Longitudinal Changes in Parkinson's Disease-Related Network Organization in REM Sleep Disorder (RBD)
Poster No:
207
Submission Type:
Abstract Submission
Authors:
Yoshikazu Nakano1, Nha Nguyen2, An Vo3, Chris Tang1, David Eidelberg4
Institutions:
1The Feinstein Instetutes for Medical Research, Manhasset, NY, 2Albert Einstein College of Medicine, New York, NY, 3Feinstein Institutes for Medical Research, Manhasset, NY, 4The Feinstein Institutes for Medical Research, Manhasset, NY
First Author:
Co-Author(s):
Introduction:
Parkinson's disease (PD)-related covariance pattern (PDRP) derived from a network analysis of resting brain images is a feasible imaging biomarker for differential diagnosis and disease progression (Perovnik, Rus, et al. 2023). Moreover, functional connectivity within the network was found to change with disease progression. In particular, assortativity, a connectivity parameter associated with unstable and inefficient flow, increased over time (Vo, Schindlbeck, et al., 2023). Isolated REM sleep behavior disorder (iRBD) is known to be a high-risk feature of prodromal PD. We have reported that the expression of the PD-related covariance pattern (PDRP), an imaging biomarker of PD, is elevated in patients with iRBD before phenoconversion to PD (Holtbernd, Gagnon et al. 2014). However, it remains unclear how the brain connectivity in the PDRP network changes longitudinally.
Methods:
Thirteen patients with iRBD (age 63.5 ± 8.4 years, 13 males) and 17 age-matched normal controls (NC) (age 59.8 ± 10.4 years, 15 males and two females) were recruited. They underwent resting metabolic positron emission tomography (PET) with [18F]-fluorodeoxyglucose (FDG), and the iRBD group underwent follow-up scans at two and four years after baseline. We identified 38 anatomical regions of interest (ROIs) as nodes corresponding to the PDRP network previously validated by voxel-wise analysis of FDG PET (Schindlbeck, Vo et al. 2020). They were classified into 20 active nodes and 18 underactive nodes on the basis of the hyper- and hypometabolism. The pairwise correlation in each node of normalized metabolic activity driven from FDG-PET data was computed for NC and each timepoint of iRBD by 100 bootstrapping iterations. The correlation matrices provided an assortativity coefficient and a degree centrality for each group. Assortativity is a correlation between node degrees across a link. In a network, increased assortativity in a subject group is reflected by a significantly higher coefficient. Assortativity is deemed reduced if the coefficient is lower. Degree centrality is the number of connections within a network or subgraph divided by total nodes. From the comparison of correlation coefficients in each time point, the edges that showed a significant increase compared to the NC group were identified as gained connections, while those showing a significant decrease were classified as lost connections. The ratio of gained to lost connections was defined as the Gaind-to-Lost Ratio (GLR). Differences between iRBD at each timepoint and NC were evaluated using ANOVA with the post hoc Bonferroni correction for multiple comparisons.
Results:
The assortativity in PDRP network showed no significant differences between NC and iRBD at baseline (P = 1.00). However, at the 2-year and 4-year follow-ups in iRBD, it was significantly elevated compared to NC (P < 0.01). Degree Centrality within the same space was higher in iRBD compared to the NC group at every timepoint (P < 0.01). Although no significant differences were observed between baseline and the 2-year follow-up (P = 1.00), a significant increase was noted at the 4-year follow-up when compared to baseline and 2-year (P < 0.036). The gained connections between active nodes showed a decrease at the 2-year and 4-year time points compared to the baseline (P < 0.001) with no significant changes between 2- and 4-year (P = 0.052). Conversely, the lost connections between active regions increased from baseline to the 2-year and 4-year follow-ups (P < 0.001). The GLR between active nodes decreased at the 2-year follow-up compared to the baseline and slightly improved at the 4-year follow-up. The GLR between active and inactive regions increased at the 2-year follow-up compared to the baseline and remained elevated at the 4-year follow-up (P < 0.001).
Conclusions:
In the context of iRBD, the connectivity within the PDRP network is observed from the early stages and gradually converges toward changes reminiscent of those seen in PD over time.
Disorders of the Nervous System:
Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s) 1
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 2
PET Modeling and Analysis
Novel Imaging Acquisition Methods:
PET
Keywords:
Computational Neuroscience
Data analysis
Degenerative Disease
Dopamine
Movement Disorder
Positron Emission Tomography (PET)
Sleep
1|2Indicates the priority used for review
Provide references using author date format
Perovnik, M. (2023), "Functional brain networks in the evaluation of patients with neurodegenerative disorders." Nat Rev Neurol 19(2): 73-90.
Schindlbeck, K. (2020), "LRRK2 and GBA Variants Exert Distinct Influences on Parkinson's Disease-Specific Metabolic Networks." Cereb Cortex 30(5): 2867-2878.
Vo, A. (2023), "Adaptive and pathological connectivity responses in Parkinson's disease brain networks." Cereb Cortex 33(4): 917-932.