Localizing Hallucination-Driven Individual Differences with Eigenvector Centrality
Poster No:
1694
Submission Type:
Abstract Submission
Authors:
Theresa Marschall1, Remco Renken1, Frans Cornelissen1, Teus van Laar1
Institutions:
1University Medical Center Groningen, Groningen, Groningen
First Author:
Co-Author(s):
Introduction:
In neurological research comparing individual cases with publicly available data can pose significant challenges, especially when scanning parameters vary. We introduce a method to detect individual differences in brain connectivity, employing rank scores on eigenvector centrality to overcome these challenges. To address discrepancies in scanning parameters across datasets, we employed a ranking system and interquartile range (IQR) normalization. Our focus on rank differences, rather than raw eigenvector centrality coefficients, allows for a more widely applicable comparison between individual cases and extensive sets of norm data.
Methods:
We applied this method to neuroimaging data from a male Parkinson's Disease (PD), with frequent visual hallucinations (VH) comparing it with two distinct datasets. The patient held a button box (for a subsequent task) which was unique to these data. All data consisted of a T1-weighted reference scan and a 10-minute resting-state fMRI. The patient's data were collected using a Siemens 3T scanner implementing online motion correction. Dataset 1, comprising 9 PD individuals without VH (3 females), was collected on the same scanner with a comparable protocol but without online motion correction. Dataset 2 included 20 male PD patients without VH, sourced from the PPMI database (Brumm et al., 2023) and collected using a different protocol and scanner. All data underwent preprocessing with fMRIPrep 23.0.1 (Esteban et al., 2019), high-pass filtering and confound regression (WM and CSF), followed by eigenvector centrality computation for 232 ROIs (Power et al. 2011) using the fECM toolbox (Wink et al., 2012). The eigenvector centrality coefficients were sorted and differences in rank between the patient and each subject, i.e. delta rank differences, were calculated. For each dataset we normalized the delta rank differences by dividing the median per ROI by its IQR.
Results:
The normalized delta rank differences between each dataset and the patient can be seen in Figure 1 and a linear regression analysis between the normalized delta ranks revealed a significant correlation (p < 0.01), with a regression coefficient of 0.656. This indicates a strong linear association between the changes in centrality ranking of the patient compared to either dataset.
·Cleveland dot plot of normalized delta ranks for 232 ROIs, sorted by Δ rank of the PPMI dataset (orange); 9 PD dataset in blue. Key divergences annotated: BB = button box, VH = visual hallucinations
Conclusions:
Notably, our method identified multiple regions in which the patient's connectivity deviated from that of both comparison datasets. These included a region associated with hand movements (annotated as BB in Figure 1), possibly associated with the patient's task of holding a button box. Of note was also a region in the visual cortex which may potentially be linked to his visual hallucinations (annotated as VH in Figure 1). We expect that our approach will facilitate personalized treatment planning and offers potential for broader applications in detecting interindividual differences, with scalability for incorporating more extensive data. Our study demonstrates the ability to reliably compare a single patient's data with different datasets, despite variations in data acquisition. This is crucial for personalized treatment strategies, such as targeted rTMS interventions, which requires the detection of brain regions that deviate from normal.
Disorders of the Nervous System:
Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s) 2
Modeling and Analysis Methods:
Exploratory Modeling and Artifact Removal 1
fMRI Connectivity and Network Modeling
Task-Independent and Resting-State Analysis
Keywords:
Data analysis
FUNCTIONAL MRI
Open Data
Other - parkinson's disease; visual hallucinations
1|2Indicates the priority used for review
Provide references using author date format
Esteban, O., Markiewicz, C. J., Blair, R. W., Moodie, C. A., Isik, A. I., Erramuzpe, A., Kent, J. D., Goncalves, M., DuPre, E., Snyder, M., Oya, H., Ghosh, S. S., Wright, J., Durnez, J., Poldrack, R. A., & Gorgolewski, K. J. (2019). fMRIPrep: a robust preprocessing pipeline for functional MRI. Nature Methods, 16(1). https://doi.org/10.1038/s41592-018-0235-4
Power, J. D., Cohen, A. L., Nelson, S. M., Wig, G. S., Barnes, K. A., Church, J. A., Vogel, A. C., Laumann, T. O., Miezin, F. M., Schlaggar, B. L., & Petersen, S. E. (2011). Functional Network Organization of the Human Brain. Neuron, 72(4). https://doi.org/10.1016/j.neuron.2011.09.006
Wink, A. M., de Munck, J. C., Van Der Werf, Y. D., Van Den Heuvel, O. A., & Barkhof, F. (2012). Fast Eigenvector Centrality Mapping of Voxel-Wise Connectivity in Functional Magnetic Resonance Imaging: Implementation, Validation, and Interpretation. Brain Connectivity, 2(5). https://doi.org/10.1089/brain.2012.0087