior findings in these other regions were not replicated. Across the PTSD literature, there is a high degree of variability associated with participant trauma exposure, length of diagnosis of PTSD, medication use, and comor- bidity. Inconsistencies between our findings and previ - ous meta-analytic results could be due to conceptual and methodological differences across the earlier studies, such as the scope of the research question exploring the neurobiology of PTSD, and the subsequent differences in inclusion/exclusion criteria that resulted in different sets Fig. 5 Functional Decoding Results. Functional decoding results for of included studies. Comparison of the included studies the ALE structural meta‑analysis (pink), rsFC (blue), and MACM (red) in this and prior VBM meta-analyses of PTSD indicated results as described by Neurosynth terms. Radar plots display the top varying degrees of overlap, including (from earliest to five terms across all three decoding analyses. The scale of the weights depends on both the GC‑LDA model weights and the input values most recent meta-analyses): 7 of 9 included studies [44], [74], thus, the scale is arbitrary and has been normalized here to 14 of 17 included studies [50], 15 of 20 included studies facilitate visualization [55], 7 of 13 included studies [7], 7 of 8 included studies [40], and 10 out of 12 included studies [79]. Beyond selection of included studies, the meta-ana- GM volume alterations among PTSD versus non-PTSD lytic approach may contribute to the source of variabil- groups identified a single node of convergent gray mat - ity across results. Previous meta-analyses used either the ter loss in the mPFC. GC-LDA-based functional decod- ALE approach [44, 50] or signed differential mapping ing of this cluster was linked to Neurosynth terms of [7, 40, 55, 79]. Consistent with the present results, the visual, emotional, memory, novel, reward, motor, self, meta-analyses by Meng et al. [55] and Klaming et al. [40] faces, learning, and face. Follow-up ALE analyses explor- also yielded a single cluster in mPFC, which used the ing GM reductions in PTSD vs. HC (non-traumatized SDM method while our current results used the ALE controls) and PTSD vs. TC (trauma-exposed controls not approach. However, of all prior meta-analyses, only the diagnosed with PTSD) yielded null findings likely due to study by Meng et al. [55] meets the current threshold of a insufficient power [23]. Subsequent analyses of the ALE- minimum of 20 contrasts for a well-powered coordinate- derived mPFC cluster were conducted to assess task-free based meta-analysis [23]. After reviewing the above prior (rsFC) and task-dependent (MACM) functional connec- meta-analytic work in comparison to our current results, tivity, identifying a consistent and widespread functional we conclude that extensive heterogeneity in the PTSD network implicated in PTSD. These results indicate that literature, combined with varying meta-analytic inclu- structural alterations in the mPFC among individu- sive/exclusion criteria, likely contributed to differences als with PTSD are possibly linked to disruptions across between our results and prior meta-analytic findings. To a larger frontoparietal network that includes the medial, our knowledge, the current meta-analysis of 25 contrasts superior, and inferior frontal gyri, PCC, parahippocam- represents the largest PTSD meta-analysis of structural pal gyri, angular gyri, superior temporal gyrus, thalamus, findings to date, with prior meta-analytic work examin - caudate, and lentiform nucleus. Functional decoding of ing 8-20 included studies. We observed that the mPFC is rsFC and MACM results indicates substantive term over- robustly associated with structural alterations in PTSD; lap with the mPFC ALE results, with additional network- however, it is important to consider how the mPFC is related terms (e.g., default, default mode network, and integrated within existing neurocircuitry models associ- intrinsic). ated with PTSD symptomology. Traditional neurocircuitry models of PTSD utilize a Structural alterations and dysfunction in PTSD fear-conditioning framework, emphasizing hyperreactiv- Our current findings suggest the mPFC appears as the ity of the amygdala in response to fear-related stimuli and most consistently reported brain region across VBM dysfunction between the mPFC and orbitofrontal cortex, neuroimaging studies exploring the impact of PTSD on as well as the hippocampus, in attention and top-down P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 11 of 16 control during threat exposure [70, 81]. However, limit- to a neurobiological theory of psychopathology [28, 56, ing consideration of the psychopathology of PTSD to 57]. The application of the tripartite model to neurobiol - focus on a single brain region (i.e., the amygdala) empha- ogy models of psychiatric disorders define dysfunction sizes fear-related brain activity while minimizing brain within and between connectivity of the DMN, SN, and circuitry implicated in the complex constellation of PTSD CEN networks and relates to a broad range psychiatric symptoms associated with response to trauma exposure, disorders [80], including PTSD [60, 63]. Overall, the cur- such as re-experiencing trauma, avoidance, negative rent meta-analysis identified a functional profile of the mood, and numbing. These additional processes