Predicting individualized gray matter atrophy in stroke patients with network-based spreading model
Poster No:
1308
Submission Type:
Abstract Submission
Authors:
Jing Yang1, Yixin Gao1, Liyuan Yang2, Yaya Jiang3, Gaolang Gong1
Institutions:
1State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University, Beijing, Beijing, 2Faculty of Psychology, Tianjin Normal University, Tianjin, Tianjin, 3Artificial Intelligence and Language Cognition Laboratory, Beijing International Studies University, Beijing, Beijing
First Author:
Jing Yang
State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University
Beijing, Beijing
State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University
Beijing, Beijing
Co-Author(s):
Yixin Gao
State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University
Beijing, Beijing
State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University
Beijing, Beijing
Yaya Jiang
Artificial Intelligence and Language Cognition Laboratory, Beijing International Studies University
Beijing, Beijing
Artificial Intelligence and Language Cognition Laboratory, Beijing International Studies University
Beijing, Beijing
Gaolang Gong
State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University
Beijing, Beijing
State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University
Beijing, Beijing
Introduction:
Gray matter volume (GMV) reduction is a major structural alteration in various brain disorders. Prior studies have identified gray matter loss in a large number of brain areas following stroke, which showed associations with cognitive impairments (Stebbins et al., 2008). However, the mechanism underlying the specificity of post-stroke spatial pattern of GMV reduction remains unclear. Given that stroke is increasingly recognized as a network disorder (Boes et al., 2015), we aim to evaluate whether 1) the brain connectome shapes the spatial distribution of GMV changes in stroke patients and 2) the individual spatial pattern of post stroke GMV change could be predicted by network spreading models together with the lesion location.
Methods:
73 patients with ischemic stroke (Corbetta et al., 2015) were included in this study. Each participant underwent T1 imaging at a minimum of two post-stroke time points. After preprocessing, changes in spatially distributed GMV and lesion map were obtained. Healthy controls from the HCP dataset were also included, with normative structural and functional connectome estimated using diffusion MRI and resting-state fMRI data.
Coordinated deformation models (CDM) (Shafiei et al., 2020) was developed to evaluate the extent to which spatial patterns of GMV changes are affected by the connectome. Specifically, GMV change at each node was estimated by the alternations of its connected neighbors and connectivity strength. (Fig.1B)
Network diffusion models (NDM) simulated the dynamic diffusion process from the lesion areas to other regions through the connectivity network (Raj et al., 2015). Notably, individual-specific vector representing the lesion map as the initial variable f0 (in Fig. 2A) of the model. Correlation between estimated and observed GMV change was computed for t ranging from 0 to 50 to find the t with maximum r values.
To evaluate the significance of CDM and NDM, three null models were applied. Null-smash and Null-spin were used to control for spatial properties of GMV changes, while Null-rewire excluded low-level network topological effects.
Coordinated deformation models (CDM) (Shafiei et al., 2020) was developed to evaluate the extent to which spatial patterns of GMV changes are affected by the connectome. Specifically, GMV change at each node was estimated by the alternations of its connected neighbors and connectivity strength. (Fig.1B)
Network diffusion models (NDM) simulated the dynamic diffusion process from the lesion areas to other regions through the connectivity network (Raj et al., 2015). Notably, individual-specific vector representing the lesion map as the initial variable f0 (in Fig. 2A) of the model. Correlation between estimated and observed GMV change was computed for t ranging from 0 to 50 to find the t with maximum r values.
To evaluate the significance of CDM and NDM, three null models were applied. Null-smash and Null-spin were used to control for spatial properties of GMV changes, while Null-rewire excluded low-level network topological effects.
Results:
The included longitudinal MRI scans were from stroke patients with a mean (SD) age of 53.8 (10.7) years, with 37 males (50.7%). As shown for an individual in Fig.1A, GMV changes were generally higher in regions near the stroke lesions and their contralateral counterparts, while also appearing in other remote regions.
Across different stages of stroke, spatial patterns of longitudinal volume changes were better captured by coordinated deformation models shaped by structural network architecture, compared with functional connectivity (Fig.1C). In this case, the prediction performance of the connectivity weighted by Fraction of Streamlines (FSe) was more accurate than that of networks in other definitions (r-mean(SD) = 0.71(0.09) for 3 months and r-mean(SD) = 0.69(0.09) for 12 months; P < 0.05 for all subjects).
