Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 May 6.
Published in final edited form as: Psychiatry Res Neuroimaging. 2021 Oct 16;318:111396. doi: 10.1016/j.pscychresns.2021.111396

Association Between Frontal Cortico-Limbic White-Matter Microstructure and Risk for Pediatric Depression

Mai Uchida a,b,**, Yuwen Hung c,*, Allison Green a, Caroline Kelberman a, James Capella c, Schuylar L Gaillard c, John DE Gabrieli c, Joseph Biederman a,b
PMCID: PMC9073702  NIHMSID: NIHMS1799710  PMID: 34695702

Abstract

This study aimed to identify white-matter microstructural characteristics associated with risk for pediatric major depressive disorder (MDD) measured by the Child Behavior Checklist (CBCL) Anxiety/Depression scores. Children (N = 32) of both sexes, aged 6-12, underwent T1-weighted whole-head anatomical and diffusion-weighted imaging. Each participant’s mean diffusion measure image was generated and thinned to create an alignment-invariant tract representation. Voxel-wise analysis on the resulting map was carried out in Track Based Spatial Statistics (TBSS) using general linear models by regressing the CBCL-Anxiety/Depression score against measures of diffusion tensor imaging (DTI). We also compared these results with prior DTI findings from the same children associated with CBCL-Emotion Dysregulation profile, an indicator for bipolar disorder. TBSS voxel-wise analysis showed a significant negative correlation between fractional anisotropy (FA) and CBCL-Anxiety/Depression scores localized in the right anterior cingulum and connected corpus callosal region. The negative FA correlations in these regions were greater in CBCL-Anxiety/ Depression scores compared to CBCL-Emotional Dysregulation scores. Reduced white-matter connectivity in the anterior cingulum and connected corpus callosal region may represent a biomarker of risk for pediatric MDD. These results may help identify brain differences associated with the development of MDD, and assist with earlier clinical identification of pediatric MDD.

Keywords: Pediatric Depression, Major Depressive Disorder, Emotional Dysregulation, Bipolar Disorder, Diffusion Tensor Imaging, White Matter, Child Behavior Checklist

1. Introduction

There is an urgent need to identify neural risk biomarkers for pediatric major depressive disorder (MDD) to promote early identification and prevention of this morbid and prevalent disorder in youth (Birmaher, 2007; Jaycox et al., 2009; Keenan-Miller et al., 2007; Perou et al., 2013). Neuroimaging studies of MDD risk have often focused on children with familial risk with parents who have the disorder (Chai et al., 2016; Chai et al., 2015; Fischer et al., 2018; Foland-Ross et al., 2015; Gotlib et al., 2010; Hung et al., 2017; Joormann et al., 2012; Singh et al., 2018). These studies, which used functional and structural magnetic resonance imaging (MRI), reported that children with familial risk, compared to children without familial risk, exhibited differences in several brain networks, including increased functional connectivity between the default mode network and subgenual anterior cingulate cortex/orbital frontal cortex (Chai et al., 2016), increased activation in the amygdala and other brain regions in response to negative mood induction (Joorman et al., 2012) and negative facial expressions (Chai et al., 2015), and reduced grey matter volumes in right fusiform gyrus (Foland-Ross et al., 2015) and amygdala (Chai et al., 2015). Finally, diffusion tensor imaging (DTI) has revealed that children with parental MDD show abnormal microstructural maturation in frontolimbic white matter pathways (Hung et al., 2017).

Although the study of risk for pediatric MDD through children with a parental diagnosis of MDD is a useful strategy because such children have increased risk for developing MDD (Beardslee et al., 2003; Elsayed et al., 2019), these studies do not address the many cases of MDD that occur in children without a parental history of MDD. Heritability of MDD is estimated at 37% as evidenced by meta-analysis of twin studies (Sullivan et al., 2000), which suggest that while depression is a familial disorder, parental MDD accounts for only a portion of the risk for MDD and other clinical risk indicators also need to be considered when searching for neural risk biomarkers for MDD. Indeed, structural and functional connectivity differences in children have been associated with the future development of (Hawkey et al., 2018; Jalbrzikowski et al., 2017; Vulser et al., 2018). While these few studies have examined neural correlates of clinical indicators of MDD risk independent of family history, more research is necessary in order to address this gap in the available literature.

Scales from the Child Behavior Checklist (CBCL) have been validated as clinical indicators of risk for both pediatric MDD and pediatric bipolar disorder (Uchida et al., 2014; Uchida et al., 2018; Yule et al., 2019). For MDD, elevated scores on the CBCL-Anxiety/Depression scale in childhood specifically predicted the development of unipolar MDD, which would manifest 10 years later in young adulthood (Uchida et al., 2018). For bipolar disorder, elevated scores on the CBCL Emotion Dysregulation profile, a combined T-score of CBCL-Anxiety/Depression, Aggression, and Attention subscale scores, were associated with the future onset of pediatric bipolar disorder (Biederman et al., 2009; Uchida et al., 2014). In addition, elevated CBCL Emotion Dysregulation scores have been associated with reduced white-matter connectivity in cingulum-callosal regions (Hung et al., 2020).

In the present study, we asked whether there would be white-matter correlates of elevated CBCL-Anxiety/Depression scores in a group of children. We recruited children with a range of CBCL-Anxiety/Depression scores and related those scores to white-matter microstructure as measured by DTI. The children were not selected in terms of either family history or current diagnosis, and therefore reflected a range of scores. Additionally, we asked whether there would be a distinction in white-matter microstructure relations to the scores most related with risk for bipolar disorder versus those most related to risk for MDD. This study aligns with the NIH Research Domain Criteria (RDoC) (Insel et al., 2010) that expand from the conventional diagnosis-driven approach, which categorizes participants as meeting or not meeting criteria for a diagnosis, to a dimensional-driven approach using clinical scores that represents varying degrees of anxiety and depression symptomatology. Identifying neural substrates underlying clinical risks for MDD is of clinical, scientific and public health significance, because such knowledge could be used to provide neurodevelopmental targets for early identification and prevention of MDD.

