Brain disorders, such as Autism Spectrum Disorder (ASD), alter brain functional (from fMRI) and structural (from diffusion MRI) connectivities at multiple levels and in varying degrees. While unraveling such alterations have been the focus of a large number of studies, morphological brain connectivity has been out of the research scope. In particular, shape-to-shape relationships across brain regions of interest (ROIs) were rarely investigated. As such, the use of networks based on morphological brain data in neurological disorder diagnosis, while leveraging the advent of machine learning, could complement our knowledge on brain wiring alterations in unprecedented ways. In this paper, we use conventional T1-weighted MRI to define morphological brain networks (MBNs), each quantifying shape relationship between different cortical regions for a specific cortical attribute at both low-order and high-order levels. While typical brain connectomes investigate the relationship between two ROIs, we propose high-order MBN which better captures brain complex interactions by modeling the morphological relationship between pairs of ROIs. For ASD identification, we present a connectomic manifold learning framework, which learns multiple kernels to estimate a similarity measure between ASD and normal controls (NC) connectional features, to perform dimensionality reduction for clustering ASD and NC subjects. We benchmark our ASD identification method against both supervised and unsupervised state-of-the-art methods, while depicting the most discriminative high- and low-order relationships between morphological regions in the left and right hemispheres.

Abstract. Previous studies have identified disordered functional (from fMRI) and structural (from diffusion MRI) brain connectivities in Autism Spectrum Disorder (ASD). However, 'shape connections' between brain regions were rarely investigated in ASD -e.g., how morphological attributes of a specific brain region (e.g., sulcal depth) change in relation to morphological attributes in other regions. In this paper, we use conventional T1-w MRI to define morphological connectivity networks, each quantifying shape similarity between different cortical regions for a specific cortical attribute at both low-order and high-order levels. For ASD identification, we present a connectomic manifold learning framework, which learns multiple kernels to estimate a similarity measure between ASD and normal controls (NC) connectomic features, to perform dimensionality reduction for clustering ASD and NC subjects. We benchmark our ASD identification method against supervised and unsupervised state-of-the-art methods, while depicting the most discriminative high-and low-order relationships between morphological regions in the left and right hemispheres.

Highlights Fine-grained and sensitive dynamic subnetwork redundancy metric for disease studies. New connectedness metrics measure network robustness and resilience. Expanded current graph theory metrics to network redundancy in addition to efficiency. Increased redundancy of the subcortical-cerebellum network could be an early sign of AD. Decreased redundancy of the medial frontal network could indicate depression.