Data Analysis and Storage
Data analytics is integral to research at LIBR, allowing researchers to test hypotheses and answer questions quantitatively. The computational power available allows complex analyses of neuroimaging and behavioral data to be done on-site with remotely accessible computing resources.
LIBR maintains an analysis cluster using the SLURM batch scheduler, which coordinates hundreds of simultaneous jobs. This centralized architecture is particularly efficient, giving all researchers access to much more computational power than would be feasible using local workstations.
Throughout 2021, researchers used over 13,000 days of CPU processing time. It would take over 30 years to do the same work on a typical personal computer.
LIBR also maintains all data on-site at using a massive network attached storage system, with replication at an off-site backup in Oklahoma City. This system provides excellent security and performance on-site as well as a disaster recovery plan should something happen to the LIBR campus.
All new studies conducted at LIBR are organized in accordance with the Brain Imaging Data Structure (BIDS). BIDS is a standard outlining best practices for the organization and description of complex datasets. Using BIDS facilitates collaboration with the neuroimaging community, which is increasingly adopting this standard.