GIFTwrap: a containerized and FAIR cloud-based implementation of the widely used GIFT toolbox: Request for supplemental funds for NIH 2R01EB006841 “Multivariate methods for identifying multi-task/mult

点击翻译按钮获取中文摘要

Abstract The GIFT software was the first software tool to implement ICA of fMRI data and group ICA which has since been widely adopted by the brain imaging community. This tool has continuously expanded over the years, now providing access to a large suite of tools including data-driven functional connectivity, over a dozen different ICA approaches, independent vector analysis, dynamic functional connectivity, spatial dynamics, connectome visu- alization, and much more that are not offered in any other single tool. It also offers fully automated ICA ap- proaches for use in individual subject prediction and classification. Though we have begun to evolve the model, GIFT is primarily still based on a standalone software development and analysis model. Modern tools have moved towards centralized analysis, comparability, analytic interaction and community development. In this sup- plement we will focus on three main goals: 1) Building of architecture improvements to facilitate FAIR principles and modernize the tools, 2) to release the GIFT tools as a brain imaging data structure app (BIDSapp) for easy use and integration into modern analysis frameworks, and to deploy GIFT in several widely used analytic plat- forms, and 3) to provide a web-based interface for individuals to run fully automated ICA analysis which requires a simple upload of data to the tools. The result will have significant impact as it will offer a large expansion of utility for a tool which already has a large user base, will open up its use to many more in the community including those focused on big data tools and deployment within a modern containerized environment, and will facilitate automation, replication, cross-study comparisons, and robustness to processing pipelines and noise via spatially constrained ICA and automated labelling of intrinsic networks (beyond just artifact versus signal).