Robust graph analysis of brain connectivity

Lead Research Organisation: University College London
Department Name: Institute of Child Health

Abstract

A broad range of neurological diseases and disorders have been linked with pathological alterations in the connectivity of the brain. However, robust methods for identifying and characterising abnormalities in connectivity, their evolution over time and response to treatments, are lacking; and addressing this omission represents a pressing clinical need. In this proposal we aim to meet this need using novel methods based on medical imaging and graph theory, a well-established mathematical framework with which to describe and characterise interconnected systems. To establish a clear baseline, we will characterise the typical "normal" connectivity network, and model statistically its variability in a healthy population. We will also investigate approaches to classifying graphs into groups, and identify common factors underlying structural and functional connectivity, to be used as new biomarkers. We will use magnetic resonance imaging (MRI) and electroencephalography (EEG) to obtain connectivity information in the living brain; and childhood epilepsy and autism spectrum disorders will be investigated as neurological disorders which can display altered brain connectivity. Our developments will also be applicable to other data which can be represented graphically. This work will provide major novel tools for graph analysis of brain connectivity, thereby driving connectivity network methods towards reliable and routine clinical application. Through our close links with clinical colleagues we will ensure that our developments are well positioned for wide-ranging applicability, elucidating connectivity failures in disease and their recovery with treatment.

Planned Impact

The primary opportunity for impact outside academia in the work described in this proposal lies in its contribution to the understanding of "disconnection syndromes". We highlight childhood epilepsy and autism spectrum disorders (ASDs) as two specific areas in which we expect a significant clinical impact, but disconnection between brain regions is thought to be a major contributing factor to a broad range of neurological deficits that collectively impose a huge health and economic burden on society. Since we also aim to elucidate the changes in connectivity in response to drug treatments, commercial exploitation of some of the developments arising from the project may also be a medium-to-long-term possibility in the pharmaceutical sector. Open access to the methods developed during the project, as well as clear summaries of the ideas embodied in it, will ensure that a very wide, international audience can benefit from them as quickly as possible.
 
Description We have developed techniques for identifying important subnetworks in the living brain using medical imaging data, and a framework for relating functional and structural measures of brain connectivity. We have discovered that EEG data is better at predicting fMRI-based functional connectivity than vice-versa, and found evidence that sophisticated structural models may be helpful in unravelling the complex relationship between structure and function. In addition, we continue to investigate the normal variability in brain connectivity features, both in adults and children.
Exploitation Route We have been involved in a number of applications in clinical and nonclinical neuroscience, demonstrating our developments' wide applicability. Areas of application that we have directly collaborated on include normal childhood development (including structural and functional differences between monolingual and bilingual children), multiple sclerosis, autism and sickle cell disease. Clinical and neurolinguistic collaborators have been very positive about the approach.
Sectors Digital/Communication/Information Technologies (including Software),Healthcare

URL http://www.homepages.ucl.ac.uk/~sejjjd2/research.html#multimodal-brain-networks
 
Description Our work on identifying important subnetworks in brain connectivity graphs has been applied to the study of functional connectivity in young children, and to structural connectivity in multiple sclerosis. Further medical applications of our brain connectivity ideas in paediatric epilepsy and sickle cell disease are in various stages of development, with the enthusiastic support of clinical colleagues. This work has also influenced new efforts to develop connectivity mapping approaches that are tailored for intra-operative neurosurgical use in nearby hospitals; again, with the collaboration of the relevant clinical specialities. Through all of these initiatives, and their gradual influence on medical policy and practice, impact is being realised in the healthcare sector.
First Year Of Impact 2013
Sector Healthcare
Impact Types Policy & public services

 
Description Research Project Grant
Amount £305,748 (GBP)
Funding ID RPG-2017-403 
Organisation The Leverhulme Trust 
Sector Charity/Non Profit
Country United Kingdom
Start 03/2018 
End 03/2021
 
Description Royal Society-Newton Mobility Grant
Amount £4,700 (GBP)
Funding ID NI160219 
Organisation The Royal Society 
Sector Charity/Non Profit
Country United Kingdom
Start 03/2017 
End 03/2018
 
Description IoE 
Organisation University College London
Department Institute of Education (IOE)
Country United Kingdom 
Sector Academic/University 
PI Contribution Originator and co-investigator on jointly held research grant (see Additional Funding)
Collaborator Contribution Lead investigator on jointly held research grant
Impact This collaboration is still in the relatively early stages, and has not produced any outputs yet. It is multidisciplinary (linguistics, education, imaging science, computer science).
Start Year 2017
 
Description IoN 
Organisation University College London
Department Institute of Neurology
Country United Kingdom 
Sector Academic/University 
PI Contribution Provision of software and expertise in tractography and network-based analysis of brain connectivity
Collaborator Contribution Access to manpower and data
Impact Several journal and conference publications have been submitted on this work. This collaboration is multidisciplinary (computer science, physics, neuroscience, medicine).
Start Year 2012
 
Title TractoR 
Description TractoR is a flexible and integrated package for medical image analysis based on the open-source R platform for statistical computing. It provides interfaces for technical and less technical users, to allow them to perform image manipulation and brain connectivity analyses on their own data. It is also the context in which our methodological work is first developed and made publicly available to the community. The software is being updated on a rolling basis during the course of the project, to include the methodological advances arising from it. 
Type Of Technology Software 
Year Produced 2014 
Open Source License? Yes  
Impact The availability and maturity of this software has been a significant factor in many established and new collaborative partnerships. 
URL http://www.tractor-mri.org.uk