Cloud-Oriented NeuroImaging with BrainForge: Auto Group ICA, Managed Study Integration, and Beyond
Poster No:
1965
Submission Type:
Abstract Submission
Authors:
Bradley Baker1, Eric Verner1, Vince Calhoun2, Helen Petropoulos1, Rajikha Raja1, Jill Fries1, Sandeep Panta3, Ravi Kalyanam1, Margaret King3
Institutions:
1Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Atlanta, GA, 2Georgia State/Georgia Tech/Emory, Atlanta, GA, 3Mind Research Network, Albuquerque, NM
First Author:
Bradley Baker, MS
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Co-Author(s):
Eric Verner
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Helen Petropoulos
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Rajikha Raja
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Jill Fries
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Ravi Kalyanam
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Tri-Institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Atlanta, GA
Introduction:
Researchers working with contemporary neuroimaging studies often manage large amounts of data that are processed through numerous analysis and preprocessing pipelines. As data from many studies accumulate over time, local storage and analysis becomes burdensome. To solve these issues, many research groups are turning to cloud-based solutions not only for data storage, but also for performing various analysis steps, and obtaining important research results. In this work, we present BrainForge, a research-oriented, BIDS-compliant web platform for the management and analysis of NeuroImaging data.
Methods:
BrainForge uses a modern web architecture, which consists of a separate front end and back end. The front end is a single-page application (SPA) written in JavaScript, specifically React JS, and the back end is an API written in Python, specifically the Django Rest Framework. We also use Docker Compose to coordinate the front end, database, reverse proxy server, and message broker.
BrainForge operates on data from multiple modalities, including structural, functional, and diffusion MRI, and utilizes dockerized images of prevalent analysis toolboxes, such as the Group Independent Component Analysis (ICA) of fMRI Toolbox (GIFT), Statistical Parametric Mapping (SPM), and Freesurfer. BrainForge also offers machine learning solutions to support streamlined preprocessing, analysis, and inference.
Unlike other scientific computing platforms, BrainForge is uniquely tailored to neuroimaging research, supporting end-to-end analysis of data using popular pipelines. The combined offering of COINS and BrainForge allows investigators to manage subjects, automatically upload data from a scanner, store data in the cloud, anonymize the data, analyze various modalities, and share the results with collaborators, all within one ecosystem.
BrainForge is the only platform which provides web-based interfaces and cloud computing for Group ICA, dynamic functional network connectivity (dFNC), and multivariate analysis of covariance (MANCOVA). Furthermore, we incorporate spatially constrained ICA as a prebuilt Auto-ICA pipeline, which uses spatially constrained ICA with the NeuroMark template to generate reliable sets of ICs and connectivity for studies of any size.
Our modular, Docker-based design allows us to quickly integrate new analysis techniques and pipelines, such as analysis techniques for genomics, electroencephalography, and magnetoencephalography data. Additionally, the use of Singularity containers enables researchers to quickly analyze their data and supports reproducibility. BrainForge currently processes data on a high-performance cluster (HPC), which is also made possible by the use of Singularity.
BrainForge operates on data from multiple modalities, including structural, functional, and diffusion MRI, and utilizes dockerized images of prevalent analysis toolboxes, such as the Group Independent Component Analysis (ICA) of fMRI Toolbox (GIFT), Statistical Parametric Mapping (SPM), and Freesurfer. BrainForge also offers machine learning solutions to support streamlined preprocessing, analysis, and inference.
Unlike other scientific computing platforms, BrainForge is uniquely tailored to neuroimaging research, supporting end-to-end analysis of data using popular pipelines. The combined offering of COINS and BrainForge allows investigators to manage subjects, automatically upload data from a scanner, store data in the cloud, anonymize the data, analyze various modalities, and share the results with collaborators, all within one ecosystem.
