Distinct Patterns of Brain Morphology associate with Specific Dimensions of Child Psychopathology
Poster No:
356
Submission Type:
Abstract Submission
Authors:
Lei Cao1, Stephanie Gorka1, K. Luan Phan1, Lei Wang1
Institutions:
1The Ohio State University, Columbus, OH
First Author:
Co-Author(s):
Introduction:
To better understand the shared also unique neural substrates underlying child psychopathology, we investigated the Child Behavior Checklist (CBCL) and decomposed CBCL into one general psychopathology factor and three specific symptom factors (Conduct Problems, ADHD and Internalizing Problems).
Following the guideline of National Institute of Mental Health (NIMH) Research Domain Criteria (RDoC) that explores brain function through a dimensional perspective, we examined the relationship between the psychopathology factors and brain morphology measures using all children's baseline data from Adolescent Brain Cognitive Development (ABCD) study regardless of their clinical diagnosis.
Following the guideline of National Institute of Mental Health (NIMH) Research Domain Criteria (RDoC) that explores brain function through a dimensional perspective, we examined the relationship between the psychopathology factors and brain morphology measures using all children's baseline data from Adolescent Brain Cognitive Development (ABCD) study regardless of their clinical diagnosis.
Methods:
The tabulated baseline data for CBCL item scores, FreeSurfer-derived (Destrieux atlas) cortical regional brain volume, surface area, and thickness from the multisite ABCD study (release 5.0) was used in the analysis. The data included a total of 11693 subjects (9.48 ± 0.50 years old, 47.7% female).
We performed factor analysis on the CBCL item scores following the methodology presented by Moore et al. (2020). Briefly, 8 items were removed due to low correlation with other items and 3 pairs of items were combined due to high correlation with each other. Exploratory factor analysis was first performed on random half of the ABCD sample, revealing 3 specific symptom factors: Conduct Problems, ADHD and Internalizing Problems. Subsequent confirmatory factor analysis was conducted on the remaining half sample defining one general psychopathology factor and 3 previously found symptom factors. All four factors were orthogonal to each other.
Separate multivariate linear mixed effects models were performed on cortical regional brain volume, surface area and thickness. All above obtained factors were included in the models with additional covariate of age, sex, race/ethnicity and scanner model to examine the relationship between each of the factors and brain morphology. Family id was included as random intercept. The resulting t-value was thresholded at p<0.05 (FDR-corrected).
Both factor analysis and multivariate linear mixed effects models take into account of the clustering of site and family by utilizing imputed raked propensity weight (Heeringa et al., 2020).
We performed factor analysis on the CBCL item scores following the methodology presented by Moore et al. (2020). Briefly, 8 items were removed due to low correlation with other items and 3 pairs of items were combined due to high correlation with each other. Exploratory factor analysis was first performed on random half of the ABCD sample, revealing 3 specific symptom factors: Conduct Problems, ADHD and Internalizing Problems. Subsequent confirmatory factor analysis was conducted on the remaining half sample defining one general psychopathology factor and 3 previously found symptom factors. All four factors were orthogonal to each other.
Separate multivariate linear mixed effects models were performed on cortical regional brain volume, surface area and thickness. All above obtained factors were included in the models with additional covariate of age, sex, race/ethnicity and scanner model to examine the relationship between each of the factors and brain morphology. Family id was included as random intercept. The resulting t-value was thresholded at p<0.05 (FDR-corrected).
Both factor analysis and multivariate linear mixed effects models take into account of the clustering of site and family by utilizing imputed raked propensity weight (Heeringa et al., 2020).
Results:
A wide-spread negative association was discovered between Conduct Problems and cortical regional volume. This negative association was mostly originated from surface area. In contrast, ADHD showed a negative correlation with cortical volume stemming from a combination of decreasing in surface area and cortical thickness. Finally, Internalizing Problems showed a pattern of positive association with cortical volume reflected in increases in surface area.
Conclusions:
Our results showed distinct patterns of association between brain morphology and different symptom factors, suggesting unique neurodevelopmental brain structure profiles associate with different child psychopathology phenotypes. This unique structural abnormality may serve as potential biomarkers for predicting the emergence of psychopathology later in life.
Disorders of the Nervous System:
Neurodevelopmental/ Early Life (eg. ADHD, autism) 1
Psychiatric (eg. Depression, Anxiety, Schizophrenia)
Novel Imaging Acquisition Methods:
Anatomical MRI 2
Keywords:
Attention Deficit Disorder
Development
Psychiatric Disorders
STRUCTURAL MRI
1|2Indicates the priority used for review
Provide references using author date format
Heeringa, S. G., & Berglund, P. A. (2020). A guide for population-based analysis of the Adolescent Brain Cognitive Development (ABCD) Study baseline data. BioRxiv, 2020-02.