UNIVERSITY PARK, Pa. — Frank Hillary, professor of psychology, has been interested in the human brain since his early days as a college student at the University of Michigan. He eventually focused his interest — and his research agenda — on magnetic resonance imaging (MRI) as a window into understanding brain function and brain injury.
In his 20 years as a neuroscientist, however, Hillary said he has realized the flaws in replicating studies and in having enough data to truly advance the science.
“I spent five years being excited, 10 years scratching my head because of the unreliability of findings in small data samples and thinking about how to make the science more reliable, and the last five years working on ways to reconfigure the tools to make them more effective,” he said.
Thanks to a recent $5 million provisional grant from the National Institute of Neurological Disorders and Stroke (NINDS), part of the National Institutes of Health, Hillary and 37 co-investigators from all over the world are hoping to change the paradigm. Titled “Advancing Secondary Data Analysis: the ENIGMA Brain Injury Data Harmonization Initiative,” the project aims to address the “replication crisis” by building a pipeline and algorithms to combine data and create novel data sharing platforms.
According to Hillary, the ENIGMA (Enhancing Neuroimaging Genetics through Meta-Analysis) consortium has supported this kind of collaboration for over a decade across a range of clinical disorders. The new grant supports the development of tools and data processing procedures for use in the ENIGMA brain injury working group and then to develop and test a workflow for harmonized processing of behavioral data as well as structural and functional MRI data.
“MRI is a very expensive tool,” said Hillary. “I have done studies with as few as eight subjects, which of course makes the result almost impossible to replicate. And then a lot of good data is simply stored away after publication, never to be used again, when the researcher moves onto the next study.”
The grant aims to change that, he said. “We have collaborators all over the world — Norway, Germany, Australia, Israel, Pakistan — who will be sharing and analyzing their data through this pipeline. The goal is to increase sample sizes and consequently increase the possibilities for combining data in unique ways that you can’t in an individual lab. This will give us greater statistical power so we can examine effects that don’t show up in any singular data set.”
The problem to solve is how to combine and aggregate the data that comes in from all over the world, Hillary explained further.
“The data are in different forms and have different protocols. Some of the data are high quality. Some are not. During the first two years of the study, we will develop algorithms to allow us to create a pipeline where all the data can come so we can make decisions about how to use it.”
The second part of the study will involve creating a virtual sharing environment.
“Labs won’t necessarily have to send us their data,” he said. “The idea is to decentralize the analysis through virtual sharing. In doing so, we can reach out to a lab, conduct an analysis, and come back with a solution without physically moving any data. This will accelerate data sharing and allow us to rapidly determine certain findings.”
“Frank continues to innovate with his research on traumatic brain injury using cutting-edge methodological and data analytic approaches,” said Kristin Buss, head of the Department of Psychology and McCourtney Professor in Children, Work, and Families. “This new grant, from the international team that he co-organized, is very exciting because of the possible questions that can be addressed with the sharing of different types of data among researchers and clinicians.”
The ultimate goal for the project, said Hillary, is not only to develop a pipeline for his work — brain injury research — but also to move beyond imaging to other types of data and to make it a model for all areas of neuroscience, including the study of depression, schizophrenia, epilepsy, and other brain injuries and disorders.
“Ideally, anybody should be able to use this,” he said. “That’s why NINDS is funding us. Five or 10 years down the road, we want to be able to share data seamlessly between labs around the world. The goal is efficient, more reliable science that ideally results in better treatments for patients.”
This research is supported by the National Institute of Neurological Disorders and Stroke of the National Institutes of Health under Award Number R61NS120249. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.