2. Methods

2.1. Participants

We recruited children of both genders ages 6 to 12 years, with low to high levels of the CBCL-Anxiety/Depression scale from the community by hospital research outreach efforts via email lists, flyers and social media. Children were excluded if they presented with neuroimaging contraindications, or acute psychiatry instability such as suicidality, psychosis, or risk of harm to others. The PRISMA flow diagram (Figure 1A) shows detailed participant enrollment and selection process for data analysis, together with the final study sample’s clinical behavioral demographics (Figure 1B). We excluded 5 children who had excessive head motion (outliers defined as exceeding over 2 standard deviations from the mean). The final study sample consists of 32 children (mean age=9.53, SD=1.83; 16 boys and 16 girls) with low to high degree of Anxiety/Depression profiles. Twenty eight children (88%) identified themselves as Caucasian, 2 children identified themselves as Asian and the remaining 2 children identified themselves as having more than one ethnicity. Children who were ineligible for the MR scan were excluded from initial recruitment; data were further excluded for incomplete scans and excessive head motions. Among the final dataset of 32 children, 10 children were supplemented from a previous study’s control cohort (Chai et al., 2016; Chai et al., 2015) who meet the current study’s eligibility criteria. The effect of participants from different sources was statistically controlled as a covariate in the analyses.

Figure 1.

Figure 1.

Open in a new tab

PRISMA diagram of the study sample and the clinical demographics. (A) Breakdown of participant recruitment, screening procedures, and final eligible data for analysis. The final study sample consists of 32 children, including 10 children supplemented by a prior study’s control cohort that meets the current study’s eligibility criteria. (B) The demographic information of the final study sample. Note: Standard deviation is shown parentheses.

Because the current study focused on a dimensional rather than a categorical approach, we did not exclude children with lifetime history of depression. As such, the effect of participants with lifetime history of depression was statistically controlled as a covariate in the analyses. This study was approved by both the Institutional Review Board of MGH and the Institutional Review Board of Massachusetts Institute of Technology (MIT).

2.2. Child Behavior Checklist (CBCL)

The Child Behavior Checklist (CBCL) is a widely used clinical tool with empirically derived scales and excellent psychometric properties (Biederman et al., 2010; Holtmann et al., 2010; Petty et al., 2008; Yule et al., 2019). The 1991 version of the CBCL for ages 6 to 18 years was completed by the participants’ parents. It characterizes the child’s behavior in the past six months by parent-report, and the data are transformed into dimensional behavioral problem standard scores (Achenbach, 1991). The CBCL-Anxiety/Depression profile is characterized by an elevated T score on the Anxious/Depressed subscale of the CBCL. Examples of the behavioral aspects captured by the CBCL-Anxiety/Depression profile are illustrated in Figure 2. The CBCL- Emotion Dysregulation profile consists of an aggregate T-score of the three subscales including Aggression, Attention and Anxiety/Depression scales. Raw scores were converted to standardized T-scores based on age and gender by a computerized program. We used the T-scores for the main neuroimaging analysis to align with how the scale is used clinically, and repeated the analysis using the raw scores.

Figure 2.

Figure 2.

Open in a new tab

Sample items of the CBCL-Anxiety/Depression and Emotional Dysregulation scales. The standard T scores computed from these two CBCL profiles are used to uncover the neural underpinnings of risk behavioral markers of depression, in comparison to risks markers for bipolar mood disorders. In clinical settings, the CBCL Anxiety/Depression profile has been consistently found to be predictive of future onset of depression in children. The CBCL Emotional Dysregulation scale is predictive of future onset of bipolar mood disorders; this scale is a composite scale constituting three subscales of CBCL (Attention Problem (top green color samples), Aggression Behavior (middle blue color), and Anxiety/Depression (bottom red color)). Higher scores of these scales indicate greater extent of problems. The figure is produced with granted permission by Dr. Achenbach (author of CBCL).

2.3. Neuroimaging

All participants underwent MRI scanning, including T1-weighted whole-head anatomical and diffusion-weighted imaging scans in the same session, at the Athinoula A. Martinos Imaging Center at the McGovern Institute for Brain Research at MIT. Imaging data were acquired on a 3 Tesla Siemens Trio scanner (Siemens, Erlangen, Germany) using a 32-channel head coil. T1 MPRAGE sequence parameters included 1.1 x 1.1 mm2 in-plane resolution, 1.0 mm slice thickness, field of view (FOV) = 247 x 247 mm2, matrix = 220 x 220, 176 slices, four-echo sequence with TE = 1.57 ms, 3.33 ms, 5.09 ms, and 6.85 ms, and TR = 2.53 s. Prospective acquisition correction (PACE) was used to mitigate artifacts due to head motion. The diffusion-weighted scan sequence included 1 non-diffusion weighted reference volume (b = 0) and 30 diffusion directions (b = 700 s/mm2) with acquisition parameters: 2.0 x 2.0 mm2 in-plane resolution, 2.0 mm slice thickness, FOV = 256 x 256 mm2 chosen for full brain coverage, matrix = 128 x 128, TE = 84 ms, and TR= 8.04 s.

2.4. Diffusion Data Processing

All diffusion data were pre-processed by DTIPrep for quality control followed by TRACULA (TRActs Constrained by UnderLying Anatomy) (Yendiki et al., 2011). Images in each diffusion weighted imaging (DWI) series were aligned to the first non-diffusion-weighted image using affine registration (Jenkinson and Smith, 2001) and the corresponding diffusion-weighting gradient vectors were reoriented accordingly; (Leemans and Jones, 2009; Rohde et al., 2004) this procedure reduces misalignment between images due to head motion and eddy currents. TRACULA utilizes prior anatomical information of pathways from a set of training subjects and computes the white-matter pathway probability given the DWI data. The TRACULA-outputted data were then processed by TBSS (Track-Based Spatial Statistics) (Smith et al., 2006; Smith et al., 2004) The fractional anisotropy (FA) data were non-linearly aligned to a common space. FMRIB58_FA image was used as the target image for a linear registration to the standard space. Each participant’s mean diffusion measure image was generated and thinned to create an alignment-invariant tract representation (e.g., the ‘mean FA skeleton') representing the centers of all tracts common to the group. All participants’ diffusion data were then aligned on the skeleton space as 4D series and set at a threshold of 0.2 before statistical testing.

