Room: Room 207
Purpose: To compare the functional connectivity in the brain before and after tumor resection by using graph theoretical measures derived from resting state fMRI data.
Methods: Pre- and post-resection (2-4 weeks) fMRI data were acquired from 5 patients with gliomas. Image pre-processing included timing and motion correction, normalization, regressing-out of covariates, and bandpass filtering (0.03-0.1Hz). ROI time courses were then extracted from AAL90 atlas to generate the connectivity matrices. Global and local efficiency, modularity, clustering coefficient, and small-worldness were calculated by thresholding to a connectivity density of 0.2. Degree centrality were calculated for a range of thresholds from 0.05 to 0.5 for the whole brain, and each hemisphere.
Results: No statistical significance was detected in each of the measures. Because the post-surgical time point only allows minimal time for functional reorganization, whole brain graph theoretical measures are likely largely preserved, with small changes requiring more patients to resolve. Interestingly, every measure decreased post-resection, which is consistent with the reduction in cognitive function seen immediately and up to 3 months after resection.By visualizing the whole brain connectivity, direct effects from the tumor resection not captured by whole brain measures can be observed. To quantify these changes would require more sophisticated measures, especially if one hopes to determine how much of the change is due to tissue loss versus functional reorganization . Lesion segmentation and tissue classification would help. By dividing the degree centrality analyses between the left and right hemisphere, the statistical power was enhanced.
Conclusion: Whole-brain functional connectivity was evaluated in brain tumor patients before and after surgery using graph theoretical analysis. Post-surgical changes were observed but the differences were not significant in the small sample of patients. These measures will be used to build models for the prediction of cognitive outcomes in a larger cohort of patients.