remain mPFC associated with connectivity between the DMN, largely unexplained in original PTSD models. However, SN, and CEN, which broadly supports a network theory more recent neurocircuitry models build from this per- of PTSD [2, 41]. spective, with increased emphasis on altered function of According to the tripartite model of brain function, the the mPFC, its role in contextualization, and how context SN is thought to mediate activity between the DMN and processing is core to the constellation of PTSD symptoms CEN networks in order to orient to external stimuli or [51, 52]. While our results indicated convergent struc- internal salient biological stimuli [57], Sripada et al. [41]. tural alterations in the mPFC, we did not observe similar Altered inter- and intra-network functional connectiv- convergence in the amygdala or other regions that have ity between the DMN, SN, and CEN has previously been been implicated in prior neurocircuitry models of PTSD implicated in PTSD [41]. Specifically, seed-based rest - [32, 42, 70, 81]. However, our results are congruent with ing state studies identified decreased connectivity within the expanded models of PTSD and we provide robust the DMN and SN, yet increased connectivity between evidence in support of the mPFC as a critical node in these two networks among PTSD patients (Sripada et al. PTSD neurocircuitry. Further, our functional decoding [89]). Furthermore, other resting state studies on PTSD results provide additional support for the contextualiza- utilizing graph theory approaches [48] and independ- tion models of PTSD. Taken together, reduced GM in the ent component analysis [102] replicated weakened con- mPFC among individuals diagnosed with PTSD supports nectivity within the DMN, SN, and CEN, yet heightened the premise that these structural alterations may contrib- connectivity between the DMN and SN [35, 89]. Taken ute to deficits in context processing and ultimately play a together, this literature suggests deficits in top-down dominant role in contributing to behaviors related to the control over heightened responses to threatening stim- constellation of symptoms in PTSD [51, 52]. uli and abnormal regulation of orienting attention to threatening stimuli [41, 48, 84, 89, 102]. Patterns from Functional profiles of structural findings in PTSD: support task-based studies reflect previous findings of weakened for the tripartite model of psychopathology connectivity between the SN and DMN and heightened rsFC and MACM analyses characterized mPFC func- connectivity between the SN and CEN [64, 87]. In a study tional connectivity as extending across widespread, among individuals with recent trauma exposure, connec- whole-brain networks engaging frontoparietal and limbic tivity between the DMN, SN, and CEN was reported to regions. These rsFC and MACM results, in conjunction be disrupted among participants who developed PTSD with functional decoding outcomes, identified a func - vs. those who do not [54, 68], providing evidence of dif- tional connectivity profile suggestive of spatial patterns ferential functional connectivity between PTSD patients associated with the default mode network (DMN) [30, and traumatized non-diagnosed individuals. Network 69], salience mode network (SN) [57, 78], and central dysfunction associated with the DMN, SN, and CEN is executive network (CEN) [17, 78]. The DMN is a system also evident in task-based studies, including cues con- of connected brain areas including the mPFC, PCC, infe- taining trauma stimuli [69], eye gaze [87], and a broad rior parietal, and temporal cortices that are often collec- range of behavioral paradigms [64]. Aberrant connectiv- tively observed as displaying anticorrelation with regions ity between and within the DMN, SN, and CEN has also actively engaged during attention-demanding tasks. been associated with PTSD symptoms, such that height- Areas of the DMN are thought to collectively contribute ened connectivity and activity of the DMN was associated to mental processes associated with introspection and with depersonalization/derealization, while weakened self-referential thought [30, 53, 93]. The SN consists of connectivity and activity of the CEN was associated with the dorsolateral ACC and bilateral insula and is involved hyperarousal and hypervigilance [2]. Additionally, weak- in saliency detection and attentional processes [57, 78]. ened inter-network connectivity between the SN and Finally, the CEN consists of the dorsolateral prefrontal DMN has been found to be positively correlated with and posterior parietal cortices and is typically involved in Clinician Administered PTSD Scale (CAPS) scores that attentionally driven cognitive functions, including goal- measure PTSD symptom severity [84, 89]. Moreover, directed behavior [87]. These three networks are central Bluhm et al. [4] found weakened spontaneous activity in Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 12 of 16 regions of the DMN; in addition, posterior cingulate con- the number of participants across each group was some- nectivity was positively correlated with self-reported dis- what unevenly distributed due to small sample sizes in sociated experiences among participants with PTSD. In the original studies. However, the current meta-analysis sum, the literature on abnormal brain function associated met the previously recommended standard of