2.5. Quality Assurance

As quality control, four DTI motion measures were derived by TRACULA (Yendiki et al., 2014), including the average translation score, rotation score, signal drop-out score (percentage of bad slices), and signal drop-out severity (Benner et al., 2011; Yendiki et al., 2014). All participants’ translation scores were below 2 mm (mean = 0.85 ± 0.47) and rotation scores below 0.01 degree (mean = 0.008 ± 0.006). Normal signal drop-out percentage (~0 %) and severity (~1) were observed across all participants. None of these motion measure showed a significant correlation with the CBCL-Anxiety/Depression score (p ≥ 0.3). A composite head motion score for each participant was computed (Yendiki et al., 2014). In the following voxel-wise statistics, this head motion score, together with data source, age, gender, and lifetime history of mood disorder (including past and current) were modeled as the covariates to statistically control for any impacts arising from these variances.

2.6. Voxel-wise Analysis

First, a voxel-wise analysis on the skeletonized FA/MD/AD/RD map was carried out in TBSS using general linear models (GLM) by regressing the CBCL-Anxiety/Depression score against each diffusion measure (FA, RD, MD, and AD, respectively) throughout the whole brain to identify brain areas significantly correlated with CBCL-Anxiety/Depression scores. Second, for any significant correlation, to compare with our previous DTI finding with the CBCL-Emotional Dysregulation profile in the same study cohort (Hung et al., 2020), a white-matter mask was created combining all significant regions from both studies. Within this mask (region-of-interest), we carried out a voxel-wise two-way ANCOVA identifying regions with significant brain-behavior interactions, where the FA and RD relationships (slopes) differed in relation to the CBCL-Anxiety/Depression profile compared to the CBCL-Emotional Dysregulation profile as an indication of differentiated neural contribution to the risk of unipolar depression versus bipolar mood disorders. In all analyses, non-parametric randomized permutation test was performed (number of permutations = 5000) (Winkler et al., 2014), correcting for multiple comparisons using the threshold-free cluster enhancement method (Smith and Nichols, 2009) and controlling for the family-wise error rate (p <0.05). The JHU DTI-based atlases were used (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/Atlases) to determine white-matter locations of significant results. Significant TBSS result images were filled into the mean FA space for presentation purposes. For any TBSS-significant findings, individual mean DWI values (FA, RD, MD, or AD) averaged across all significant voxels in the standard space were calculated and shown in scatter plot for visualization purposes.

2.7. Tractography

To visualize the white-matter pathways associated with the TBSS-identified areas, probabilistic tractography was performed using the TBSS-significant regions as the tractography seeds. The FMRIB software library (FSL) tractography toolbox (FDT) (Behrens et al., 2003) was used for automated probabilistic reconstruction of major white-matter pathways from individual DWIs in native space. This method repetitively samples from the distributions on voxel-wise principal diffusion directions and computes a streamline through these local samples to generate a probabilistic streamline (sample) from the distribution on the location of the true streamline. The local diffusion directions were calculated using the BEDPOSTX toolbox that allowed modeling multiple fiber orientations per voxel. The output images of individual connectivity paths were corrected by individual total counts of established samples by dividing each voxel value by the waytotal number, correcting for individual variations. Lastly, the individual path images were nonlinearly registered to the standard space for group averaging for visualization.

3. Results

The demographic information of the final sample is shown in Figure 1B. The mean age was 9.53 years (SD = 1.83). Seventeen males and 15 females were included. The mean CBCL-Anxiety/Depression T-score was 59.1 (SD = 11.6) and the mean raw score was 4.96 (SD = 5.19).

TBSS voxel-wise analysis of all 32 children showed a significant negative correlation between FA and CBCL-Anxiety/Depression scores (P < 0.05, corrected for multiple comparisons) localized in the right anterior cingulum (aCG) and connected corpus callosal (aCC) region (Figure 3A). Individual mean FA extracted from the TBSS-significant regions visualized the significant negative relationship between FA values and CBCL-Anxiety/Depression scores and (Figure 3B; Spearman’s Rho = 0.45). No other significant results were found with MD, RD, or AD measures. These results were replicated using CBCL Anxiety/Depression raw scores (Supplementary Figure 1A, B).

Figure 3.

Figure 3.

Open in a new tab

TBSS whole-brain result and tractography reconstructed path associated with CBCL-Anxiety/Depression scores in children. (A) TBSS localized the right anterior cingulum (aCG) and anterior corpus callosal (aCC) white matter regions (red clusters) that were significantly and negatively correlated with CBCL-Anxiety/Depression scores.. Probabilistic tractography reconstructed the connection path (in yellow) from the TBSS-identified seed region, identifying the anterior forceps (AF) and connected anterior corona radiata (aCR) extending into prefrontal cortex. Image threshold at P < .05 voxel-wise. Findings shown on FMRIB58 1mm FA template and in MNI coordinate system.. (B) The scatter plot shows the significant negative relationship between the individual mean FA values extracted from the TBSS-identified and filled FA region (Y-axis) plotted CBCL Anxious/Depression T scores (X-axis, with 50 being the lowest cutoff score), with the prediction trend line (middle line) and 95% confidence interval (between the curved lines). (C, D) 3D rendering of white matter path reconstructed from the TBSS associations with CBCL Anxious/Depression T scores. Probabilistic tractography visualizes the connectivity distribution path (in yellow) extending from the TBSS-identified region (anterior cingulum/aCG, anterior corpus callosum/aCC) primarily into prefrontal cortex (PFC). Images are overlaid on the standard 3D FA map (C: FMRIB58 1mm FA template) and superimposed over the standard 3D T1 map (D: MNI152 1mm T1 template). Reconstructed path images were thresholded by a minimum of 30 % of the maximum established total streamline connections.

