Cognitive, Clinical, and Biomarker Correlates of Insight in Obsessive–Compulsive Disorder: preliminary results of a cross-sectional study

Obsessive–compulsive disorder (OCD) is a disabling psychiatric condition characterized by intrusive thoughts and repetitive behaviors. Increasing evidence suggests that its clinical heterogeneity reflects dysfunctions across cognitive, neurobiological, and neuroplastic domains. This study examined the associations among insight, cognitive performance, sleep architecture, and peripheral biomarkers in OCD, integrating dimensional and machine-learning approaches. A multimodal dataset including psychometric scales (Y-BOCS, BABS), MATRICS-defined cognitive tests, polysomnography, and serum biomarkers (BDNF, TNF-α, CRP) was analyzed in adult OCD subgroups characterized by late onset (LO) and poor insight (PI). Principal Component Analysis and supervised classification revealed two main latent dimensions—one dominated by cognitive and insight measures, and another by sleep–arousal indices. Poorer insight, particularly reduced ability to explain differing views, was associated with lower social cognition performance, while no robust associations emerged between biomarkers and cognition after correction for multiple testing. Although unsupervised clustering failed to separate subtypes clearly, supervised models achieved high accuracy (F1 = 0.9), identifying BABS conviction as the strongest discriminator. These findings highlight metacognitive flexibility and sleep physiology as key elements of interindividual variability in OCD, supporting a multilevel framework linking cognition, psychopathology, and biological plasticity toward personalized phenotyping.