at least with PTSD points to a pattern of results suggesting that 20 experimental contrasts required to conduct a well- symptoms are related to aberrant connectivity within and powered meta-analysis [23]. Second, much heterogeneity between the DMN, SN, and CEN. In a recent review of exists across the studies included in our meta-analysis. the neuroimaging literature on PTSD, Lanius et  al. [47] For example, many of the studies had diagnostic criteria summarized this work to reflect that dysfunction in the for PTSD using different clinical measures and reported DMN is associated with an altered sense of self, dysfunc- different instances of the duration of PTSD (e.g., lifetime tion in the SN is associated with hyperarousal and hyper- vs. first onset). Substantial variability was also present in vigilance, and dysfunction in the CEN is associated with the type of trauma and duration of exposure to trauma cognitive dysfunction, including memory and cognitive within the different groups for this study. Given these control deficits. issues, we were unable to classify PTSD subtypes across The results from the current meta-analysis provide the included studies and thus have reported results that a robust mPFC-centric model of PTSD that is aligned relate to generalized PTSD. Many of the original stud- with the extant literature and compliments the tripartite ies were not able to clearly disentangle comorbidity of model of psychopathology. The mPFC, a core region of PTSD with other psychiatric disorders (e.g., depression, the DMN [30, 69], is often disrupted in individuals with anxiety) or report instances of medication and drug PTSD [15, 68]. The results of the present meta-analysis abuse. Furthermore, studies relied on various neuro- suggest alterations in mPFC structure, and related func- imaging acquisition and analysis methods, which likely tion, may play a crucial role in the underlying neurobi- introduced additional variability associated with meth- ology of PTSD. Dysfunction of the mPFC is thought to odological flexibility [6, 9]. However, the goal of neuro- be associated with poorer regulation of contextualization imaging meta-analysis was to examine consensus despite of PTSD symptoms. Prior literature indicates weakened such variability in the literature. With this in mind, we integration of the DMN and disrupted inter-network are confident that the mPFC is a significant brain region connectivity with the SN and CEN, representing aberrant linked to GM reductions in PTSD, as well as a robust dysfunction of these tripartite networks in the psychopa- node of the DMN that plays an important role in toggling thology of PTSD [73]. Most of the prior functional and between the DMN, SN, and CEN. Future transdiagnostic structural work involved varying analytic approaches, and meta-analytic work is needed to identify similar and examined heterogeneous populations, and utilized region unique neurobiological mechanisms of PTSD in compar- of interest approaches or a priori hypotheses. The cur - ison to other related disorders, including complementary rent application of advanced meta-analytic techniques disease-decoding or structural covariance analysis, which allowed for a whole-brain assessment of structural altera- would further advance clinical insight. tions associated with PTSD and the associated functional profiles of the mPFC. Future work in PTSD should con - Conclusions sider integrating network-based analytic approaches The present study utilized coordinate-based meta-ana - with an mPFC-centric tripartite model to investigate lytic techniques to determine that reduced mPFC GM is differences in neuropathology of PTSD subtypes (e.g., consistently found among individuals with PTSD. Com- trauma experiences, duration of exposures), characteriz- plementary analyses of rsFC and MACM functional ing heterogeneous presentations of PTSD symptoms, and connectivity provided novel insight into how structural potential predispositional developmental effects among alterations may have potential functional consequences. youth, adolescent, and adult populations. Our results indicated that decreases in mPFC GM may be linked to widespread functional systems that are Limitations implicated in behavioral deficits and cluster symptoma - Our study is limited by several considerations. First, the tology of PTSD. Specifically, consensus-based func - present meta-analysis is limited by the small number tional profiles, across task-free and task-based domains, of studies included. The studies that met the standards emphasized brain regions associated with the tripartite of inclusion for this study were considered to reduce model of psychiatric disorders where inter- and intra- instances of variance and consider reliability of study network connectivity involving the DMN, SN, and CEN findings (inclusion and exclusion criteria are shown in are core to PTSD dysfunction. Overall, these results may Fig.  2). By considering the inclusion of trauma-exposed be important in providing a more comprehensive under- controls, healthy controls, and individuals with PTSD, standing of the neurobiological bases of PTSD, which is P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 13 of 16 3. Bao W, Gao Y, Cao L, Li H, Liu J, Liang K, Hu X, Zhang L, Hu X, Gong Q, needed to understand the varying diagnosis, symptoma- Huang X. Alterations in large‑scale functional networks in adult post ‑ tology, and treatment of PTSD, as well as enhanced tar- traumatic stress disorder: a systematic review and meta‑analysis of geting of treatment towards heterogeneous classification resting‑state functional connectivity studies. Neurosci Biobehav Rev. 2021;131:1027–36. and symptom clusters of PTSD. 4. Bluhm RL, Williamson PC, Osuch EA, Frewen PA, Stevens TK, Boksman K, et al. Alterations in default network connectivity in posttraumatic Supplementary Information stress disorder related to early‑life trauma. J Psychiatry Neurosci JPN. 2009;34:187–94. The online version contains supplementary material available at https:// doi. 5. Bossini L, Santarnecchi E, Casolaro I, Koukouna D, Caterini C, Cecchini org/ 10. 