Probabilistic tractography successfully reconstructed the white matter pathways connected with the TBSS-significant anterior cingulum-callosal seed region, visualizing the prefrontal white matter tract including the anterior forceps (AF) and connected anterior corona radiata (aCR) which extended into the prefrontal cortex (Figure 3A, C, D).

The current finding of negative correlation between CBCL-Anxiety/Depression scores and FA at the anterior cingulum-callosal area (Figure 4A) overlaps with our previous study where we found a significant negative correlation between the CBCL-Emotional Dysregulation scores and FA in the same study cohort (Hung et al., 2020). We created a white-matter mask combining all significant voxels in the current and the previous analyses ( Figure 4B), and within this mask of region-of-interest (ROI), we carried out a voxel-wise 2-way ANCOVA to identified differential brain-behavior relationships associated with the CBCL-Anxiety/Depression versus the Emotional Dysregulation scores as an indication of differentiated neural contribution of risk of unipolar depression versus bipolar mood disorders. The result showed a significant dissociation in brain-behavior relationships, in that a significantly stronger negative association was found between FA and CBCL-Anxiety/Depression scores compared with the association between FA and CBCL-Emotional Dysregulation scores, localized to the right anterior cingulum-callosal conjunction (aCC+aCG; Figure 4C), which overlapped with our main report above. In addition, the ROI analysis on the RD revealed a significantly stronger positive association between RD and CBCL-Anxiety/Depression scores than that between RD and CBCL-Emotional Dysregulation scores, localized to the same region in the aCC+aCG region (Figure 4D).

Figure 4.

Figure 4.

Open in a new tab

Comparisons of brain-behavior associations. (A) The CBCL Anxiety/Depression scores were significantly and negatively correlated with FA, localized to the anterior cingulum-callosal (CC-CG) region in the current study (red cluster). (B) In our prior study, CBCL Emotional Dysregulation (ED) scores were significantly and negatively correlated with the FA and positively with the RD, localized in the CC-CG bundles (yellow clusters), with primary peaks located in both the anterior and posterior subdivisions of the CC-CG bundles (connected to adjacent regions) (Hung et al., 2020). (C) In the current study, within all the significant regions from the current and the prior studies (shown as the semi-transparent yellow-color mask), two-way ANCOVA further reveals that the anterior CC-CG regions (red cluster) shows a significantly lower degree of FA association with the CBCL Anxiety/Depression score compared with the Emotional Dysregulation score, suggesting a unique role of the anterior subdivision of the CC-CG pathways contributing to the risk of depression as assessed by CBCL Anxiety/Depression profile. (D) Consistently, the anterior CC-CG (blue cluster) also shows significantly lower degree of RD association with CBCL Anxiety/Depression profile. Bottom brain image: FMRIB58 1mm standard FA template image. Green colors lines: group-invariant white matter skeleton map. The JHU DTI-based atlases are used to determine white-matter locations of significant results (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/Atlases).

4. Discussion

Our study aimed to examine whether neural biomarkers signifying risk for pediatric MDD could be identified. To this end, we examined the association between microstructural white-matter characteristics and scores on the CBCL-Anxiety/Depression scale which are known to longitudinally predict the development of pediatric MDD (Uchida et al., 2018). There were significant negative correlations between higher (worse) CBCL-Anxiety/Depression scores and lower FA values in the anterior cingulum and connected corpus callosal region. FA measures the strength of directionality and is often interpreted as reflecting the magnitude of neural connectivity along local white-matter tract. Decreases in FA coupled with increases in RD implicates potential underlying myelination-related problems (Song et al., 2002). The current findings suggest that weaker structural connectivity in the anterior CC-CG region is associated with elevated risk for developing pediatric MDD.

The present findings align with other studies by indicating that risk for depression is associated with structural variation in the anterior fronto-limbic system. Children at familial risk for depression exhibited atypical cross-sectional development of white-matter in this region (Hung et al., 2016). Lower FA in tracts projecting from the corpus callosum to the anterior cingulate cortex were associated with higher risk of transition to depression in adolescents (Vulser et al., 2018). Adolescent daughters of mothers with recurrent depression exhibited abnormal gray matter development, as measured by cortical thickness, in the anterior cingulate cortex and anterior insula regions (Foland-Ross et al., 2015).

The associations between white matter microstructure and CBCL risk scores identified in the current report can be interpreted in the context of prior findings about the white-matter pathways where the associations were localized. The cingulum and callosal bundles constitute the largest white matter tracts in the brain, connecting dorsal and ventral cortical regions, particularly in the limbic system (Catani and Thiebaut de Schotten, 2012; Schmahmann and Pandya, 2006). These neural pathways are implicated in cognitive attribution and regulation of emotional processing, which plays a key role in emotional regulation (Buhle et al., 2014; Ochsner and Gross, 2005; Vandekerckhove et al., 2020). The callosal fibers interdigitate and connect with the cingulum fibers (Bubb et al., 2018), which receive neuronal inputs from the cingulate cortex (Catani and Thiebaut de Schotten, 2012), the brain region implicated in the cognitive regulation of emotional processes that are altered in mood disorders (Bush et al., 2000; Giuliani et al., 2011; Versace et al., 2015; Wessa and Linke, 2009). The anterior CC-CG fibers receive axonal input from the anterior cingulate cortex (Catani et al., 2002), which is a brain region involved in regulation and control of emotional processes (Bush et al., 2000) and is related to psychotherapy outcome for MDD (Carl et al., 2016). Therefore, the current finding, suggests that variation in CC-CG microstructure which was linked to increased risk of MDD, may involve impairment of a cognitive-emotional regulatory mechanism.

