Multimodal Hyper-connectivity Networks for MCI Classification

Abstract

Hyper-connectivity network is a network where every edge is connected to more than two nodes, and can be naturally denoted using a hyper-graph. Hyper-connectivity brain network, either based on structural or functional interactions among the brain regions, has been used for brain disease diagnosis. However, the conventional hyper-connectivity network is constructed solely based on single modality data, ignoring potential complementary information conveyed by other modalities. The integration of complementary information from multiple modalities has been shown to provide a more comprehensive representation about the brain disruptions. In this paper, a novel multimodal hyper-network modelling method was proposed for improving the diagnostic accuracy of mild cognitive impairment (MCI). Specifically, we first constructed a multimodal hyper-connectivity network by simultaneously considering information from diffusion tensor imaging and resting-state functional magnetic resonance imaging data. We then extracted different types of network features from the hyper-connectivity network, and further exploited a manifold regularized multi-task feature selection method to jointly select the most discriminative features. Our proposed multimodal hyper-connectivity network demonstrated a better MCI classification performance than the conventional single modality based hyper-connectivity networks.

Construction and Multiple Feature Classification Based on a High-Order Functional Hypernetwork on fMRI Data

“…Network construction is critical for the classification of brain networks based on hypergraphs. Hypernetwork construction methods have been proposed previous publications; however, most of the existing functional hypernetwork models were characterized by the interaction of multi-regions in a static form ( Jie et al, 2016 ; Li et al, 2017 ; Zu et al, 2018 ; Li Y. et al, 2019 ). However, research has shown that even in the resting state, brain neural activity still reveals transient and subtle dynamic changes.…”

“…In a previous study, Jie et al (2016) used the Least absolute shrinkage and selection operator (LASSO) method to create a hyper-network model and applied this to the diagnosis of brain diseases. In another study, Yang et al ( Li et al, 2017 ) adopted the star expansion method to construct structural hyper-networks and functional hyper-networks, respectively, to then perform classification analysis. Taking into account the group effecting problem within brain networks, Guo et al (2018a) proposed the elastic net and group LASSO method to improve the establishment of hyper-network models to facilitate brain disease classification research.…”

“…In previous studies of brain functional hypernetworks, researchers usually used a single type of quantifiable property, for example, cluster coefficients ( Jie et al, 2016 ; Li et al, 2017 ; Guo et al, 2018a ; Wang et al, 2018 ; Li Y. et al, 2019 ). Although only a single property is used to obtain better classification performance, this method ignores the role of other properties in the hypernetwork.…”

Applications of hypergraph-based methods in classifying and subtyping psychiatric disorders: a survey

“…Multi-kernel SVM has been used to classify patients with AD and NCs [93]. Subsequent studies have improved these methods and proposed multimodal hypernetwork modeling techniques that leverage the richer information present in multimodal data [94][95][96]. Moreover, hypergraphs have been used to identify connectivity relationships and optimize the weights of hyperedges, to better represent the relationships among brain regions [97,98].…”