1186/ s12993‑ 022‑ 00196‑2. F, et al. Morphovolumetric changes after EMDR treatment in drug‑ naïve PTSD patients. Riv Psichiatr. 2017;52:24–31. https:// doi. org/ 10. Additional file 1: Table S1. Summary of the demographic and clinical 1708/ 2631. 27051. variables for the voxel‑based morphometry studies included in the 6. Botvinik‑Nezer R, Holzmeister F, Camerer CF, et al. Variability in the meta‑analysis analysis of a single neuroimaging dataset by many teams. Nature. 2020. https:// doi. org/ 10. 1038/ s41586‑ 020‑ 2314‑9. 7. Bromis K, Calem M, Reinders AATS, Williams SCR, Kempton MJ. Meta‑ Acknowledgements analysis of 89 structural MRI studies in posttraumatic stress disorder The authors would like to thank the FIU Instructional & Research Comput‑ and comparison with major depressive disorder. Am J Psychiatry. ing Center (IRCC, http:// ircc. fiu. edu) for providing the HPC and computing 2018;175:989–98. https:// doi. org/ 10. 1176/ appi. ajp. 2018. 17111 199. resources that contributed to the research results reported within this paper. 8. Burgess GC, Kandala S, Nolan D, Laumann TO, Power JD, Adeyemo B, et al. Evaluation of denoising strategies to address motion‑ correlated Author contributions artifacts in resting‑state functional magnetic resonance imaging data BP and IC collected and prepared data for meta‑analysis. MCR and JEB from the human connectome Project. Brain Connect. 2016;6:669–80. analyzed data. MCR, JEB, LDHB, and TS contributed scripts and pipelines. BP, https:// doi. org/ 10. 1089/ brain. 2016. 0435. MRC, and ARL wrote the paper. All authors contributed to the revisions and 9. Carp J. On the plurality of (methodological) worlds: estimating the approved the final version. All authors read and approved the final manuscript. analytic flexibility of fMRI experiments. Front Neurosci. 2012. https:// doi. org/ 10. 3389/ fnins. 2012. 00149. Funding 10. Chao LL, Lenoci M, Neylan TC. Eec ff ts of post ‑traumatic stress Funding for this project was provided by NSF 1631325, NIH R01 DA041353, disorder on occipital lobe function and structure. NeuroReport. and NIH U01 DA041156. 2012;23:412–9. https:// doi. org/ 10. 1097/ WNR. 0b013 e3283 52025e. 11. Chen S, Li L, Xu B, Liu J. Insular cortex involvement in declarative Availability of data and materials memory deficits in patients with post ‑traumatic stress disorder. BMC Data and materials are available in a GitHub repository (https:// github. com/ Psychiatry. 2009;9:39. https:// doi. org/ 10. 1186/ 1471‑ 244X‑9‑ 39. NBCLab/ meta‑ analy sis_ ptsd), including the meta‑analytic coordinate files, 12. Chen Y, Fu K, Feng C, Tang L, Zhang J, Huan Y, et al. Different regional data analysis scripts (i.e., code), image‑based results (i.e., ALE, rsFC, and MACM gray matter loss in recent onset PTSD and non PTSD after a single images), and functional decoding results. rsFC analyses used the Human Con‑ prolonged trauma exposure. PLoS ONE. 2012;7:e48298. https:// doi. nectome Project’s [90] Young Adult Study S1200 Data Release (March 1, 2017), org/ 10. 1371/ journ al. pone. 00482 98. which is available at db.humanconnectome.org. 13. Cheng B, Huang X, Li S, Hu X, Luo Y, Wang X, et al. Gray matter altera‑ tions in post‑traumatic stress disorder, obsessive ‑ compulsive disor‑ der, and social anxiety disorder. Front Behav Neurosci. 2015;9:219. Declarations https:// doi. org/ 10. 3389/ fnbeh. 2015. 00219. 14. Corbo V, Clément M‑H, Armony JL, Pruessner JC, Brunet A. Size versus Ethics approval and consent to participate shape differences: contrasting voxel‑based and volumetric analyses This secondary data analysis was approved by the Institutional Review Board of the anterior cingulate cortex in individuals with acute posttrau‑ of Florida International University. matic stress disorder. Biol Psychiatry. 2005;58:119–24. https:// doi. org/ 10. 1016/j. biops ych. 2005. 02. 032. Consent for publication 15. DiGangi JA, Tadayyon A, Fitzgerald DA, Rabinak CA, Kennedy A, Not applicable. Klumpp H, et al. Reduced default mode network connectivity follow‑ ing combat trauma. Neurosci Lett. 2016;615:37–43. https:// doi. org/ Competing interests 10. 1016/j. neulet. 2016. 01. 010. The authors declare no competing interests. 16. Dogan I, Eickhoff CR, Fox PT, Laird AR, Schulz JB, Eickhoff SB, Reetz K. Functional connectivity modeling of consistent cortico‑striatal Author details degeneration in HD. Neuroimage Clin. 2015;7:640–52. Department of Psychology, Florida International University, Miami, FL, USA. 17. Dosenbach NUF, Fair DA, Miezin FM, Cohen AL, Wenger KK, Dosen‑ Department of Physics, Florida International University, Miami, FL, USA. bach RAT, et al. Distinct brain networks for adaptive and stable task Department of Psychology, Old Dominion University, Norfolk, VA, USA. control in humans. Proc Natl Acad Sci. 2007;104:11073–8. https:// doi. org/ 10. 