Wammes, J. D., Norman, K. A., & Turk-Browne, N. B. (2021). Increasing stimulus similarity drives nonmonotonic representational change in hippocampus. eLife.
This repository contains the code used for the analyses of this manuscript (IN PROGRESS).
Data are available via Dryad.
To starthis repository assumes that the 41 subjects' zip files, as well as fieldmaps.zip, design.zip and sorting_data.zip are housed in the folder differint/data/zip
To start, you can either run . arrange_all.sh to put all participants in the correct folder organization for the code, or run them one by one by running . arrange_data sub-${n} for each of the 41 participant numbers. From there, the code in the github repository will walk you through preprocessing and analyzing the data.
The scan sequences are as follows:
- T1 MPRAGE: TR = 2300 ms, TE = 2.27 ms, flip angle = 8 degrees, matrix = 256 x 256, slices = 208, resolution = 1 mm isotropic, GRAPPA acceleration factor=3
- T2 TSE: TR = 11390 ms, TE = 90 ms, flip angle = 150 degrees, matrix = 384 x 384, slices = 54, perpindicular to hippocampal long axis, resolution = 0.44 x 0.44 x 1.5 mm
- EPI: TR = 1.5 s, TE = 32.6 ms, flip angle = 71 degrees, matrix = 128 x 128, slices = 90, resolution = 1.5 mm isotropic
Each participants' entire analysis, once their data is places in the appropriate place (see above), is managed by a script in their directory, called analyze.sh
- Move to a participants directory (e.g.
subjects/sub-05) - Submit the analysis.sh script as a batch job (e.g.
sbatch analyze.sh) - This script will call on various others that fit GLMs to the data for each stimulus and run, run ASHS and Freesurfer, and create ROImasks from the ouput of these.
- It will also call on a script to run representational siilarity analyses, and searchlights.
Once these have run, you can move on to the jupyter notebooks (house in analysis/ipynb), which will run the analyses reported in the paper.