Network diffusion modelling, based on FSe-weighted networks, predicted the changes of each region to be positively correlated with longitudinal real GMV fluctuations (r-mean(SD) = 0.47(0.16); 86.9% of subjects P < 0.05 for 3 months, and r-mean(SD) = 0.55(0.17); 90.3% of P < 0.05 for 12 months). The same model, using networks defined by fiber length, also robustly estimated longitudinal volume changes for 3 months (r-mean(SD) = 0.48(0.17); 89.9% of P < 0.05) and 12 months (r-mean(SD) = 0.56(0.17); 88.7% of P < 0.05).(Fig. 2)
Across different stages of stroke, spatial patterns of longitudinal volume changes were better captured by coordinated deformation models shaped by structural network architecture, compared with functional connectivity (Fig.1C). In this case, the prediction performance of the connectivity weighted by Fraction of Streamlines (FSe) was more accurate than that of networks in other definitions (r-mean(SD) = 0.71(0.09) for 3 months and r-mean(SD) = 0.69(0.09) for 12 months; P < 0.05 for all subjects).
Network diffusion modelling, based on FSe-weighted networks, predicted the changes of each region to be positively correlated with longitudinal real GMV fluctuations (r-mean(SD) = 0.47(0.16); 86.9% of subjects P < 0.05 for 3 months, and r-mean(SD) = 0.55(0.17); 90.3% of P < 0.05 for 12 months). The same model, using networks defined by fiber length, also robustly estimated longitudinal volume changes for 3 months (r-mean(SD) = 0.48(0.17); 89.9% of P < 0.05) and 12 months (r-mean(SD) = 0.56(0.17); 88.7% of P < 0.05).(Fig. 2)
Conclusions:
These findings illustrate the importance of structural connectome as conduits for the propagation of pathology throughout stroke progression, resembling the transneuronal "prion-like" theory proposed in neurodegenerative diseases (Jucker & Walker, 2011). The network diffusion model demonstrated its ability to predict individual post-stroke gray matter volume changes from their lesion map and therefore could be taken as a potential prognostic tool for stroke assessment and treatment.
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural)
Diffusion MRI Modeling and Analysis 1
fMRI Connectivity and Network Modeling 2
Keywords:
FUNCTIONAL MRI
Modeling
Neurological
STRUCTURAL MRI
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
Other - Network
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
2.Corbetta, M., Ramsey, L., Callejas, A., Baldassarre, A., Hacker, C. D., Siegel, J. S., Astafiev, S. V., Rengachary, J., Zinn, K., Lang, C. E., Connor, L. T., Fucetola, R., Strube, M., Carter, A. R., & Shulman, G. L. (2015). Common behavioral clusters and subcortical anatomy in stroke. Neuron, 85(5), 927–941.
3. Jucker, M., & Walker, L. C. (2011). Pathogenic protein seeding in alzheimer disease and other neurodegenerative disorders. Annals of Neurology, 70(4), 532–540.
4. Raj, A., LoCastro, E., Kuceyeski, A., Tosun, D., Relkin, N., & Weiner, M. (2015). Network Diffusion Model of Progression Predicts Longitudinal Patterns of Atrophy and Metabolism in Alzheimer’s Disease. Cell Reports, 10(3), 359–369.
5. Shafiei, G., Markello, R. D., Makowski, C., Talpalaru, A., Kirschner, M., Devenyi, G. A., Guma, E., Hagmann, P., Cashman, N. R., Lepage, M., Chakravarty, M. M., Dagher, A., & Mišić, B. (2020). Spatial Patterning of Tissue Volume Loss in Schizophrenia Reflects Brain Network Architecture. Biological Psychiatry, 87(8), 727–735.
6. Stebbins, G. T., Nyenhuis, D. L., Wang, C., Cox, J. L., Freels, S., Bangen, K., deToledo-Morrell, L., Sripathirathan, K., Moseley, M., Turner, D. A., Gabrieli, J. D. E., & Gorelick, P. B. (2008). Gray Matter Atrophy in Patients With Ischemic Stroke With Cognitive Impairment. Stroke, 39(3), 785–793.