The white-matter tract differences associated with CBCL Anxiety/Depression scores could reflect initial vulnerabilities to develop anxiety or depression, or the impact of experience with subclinical or clinical anxiety or depression, or both. Because children presenting with subclinical elevations of CBCL Anxiety/Depression scores have been found to have increased risks of developing depression in the future (Uchida et al., 2018), the CBCL Anxiety/Depression score elevation is considered a clinical risk indicator for future depression. Therefore, regardless as whether the DTI differences were a cause or a consequence of anxiety and depression symptoms, the neural findings associated with elevations of CBCL Anxiety/Depression scores (a clinical risk indicator) could be considered as a potential biomarker of the risk for depression.

There is an inherent overlap between the CBCL predictors of MDD and bipolar disorder because the CBCL-Anxiety/Depression score associated with the future development of MDD is part of the CBCL Emotional Dysregulation profile (aggregate of CBCL Anxiety/Depression, Aggression, and Attention scales), which is associated with the future development of bipolar disorder. The present study reveals that decreases in structural connectivity in the anterior cingulate region is more strongly associated with the CBCL-Anxiety/Depression scores than with the Emotional Dysregulation profile, further suggesting that this could be a neural marker for the risk for depression. Relatedly, a study that examined the differentiation of brain activation in pediatric bipolar disorder and MDD found that children with MDD had reduced activation in anterior cingulate cortex compared to those with bipolar disorder (Burger et al., 2017). Considering the clinical overlap of major depression and bipolar disorder symptomatology, the partial similarities in their neural underpinnings are excepted.

Another result that separated our current findings associated with the CBCL-Anxiety/Depression scale and our previous findings associated with the CBCL-Emotion Dysregulation profile was that the DTI findings associated with CBCL-Emotion Dysregulation profile, which documented negative correlations between those scores and FA in both the posterior cingulum and posterior cerebral cortex regions, were not seen in our current study examination of children with CBCL-Anxiety/Depression elevation. If these results are replicated, these findings could suggest that there is a common neural biomarker signifying risk for both MDD and bipolar disorder associated with perturbations of the anterior fronto-limbic system. However, the risk for pediatric bipolar disorder may be uniquely associated with additional posterior cortical-limbic system perturbation of white matter tracts.

In our previous work, we found that the posterior CC-CG brain region (as the peak location of the results) was associated with the risk for pediatric bipolar disorders as assessed by the CBCL-ED profile. In the current study, when comparing with the risk measure of unipolar disorders, this posterior CC-CG location was not differentially associated with risk of bipolar disorders. This may be due to the inherent fact that the diagnosis of bipolar disorders include symptoms (episodes) of unipolar disorders. At the clinical course and diagnostic level, the two disorders overlap and are not exclusively dissociated with each other, which limits the ability of dissociating risk neuromarkers based on diagnoses. Nevertheless, multiple imaging studies comparing pediatric bipolar disorder to unipolar depression have found that abnormalities in the posterior cingulate gyrus and/or the cingulate cortex were associated with pediatric bipolar disorder, but not pediatric MDD. One study found that youth with unipolar and bipolar depression with higher scores on the Bipolarity Index had less default mode network resting-state connectivity in the postcentral gyrus and posterior cingulate cortex (Ford et al., 2013). Similarly, another study found reduced FA in the posterior cingulate cortex in bipolar disorder patients compared MDD patients (Repple et al., 2017). Anatomically, the posterior CC-CG location connects the posterior cingulate with posterior cortical systems (including the motor and parietal attention systems) with the limbic system (e.g., via the cingulum’s posterior-to-inferior, angular bundle). These brain circuits potentially contribute to an attentional-driven and motor-based cognitive-emotional regulatory mechanism, relative to the frontal-based, executive regulatory top-down control mechanism underlying mood disorders. Further research efforts on finding predictive neural biomarkers for mood disorders may benefit from considering a transdiagnostic method to identify more sensitive neuromarkers that distinguish between the unipolar depression versus bipolar disorders, or more specifically, between the depressive and manic episodes that constitute the two differing disorders.

The identification of a neural biomarker to assist in the differentiation between developing MDD versus bipolar disorder may provide the field with a novel method to assist clinicians in their diagnosis and treatment of pediatric mood disorder patients. Earlier and correct identification of pediatric mood disorders can prevent adverse outcomes, increase patient recovery rates, and improve a patient’s overall quality of life.

Our results need to be viewed with the consideration of some methodological limitations. The sample size was relatively small limiting our statistical power to fully examine all brain differences. The sample was overwhelmingly Caucasian and further studies are needed to examine whether these findings generalize to other races and ethnicities. The CBCL is a parent reported scale and could be subjected bias based on the reporters’ impression of the child’s condition (although the CBCL scales have been linked to specific outcomes of MDD and bipolar disorder). Parental education and family income were not taken into consideration in our examination. While clinical data suggests that subthreshold CBCL-Anxiety/Depression scores predicted the development of MDD, in this study, due to the issue of power, we analyzed all scores of CBCL-Anxiety/Depression. DTI measures used did not fully characterize crossing fibers, thus we are limited to interpreting the regions that could intersect with corona radiata and cortical spinal tract. Future research using more advanced microstructural measures (e.g., diffusion spectrum imaging) may further characterize potential attribution of brain regions of risk for depression.

Despite these considerations, our study documents that symptoms identified through the CBCL-Anxiety/Depression scale, a clinical risk indicator for development of depression, are associated with weaker structural connectivity in the anterior cingulate and callosal areas extending into prefrontal cortical tracts. These results together with our previous finding examining the neural underpinning associated with the risk for pediatric bipolar disorder suggest that dysfunction in these frontal regions of the brain could represent biomarkers of risk for pediatric mood disorders worthy of further investigation.

