A Graph Framework for Multimodal Medical Information Processing

TL;DR

This paper introduces a graph-based framework for multimodal medical data processing, enabling improved data fusion and diagnosis accuracy through a multilayer graph model and co-author graph ranking.

Contribution

It presents a novel multilayer graph framework for medical information retrieval, analysis, and storage, with an application to frailty assessment using Neo4j.

Findings

Effective data fusion via multilayer graphs

Enhanced diagnosis accuracy through graph analysis

Successful implementation with Neo4j for frailty case study

Abstract

Multimodal medical information processing is currently the epicenter of intense interdisciplinary research, as proper data fusion may lead to more accurate diagnoses. Moreover, multimodality may disambiguate cases of co-morbidity. This paper presents a framework for retrieving, analyzing, and storing medical information as a multilayer graph, an abstract format suitable for data fusion and further processing. At the same time, this paper addresses the need for reliable medical information through co-author graph ranking. A use case pertaining to frailty based on Python and Neo4j serves as an illustration of the proposed framework.