Multimodal Cortical Parcellation Based on Anatomical and Functional Brain Connectivity
Reliable cortical parcellation is a crucial step in human brain network analysis since incorrect definition of nodes may invalidate the inferences drawn from the network. Cortical parcellation is typically cast as an unsupervised clustering problem on fun
- PDF / 477,302 Bytes
- 8 Pages / 439.363 x 666.131 pts Page_size
- 53 Downloads / 219 Views
Abstract. Reliable cortical parcellation is a crucial step in human brain network analysis since incorrect definition of nodes may invalidate the inferences drawn from the network. Cortical parcellation is typically cast as an unsupervised clustering problem on functional magnetic resonance imaging (fMRI) data, which is particularly challenging given the pronounced noise in fMRI acquisitions. This challenge manifests itself in rather inconsistent parcellation maps generated by different methods. To address the need for robust methodologies to parcellate the brain, we propose a multimodal cortical parcellation framework based on fused diffusion MRI (dMRI) and fMRI data analysis. We argue that incorporating anatomical connectivity information into parcellation is beneficial in suppressing spurious correlations commonly observed in fMRI analyses. Our approach adaptively determines the weighting of anatomical and functional connectivity information in a data-driven manner, and incorporates a neighborhood-informed affnity matrix that was recently shown to provide robustness against noise. To validate, we compare parcellations obtained via normalized cuts on unimodal vs. multimodal data from the Human Connectome Project. Results demonstrate that our proposed method better delineates spatially contiguous parcels with higher test-retest reliability and improves inter-subject consistency. Keywords: Anatomical connectivity, brain network analysis, clustering, functional connectivity, multimodal cortical parcellation.
1
Introduction
Brain cortical parcellation refers to subdividing the cerebral cortex into regions that exhibit internal homogeneity in certain properties [1]. Connectivity-based parcellation (CBP) is a promising method for brain cortical parcellation, where voxels in the brain are divided into a coarser collection of functionally or anatomically homogeneous brain regions by clustering voxels with similar connectivity profiles [2]. Among CBP techniques, functional connectivity (FC) based parcellation is the predominant approach, where voxels are grouped according to the statistical dependencies between their functional magnetic resonance imaging (fMRI) time courses [1]. However, even with advanced clustering methods, reliable brain parcellation remains challenging due to the notoriously low signal-tonoise ratio (SNR) of fMRI data. Given the close relationship between anatomical c Springer International Publishing Switzerland 2015 N. Navab et al. (Eds.): MICCAI 2015, Part III, LNCS 9351, pp. 21–28, 2015. DOI: 10.1007/978-3-319-24574-4_3
22
C. Wang, B. Yoldemir, and R. Abugharbieh
connectivity (AC) and FC in the brain [3], we argue that simultaneous analysis of AC and FC information for CBP is beneficial. Despite its promising potential, multimodal CBP remains a relatively unexplored direction. To the best of our knowledge, the only existing approach to multimodal CBP uses the overlap among probabilistic parcellation maps derived from different modalities to generate a consensus map [4,5]. A major disadvantage of
Data Loading...
