The Memory Index Score (MIS) of the MoCA in Parkinson's disease: diagnostic accuracy and comparison with standard memory tests

Background. The Memory Index Score (MIS) is a brief, memory-specific index derived from of the routinely administered Montreal Cognitive Assessment (MoCA). It was developed to detect the encoding-type memory deficit characteristic of amnestic mild cognitive impairment (MCI) and Alzheimer's disease (AD). Memory impairment in Parkinson's disease (PD) is common but heterogeneous and frequently retrieval-based, and the MoCA-MIS has not previously been evaluated in this population, leaving its clinical utility and appropriate cut-offs in PD unknown. Objective. To evaluate the MoCA-MIS as a rapid memory-screening tool in PD, benchmarking it against a comprehensive neuropsychological memory battery (RAVLT immediate and delayed recall, RAVLT Memory Efficiency Index, Word-Pair Association Test, Babcock Story Recall, and Rey- Osterrieth Complex Figure delayed recall). Methods. In this retrospective, cross-sectional study, clinical and neuropsychological data were drawn from 847 participants (627 PD patients, 220 control subjects) assessed at the Neurology Unit of Padua University Hospital. Participants were classified, following Movement Disorder Society criteria, as PD with normal cognition (PD-NC), PD with mild cognitive impairment (PD-MCI), or PD dementia (PDD), and non-PD controls as cognitively unimpaired (CU) or MCI. Group differences were examined with analysis of variance (ANOVA) on demographically adjusted scores; associations between the MoCA-MIS and memory battery were examined using with Spearman correlations; the prevalence of impairment with chi-square tests; and diagnostic accuracy with receiver operating characteristic (ROC) analyses, with optimal cut-offs derived by Youden's index. Results. The MoCA-MIS correlated significantly and positively with every battery measure (rs = .36–.54), supporting its convergent validity. It discriminated between cognitive-status groups, but with smaller effect sizes than the dedicated memory measures, and showed a modest but stable diagnostic accuracy across disease stages (AUC ≈ .71–.77). Its sensitivity was low in the milder contrasts and increased as impairment worsened (from 46.8% for CU vs MCI, to 73.4% for PD-MCI vs PDD), with correspondingly high specificity in the milder stages. The Alzheimer-derived cut-offs did not apply to PD. Conclusions. The MoCA-MIS captures clinically relevant episodic memory impairment in PD but is not adequate as a standalone memory screen, due to its low sensitivity in earlier disease stages. Nevertheless, it may provide a useful supplementary memory-specific index, particularly at more advanced stages. Its lower sensitivity at earlier cognitive stages may reflect the evolving nature of memory impairment in PD, where retrieval-based deficits may predominate initially and encoding/storage deficits become more prominent with increasing cognitive severity.