Objective: To analyse seed-based metabolic networks and time series variation in Parkinson’s disease with and without mild cognitive impairment based on functional PET (fPET)

Background: Seed-based network analysis (SBN) allows analysing a defined region’s connectivity map. Until recently, SBN could be done only on group level in [¹⁸F]-FDG PET studies utilising a static scan per subject. Several studies indicate altered SBN maps in Parkinson’s disease (PD) on group level using this approach[1]. Here we propose a new technique by utilising functional [¹⁸F]-FDG PET (fPET) data which provide metabolic time series to derive SBN maps on subject-level. Another novelty of the proposed approach is time series activity variation per subject, which has been recently related to cognition in humans[2].

Method: We utilised multimodal data from 13 matched controls and 14 mid-stage PD patients acquired at the Philipps-University of Marburg. Dynamic fPET data were acquired with list-mode acquisition over 90-minutes with constant infusion of [¹⁸F]-FDG. After standardised preprocessing in SPM12, normalised ¹⁸F-FDG uptake values extracted from seed regions using the MarsBaR toolbox [figure1]. Extracted time series were utilised as regressor in a voxel-wise subject-level regression analysis. Resulting subject maps were converted into z-maps and compared to fMRI maps. Time series variation was calculated as variation coefficient and compared between groups using permutation tests in R.

Results: A higher variation coefficient was observed in the left posterior parietal cortex (p = 0.004) and the superior sensorimotor cortex (p= 0.036) in PD. Higher variation coefficients in the superior sensorimotor cortex were inversely correlated with cognitive performance, measured by cognitive z-scores (p= 0.022). SBN maps revealed the default-mode network and sensorimotor network structure in all groups with a more seed-confined structure in PD with cognitive impairment, especially in the fPET modality.

Conclusion: We report an SBN and time series variation approach based on metabolic time series that shows disease-related alterations and is able to derive metabolic networks on subject-level. The approach offers an interesting framework for targeted connectivity studies e.g. for studying the connectivity profile of the stimulation sites in patients with deep-brain stimulation.

figure1

References: 1. Sala, A. et al. Altered brain metabolic connectivity at multiscale level in early Parkinson’s disease. Sci. Rep. 7, 4256 (2017).
2. Boylan, M. A. et al. Greater BOLD Variability is Associated With Poorer Cognitive Function in an Adult Lifespan Sample. Cereb. Cortex 31, 562–574 (2021).

« Back to 2025 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/seed-based-metabolic-networks-and-time-series-variation-based-on-functional-18f-fdg-pet-in-parkinsons-disease/