Supplementary Material

Raw Score Analysis

Highlights.

  • Weaker connectivity in the anterior cingulum is associated with risk for depression

  • Weaker connectivity in the corpus callosum is also correlated with depression risk

  • This dysfunction represents a potential biomarker of risk for pediatric depression

  • These biomarkers may assist with early identification and treatment of depression

Funding

This work was supported by the Brain and Behavior Research Foundation, the Canadian Institutes of Health Research (CIHR), and the Poitras Center for Psychiatric Disorders Research at the McGovern Institute for Brain Research at MIT. This work was further supported by the Dupont Warren Fellowship, Livingston Fellowship, William F. Milton Fund, and the Mass General for Children Pilot and Feasibility awards.

Declaration of Interests

Dr. Joseph Biederman is currently receiving research support from the following sources: AACAP, Feinstein Institute for Medical Research, Food & Drug Administration, Genentech, Headspace Inc., NIDA, Pfizer Pharmaceuticals, Roche TCRC Inc., Shire Pharmaceuticals Inc., Sunovion Pharmaceuticals Inc., Tris, and NIH. Dr. Biederman’s program has received departmental royalties from a copyrighted rating scale used for ADHD diagnoses, paid by Bracket Global, Ingenix, Prophase, Shire, Sunovion, and Theravance; these royalties were paid to the Department of Psychiatry at MGH. In 2020: Through MGH corporate licensing, Dr. Biederman has a US Patent (#14/027,676) for a non-stimulant treatment for ADHD, a US Patent (#10,245,271 B2) on a treatment of impaired cognitive flexibility, and a patent pending (#61/233,686) on a method to prevent stimulant abuse. He receives honoraria from the MGH Psychiatry Academy for tuition-funded CME courses. In 2019, Dr. Biederman was a consultant for Akili, Avekshan, Jazz Pharma, and Shire/Takeda. He received research support from Lundbeck AS and Neurocentria Inc. Through MGH CTNI, he participated in a scientific advisory board for Supernus. He received honoraria from the MGH Psychiatry Academy for tuition-funded CME courses. In 2018, Dr. Biederman was a consultant for Akili and Shire. He received honoraria from the MGH Psychiatry Academy for tuition-funded CME courses.

Dr. Mai Uchida is partially supported by a K award, grant number 1K23MH122667-01.

Dr. Hung, Ms. Allison Green, Ms. Caroline Kelberman, Mr. James Capella, Ms. Gaillard, and Dr. John D. Gabrieli do not have any financial relationships to disclose.