1073/ pnas. 07043 20104. Received: 8 April 2022 Accepted: 27 August 2022 18. Duval E, Liberzon I, Javanbakht A. Neural circuits in anxiety and stress disorders: a focused review. Ther Clin Risk Manag. 2015. https:// doi. org/ 10. 2147/ TCRM. S48528. 19. Eckart C, Stoppel C, Kaufmann J, Tempelmann C, Hinrichs H, Elbert T, et al. Structural alterations in lateral prefrontal, parietal and posterior References midline regions of men with chronic posttraumatic stress disorder. J 1. Agaibi CE, Wilson JP. Trauma, PTSD, and resilience: a review of the litera‑ Psychiatry Neurosci JPN. 2011;36:176–86. https:// doi. org/ 10. 1503/ jpn. ture. Trauma Violence Abuse. 2005;6:195–216. https:// doi. org/ 10. 1177/ 15248 38005 277438. 20. Ehlers A, Clark DM. A cognitive model of posttraumatic stress disor‑ 2. Akiki TJ, Averill CL, Abdallah CG. A Network‑based neurobiological der. Behav Res Ther. 2000;38:319–45. https:// doi. org/ 10. 1016/ S0005‑ model of PTSD: evidence from structural and functional neuroimag‑ 7967(99) 00123‑0. ing studies. Curr Psychiatry Rep. 2017;19:81. https:// doi. org/ 10. 1007/ s11920‑ 017‑ 0840‑4. Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 14 of 16 21. Eickhoff SB, Bzdok D, Laird AR, Kurth F, Fox PT. Activation likelihood esti‑ 40. Klaming R, Harlé KM, Infante MA, Bomyea J, Kim C, Spadoni AD. Shared mation meta‑analysis revisited. Neuroimage. 2012;59:2349–61. https:// gray matter reductions across alcohol use disorder and posttraumatic doi. org/ 10. 1016/j. neuro image. 2011. 09. 017. stress disorder in the anterior cingulate cortex: a dual meta‑analysis. 22. Eickhoff SB, Laird AR, Grefkes C, Wang LE, Zilles K, Fox PT. Coordinate ‑ Neurobiol Stress. 2019;10:100132. https:// doi. org/ 10. 1016/j. ynstr. 2018. based activation likelihood estimation meta‑analysis of neuroimaging 09. 009. data: a random‑ effects approach based on empirical estimates of 41. Koch SBJ, van Zuiden M, Nawijn L, Frijling JL, Veltman DJ, Olff M. spatial uncertainty. Hum Brain Mapp. 2009;30:2907–26. https:// doi. org/ Aberrant resting‑state brain activity in posttraumatic stress disorder: a 10. 1002/ hbm. 20718. meta‑analysis and systematic review: theoretical review—brain activity 23. Eickhoff SB, Nichols TE, Laird AR, Hoffstaedter F, Amunts K, Fox PT, et al. in PTSD during rest. Depress Anxiety. 2016;33:592–605. https:// doi. org/ Behavior, sensitivity, and power of activation likelihood estimation char‑10. 1002/ da. 22478. acterized by massive empirical simulation. Neuroimage. 2016;137:70– 42. Koenigs M, Grafman J. Posttraumatic stress disorder: the role of medial 85. https:// doi. org/ 10. 1016/j. neuro image. 2016. 04. 072. prefrontal cortex and amygdala. Neuroscientist. 2009;15:540–8. https:// 24. Etkin A, Wager TD. Functional neuroimaging of anxiety: a meta‑analysis doi. org/ 10. 1177/ 10738 58409 333072. of emotional processing in PTSD, social anxiety disorder, and specific 43. Kroes MCW, Rugg MD, Whalley MG, Brewin CR. Structural brain abnor‑ phobia. Am J Psychiatry. 2007;164:1476–88. malities common to posttraumatic stress disorder and depression. J 25. Feinberg DA, Moeller S, Smith SM, Auerbach E, Ramanna S, Glasser MF, Psychiatry Neurosci JPN. 2011;36:256–65. https:// doi. org/ 10. 1503/ jpn. et al. Multiplexed echo planar imaging for sub‑second whole brain 100077. FMRI and fast diffusion imaging. PLoS ONE. 2010;5:e15710. https:// doi. 44. Kühn S, Gallinat J. Gray matter correlates of posttraumatic stress disor‑ org/ 10. 1371/ journ al. pone. 00157 10. der: a quantitative meta‑analysis. Biol Psychiatry. 2013;73:70–4. https:// 26. Felmingham K, Williams LM, Whitford TJ, Falconer E, Kemp AH, Peduto doi. org/ 10. 1016/j. biops ych. 2012. 06. 029. A, et al. Duration of posttraumatic stress disorder predicts hippocampal 45. Laird AR, Robinson JL, McMillan KM, Tordesillas‑ Gutiérrez D, Moran ST, grey matter loss. NeuroReport. 2009;20:1402–6. https:// doi. org/ 10. Gonzales SM, et al. Comparison of the disparity between Talairach and 1097/ WNR. 0b013 e3283 300fbc. MNI coordinates in functional neuroimaging data: validation of the 27. Glasser MF, Smith SM, Marcus DS, Andersson JLR, Auerbach EJ, Behrens Lancaster transform. Neuroimage. 2010;51:677–83. https:// doi. org/ 10. TEJ, et al. The Human Connectome Project’s neuroimaging approach. 1016/j. neuro image. 2010. 02. 048. Nat Neurosci. 2016;19:1175–87. https:// doi. org/ 10. 1038/ nn. 4361. 46. Lancaster JL, Tordesillas‑ Gutiérrez D, Martinez M, Salinas F, Evans A, Zilles 28. Goodkind M, Eickhoff SB, Oathes DJ, Jiang Y, Chang A, Jones‑Hagata LB, K, et al. Bias between MNI and Talairach coordinates analyzed using the et al. Identification of a common neurobiological substrate for mental ICBM‑152 brain template. Hum Brain Mapp. 2007;28:1194–205. https:// illness. JAMA Psychiat. 2015;72:305. https:// doi. org/ 10. 1001/ jamap sychi doi. org/ 10. 1002/ hbm. 20345. atry. 2014. 2206. 47. Lanius RA, Frewen PA, Tursich M, Jetly R, McKinnon MC. Restoring 29. Gorgolewski K, Burns CD, Madison C, Clark D, Halchenko YO, Waskom large‑scale brain networks in PTSD and related disorders: a proposal for ML, et al. Nipype: a flexible, lightweight and extensible neuroimaging neuroscientifically‑informed treatment interventions. Eur J Psychotrau‑ data processing framework in python. Front Neuroinformatics. 