Advanced brain aging, selective vulnerability in gray matter, and cognition in Parkinson's disease

Abstract

Background: To identify the most vulnerable brain regions in gray matter attributable to advanced brain aging and examine the cognitive correlates of advanced brain aging in Parkinson's disease (PD).

Methods: One hundred twenty-five early-stage PD patients with both structural, diffusion MRI and DAT-SPECT data available were included at baseline (year 0) from Parkinson's Progression Markers Initiative, with neuroimaging follow-up at year 1, 2, 4. Annual assessment of cognition was performed in 5 years. The relationships between brain-predicted age difference (PAD), free water (FW) in cortical and subcortical gray matter, and cognition were examined with linear regression and linear mixed-effects model. Cox proportional hazards model was used to investigate the relation between brain PAD and the risk of conversion to mild cognitive impairment (MCI).

Results: One hundred twenty-five PD patients with a mean (SD) chronological age of 60.99 (9.50) years and 82 (65.6%) were men. Brain PAD followed a non-linear progression pattern over time (P = .028). Brain PAD was differentially associated with FW in cortical and subcortical gray matter, with the most preferentially vulnerable regions identified as temporal cortex, striatum, hippocampus, and cholinergic basal forebrain. Baseline brain PAD was associated with cognitive deficits and the conversion to MCI during the 5-year follow-up.

Conclusions: Our findings suggest that brain PAD offers potential in pinpointing regions most susceptible to accelerated brain aging and identifying patients with Parkinson's disease who are at an increased risk of converting to mild cognitive impairment. .

Keywords: Parkinson’s disease; brain age; mild cognitive impairment; neuroimaging; regional vulnerability.

Figure 1.

Outline of the study. Top left, study population from PPMI and yearly cognitive assessment; top right, neuroimaging pipeline; Bottom, the main analyses performed in the study. DWI, diffusion weighted imaging; PAD, predicted age difference; DAT-SPECT, dopamine transporter single-photon emission computed tomography; MCI, mild cognitive impairment; ROI, region of interest; SBR, striatal binding ratio; PD, Parkinson’s disease; DK, Desikan–Killiany (DK) atlas.

Figure 2.

Features of brain PAD and PAD trajectory over time. A-C, Inter-relationship between chronological age, brain PAD, and brain age. D, Plotted brain PAD at each time point, while the red trend line denotes the average slope of all data points (linear: y = 1.6 + 0.04 × time). E, Estimated progression trend of brain PAD over 4 years (quadratic: y = 0.81 + 0.35 × time − 0.09 × time²).

Figure 3.

The relation between brain PAD and free water in cortical regions at baseline. A, C, The relation between brain PAD and FW in cortical regions, adjusted for age, sex and study sites. B, D, The relation between brain PAD and FW in cortical regions, adjusted for age, sex, study sites, average cortical thickness. ctx, cortex. The color bar in A, B denotes the value of standardized beta coefficient (std.beta) for the relation between brain PAD and cortical FW in each region. The regions that are significantly associated with brain PAD were visualized in A, B. *P <.05, **P <.01.

Figure 4.

The relation between brain PAD and FW in subcortical regions at baseline. A, At baseline, higher brain PAD was associated with higher FW in caudate, putamen, hippocampus, cBF. B, Higher brain PAD at baseline was significantly associated with faster progression rate (ie, steeper slope) of FW in caudate, thalamus, hippocampus, cholinergic basal forebrain. cBF, cholinergic basal forebrain.