References

  1. Achenbach TM, 1991. Manual for the Child Behavior Checklist/4-18 and the 1991 Profile. University of Vermont, Department of Psychiatry, Burlington, VT. [Google Scholar]
  2. Achenbach TM, Rescorla LA, 2001. Manual for ASEBA School-Age Forms & Profiles. University of Vermont, Research Center for Children, Youth, & Families, Burlington, VT. [Google Scholar]
  3. Beardslee WR, Gladstone TR, Wright EJ, Cooper AB, 2003. A family-based approach to the prevention of depressive symptoms in children at risk: evidence of parental and child change. Pediatrics 112, e119–131. [DOI] [PubMed] [Google Scholar]
  4. Benner T, van der Kouwe AJ, Sorensen AG, 2011. Diffusion imaging with prospective motion correction and reacquisition. Magn Reson Med 66, 154–167. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Biederman J, Petty CR, Fried R, Wozniak J, Micco JA, Henin A, Doyle R, Joshi G, Galdo M, Kotarski M, Caruso J, Yorks D, Faraone SV, 2010. Child behavior checklist clinical scales discriminate referred youth with autism spectrum disorder: a preliminary study. J Dev Behav Pediatr 31, 485–490. [DOI] [PubMed] [Google Scholar]
  6. Biederman J, Petty CR, Monuteaux MC, Evans M, Parcell T, Faraone SV, Wozniak J, 2009. The child behavior checklist-pediatric bipolar disorder profile predicts a subsequent diagnosis of bipolar disorder and associated impairments in ADHD youth growing up: a longitudinal analysis. J Clin Psychiatry 70, 732–740. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Birmaher B, 2007. Longitudinal Course of Pediatric Bipolar Disorder. Am J Psychiatry 164, 537–539. [DOI] [PubMed] [Google Scholar]
  8. Bubb EJ, Metzler-Baddeley C, Aggleton JP, 2018. The cingulum bundle: Anatomy, function, and dysfunction. Neurosci Biobehav Rev 92, 104–127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Burger C, Redlich R, Grotegerd D, Meinert S, Dohm K, Schneider I, Zaremba D, Forster K, Alferink J, Bolte J, Heindel W, Kugel H, Arolt V, Dannlowski U, 2017. Differential Abnormal Pattern of Anterior Cingulate Gyrus Activation in Unipolar and Bipolar Depression: an fMRI and Pattern Classification Approach. Neuropsychopharmacology 42, 1399–1408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bush G, Luu P, Posner MI, 2000. Cognitive and emotional influences in anterior cingulate cortex. Trends Cogn Sci 4, 215–222. [DOI] [PubMed] [Google Scholar]
  11. Carl H, Walsh E, Eisenlohr-Moul T, Minkel J, Crowther A, Moore T, Gibbs D, Petty C, Bizzell J, Dichter GS, Smoski MJ, 2016. Sustained anterior cingulate cortex activation during reward processing predicts response to psychotherapy in major depressive disorder. J Affect Disord 203, 204–212. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Catani M, Howard RJ, Pajevic S, Jones DK, 2002. Virtual in vivo interactive dissection of white matter fasciculi in the human brain. Neuroimage 17, 77–94. [DOI] [PubMed] [Google Scholar]
  13. Catani M, Thiebaut de Schotten M, 2012. Atlas of human brain connections. Oxford University Press, Oxford; ; New York. [Google Scholar]
  14. Chai XJ, Hirshfeld-Becker D, Biederman J, Uchida M, Doehrmann O, Leonard JA, Salvatore J, Kenworthy T, Brown A, Kagan E, de Los Angeles C, Gabrieli JD, Whitfield-Gabrieli S, 2016. Altered Intrinsic Functional Brain Architecture in Children at Familial Risk of Major Depression. Biol Psychiatry 80, 849–858. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Chai XJ, Hirshfeld-Becker D, Biederman J, Uchida M, Doehrmann O, Leonard JA, Salvatore J, Kenworthy T, Brown A, Kagan E, de Los Angeles C, Whitfield-Gabrieli S, Gabrieli JD, 2015. Functional and structural brain correlates of risk for major depression in children with familial depression. Neuroimage Clin 8, 398–407. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Elsayed NM, Fields KM, Olvera RL, Williamson DE, 2019. The role of familial risk, parental psychopathology, and stress for first-onset depression during adolescence. J Affect Disord 253, 232–239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Fischer AS, Camacho MC, Ho TC, Whitfield-Gabrieli S, Gotlib IH, 2018. Neural Markers of Resilience in Adolescent Females at Familial Risk for Major Depressive Disorder. JAMA Psychiatry 75, 493–502. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Foland-Ross LC, Gilbert BL, Joormann J, Gotlib IH, 2015. Neural markers of familial risk for depression: An investigation of cortical thickness abnormalities in healthy adolescent daughters of mothers with recurrent depression. J Abnorm Psychol 124, 476–485. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Ford KA, Theberge J, Neufeld RJ, Williamson PC, Osuch EA, 2013. Correlation of brain default mode network activation with bipolarity index in youth with mood disorders. J Affect Disord 150, 1174–1178. [DOI] [PubMed] [Google Scholar]
  20. Giuliani NR, Drabant EM, Gross JJ, 2011. Anterior cingulate cortex volume and emotion regulation: is bigger better? Biol Psychol 86, 379–382. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Gotlib IH, Hamilton JP, Cooney RE, Singh MK, Henry ML, Joormann J, 2010. Neural processing of reward and loss in girls at risk for major depression. Arch Gen Psychiatry 67, 380–387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hawkey EJ, Tillman R, Luby JL, Barch DM, 2018. Preschool Executive Function Predicts Childhood Resting-State Functional Connectivity and Attention-Deficit/Hyperactivity Disorder and Depression. Biol Psychiatry Cogn Neurosci Neuroimaging 3, 927–936. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Holtmann M, Buchmann AF, Esser G, Schmidt MH, Banaschewski T, Laucht M, 2010. The Child Behavior Checklist-Dysregulation Profile predicts substance use, suicidality, and functional impairment: a longitudinal analysis. J Child Psychol Psychiatry 52, 139–147. [DOI] [PubMed] [Google Scholar]
  24. Hung Y, Saygin ZM, Biederman J, Hirshfeld-Becker D, Uchida M, Doehrmann O, Han M, Chai XJ, Kenworthy T, Yarmak P, Gaillard SL, Whitfield-Gabrieli S, Gabrieli JDE, 2017. Impaired Frontal-Limbic White Matter Maturation in Children at Risk for Major Depression. Cereb Cortex 27, 4478–4491. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Hung Y, Uchida M, Gaillard SL, Woodworth H, Kelberman C, Capella J, Kadlec K, Goncalves M, Ghosh S, Yendiki A, Chai XJ, Hirshfeld-Becker DR, Whitfield-Gabrieli S, Gabrieli JDE, Biederman J, 2020. Cingulum-Callosal white-matter microstructure associated with emotional dysregulation in children: A diffusion tensor imaging study. Neuroimage Clin 27, 102266. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Jalbrzikowski M, Larsen B, Hallquist MN, Foran W, Calabro F, Luna B, 2017. Development of White Matter Microstructure and Intrinsic Functional Connectivity Between the Amygdala and Ventromedial Prefrontal Cortex: Associations With Anxiety and Depression. Biol Psychiatry 82, 511–521. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jaycox LH, Stein BD, Paddock S, Miles JN, Chandra A, Meredith LS, Tanielian T, Hickey S, Burnam MA, 2009. Impact of teen depression on academic, social, and physical functioning. Pediatrics 124, e596–605. [DOI] [PubMed] [Google Scholar]