2011. matology. 2015;6:27313. https:// doi. org/ 10. 3402/ ejpt. v6. 27313. https:// doi. org/ 10. 3389/ fninf. 2011. 00013. 48. Lei D, Li K, Li L, Chen F, Huang X, Lui S, et al. Disrupted functional brain 30. Greicius MD, Krasnow B, Reiss AL, Menon V. Functional connectivity in connectome in patients with posttraumatic stress disorder. Radiology. the resting brain: a network analysis of the default mode hypothesis. 2015;276:818–27. https:// doi. org/ 10. 1148/ radiol. 15141 700. Proc Natl Acad Sci. 2003;100:253–8. https:// doi. org/ 10. 1073/ pnas. 01350 49. Li L, Chen S, Liu J, Zhang J, He Z, Lin X. Magnetic resonance imaging 58100. and magnetic resonance spectroscopy study of deficits in hippocampal 31. Hakamata Y, Matsuoka Y, Inagaki M, Nagamine M, Hara E, Imoto S, et al. structure in fire victims with recent ‑ onset posttraumatic stress disorder. Structure of orbitofrontal cortex and its longitudinal course in cancer‑ Can J Psychiatry. 2006;51:431–7. https:// doi. org/ 10. 1177/ 07067 43706 related post‑traumatic stress disorder. Neurosci Res. 2007;59:383–9. 05100 704. https:// doi. org/ 10. 1016/j. neures. 2007. 08. 012. 50. Li L, Wu M, Liao Y, Ouyang L, Du M, Lei D, et al. Grey matter reduction 32. Hamner MB. Potential role of the anterior cingulate cortex in PTSD: associated with posttraumatic stress disorder and traumatic stress. review and hypothesis. Depress Anxiety. 1999;9:14. Neurosci Biobehav Rev. 2014;43:163–72. https:// doi. org/ 10. 1016/j. neubi 33. Hayes JP, Hayes SM, Mikedis AM. Quantitative meta‑analysis of neural orev. 2014. 04. 003. activity in posttraumatic stress disorder. Biol Mood Anxiety Disord. 51. Liberzon I, Abelson JL. Context processing and the neurobiology of 2012;2:9. post‑traumatic stress disorder. Neuron. 2016;92:14–30. https:// doi. org/ 34. Herringa R, Phillips M, Almeida J, Insana S, Germain A. Post‑traumatic 10. 1016/j. neuron. 2016. 09. 039. stress symptoms correlate with smaller subgenual cingulate, caudate, 52. Liberzon I, Garfinkel SN. Functional neuroimaging in post ‑traumatic and insula volumes in unmedicated combat veterans. Psychiatry Res. stress disorder. In: LeDoux JE, Keane T, Shiromani P, editors. Post‑ 2012;203:139–45. https:// doi. org/ 10. 1016/j. pscyc hresns. 2012. 02. 005. Traumatic stress disorder: basic science and clinical practice. Totowa NJ: 35. Holmes SE, Scheinost D, DellaGioia N, Davis MT, Matuskey D, Pietrzak Humana Press; 2009. p. 219–317 (10.1007/978‑1‑60327‑329‑9). RH, et al. Cerebellar and prefrontal cortical alterations in PTSD: structural 53. Liberzon I, Shulman GL. A default mode of brain function. Proc Natl and functional evidence. Chronic Stress. 2018;2:247054701878639. Acad Sci. 2001;98:676–82. https:// doi. org/ 10. 1073/ pnas. 98.2. 676. https:// doi. org/ 10. 1177/ 24705 47018 786390. 54. Liu Y, Li L, Li B, Feng N, Li L, Zhang X, et al. Decreased triple network 36. Jenkinson M, Beckmann CF, Behrens TEJ, Woolrich MW, Smith SM. FSL. connectivity in patients with recent onset post‑traumatic stress disor ‑ Neuroimage. 2012;62:782–90. https:// doi. org/ 10. 1016/j. neuro image. der after a single prolonged trauma exposure. Sci Rep. 2017;7:12625. 2011. 09. 015.https:// doi. org/ 10. 1038/ s41598‑ 017‑ 12964‑6. 37. Kamalian A, Khodadadifar T, Saberi A, Masoudi M, Camilleri JA, Eickhoff 55. Meng Y, Qiu C, Zhu H, Lama S, Lui S, Gong Q, et al. Anatomical deficits in CR, Zarei M, Pasquini L, Laird AR, Fox PT, Eickhoff SB, Tamasian M. Con‑ adult posttraumatic stress disorder: a meta‑analysis of voxel‑based mor ‑ vergent regional brain abnormalities in behavioral variant frontotempo‑ phometry studies. Behav Brain Res. 2014;270:307–15. https:// doi. org/ 10. ral dementia: a neuroimaging meta‑analysis of 73 studies. Alzheimer’s 1016/j. bbr. 2014. 05. 021. Dement: Diagn, Assess Dis Monit. 2022;14:e12318. 56. Menon V. Large‑scale brain networks and psychopathology: a unifying 38. Kasai K, Yamasue H, Gilbertson MW, Shenton ME, Rauch SL, Pitman RK. triple network model. Trends Cogn Sci. 2011;15:483–506. https:// doi. Evidence for acquired pregenual anterior cingulate gray matter loss org/ 10. 1016/j. tics. 2011. 08. 003. from a twin study of combat‑related posttraumatic stress disorder. Biol 57. Menon V, Uddin LQ. Saliency, switching, attention and control: a Psychiatry. 2008;63:550–6. https:// doi. org/ 10. 1016/j. biops ych. 2007. 06. network model of insula function. Brain Struct Funct. 2010;214:655–67. 022.https:// doi. org/ 10. 1007/ s00429‑ 010‑ 0262‑0. 39. Kirkpatrick HA, Heller GM. Post‑traumatic stress disorder: theory and 58. Nardo D, Högberg G, Looi JCL, Larsson S, Hällström T, Pagani M. Gray treatment update. Int J Psychiatry Med. 2014;47:337–46. https:// doi. matter density in limbic and paralimbic cortices is associated with org/ 10. 2190/ PM. 47.4.h. P ankey et al. Behavioral and Brain Functions (2022) 18:9 Page 15 of 16 trauma load and EMDR outcome in PTSD patients. J Psychiatr Res. 77. Salo T, Yarkoni T, Nichols TE, Poline J‑B, Bilgel M, Bottenhorn KL, 2010;44:477–85. https:// doi. org/ 10. 1016/j. jpsyc hires. 2009. 10. 014. Jarecka D, Kent JD, Kimbler A, Nielson DM, Oudyk KM, Peraza JA, 59. National Center for PTSD (2019). PTSD: National Center for PTSD. Com‑ Pérez A, Reeders PC, Yanes JA, Laird AR. NiMARE: neuroimaging meta‑ mon PTSD Adults. Available at: https:// www. ptsd. va. gov/ under stand/ analysis research environment. NeuroLibre. 