Estimation of gender-specific connectional brain templates using joint multi-view cortical morphological network integration

Abstract

The estimation of a connectional brain template (CBT) integrating a population of brain networks while capturing shared and differential connectional patterns across individuals remains unexplored in gender fingerprinting. This paper presents the first study to estimate gender-specific CBTs using multi-view cortical morphological networks (CMNs) estimated from conventional T1-weighted magnetic resonance imaging (MRI). Specifically, each CMN view is derived from a specific cortical attribute (e.g. thickness), encoded in a network quantifying the dissimilarity in morphology between pairs of cortical brain regions. To this aim, we propose Multi-View Clustering and Fusion Network (MVCF-Net), a novel multi-view network fusion method, which can jointly identify consistent and differential clusters of multi-view datasets in order to capture simultaneously similar and distinct connectional traits of samples. Our MVCF-Net method estimates a representative and well-centered CBTs for male and female populations, independently, to eventually identify their fingerprinting regions of interest (ROIs) in four main steps. First, we perform multi-view network clustering model based on manifold optimization which groups CMNs into shared and differential clusters while preserving their alignment across views. Second, for each view, we linearly fuse CMNs belonging to each cluster, producing local CBTs. Third, for each cluster, we non-linearly integrate the local CBTs across views, producing a cluster-specific CBT. Finally, by linearly fusing the cluster-specific centers we estimate a final CBT of the input population. MVCF-Net produced the most centered and representative CBTs for male and female populations and identified the most discriminative ROIs marking gender differences. The most two gender-discriminative ROIs involved the lateral occipital cortex and pars opercularis in the left hemisphere and the middle temporal gyrus and lingual gyrus in the right hemisphere.

Comparative Survey of Multigraph Integration Methods for Holistic Brain Connectivity Mapping

“…We present the first review paper which provides an insightful survey of the existing integration models promoted with a comparative study to evaluate their performance across extensive experiments in terms of producing the most centered templates, recapitulating unique traits of populations, and , similarity network fusion (SNF) (Wang et al, 2014) and supervised multi-topology network cross-diffusion (SM-netFusion) ; and b) multi-graph integration methods for a population of multi-view brain connectivity dataset: multi-view networks normalizer (netNorm) (Dhifallah et al, 2020), cluster-based network fusion (SCA) (Dhifallah et al, 2019), multiview clustering and fusion (MVCF-Net) (Chaari et al, 2020) and cluster-based multi-graph integrator networks (cMGI-Net) (Gurbuz and Rekik, 2020).…”

“…The second group represents multi-view graphs integration methods that fuse populations of multi-view networks into a single connectional template. For this category, we review five multigraph fusion methods: netNorm (Dhi-fallah et al, 2020), SCA (Dhifallah et al, 2019), MVCF-Net (Chaari et al, 2020), cMGI-Net (Demir et al, 2020), and DGN (Gurbuz and Rekik, 2020). Multigraph fusion methods can be sub-categorized into two big classes: machine learning-based and deep learning-based models.…”

“…MVCF-Net. More recently, (Chaari et al, 2020) proposed multi-view clustering and fusion (MVCF-Net), a graph-based clustering method, to fuse a population of multi-view networks. This method is rooted in the identification of consistent and differential clusters across brain views to generate a connectional brain template for a given population.…”

“…We evaluate the performance of the single-view fusion methods set including SNF (Wang et al, 2014), NAG-FS , and SM-netFusion , and the multigraph integration methods set including DGN (Gurbuz and Rekik, 2020), cMGI-Net (Demir et al, 2020), netNorm (Dhifallah et al, 2020), MVCF-Net (Chaari et al, 2020), and SCA (Dhifallah et al, 2019), separately. For CBT generation, we use as input dataset single-view networks for single-view CBT estimation and multi-view networks composed of 4 cortical morphological brain graphs for multigraph integration methods into a holistic CBT.…”

“…Human neuroscience studies have made significant progress on the path towards estimating brain network templates for a population of connectomes since the inception of the seed paper introducing the concept of 'populationdriven network atlas' in (Rekik et al, 2017). Several integration methods were proposed to form the integral representation for a population of unimodal (single-view) (Wang et al, 2014; as well as multi-modal (multi-view) (Dhifallah et al, 2019(Dhifallah et al, , 2020Chaari et al, 2020;Demir et al, 2020;Gurbuz and Rekik, 2020) brain networks. Nevertheless, single-view integration methods were limited to fuse single-view networks, thereby, overlooking the complementary and richness of multi-view brain network populations.…”