Abstract
Repeated neuropsychological assessments are often conducted in clinical and research settings to track cognitive changes over single or multiple intervals in patients with Parkinson's disease (PD). Yet few studies have documented test-retest reliability in PD. To address this gap, we used data from the Parkinson's Progression Markers Initiative (PPMI) to investigate the reliability of five well-known neuropsychological tests over a 3-year follow-up assessment in early-stage PD with either normal (PD-NC; N = 158) or abnormal (PD-AC; N = 39) cognitive screening, categorized based on recommended cutoffs for the Montreal Cognitive Assessment (MoCA), and healthy older adults (HOA; N = 102). All participants analyzed maintained the same cognitive status category across the assessment points. Intraclass correlation coefficients (ICCs) estimated reliability. The overall ICCs calculated across time points were as follows: Judgment of Line Orientation (PD-NC = .47, PD-AC = .50, HOA = .59); Letter-Number Sequencing (PD-NC = .64, PD-AC = .64, HOA = .65); Semantic Fluency (PD-NC = .69, PD-AC = .89, HOA = .77); Symbol Digit Modalities Test (PD-NC = .67, PD-AC = .83, HOA = .71). For the two primary components of the Hopkins Verbal Learning Test-Revised, we found the following ICCs: immediate recall (PD-NC = .46, PD-AC = .57, HOA = .58); delayed recall (PD-NC = .42, PD-AC = .57, HOA = .54). Findings from this study provide useful information for clinicians and researchers toward selecting suitable neuropsychological tests to monitor cognition at two or more time points among newly diagnosed individuals with PD and HOA.