Discriminant subgraph learning from functional brain sensory data

Abstract

The human brain is a complex system with many functional units interacting with each other. This interacting relationship, known as the Functional Connectivity Network (FCN), is critical for brain functions. To learn the FCN, machine learning algorithms can be built based on brain signals captured by sensing technologies such as EEG and fMRI. In neurological diseases, past research has revealed that the FCN is altered. Also, focusing on a specific disease, some part of the FCN, i.e., a sub-network can be more susceptible than other parts. However, the current knowledge about disease-specific sub-networks is limited. We propose a novel Discriminant Subgraph Learner (DSL) to identify a functional sub-network that best differentiates patients with a specific disease from healthy controls based on brain sensory data. We develop an integrated optimization framework for DSL to simultaneously learn the FCN of each class and identify the discriminant sub-network. Further, we develop tractable and converging algorithms to solve the optimization. We apply DSL to identify a functional sub-network that best differentiates patients with episodic migraine from healthy controls based on a fMRI dataset. DSL achieved the best accuracy compared to five state-of-the-art competing algorithms.

Keywords:

• Statistical models; machine learning; brain sensory data; health care

Additional information

Funding

This work was partially supported by NIH K23NS070891, NSF CMMI CAREER 1149602, and NSF DMS-1903135.

Notes on contributors

Lujia Wang

Lujia Wang is a PhD student in the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Institute of Technology. Her research interests include machine learning and biomedical imaging analytics. She is a member of IISE, INFORMS, and IEEE.

Todd J. Schwedt

Todd J. Schwedt, MD, is a professor of neurology at Mayo Clinic Arizona. His research investigates the mechanisms, classification, and treatment of migraine, post-traumatic headache, and other headaches. A core research goal is to use advanced magnetic resonance imaging techniques to identify biomarkers that will help with the diagnosis and treatment of headache. He has published over 90 manuscripts, lectures nationally and internationally, and serves on the Board of Directors for the American Headache Society and the Board of Trustees for the International Headache Society.

Catherine D. Chong

Catherine D. Chong, PhD, is an associate professor in the Department of Neurology at Mayo Clinic Arizona. She completed her PhD in neuroscience at the University of Utah in 2002. Her research interests focus on using structural and functional neuroimaging techniques for delineating the neuropathology associated with migraine.

Teresa Wu

Teresa Wu, PhD, is professor in industrial engineering at Arizona State University. She received her PhD from the University of Iowa in 2001. Her research interests are health informatics and distributed decision supports. She is a recipient of an NSF CAREER award. She is a member of IISE and INFORMS.

Jing Li

Jing Li, PhD, is Harold E. Smalley Professor in the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Institute of Technology. She received her PhD in industrial and operations engineering from the University of Michigan in 2007. Her research interests are statistical modeling and machine learning for health care applications. She is a recipient of an NSF CAREER award. She is a member of IISE, INFORMS, and IEEE.