2022;1:7. https:// doi. org/ common/ common_ adults. asp. Accessed 31 October 2020.10. 55458/ neuro libre. 00007. 60. Nicholson AA, Harricharan S, Densmore M, Neufeld RWJ, Ros T, McKin‑ 78. Seeley WW, Menon V, Schatzberg AF, Keller J, Glover GH, Kenna H, non MC, et al. Classifying heterogeneous presentations of PTSD via the et al. Dissociable intrinsic connectivity networks for salience process‑ default mode, central executive, and salience networks with machine ing and executive control. J Neurosci. 2007;27:2349–56. https:// doi. learning. NeuroImage Clin. 2020;27:102262. https:// doi. org/ 10. 1016/j. org/ 10. 1523/ JNEUR OSCI. 5587‑ 06. 2007. nicl. 2020. 102262. 79. Serra‑Blasco M, Radua J, Soriano ‑Mas C, Gómez‑Benlloch A, Porta‑ 61. O’Doherty DCM, Tickell A, Ryder W, Chan C, Hermens DF, Bennett MR, Casteràs D, Carulla‑Roig M, et al. Structural brain correlates in major et al. Frontal and subcortical grey matter reductions in PTSD. Psychiatry depression, anxiety disorders and post‑traumatic stress disorder: a Res Neuroimaging. 2017;266:1–9. https:// doi. org/ 10. 1016/j. pscyc hresns. voxel‑based morphometry meta‑analysis. Neurosci Biobehav Rev. 2017. 05. 008. 2021;129:269–81. https:// doi. org/ 10. 1016/j. neubi orev. 2021. 07. 002. 62. Pai A, Suris A, North C. Posttraumatic stress disorder in the DSM‑5: con‑ 80. Sha Z, Wager TD, Mechelli A, He Y. Common dysfunction of large‑ troversy, change, and conceptual considerations. Behav Sci. 2017;7:7. scale neurocognitive networks across psychiatric disorders. Biol https:// doi. org/ 10. 3390/ bs701 0007. Psychiatry. 2019;85:379–88. https:// doi. org/ 10. 1016/j. biops ych. 2018. 63. Patel R, Spreng RN, Shin LM, Girard TA. Neurocircuitry models of post‑11. 011. traumatic stress disorder and beyond: a meta‑analysis of functional 81. Shin LM, Rauch SL, Pitman RK. Amygdala, medial prefrontal cortex, neuroimaging studies. Neurosci Biobehav Rev. 2012;36:2130–42. and hippocampal function in PTSD. Ann NY Acad Sci. 2006;1071:67– https:// doi. org/ 10. 1016/j. neubi orev. 2012. 06. 003. 79. https:// doi. org/ 10. 1196/ annals. 1364. 007. 64. Pitman RK, Rasmusson AM, Koenen KC, Shin LM, Orr SP, Gilbertson 82. Smith SM, Beckmann CF, Andersson J, Auerbach EJ, Bijsterbosch MW, et al. Biological studies of post‑traumatic stress disorder. Nat Rev J, Douaud G, et al. Resting‑state fMRI in the Human Connectome Neurosci. 2012;13:769–87. https:// doi. org/ 10. 1038/ nrn33 39. Project. Neuroimage. 2013;80:144–68. https:// doi. org/ 10. 1016/j. neuro 65. PTSD Alliance (2018). Traumatic stress disorder fact sheet. Available at: image. 2013. 05. 039. http:// www. sidran. org/ wp‑ conte nt/ uploa ds/ 2018/ 11/ Post T‑raum atic‑ 83. Sripada RK, King AP, Welsh RC, Garfinkel SN, Wang X, Sripada CS, et al. Stress‑ Disor der‑ Fact‑ Sheet‑. pdf . Accessed 31 October 2020. Neural dysregulation in posttraumatic stress disorder: evidence for 66. Qi R, Luo Y, Zhang L, Weng Y, Surento W, Jahanshad N, et al. Social sup‑ disrupted equilibrium between salience and default mode brain port modulates the association between PTSD diagnosis and medial networks. Psychosom Med. 2012;74:904–11. https:// doi. org/ 10. 1097/ frontal volume in Chinese adults who lost their only child. Neurobiol PSY. 0b013 e3182 73bf33. Stress. 2020;13:100227. https:// doi. org/ 10. 1016/j. ynstr. 2020. 100227. 84. Sui SG, Wu MX, King ME, Zhang Y, Ling L, Xu JM, et al. Abnormal grey 67. Qin L. A preliminary study of alterations in default network connectivity matter in victims of rape with PTSD in Mainland China: a voxel‑based in post‑traumatic stress disorder patients following recent trauma. Brain morphometry study. Acta Neuropsychiatr. 2010;22:118–26. https:// Res. 2012. https:// doi. org/ 10. 1016/j. brain res. 2012. 09. 029.doi. org/ 10. 1111/j. 1601‑ 5215. 2010. 00459.x. 68. Rabellino D, Tursich M, Frewen PA, Daniels JK, Densmore M, Théberge J, 85. Tavanti M, Battaglini M, Borgogni F, Bossini L, Calossi S, Marino D, et al. Intrinsic Connectivity Networks in post‑traumatic stress disorder et al. Evidence of diffuse damage in frontal and occipital cortex in during sub‑ and supraliminal processing of threat ‑related stimuli. Acta the brain of patients with post‑traumatic stress disorder. Neurol Sci. Psychiatr Scand. 2015;132:365–78. https:// doi. org/ 10. 1111/ acps. 12418. 2012;33:59–68. https:// doi. org/ 10. 1007/ s10072‑ 011‑ 0659‑4. 69. Raichle ME. The brain’s default mode network. Annu Rev 86. Thome J, Frewen P, Daniels JK, Densmore M, Lanius RA. Altered con‑ Neurosci. 2015;38:433–47. https:// doi. org/ 10. 1146/ annur nectivity within the salience network during direct eye gaze in PTSD. ev‑ neuro‑ 071013‑ 014030. Borderline Personal Disord Emot Dysregul. 2014;1:17. https:// doi. org/ 70. Rauch SL, Shin LM, Phelps EA. Neurocircuitry models of posttraumatic 10. 1186/ 2051‑ 6673‑1‑ 17. stress disorder and extinction: human neuroimaging research—past, 87. Turner BM, Rodriguez CA, Liu Q, Molloy MF, Hoogendijk M, McClure present, and future. Biol Psychiatry. 2006;60:376–82. https:// doi. org/ 10. SM. On the neural and mechanistic bases of self‑ control. Cereb 1016/j. biops ych. 2006. 06. 004. Cortex. 2019;29:732–50. https:// doi. org/ 10. 1093/ cercor/ bhx355. 