jmtyszka / citatlaskit

Tools for applying the CIT168 atlas to individual MRI data

citatlaskit

Tools for applying the CIT168 atlas to individual MRI data

Introduction

The bash scripts provided here are wrappers for a combination of FSL and ANTs commands which map individual T1w and/or T2w structural images to most high SNR midspace templates, including MNI/ICBM152 and CIT168/HCP.

Requirements

These bash scripts wrap ANTs and FSL commands and provide a convenient starting point for registering the CIT168 atlas (or any atlas for that matter) to an individual space. The original script development was under MacOS 10.12 (Sierra), but they should work fine under most Linux and Unix variants.

Software	Description	Version	Link
FSL	Neuroimaging analysis	5.0.9+	https://fsl.fmrib.ox.ac.uk
ANTs	Image warp registration	2.1.0+	https://github.com/stnava/ANTs
CIT168	Templates and amygdala atlas	1.0.1+	http://evendim.caltech.edu/amygdala-atlas/

Registering the CIT168 atlas to an individual brain

The CIT168 atlas provides both a T1w and T2w template. Both are accurately registerd to each other (ie they're in the same space) so you can register the atlas to just an individual T1w image or to a similar pair of T1w and T2w images from an individual. If you only have T1w 3D structural images, use the tmp2ind_T1.sh script. If you have both T1w and T2w 3D structural images for an individual participant, use the tmp2ind_T1T2.sh script which makes use of both contrasts during the registration.

Run either script with no arguments for help on usage.

Creating a midspace template from your own data

We've included a pair of scripts for convenience that simply wrap the antsMultivariateTemplateConstruction.sh command. midspace_T1.sh constructs a minimum deformation midspace template from a set of T1w individual images only, and midspace_T1T2.sh constructs a T1w and T2w template pair from accurately registered T1w and T2w individual images.

To run either script:

1. Create a directory containing all your individual images.

2. If you have both T1w and T2w images for each individual, make sure that

(a) They are accurately registered and resampled to the same resolution and field of view. This can be achieved using any linear registration command that supports a mutual information metric (eg FLIRT).

(b) The subject ID appears before the contrast type (T1 or T2) in the filename. For example the image pair for subject 123456 could be named sub-123456_T1w_brain.nii.gz and sub-123456_T2w_brain.nii.gz.

3. Copy the required script to the image directory. Make sure it is executable using chmod +x *.sh and run from the image directory.

Template generation could take from several hours to several days depending on the number of individual images, image dimensions and computing resources available.

Bugs, Issues, Feature Requests

If you encounter any problems using these scripts or the CIT168 atlas, let us know by raising a new issue for this repository through https://github.com/jmtyszka/citatlaskit/issues

citatlaskit's Issues

tmp2ind_T1.sh

Hi Dr. Tyszka,
First, I want to say I am very excited about this atlas.

I am currently trying to warp an individual structural scan to the CIT168_T1w_700um_MNI.nii.gz template using tmp2ind_T1.sh. The third argument calls for a prob atlas. Is this the CIT168_iAmyNuc_700um_MNI.nii.gz file included in the atlas download? What is the difference between the "i" and "p" labeled "AmyNuc" nifti files?

Ultimately, I am attempting to warp all of my individual structural scans to your T1w 700um atlas and export the individual amygdala subfields as text files to be brought up in a separate software. How might I access individual subfield NIFTI files from the atlas download in order to convert to text files? I plan on using fsl's fsl2ascii in order to convert the nifti's but any insight on this process would be extremely helpful!

Thank you so much in advance and I hope to hear from you soon.

Sincerely,
Steven Granger