ADAPTIVE MULTIVIEW COMMUNITY-PRESERVED GRAPH CONVOLUTIONAL NETWORK FOR MULTIATLAS-BASED FUNCTIONAL CONNECTIVITY ANALYSIS

Wei Wang, Xingcan Hu, Li Xiao, Yu-Ping Wang

Abstract: Recently, functional connectivity network (FCN) analysis vi- a graph convolutional networks (GCNs) has greatly boost- ed diagnostic performance of brain diseases on a population graph for subject classification. However, most existing meth- ods only focus on FCNs based on a single brain atlas (ig- noring complementary information among multiatlas-based FCNs), and the population graph structure is preconstruct- ed and fixed during the GCN training (not truly reflecting the relation between subjects). In this paper, we propose an adaptive multiview community-preserved graph convolution- al network (CP-GCN) method to accommodate multiatlas- based FCNs. Specifically, we first introduce a multiview FCN fusion module to obtain multiatlas FCN embeddings via con- catenating both intra- and inter-atlas embeddings that are ex- tracted separately from fully connected layers. We then devel- op a multihead similarity learning module to adaptively learn the population graph structure, best serving GCN for node classification. Finally, under the learned graph structure, a CP-GCN based node classification module is applied for sub- ject classification through designing a community-preserved constraint on the GCN. Experimental results on the ABIDE validate the effectiveness of our method for autism identifica- tion, and our findings related to autism can be easily traced back with biological interpretability.