Graph data widely exist in many high-impact applications. Inspired by the success of deep learning in grid-structured data, graph neural network models have been proposed to learn powerful node-level or graph-level representation. However, most of the existing graph neural networks suffer from the following limitations: (1) there is limited analysis regarding the graph convolution properties, such as seed-oriented, degree-aware and order-free; (2) the node's degreespecific graph structure is not explicitly expressed in graph convolution for distinguishing structure-aware node neighborhoods;(3) the theoretical explanation regarding the graph-level pooling schemes is unclear.To address these problems, we propose a generic degree-specific graph neural network named DEMO-Net motivated by Weisfeiler-Lehman graph isomorphism test that recursively identifies 1-hop neighborhood structures. In order to explicitly capture the graph topology integrated with node attributes, we argue that graph convolution should have three properties: seed-oriented, degree-aware, order-free. To this end, we propose multi-task graph convolution where each task represents node representation learning for nodes with a specific degree value, thus leading to preserving the degreespecific graph structure. In particular, we design two multi-task learning methods: degree-specific weight and hashing functions for graph convolution. In addition, we propose a novel graph-level pooling/readout scheme for learning graph representation provably lying in a degree-specific Hilbert kernel space. The experimental results on several node and graph classification benchmark data sets demonstrate the effectiveness and efficiency of our proposed DEMO-Net over state-of-the-art graph neural network models.
A viability-time threshold for H-I injury is Sc(O2) of 35% for 2-3 h, heralded by abnormalities in NIRS and CFM during reperfusion. These findings suggest that NIRS and CFM might be used together to predict neurological outcome, and illustrate that there is a several hour window of opportunity during H-I to prevent neurological injury.
OBJECT Cerebral arteriovenous malformations (AVMs) are congenital malformations that may grow in the language cortex but usually do not lead to aphasia. In contrast, language dysfunction is a common presentation for patients with a glioma that involves language areas. The authors attempted to demonstrate the difference in patterns of language cortex reorganization between cerebral AVMs and gliomas by blood oxygen level-dependent (BOLD) functional MRI (fMRI) evaluation. METHODS The authors retrospectively reviewed clinical and imaging data of 63 patients with an unruptured cerebral AVM (AVM group) and 38 patients with a glioma (glioma group) who underwent fMRI. All the patients were right handed, and all their lesions were located in the left cerebral hemisphere. Patients were further categorized into 1 of the 2 following subgroups according to their lesion location: the BA subgroup (overlying or adjacent to the inferior frontal or the middle frontal gyri [the Broca area]) and the WA subgroup (overlying or adjacent to the supramarginal, angular, or superior temporal gyri [the Wernicke area]). Lateralization indices of BOLD signal activations were calculated separately for the Broca and Wernicke areas. Statistical analysis was performed to identify the difference in patterns of language cortex reorganization between the 2 groups. RESULTS In the AVM group, right-sided lateralization of BOLD signal activations was observed in 23 patients (36.5%), including 6 with right-sided lateralization in the Broca area alone, 12 in the Wernicke area alone, and 5 in both areas. More specifically, in the 34 patients in the AVM-BA subgroup, right-sided lateralization of the Broca area was detected in 9 patients (26.5%), and right-sided lateralization of the Wernicke area was detected in 4 (11.8%); in the 29 patients in the AVM-WA subgroup, 2 (6.9%) had right-sided lateralization of the Broca area, and 13 (44.8%) had right-sided lateralization of the Wernicke area. In the glioma group, 6 patients (15.8%) showed right-sided lateralization of the Wernicke area, including 2 patients in the glioma-BA subgroup and 4 patients in the glioma-WA subgroup. No patient showed right-sided lateralization of the Broca area. Moreover, although the incidence of right-sided lateralization was higher in cases of low-grade gliomas (5 in 26 [19.2%]) than in high-grade gliomas (1 in 12 [8.3%]), no significant difference was detected between them (p = 0.643). Compared with the AVM group, the incidence of aphasia was significantly higher (p < 0.001), and right-sided lateralization of language areas was significantly rarer (p = 0.026) in the glioma group. CONCLUSIONS Right-sided lateralization of BOLD signal activations was observed in patients with a cerebral AVM and in those with a glioma, suggesting that language cortex reorganization may occur with both diseases. However, the potential of reorganization in patients with gliomas seems to be insufficient compared with patients AVMs, which is suggested by clinical manifestations and the fMRI findings. More...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.