(d) Decompose surface into multi-resolution surfaces [1].

Fig. 3. An example of multi-resolution decomposition for a cortical surface. The ROIs are mapped onto all resolutions respectively with certain colors.

(e) Map ROIs on multi-resolution surfaces. (f) Compute folding patterns.To compute the folding pattern of a surface patch, a parametric folding descriptor using polynomials [2] is applied:

where, a and b describe the mirror symmetric components of the patch along x axis and y axis respectively, while c and d represent the rotational symmetric components [2].(g) Generate meso-scale attributes for each ROI.

Fig. 4. Visualization of meso-scale attributes.

(h) Atlas labeling [3]. (i)Generate global-scale attribute for each ROI.

•Meso-scale Attribute AnalysisWe performed cross comparison between the attribute of each ROI and simulate the Gaussian model of the attribute (Fig. 6). The results indicate that the folding pattern attributes of ROIs is distinctive on certain resolutions of surface and unique in a relatively small region.•Global scale attributes analysisThe result further indicates that the level of regularity and variability of ROI folding patterns is very ROI-specific.The folding patterns of some ROIs such as ROI #2 and #9 is relatively constant across subjects in both meso and global scale.

Regularity and Variability of Cortical Folding PatternsHanbo Chen

Computer Science DepartmentThe University of Georgia

cojoc@uga.edu

IntroductionCortical folding patterns are believed to be good predictors of brain cytoarchitecture and function. For instance, neuroscientists frequently apply their domain knowledge to identify brain Regions of Interests (ROIs) based on cortical folding patterns. However, quantitative mapping of cortical folding pattern and brain function has not been established yet in the literature. This poster presents our initial effort in quantification of the regularity and variability of cortical folding pattern features for the ROIs identified by task-based fMRI, which is widely accepted as a standard approach to localize functionally-specialized brain regions. Specifically, we developed a set of novel shape attributes for each ROI base on multiple resolution decomposition of cortical surfaces, and described the meso-scale folding pattern via a polynomial-based approach. We also applied brain atlas label distribution as global-scale description of ROI folding pattern.

ApproachThe procedure of our approach is described in the figure below.

Fig. 1. The flowchart of your approach.

(a)Obtain working memory ROIs from task-based fMRI. (b) Reconstruct cortical surface from DTI data. (c) Map ROIs onto cortical surface.

ROI Prediction

Discussion and ContributionsOur studies of these attributes for working memory ROIs across subjects suggest that: (1) There is deep-rooted regularity of cortical folding patterns for certain working memory ROIs across subjects. This regularity is significantly - prominent in specific resolutions of cortical surface representation. (2) There is tremendous inter-subject variability of cortical folding patterns in different resolution representations of cortical surfaces. These results suggest that cortical folding pattern attributes could be useful for the characterization, recognition and prediction of certain ROIs, if extracted and applied in a proper way.This kind of technique, if fully developed, can be applied in computer aided diagnose in neuron disease.

References1.Yeo, B.T.T., Yu, P., Grant, P. E., Fischl, B. & Golland, P. Shape Analysis with Overcomplete Spherical Wavelets. MICCAI, 468-476 (2008)2.Tuo Zhang, et al.. Parametric Representation of Cortical Surface Folding based on Polynomials. MICCAI, 184-191 (2009)3.D. Shen, et al., HAMMER: hierarchical attribute matching mechanism for elastic registration. IEEE Trans Med Imaging, 21(11), 1421-39, 2002.

AcknowledgmentsThe work is directed by Dr. Tianming Liu. Tuo Zhang, Kaiming Li, Xi Jiang, and Xintao Hu also contributed to this work.

Fig. 2. ROI indices and names. ROIs are mapped on the surface respectively.

3322 dYcXbYaXZ

Fig. 5. Examples of global-scale ROI attribute based on MNI label. Left panel is an MNI labeled cortical surface. Top-right and bottom-right panels are MNI label distribution of ROI #2 and #15 for all subjects respectively.

Fig. 6. Gaussian model simulation. Fig. 7. Similarity of global ROI attributes.

Fig. 8. (a) Prediction results. Benchmark ROIs and predicted ROIs are presented by green and red bubbles respectively and connected by yellow line for each pair.(b) Euclidean distance between 15 predicted ROIs and benchmark ROIs within 5 subjects.