Mental health diagnoses group diverse symptom presentations under broad labels, yet individuals with the same diagnosis often differ markedly in symptoms, underlying causes, and treatment response. This heterogeneity suggests that categorical systems, while clinically useful, do not fully capture the complexity of mental health variation. Emerging frameworks—such as network models, dimensional approaches (e.g., p-factor, HiTOP), and RDoC—treat variability as informative, offering structured hypotheses about mechanisms and moving beyond traditional diagnostic categories.
Although dimensional constructs show some links to neurocognitive measures, findings are inconsistent and rarely compare frameworks directly. It remains unclear whether these models—or integrated profiles combining brain, cognition, and environmental risk factors—offer meaningful explanatory advantages over diagnoses. Addressing this gap is critical for advancing theory and improving clinical practice.
This project will provide one of the first large-scale, systematic comparisons of diagnostic, dimensional, and symptom-network frameworks in relation to neurocognitive variation. Using UK Biobank and replication in the Adolescent Brain Cognitive Development (ABCD) Study, we will apply advanced multivariate analyses to test which models best capture mental health complexity and its links to brain and cognitive measures. Analyses will draw on harmonised multimodal data across cohorts to ensure robust comparisons and support generalisable findings. Ultimately, this work aims to identify models that move beyond categorical diagnoses toward mechanism-based understanding and clinically relevant stratification.
Aims and Objectives
Objective:
To compare diagnostic, dimensional, and symptom-network models of psychopathology, and test whether integrated neurocognitive–risk profiles improve explanatory power in UK Biobank, with replication in ABCD.
Specific Aims:
1. Map brain and cognitive correlates of transdiagnostic dimensions (p-factor, spectra) versus symptom-network features.
2. Test environmental risk factors as mediators or moderators of brain–cognition–symptom links.
3. Compare the explanatory value of profiles integrating brain and cognitive measures with risk factors relative to diagnoses and dimensions using multivariate and model comparison approaches.
4. Replicate analyses in ABCD and use reproducible workflows to harmonise multimodal data across cohorts for transparent and scalable analysis.