Mapping Extreme Deviations in Cognition to Transdiagnostic Compulsivity: A Normative Modelling Approach

Transdiagnostic compulsivity has been linked to reductions in model-based planning, cognitive flexibility and working memory. However, it’s unclear if these effects reflect dimensional variation or asymmetric effects at the extremes. To address this, we used normative modelling to identify extreme positive and negative deviations in cognitive performance using data from N=10,328 participants who completed three smartphone-based tasks. Individuals were stratified into model-derived cognition categories (extreme low: <10th percentile, moderate: 10th-90th percentile, extreme high: >90th percentile), and differences in compulsivity scores across groups were assessed via regression. We report disproportionately elevated compulsivity scores among participants with extreme negative deviations, but no commensurate decrease in those with extreme positive deviations. Compulsivity norms revealed disproportionally reduced cognitive performance among participants with extreme positive deviations, but no corresponding increase in those with extreme negative deviations. These findings provide a nuanced understanding of directional differences driving clinical-cognitive relationships, further refining mechanistic accounts of compulsivity.