Abstract
BACKGROUND: Cognitive impairment is common at all stages of Parkinson's disease (PD), but there is no consensus on which neuropsychological tests to use or how to interpret cognitive battery results. A cognitive summary score (CSS) combines the richness of a neuropsychological battery with the simplicity of a single score. OBJECTIVE: The objective of this study was to determine whether a CSS created using robust norming can detect early cognitive deficits in de novo, untreated PD. METHODS: Baseline cognitive data from PD participants and healthy control participants (HCs) in the Parkinson's Progression Markers Initiative were used to (1) create a robust HC subgroup without cognitive decline, (2) generate regression-based z scores for six cognitive measures using this subgroup, and (3) create a CSS by averaging all z scores. RESULTS: PD participants scored worse than HCs on all cognitive tests, with larger effects when compared with the robust HC subgroup rather than all HCs. Applying internally derived norms, the largest effects were for processing speed/working memory (Cohen's d = -0.55) and verbal episodic memory (Cohen's d = -0.48 and -0.52). Robust norming shifted PD performance from average (CSS z score = -0.01) to low average (CSS z score = -0.40), with a larger effect for the CSS (PD vs. robust HC subgroup; Cohen's d = -0.60) compared with individual tests. CONCLUSIONS: Patients with PD perform worse cognitively than HCs, particularly in processing speed and verbal memory. Robust norming increases effect sizes and decreases PD scores to expected levels. The CSS outperformed individual tests and may detect cognitive changes in early PD, making it a useful outcome measure in clinical research. © 2025 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.