BrainForge is the only platform which provides web-based interfaces and cloud computing for Group ICA, dynamic functional network connectivity (dFNC), and multivariate analysis of covariance (MANCOVA). Furthermore, we incorporate spatially constrained ICA as a prebuilt Auto-ICA pipeline, which uses spatially constrained ICA with the NeuroMark template to generate reliable sets of ICs and connectivity for studies of any size.
Our modular, Docker-based design allows us to quickly integrate new analysis techniques and pipelines, such as analysis techniques for genomics, electroencephalography, and magnetoencephalography data. Additionally, the use of Singularity containers enables researchers to quickly analyze their data and supports reproducibility. BrainForge currently processes data on a high-performance cluster (HPC), which is also made possible by the use of Singularity.
·The COINS managed study platform and data are stored in AWS, while the BrainForge web servers and HPC cluster are on premise at GSU. BrainForge and COINS communicate via an API to relay metadata on ne
·We show the workflow and some example output for the Auto-ICA pipeline. First, structural or functional MRI is preprocessed with voxel-based morphometry or fMRI preprocessing prior. Then, GIFT is used
Results:
BrainForge supports integration a managed study platform and is the first analysis platform to interface with the Collaborative Informatics and Neuroimaging Suite (COINS) data management platform. It also supports Amazon Web Services (AWS) and SLURM cluster storage and computation, making it a powerful and flexible platform for performing neuroimaging research without the hassle of local data management. A high-level architectural diagram is attached.
BrainForge is doing initial deployment for dozens of studies across multiple institutions. It is hosted on infrastructure within Georgia State University and utilizes a cluster of 20 nodes to perform processing, including some GPU nodes. BrainForge will be the default analysis platform for studies from the Center for Advanced Brain Imaging (CABI).
BrainForge is doing initial deployment for dozens of studies across multiple institutions. It is hosted on infrastructure within Georgia State University and utilizes a cluster of 20 nodes to perform processing, including some GPU nodes. BrainForge will be the default analysis platform for studies from the Center for Advanced Brain Imaging (CABI).
Conclusions:
BrainForge is a cloud-oriented, neuroimaging platform that allows researchers to quickly and easily process and share data in multiple modalities. It is the first web platform to feature GIFT and Auto-ICA, and its dockerized design supports rapid deployment of new analyses and supports reproducible science.
Modeling and Analysis Methods:
Methods Development
Neuroinformatics and Data Sharing:
Databasing and Data Sharing
Workflows 1
Informatics Other 2
Keywords:
Computational Neuroscience
Data analysis
Data Organization
Design and Analysis
FUNCTIONAL MRI
Machine Learning
Statistical Methods
STRUCTURAL MRI
Workflows
Other - Analysis Tools
1|2Indicates the priority used for review
My abstract is being submitted as a Software Demonstration.
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Are you Internal Review Board (IRB) certified? Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
Was any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Was any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
Provide references using author date format
Gorgolewski K. (2017), “Openneuro—a free online platform for sharing and analysis of neuroimaging data,” Organization for Human Brain Mapping. Vancouver, Canada, vol. 1677.
Scott, A. (2011), "COINS: an innovative informatics and neuroimaging tool suite built for large heterogeneous datasets." Frontiers in neuroinformatics 5, pp 33.
Rachakonda, S. (2007) "Group ICA of fMRI toolbox (GIFT) manual." http://www. nitrc. org/docman/view. php/55/295/v1. 3d_ GIFTManual
Chang L. (2018), "Crowdsourced development and validation of neurocomputational models of psychological processes", Organization for Human Brain Mapping, vol. 1678.
Avesani P. (2019), “The open diffusion data derivatives, brain data upcycling via integrated publishing of derivatives and reproducible open cloud services,” Scientific data , vol. 6, no. 1, p. 69.
Sherif T. (2014), “Cbrain: a web-based, distributed computing platform for collaborative neuroimaging research,” Frontiers in Neuroinformatics, vol. 8, p. 54.