Intra-hemishperic functional reorganization in ischemic stroke
Poster No:
1526
Submission Type:
Abstract Submission
Authors:
Cemal koba1, Joan Falco-Roget2, Alessandro Crimi2
Institutions:
1Sano Centre for Computational Personalized Medicine, Krakow, Lesser Poland, 2Sano Center for Computational Personalized Medicine, Krakow, Lesser Poland
First Author:
Co-Author(s):
Introduction:
Ischemic stroke is a case of lack of or interrupted blood flow to arteries in the brain. If the abnormality lasts long enough, neural death is unavoidable. The region that suffers from the lack of blood flow will be infarcted, and there is a chance that the surrounding region will be swollen. The change in the metabolism of the brain is associated with a wide range of neural and behavioral dysfunctions e.g. motor dysfunction and hemispheric imbalance in the motor activity during the relevant behavioral task. There are also many functional connectivity differences reported in resting-state fMRI. Neural dynamics from the damaged regions are naturally affected by stroke, due to the structural and metabolic changes. However, functional change in the intact hemisphere can be a more accurate indicator of neural plasticity. In this study, we examine the functional-structural organization of the intact hemispheres, for the right and left hemispheres separately.
Methods:
Methods:
104 stroke patients in the acute stage (within 2 weeks after the stroke) and 26 healthy controls were included in this study. Each subject went through a resting-state fMRI scan of 256 scans 1-7 times. In total, 130 subjects and their 875 resting state scans were used in the study. Each run was preprocessed with fmriprep (Esteban et al., 2019), and the regressors of no interest were removed with the 36P strategy described in Satterwaite et al. (2013). In addition, mean framewise displacement and a linear trend were acquainted. After band-passing the time series between 0.01 and 0.1 Hz, mean time series were extracted based on Schaefer's 400 region parcellation. The mean time series were used to generate the correlation matrices between the regions. The mean correlation matrix for each subject was calculated, resulting in one matrix for each subject (104 for stroke subjects and 26 for controls). The functional-structural organization was calculated via the functional connectivity gradients method explained in Marguiles et al.(2016). For each subject, the first 3 gradients that explain the most variance in connectivity both within the right and left hemispheres were calculated separately. For stroke subjects, this procedure included only the intact hemispheres, leaving 48 stroke subjects for within-left hemisphere connectivity maps (due to their damage on the right hemisphere) and 56 within-right connectivity maps. The connectivity gradients approach represents the functional distribution of the regions in the embedding space, thanks to the positions of the first 3 gradients. Leveraging this advantage, for each stroke subject, the Euclidean distance between the position of each region to the mean control distribution in both hemispheres was calculated. The mean Euclidean distance for each subject was saved. Mean Euclidean distances for the right and left hemispheres were compared via an independent sample t-test. The same analysis was repeated by calculating the Euclidean distances for each 3 gradients separately in order to identify their contribution.
104 stroke patients in the acute stage (within 2 weeks after the stroke) and 26 healthy controls were included in this study. Each subject went through a resting-state fMRI scan of 256 scans 1-7 times. In total, 130 subjects and their 875 resting state scans were used in the study. Each run was preprocessed with fmriprep (Esteban et al., 2019), and the regressors of no interest were removed with the 36P strategy described in Satterwaite et al. (2013). In addition, mean framewise displacement and a linear trend were acquainted. After band-passing the time series between 0.01 and 0.1 Hz, mean time series were extracted based on Schaefer's 400 region parcellation. The mean time series were used to generate the correlation matrices between the regions. The mean correlation matrix for each subject was calculated, resulting in one matrix for each subject (104 for stroke subjects and 26 for controls). The functional-structural organization was calculated via the functional connectivity gradients method explained in Marguiles et al.(2016). For each subject, the first 3 gradients that explain the most variance in connectivity both within the right and left hemispheres were calculated separately. For stroke subjects, this procedure included only the intact hemispheres, leaving 48 stroke subjects for within-left hemisphere connectivity maps (due to their damage on the right hemisphere) and 56 within-right connectivity maps. The connectivity gradients approach represents the functional distribution of the regions in the embedding space, thanks to the positions of the first 3 gradients. Leveraging this advantage, for each stroke subject, the Euclidean distance between the position of each region to the mean control distribution in both hemispheres was calculated. The mean Euclidean distance for each subject was saved. Mean Euclidean distances for the right and left hemispheres were compared via an independent sample t-test. The same analysis was repeated by calculating the Euclidean distances for each 3 gradients separately in order to identify their contribution.
Results:
The comparison between the Euclidean distances of the intact right and left hemispheres to their corresponding control distribution in 3D space showed a higher mean Euclidean distance value in the right hemisphere (t(398)=-3.43, p=0.0006). Repeating the same analysis by using the Euclidean distances of each 3-gradient separately showed that the difference is specific to the second gradient (Difference for the 1st, 2nd, and 3rd gradients respectively: t(398)=-0.88, -2.39, -1.61, p=0.37, 0.01, 0.11). Relaying the mean distance on the surface showed that the highest difference on the 2nd gradient is placed on the visual and sensorimotor areas.
·Figure 1: Functional distribution of the regions in 3D space for separate hemispheres of the control and stroke subjects.
·Figure 2: Distribution of the 2nd gradient for control and stroke subjects and the mean distance between them.
Conclusions:
The right hemisphere is more sensitive to ischemic stroke damage on the contralateral hemisphere. This difference is the sensitivity comes from the second gradient, where the brain function is clustered between visual and sensorimotor areas.
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 1
fMRI Connectivity and Network Modeling
Multivariate Approaches 2
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
Anatomy and Functional Systems
Novel Imaging Acquisition Methods:
BOLD fMRI
Keywords:
FUNCTIONAL MRI
Plasticity
Other - stroke, ischemic, gradient
1|2Indicates the priority used for review
Provide references using author date format
Margulies, D. S., Ghosh, S. S., Goulas, A., Falkiewicz, M., Huntenburg, J. M., Langs, G., ... & Smallwood, J. (2016). Situating the default-mode network along a principal gradient of macroscale cortical organization. Proceedings of the National Academy of Sciences, 113(44), 12574-12579.
Satterthwaite, T. D., Elliott, M. A., Gerraty, R. T., Ruparel, K., Loughead, J., Calkins, M. E., ... & Wolf, D. H. (2013). An improved framework for confound regression and filtering for control of motion artifact in the preprocessing of resting-state functional connectivity data. Neuroimage, 64, 240-256.