Evaluating the Viability of Neurocognition as a Transdiagnostic Construct Using Both Latent Variable Models and
Hana-May Eadeh1, Kristian E Markon2, Joel T Nigg3
1Department of Psychological and Brain Sciences, G60 Psychological and Brain Sciences Bldg, University of Iowa, Iowa City, IA, US. hana-may-eadeh@uiowa.edu.
Research on Child and Adolescent Psychopathology
|February 3, 2021
Summary
This study compared latent variable modeling and network analysis to understand child psychopathology, finding both methods complementarily reveal distinct neurocognitive domains linked to internalizing and externalizing symptoms.
Area of Science:
- Child and Adolescent Psychiatry
- Quantitative Psychology
- Neurodevelopmental Disorders
Background:
- Understanding the relational structure of psychological symptoms is vital for mechanistic and causal research in psychopathology.
- Latent variable modeling and network analyses are dominant methodological approaches, but their simultaneous use in child psychopathology is rare.
- Transdiagnostic constructs like neurocognition can enrich nosological structures but are often excluded from analyses.
Purpose of the Study:
- To simultaneously apply latent variable modeling and network analyses to examine symptom structure and neurocognitive domains in youth.
- To investigate the utility of transdiagnostic neurocognitive constructs within child and adolescent psychopathology research.
- To compare the findings from latent variable and network approaches regarding psychopathology structure.
Main Methods:
- Utilized a sample of 498 youth (ages 6-17) enriched for Attention-Deficit/Hyperactivity Disorder (ADHD).
- Applied latent variable modeling and network analyses to model symptom structure and seven neurocognitive domains.
- Examined the structure of psychopathology and the role of neurocognition as a transdiagnostic feature.
Main Results:
- Both latent variable and network analyses revealed similar psychopathology structures, including internalizing, externalizing, and neurocognitive performance clusters.
- Neurocognition emerged as a distinct domain, showing moderate associations with internalizing/externalizing domains in latent variable models and high connectivity in network analyses.
- Latent variable and network approaches yielded convergent, rather than discriminant, findings.
Conclusions:
- Latent variable and network analyses are complementary tools for understanding psychopathology.
- Neurocognitive domains represent important transdiagnostic features that can be effectively modeled using both statistical approaches.
- The findings support the integration of neurocognition into research on the mechanistic architecture of child and adolescent psychopathology.