  28. Jenkinson M, Smith S, 2001. A global optimisation method for robust affine registration of brain images. Med Image Anal 5, 143–156. [DOI] [PubMed] [Google Scholar]
  29. Joormann J, Cooney RE, Henry ML, Gotlib IH, 2012. Neural correlates of automatic mood regulation in girls at high risk for depression. J Abnorm Psychol 121, 61–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Keenan-Miller D, Hammen CL, Brennan PA, 2007. Health outcomes related to early adolescent depression. J Adolesc Health 41, 256–262. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Leemans A, Jones DK, 2009. The B-matrix must be rotated when correcting for subject motion in DTI data. Magn Reson Med 61, 1336–1349. [DOI] [PubMed] [Google Scholar]
  32. Perou R, Bitsko RH, Blumberg SJ, Pastor P, Ghandour RM, Gfroerer JC, Hedden SL, Crosby AE, Visser SN, Schieve LA, Parks SE, Hall JE, Brody D, Simile CM, Thompson WW, Baio J, Avenevoli S, Kogan MD, Huang LN, 2013. Mental Health Surveillance Among Children - United States, 2005-2011. Morbidity and Mortality Weekly Report 63, 1–35. [PubMed] [Google Scholar]
  33. Petty CR, Rosenbaum JF, Hirshfeld-Becker DR, Henin A, Hubley S, LaCasse S, Faraone SV, Biederman J, 2008. The child behavior checklist broad-band scales predict subsequent psychopathology: A five-year follow-up. J Anxiety Disord 22, 532–539. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Repple J, Meinert S, Grotegerd D, Kugel H, Redlich R, Dohm K, Zaremba D, Opel N, Buerger C, Forster K, Nick T, Arolt V, Heindel W, Deppe M, Dannlowski U, 2017. A voxel-based diffusion tensor imaging study in unipolar and bipolar depression. Bipolar Disord 19, 23–31. [DOI] [PubMed] [Google Scholar]
  35. Rohde GK, Barnett AS, Basser PJ, Marenco S, Pierpaoli C, 2004. Comprehensive approach for correction of motion and distortion in diffusion-weighted MRI. Magn Reson Med 51, 103–114. [DOI] [PubMed] [Google Scholar]
  36. Schmahmann JD, Pandya DN, 2006. Fiber pathways of the brain. Oxford University Press, Oxford; ; New York. [Google Scholar]
  37. Singh MK, Leslie SM, Packer MM, Weisman EF, Gotlib IH, 2018. Limbic Intrinsic Connectivity in Depressed and High-Risk Youth. J Am Acad Child Adolesc Psychiatry 57, 775–785 e773. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Smith SM, Jenkinson M, Johansen-Berg H, Rueckert D, Nichols TE, Mackay CE, Watkins KE, Ciccarelli O, Cader MZ, Matthews PM, Behrens TE, 2006. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 31, 1487–1505. [DOI] [PubMed] [Google Scholar]
  39. Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TE, Johansen-Berg H, Bannister PR, De Luca M, Drobnjak I, Flitney DE, Niazy RK, Saunders J, Vickers J, Zhang Y, De Stefano N, Brady JM, Matthews PM, 2004. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 23 Suppl 1, S208–219. [DOI] [PubMed] [Google Scholar]
  40. Smith SM, Nichols TE, 2009. Threshold-free cluster enhancement: addressing problems of smoothing, threshold dependence and localisation in cluster inference. Neuroimage 44, 83–98. [DOI] [PubMed] [Google Scholar]
  41. Song SK, Sun SW, Ramsbottom MJ, Chang C, Russell J, Cross AH, 2002. Dysmyelination revealed through MRI as increased radial (but unchanged axial) diffusion of water. Neuroimage 17, 1429–1436. [DOI] [PubMed] [Google Scholar]
  42. Sullivan PF, Neale MC, Kendler KS, 2000. Genetic epidemiology of major depression: review and meta-analysis. American Journal of Psychiatry 157, 1552–1562. [DOI] [PubMed] [Google Scholar]
  43. Uchida M, Faraone SV, Martelon M, Kenworthy T, Woodworth KY, Spencer TJ, Wozniak JR, Biederman J, 2014. Further evidence that severe scores in the aggression/anxiety-depression/attention subscales of child behavior checklist (severe dysregulation profile) can screen for bipolar disorder symptomatology: a conditional probability analysis. J Affect Disord 165, 81–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Uchida M, Fitzgerald M, Woodworth H, Carrellas N, Kelberman C, Biederman J, 2018. Subsyndromal Manifestations of Depression in Children Predict the Development of Major Depression. J Pediatr 201, 252–258. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Versace A, Acuff H, Bertocci MA, Bebko G, Almeida JR, Perlman SB, Leemans A, Schirda C, Aslam H, Dwojak A, Bonar L, Travis M, Gill MK, Demeter C, Diwadkar VA, Sunshine JL, Holland SK, Kowatch RA, Birmaher B, Axelson D, Horwitz SM, Frazier TW, Arnold LE, Fristad MA, Youngstrom EA, Findling RL, Phillips ML, 2015. White matter structure in youth with behavioral and emotional dysregulation disorders: a probabilistic tractographic study. JAMA Psychiatry 72, 367–376. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Vulser H, Paillere Martinot ML, Artiges E, Miranda R, Penttila J, Grimmer Y, van Noort BM, Stringaris A, Struve M, Fadai T, Kappel V, Goodman R, Tzavara E, Massaad C, Banaschewski T, Barker GJ, Bokde ALW, Bromberg U, Bruhl R, Buchel C, Cattrell A, Conrod P, Desrivieres S, Flor H, Frouin V, Gallinat J, Garavan H, Gowland P, Heinz A, Nees F, Papadopoulos-Orfanos D, Paus T, Poustka L, Rodehacke S, Smolka MN, Walter H, Whelan R, Schumann G, Martinot JL, Lemaitre H, Consortium I, 2018. Early Variations in White Matter Microstructure and Depression Outcome in Adolescents With Subthreshold Depression. Am J Psychiatry 175, 1255–1264. [DOI] [PubMed] [Google Scholar]
  47. Wessa M, Linke J, 2009. Emotional processing in bipolar disorder: behavioural and neuroimaging findings. Int Rev Psychiatry 21, 357–367. [DOI] [PubMed] [Google Scholar]
  48. Winkler AM, Ridgway GR, Webster MA, Smith SM, Nichols TE, 2014. Permutation inference for the general linear model. Neuroimage 92, 381–397. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Yendiki A, Koldewyn K, Kakunoori S, Kanwisher N, Fischl B, 2014. Spurious group differences due to head motion in a diffusion MRI study. Neuroimage 88, 79–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Yendiki A, Panneck P, Srinivasan P, Stevens A, Zollei L, Augustinack J, Wang R, Salat D, Ehrlich S, Behrens T, Jbabdi S, Gollub R, Fischl B, 2011. Automated probabilistic reconstruction of white-matter pathways in health and disease using an atlas of the underlying anatomy. Front Neuroinform 5, 23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Yule A, Fitzgerald M, Wilens T, Wozniak J, Woodworth KY, Pulli A, Uchida M, Faraone SV, Biederman J, 2019. Further evidence of the Diagnostic Utility of the Child Behavior Checklist for identifying pediatric Bipolar I Disorder. SJCAPP 7, 1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Raw Score Analysis
NIHMS1799710-supplement-Raw_Score_Analysis.docx (83.6KB, docx)

RESOURCES