A Shared Multivariate Brain-Behavior Relationship in a Transdiagnostic Sample of Adolescents

Johannah Bashford-Largo, Hajer Nakua, R. James R. Blair, Ahria Dominguez, Melissa Hatch, Karina S. Blair, Matthew Dobbertin, Stephanie Ameis, Sahil Bajaj

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Background: Internalizing and externalizing psychopathology typically present in early childhood and can have negative implications on general functioning and quality of life. Prior work has linked increased psychopathology symptoms with altered brain structure. Multivariate analysis such as partial least squares correlation can help identify patterns of covariation between brain regions and psychopathology symptoms. This study examined the relationship between gray matter volume (GMV) and psychopathology symptoms in adolescents with various psychiatric diagnoses. Methods: Structural magnetic resonance imaging data were collected from 490 participants with various internalizing and externalizing diagnoses (197 female/293 male; age = 14.68 ± 2.35 years; IQ = 104.05 ± 13.11). Cortical and subcortical volumes were parcellated using the Desikan-Killiany atlas. Partial least squares correlation was used to identify multivariate linear relationships between GMV and the Strength and Difficulties Questionnaire difficulties domains (emotional, peer, conduct, and hyperactivity issues). Resampling approaches were used to determine significance (permutation test), stability (bootstrap resampling), and reproducibility (split-half resampling) of identified relationships. Results: We found a significant, stable, and largely reproducible dimension that linked lower Strength and Difficulties Questionnaire scores (less impairment) across all difficulties domains with greater widespread GMV (singular value = 1.17, accounts for 87.1% of the covariance; p < .001). This dimension emphasized the relationship between lower conduct problems and greater GMV in frontotemporal regions. Conclusions: Our results indicate that the most significant and stable brain-behavior relationship in a transdiagnostic sample is a domain-general relationship, linking lower psychopathology symptom scores to greater global GMV. This finding suggests that a shared brain-behavior relationship may be present across adolescents with and without clinically significant psychopathology symptoms.

Original languageEnglish (US)
Pages (from-to)377-386
Number of pages10
JournalBiological Psychiatry: Cognitive Neuroscience and Neuroimaging
Volume9
Issue number4
DOIs
StatePublished - Apr 2024
Externally publishedYes

Keywords

  • Adolescents
  • Morphometry
  • MRI
  • Multivariate
  • Psychopathology
  • Transdiagnostic

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Cognitive Neuroscience
  • Clinical Neurology
  • Biological Psychiatry

Fingerprint

Dive into the research topics of 'A Shared Multivariate Brain-Behavior Relationship in a Transdiagnostic Sample of Adolescents'. Together they form a unique fingerprint.

Cite this