71. Reetz K, Dogan I, Rolfs A, Binkofski F, Schulz JB, Laird AR, Fox PT, Eickhoff 88. Tursich M, Ros T, Frewen PA, Kluetsch RC, Calhoun VD, Lanius RA. Dis‑ SB. Investigating function and connectivity of morphometric find‑ tinct intrinsic network connectivity patterns of post‑traumatic stress ings—Exemplified on cerebellar atrophy in spinocerebellar ataxia 17 disorder symptom clusters. Acta Psychiatr Scand. 2015;132:29–38. (SCA17). Neuroimage. 2012;62:1354–66.https:// doi. org/ 10. 1111/ acps. 12387. 72. Rocha‑Rego V, Pereira MG, Oliveira L, Mendlowicz MV, Fiszman A, 89. Uğurbil K, Xu J, Auerbach EJ, Moeller S, Vu AT, Duarte‑ Carvajalino Marques‑Portella C, et al. Decreased premotor cortex volume in JM, et al. Pushing spatial and temporal resolution for functional victims of urban violence with posttraumatic stress disorder. PLoS ONE. and diffusion MRI in the Human Connectome Project. Neuroimage. 2012;7:e42560. https:// doi. org/ 10. 1371/ journ al. pone. 00425 60. 2013;80:80–104. https:// doi. org/ 10. 1016/j. neuro image. 2013. 05. 012. 73. Ross MC, Cisler JM. Altered large‑scale functional brain organization in 90. Van Essen DC, Smith SM, Barch DM, Behrens TEJ, Yacoub E, Ugurbil K. posttraumatic stress disorder: a comprehensive review of univariate The WU‑Minn Human Connectome Project: an overview. Neuroim‑ and network‑level neurocircuitry models of PTSD. NeuroImage Clin. age. 2013;80:62–79. https:// doi. org/ 10. 1016/j. neuro image. 2013. 05. 2020;27:102319. https:// doi. org/ 10. 1016/j. nicl. 2020. 102319. 041. 74. Rubin TN, Koyejo O, Gorgolewski KJ, Jones MN, Poldrack RA, 91. Van Essen DC, Ugurbil K, Auerbach E, Barch D, Behrens TEJ, Bucholz R, Yarkoni T. Decoding brain activity using a large‑scale probabilistic et al. The Human Connectome Project: a data acquisition perspec‑ functional‑anatomical atlas of human cognition. PLOS Comput Biol. tive. Neuroimage. 2012;62:2222–31. https:// doi. org/ 10. 1016/j. neuro 2017;13:e1005649. https:// doi. org/ 10. 1371/ journ al. pcbi. 10056 49.image. 2012. 02. 018. 75. Salimi‑Khorshidi G, Douaud G, Beckmann CF, Glasser MF, Griffanti L, 92. Wang T, Liu J, Zhang J, Zhan W, Li L, Wu M, Huang H, Zhu H, Kemp GJ, Smith SM. Automatic denoising of functional MRI data: combining Gong Q. Altered resting‑state functional activity in posttraumatic stress independent component analysis and hierarchical fusion of classifiers. disorder: a quantitative meta‑analysis. Sci Rep. 2016;6:27131. Neuroimage. 2014;90:449–68. https:// doi. org/ 10. 1016/j. neuro image. 93. Whitfield‑ Gabrieli S, Ford JM. Default mode network activity and con‑ 2013. 11. 046. nectivity in psychopathology. Annu Rev Clin Psychol. 2012;8:49–76. 76. Salimi‑Khorshidi G, Smith SM, Keltner JR, Wager TD, Nichols TE. Meta‑https:// doi. org/ 10. 1146/ annur ev‑ clinp sy‑ 032511‑ 143049. analysis of neuroimaging data: a comparison of image‑based and 94. Winkler AM, Ridgway GR, Webster MA, Smith SM, Nichols TE. Permuta‑ coordinate‑based pooling of studies. Neuroimage. 2009;45:810–23. tion inference for the general linear model. Neuroimage. 2014;92:381– https:// doi. org/ 10. 1016/j. neuro image. 2008. 12. 039. 97. https:// doi. org/ 10. 1016/j. neuro image. 2014. 01. 060. Pankey et al. Behavioral and Brain Functions (2022) 18:9 Page 16 of 16 95. Woo C‑ W, Krishnan A, Wager TD. Cluster‑ extent based threshold‑ ing in fMRI analyses: pitfalls and recommendations. Neuroimage. 2014;91:412–9. https:// doi. org/ 10. 1016/j. neuro image. 2013. 12. 058. 96. Worsley KJ. Statistical analysis of activation images. In: Jezzard P, Mat‑ thews PM, Smith SM, editors. Functional magnetic resonance imaging. Oxford: Oxford University Press; 2001. p. 251–70 (10.1093/acprof: oso/9780192630711.003.0014). 97. Yamasue H, Kasai K, Iwanami A, Ohtani T, Yamada H, Abe O, et al. Voxel‑ based analysis of MRI reveals anterior cingulate gray‑matter volume reduction in posttraumatic stress disorder due to terrorism. Proc Natl Acad Sci USA. 2003;100:9039–43. https:// doi. org/ 10. 1073/ pnas. 15304 98. Yarkoni T, Poldrack RA, Nichols TE, Van Essen DC, Wager TD. Large‑scale automated synthesis of human functional neuroimaging data. Nat Methods. 2011;8:665–70. https:// doi. org/ 10. 1038/ nmeth. 1635. 99. Yehuda R, Hoge CW, McFarlane AC, Vermetten E, Lanius RA, Niev‑ ergelt CM, et al. Post‑traumatic stress disorder. Nat Rev Dis Primer. 2015;1:15057. https:// doi. org/ 10. 1038/ nrdp. 2015. 57. 100. Zhang H, Chen X, Chen S, Li Y, Chen C, Long Q, et al. Facial expression enhances emotion perception compared to vocal prosody: behavioral and fMRI studies. Neurosci Bull. 2018;34:801–15. https:// doi. org/ 10. 1007/ s12264‑ 018‑ 0231‑9. 101. Zhang J, Tan Q, Yin H, Zhang X, Huan Y, Tang L, et al. Decreased gray matter volume in the left hippocampus and bilateral calcarine cortex in coal mine flood disaster survivors with recent onset PTSD. Psychiatry Res. 2011;192:84–90. https:// doi. org/ 10. 1016/j. pscyc hresns. 2010. 09. 001. 102. Zhang Y, Liu F, Chen H, Li M, Duan X, Xie B, et al. Intranetwork and internetwork functional connectivity alterations in post‑traumatic stress disorder. J Aec ff t Disord. 2015;187:114–21. https:// doi. org/ 10. 1016/j. jad. 2015. 08. 043. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions

Journal

Behavioral and Brain FunctionsSpringer Journals

Published: Sep 13, 2022

Keywords: Post-traumatic stress disorder; Meta-analysis; Voxel-based morphometry; Functional connectivity

References