Overview

The goal of this course was to provide you with the necessary knowledge to process neuroimaging data using accepted best practices, including BIDS, High-Performance Computing, Containerized applications, and Datalad. Neuroimaging is a rapidly evolving discipline, and there is no way to introduce everything. However, I wanted to make sure you had the hands-on experience with the relevant computational tools to get started.

Expected Learning Outcomes

You should now understand why BIDS is a good choice.
You should be able to convert your raw DICOM data to BIDS, and understand the JSON sidecars and how to manipulate them if necessary.
You should be able to work on the HPC, including configuring your environment, transferring data, working with OOD, understanding the SLURM job system and logs, working with apptainer, using the interactive desktop, and using Matlab.
You got an introduction to Neuroimaging repositories, and using DataLad to access large datasets efficiently.
You should be familiar with several standard BIDS apps which will continue to be valuable as you work on your own data:
- BIDsonym for defacing and identifying data
- MRIQC for quality assessment
- fMRIprep for preprocessing fMRI and T1w data
- QSIprep for preprocessing and reconstructing dMRI data.
You also got an introduction to GIFT ICA for fMRI analysis. GIFT is a well-accepted tool, and Vince Calhoun's group continues to develop more approaches to the analysis of fMRI data.

Class Summary (5 pts)