Hippocampal subfield diffusion abnormalities in Parkinson's disease with mild cognitive impairment: a volumetric and...

To investigate whether patients with Parkinson's disease with mild cognitive impairment exhibit hippocampal subfield-specific macrostructural and microstructural alterations and to determine whether these imaging abnormalities are associated with global cognitive performance.

This retrospective cross-sectional study included 34 patients with Parkinson's disease and mild cognitive impairment (PD-MCI), 34 patients with Parkinson's disease and preserved cognitive function (PD-NC), and 18 healthy controls (HCs). Bilateral hippocampal subfield volumes were quantified from T1-weighted magnetic resonance images, while fractional anisotropy (FA), axial diffusivity (AD), mean diffusivity (MD), and radial diffusivity (RD) were derived from diffusion tensor imaging. Between-group differences were assessed using analysis of covariance. Diffusion models were adjusted for age, sex, and years of education, whereas volumetric models were additionally adjusted for estimated total intracranial volume. The Benjamini-Hochberg false discovery rate (FDR) procedure was applied across 182 imaging outcomes, and outcomes surviving global FDR correction were followed by Tukey-adjusted pairwise comparisons. Exploratory partial correlations assessed associations between FDR-positive diffusion measures and MoCA scores in the combined PD cohort after adjustment for age, sex, and years of education. BH-FDR correction was applied across the 42 correlation tests, and an additional sensitivity analysis further adjusted for PD-MCI/PD-NC group membership.

None of the 38 volumetric measures survived global FDR correction. In contrast, significant group effects were identified in 42 diffusion measures, comprising 14 AD, 2 FA, 12 MD, and 14 RD outcomes. The strongest effects for the diffusivity measures were observed in the left GC-ML-DG-head (AD: F (2,80) = 11.87, q = 0.002, partial η² = 0.229; MD: F (2,80) = 11.89, q = 0.002, partial η² = 0.229; RD: F (2,80) = 12.34, q = 0.002, partial η² = 0.236). Tukey-adjusted pairwise comparisons identified 38 significant PD-NC versus PD-MCI contrasts and 34 HC versus PD-MCI contrasts, whereas only one HC versus PD-NC contrast was significant. This involved FA in the left GC-ML-DG-head, which was lower in both PD groups than in HC. In exploratory pooled-PD analyses, 39 of the 42 diffusion measures were associated with MoCA scores after FDR correction; however, none remained significant after additional adjustment for PD-MCI/PD-NC group membership.

PD-MCI was characterized by distributed hippocampal diffusivity alterations despite the absence of detectable volumetric differences. Subfield-specific diffusion measures may provide complementary information on cognition-related hippocampal involvement beyond conventional volumetry, but their diagnostic and prognostic value requires longitudinal and external validation.