Code to run the analysis of the AV attention fMRI experiment (7T)

This still needs more commenting and documenting (I am still learning) but do reach out if you need any help with it.

A lot of the code is similar to that of the better documented AVT experiment analysis.

Beta values extracted from our layers / ROIs for each participant as well as the summary data necessary to reproduce the figures from the paper have been uploaded as CSV or mat files on the open-science framework

The raw data (in a BIDS compatible format) of this project are available upon request: we are still figuring out if the ethics under which this data was acquired covers open data sharing.

Group average statistical maps are available in an NIDM format from neurovault.

The results of the quality control MRIQC pipeline on the BOLD data as well as additional about motion and framewise displacement during scanning is also available from the same repository.

You will need the following softwares to run part of the analysis.

Many extra matlab functions from github and the mathwork file exchange are needed and are added to the path by the function code/subfun/Get_dependencies. Yeah this is tiring and cumbersome but that's matlab weirdness for you (“And this why we can’t have nice things. Have you heard of python?”)

You should be able to reproduce the laminar profile figures from the paper by using the following scripts on the CSV files available on OSF.

Small script to print out the results of the LMM and run the step down approach (requires the output from BOLDProfiles/Surface/make_figures_rasters.m or MVPA/make_figures_MVPA.m) also outputs tables.

You just need to specify at the top of some of those scripts where you put the .csv files from OSF and where you put the code from this repository.

The linear mixed model are estimated by make_figures_BOLD.m and make_figures_MVPA.m by calling AV-Attention-7T_code/SubFun/linear_mixed_model.m. The contrasts of the LMM are then run by display_lmm_results.m.

I indicate here the different folders where the code is kept. I try to indicate and in which order the scripts (or other manual interventions) have to be run.

Preprocessing of EPIs: code/preprocess/

1. Preprocess_01_CreateVDM.m : creates the voxel displacement map using the fieldmap
2. Preprocess_02_RealignAndUnwarp.m : realign and unwarp the EPIs

Running subject level GLM: code/ffx/

1. Analysis_FFX_Block.m : runs the subject level GLM. It must first be run a first time on smoothed images to get an inclusive mask (GLM-mask) that will be used for a second pass.

Preprocessing anatomical: code/cbs/ or sub-xx/code/cbs/ segment-layer.LayoutXML : high-res segmention and layering using the CBS tools