paperKB
coga / coga-kb
Processing
Help
Sign in

Chunk #16 — Results — Clinical variables and sensitivity analysis

Source
Structural neuroimaging biomarkers for obsessive-compulsive disorder in the ENIGMA-OCD consortium: medication matters.
Embedded
yes

Text

To assess the influence of different clinical variables on classification performance, we repeated the analysis for specific subgroups split according to medication use, AO, disease duration, and severity. A complete overview is provided in supplementary Tables S6(a–d), and results using RFC on combined data with age, sex and site as covariates are reported below. Medicated OCD vs. HC classification resulted in 0.69 AUC (95% CI = 0.63–0.75; pcorrected < 0.001), unmedicated OCD vs. HC in 0.60 (95% CI = 0.54–0.67; pcorrected = 0.03), and medicated vs. unmedicated OCD in 0.78 (95% CI = 0.72–0.85; pcorrected < 0.001) (see Fig. 3). XGB performance was notably higher for medicated vs. unmedicated OCD classification with an AUC of 0.86 (95% CI = 0.78–0.9; pcorrected < 0.001). Early AO OCD vs. HC classification resulted in 0.68 AUC (95% CI = 0.62–0.75; pcorrected < 0.001), late AO OCD vs. HC in 0.73 (95% CI = 0.67–0.79; pcorrected < 0.001), and early vs. late AO in 0.81 (95% CI = 0.74–0.88; pcorrected < 0.001). As no late AO patients were present in pediatric samples, classifications were