B-92: Performance-Based and Self-Rated Working Memory in Schizophrenia, Cannabis Use Disorder, and Their Comorbidity: Associations With Structural Covariance
Abstract
Objective: Working memory (WM) deficits are seen in schizophrenia (SCZ), cannabis use disorder (CUD), and their comorbidity (SCZ/CUD). Structural covariance abnormalities are reported in SCZ and CUD, but impact on WM is unknown. We used network analysis to examine associations between WM and structural covariance in SCZ, CUD, and SCZ/CUD. Method/Data Selection: Persons with SCZ (n=18), CUD (n=54), SCZ/CUD (n=33), and healthy controls (CON; n=19) completed MRI scan, Letter-Number Sequencing (LNS), and BRIEF-A. Cannabis users were abstinent. Volumes of 200 cortical regions of interest (ROI) were derived using Schaefer 2018 parcellation. Structural covariance was calculated among top regions associated with WM. Results/Data Synthesis: LNS was worse for SCZ (p< 0.01), CUD (p=0.05), and SCZ/CUD (p< 0.01) versus CON. Self-rated WM was poorer than CON for SCZ (p=0.05) and SCZ/CUD (p< 0.01). WM was equivalent between SCZ groups. In both CUD groups, LNS was associated with lower covariance between ROIs within default mode, frontoparietal, limbic, somatomotor, and visual networks. In contrast, SCZ showed higher covariance among regions associated with LNS. Poorer self-rated WM related to lower covariance in frontoparietal, somatomotor, dorsal attention, and ventral attention networks across CUD groups. Self-rated WM and covariance unrelated in SCZ. Conclusions: Performance-based WM deficits were seen in all patient groups, while self-rated WM was worse in SCZ with/without CUD. In both CUD groups, performance-based and self-rated WM was associated with lower covariance in several networks. In contrast, in SCZ-only, performance-based WM (but not self-rated) was related to higher covariance. These findings suggest that the neural correlates of WM deficits differ in the presence of CUD in those with SCZ.