Mediating Role of Emotional States Linking Parkinson’s Disease Severity and Cognitive Function: A Cross-Instrument Comparison Using Two Cognitive Assessment Tools

Objective This cross-sectional study aimed to explore the relationships between disease severity and cognitive function in Parkinson’s disease (PD), statistical indirect associations via anxiety and depression, and their stability across cognitive measures. Methods A total of 255 PD patients were recruited. Disease severity was assessed using the Hoehn-Yahr (H-Y) stage. Cognitive function was evaluated with the Chinese Mini-Mental Status (CMMS) and Montreal Cognitive Assessment (MoCA). Anxiety and depressive symptoms were measured using the Hamilton scales (HAMA/HAMD). Spearman’s rank correlation and partial correlation analyses were used. Mediation analyses were performed using PROCESS, including separate single-mediator models and a parallel multiple-mediator model. Results H-Y stage correlated negatively with CMMS ( r s = −0.20, P = 0.001) and MoCA ( r s = −0.30, P < 0.001), and positively with HAMA ( r s = 0.22, P < 0.001) and HAMD ( r s = 0.28, P < 0.001). In separate single-mediator models, HAMA showed no significant indirect association with CMMS ( b = −0.10, 95% CI [−0.24, 0.01]), while HAMD accounted for 23.74% of the association ( b = −0.19, 95% CI [−0.39, −0.02]). With MoCA, both HAMA ( b = −0.21, 95% CI [−0.39, −0.06]) and HAMD ( b = −0.45, 95% CI [−0.73, −0.21]) showed significant statistical indirect associations, with HAMD accounting for 35.12% and HAMA accounting for 16.63% of the total association. In the parallel multiple-mediator model with both scales entered simultaneously, only the indirect association via HAMD remained significant ( b = −0.53, 95% CI [−0.94, −0.17]), whereas the indirect association via HAMA did not ( b = 0.06,95% CI [−0.16, 0.29]). Conclusion Greater PD severity was associated with poorer cognitive performance. These cross-sectional findings cannot establish causality. In separate single-mediator models, depressive symptoms showed consistent statistical indirect associations across both cognitive tools, while anxiety exhibited a significant indirect association only for MoCA. In the parallel model, the indirect pathway through depression remained significant with MoCA but not with CMMS. MoCA appeared more sensitive for capturing cognition associated with mood symptoms. Our results highlight the importance of comprehensive cognitive assessment and routine screening for depressive symptoms in PD.