Poster No:
1671
Submission Type:
Abstract Submission
Authors:
Mariya Chepisheva1, Xilin Shen2, Jenna Appleton3, Sacit Omay3, Amit Mahajan2, Emily Gilmore1, Todd Constable2, Jennifer Kim1
Institutions:
1Neurocritical Care and Emergency Neurology, Yale University, New Haven, CT, 2Radiology and Biomedical Imaging, Yale University, New Haven, CT, 3Neurosurgery, Yale-New Haven Hospital, New Haven, CT
First Author:
Mariya Chepisheva
Neurocritical Care and Emergency Neurology, Yale University
New Haven, CT
Co-Author(s):
Xilin Shen
Radiology and Biomedical Imaging, Yale University
New Haven, CT
Sacit Omay
Neurosurgery, Yale-New Haven Hospital
New Haven, CT
Amit Mahajan
Radiology and Biomedical Imaging, Yale University
New Haven, CT
Emily Gilmore
Neurocritical Care and Emergency Neurology, Yale University
New Haven, CT
Todd Constable
Radiology and Biomedical Imaging, Yale University
New Haven, CT
Jennifer Kim
Neurocritical Care and Emergency Neurology, Yale University
New Haven, CT
Introduction:
Traumatic brain injury (TBI) is considered a leading cause of long-term disability. While bedside clinical scores like the Glasgow Coma Scale (GCS) and the modified Rankin Scale (mRS) are routine clinical tests for injury severity and outcome, there is limited understanding of the brain basis of these assessments. One barrier to the acquisition of studies able to address this gap, is their high cost. However, with increasing recognition of the importance of advanced neuroimaging techniques such as resting state functional MRI, an increasing number of clinical sites have started exploring their use. If such neuroimaging findings could be turned into tools with practical utility, this would counterbalance the high costs for the healthcare service. Here, we present a connectome based predictive modelling (CPM) approach that used functional connectivity maps to establish a predictive model for GCS at admission or mRS at 3months.
Methods:
We retrospectively (2018-2022) assembled resting fMRI and clinical data from 108 patients (46.4 ± 20.1yrs) across all TBI severities. Patients were scanned acutely (within 31days) or chronically (up to 2yrs) and presented as a first or a repeat (i.e. not first) TBI case. The latter criteria guided the separation of the patients for this analysis into six testing groups. Anatomical and functional preprocessing was performed using standard procedures alongside regression of mean time-courses in white matter, grey matter and CSF; global signal regression and low-pass Gaussian filtering (cut-off 0.12 Hz). We used a 268-node functionally derived atlas to calculate the mean time course of each node and the correlation for each edge. Fisher transformed Z-scores for 35,778 unique edges were derived. Due to non-normal distribution of the behavioral data, we selected a Spearman correlation and corrected for the presence of motion noise and age. The performance of the CPM model (programmed in Matlab, Mathworks) was evaluated using the correlation between the predicted and the observed value. One hundred iterations of 10-fold cross-validations, and 1000 permutation tests were used to derive the statistical importance of the model.
Results:
We were able to construct 3 successful models. [1st model] Acute first TBI (n = 58) patients, established a working model with median r = 0.2485, and p-value from permutation testing P=0.014, when predicting GCS at admission. In comparison, [2nd model], all first TBI patients (i.e. acute + chronic) predicting GCS at admission had an even higher median r = 0.3919 and p -value from permutation testing P=0.001. Positive correlations contributing to both models were (1) within the Motor network, (2) Motor – Salience, (3) within the Fronto-Parietal (FP) and (4) FP – Medial-Frontal (MF) networks. The "Subcortical – Cerebellum" correlation was unique to the acute first CPM model, whereas "FP – DMN" – to all first CPM model. Regarding negative correlations, those common to both models were (1) Motor – FP, (2) Motor – DMN, (3) Motor – MF and (4) Salience – FP. Additional negative correlations were present for all first TBIs: (1) FP – MF and (2) Salience – MF. Next, the final model we established [3rd model] was in the acute first TBI group trying to predict mRS at 3months with median r = 0.2525 and p -value from permutation testing P=0.02. A significant number of negative correlations were noted (1) within the Motor network, (2) Motor – MF, (3) Motor – FP, (4) Motor – Visual Association and (5) within the Cerebellum network. Positive correlations were also present (1) Motor – FP, (2) Motor – DMN and (3) between Motor – Cerebellum. Results for other groupings based on TBI frequency did not yield successful model constructs.
Conclusions:
In this study, we obtained a moderate predictive performance based on the TBI functional MRI connectome. This provided a generalizable brain-behaviour relationship that could be useful in the development of translational tools with clinical utility.
Modeling and Analysis Methods:
Classification and Predictive Modeling 2
fMRI Connectivity and Network Modeling
Task-Independent and Resting-State Analysis 1
Novel Imaging Acquisition Methods:
BOLD fMRI
Keywords:
DISORDERS
FUNCTIONAL MRI
Machine Learning
Modeling
MRI
Neurological
Trauma
Other - MRS - Modified Rankin Scale, GCS - Glasgow Coma Scale, clinical assessments
1|2Indicates the priority used for review
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.
I accept
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:
I do not want to participate in the reproducibility challenge.
Please indicate below if your study was a "resting state" or "task-activation” study.
Resting state
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Patients
Was this research conducted in the United States?
Yes
Are you Internal Review Board (IRB) certified?
Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
Yes, I have IRB or AUCC approval
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.
Yes
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.
Not applicable
Please indicate which methods were used in your research:
Functional MRI
Behavior
Computational modeling
Other, Please specify
-
GCS, mRS
For human MRI, what field strength scanner do you use?
3.0T
Which processing packages did you use for your study?
SPM
FSL
Other, Please list
-
Yale BioImage Suite
Provide references using APA citation style.
(1) Shen, X., Finn, E., Scheinost, D. et al. Using connectome-based predictive modeling to predict individual behavior from brain connectivity. Nat Protoc 12, 506–518 (2017).
(2) Shen, X., Tokogly, F., Papademetris, X., & Constable, R. T. (2013). Groupwise whole-brain parcellation from resting-state fMRI data for network node identification. Neuroimage, 0, 403–415.
(3) Finn, E. S., Shen, X., Scheinost, D., Rosenberg, M. D., Huang, J., Chun, M. M., Papademetris, X., & Constable, R. T. (2015). Functional connectome fingerprinting: Identifying individuals using patterns of brain connectivity. Nature Neuroscience, 18(11), 1664–1671.
(4) Noble, S., Spann, M. N., Tokoglu, F., Shen, X., Constable, R. T., & Scheinost, D. (2017). Influences on the Test-Retest Reliability of Functional Connectivity MRI and its Relationship with Behavioral Utility. Cerebral Cortex, 27(11), 5415–5429.
No