This repo is designed for a seamless experience of pRF mapping and is closing the gap between preprocessing (using e.g. fmriprep) and the pRF analysis (using github.com/vistalab/PRFmodel).

You can pull the latest docker: docker pull davidlinhardt/prfprepare:1.4.0 Or as singularity image: singularity build prfprepare_1.4.0.sif docker://davidlinhardt/prfprepare:1.4.0

The docker is built up in three stages:

1. build the stimuli from stimulus image and vistadisp log files (if vistadisp was used for stimulus presentation)

2. create a mask in fsnative/volume space containing all specified areas (benson, wang atlas) and output corresponding information, apply this mask and convert the preprocessed surface/volume bold files to 2D NIFIT2 files as preparation for prfanalyze

3. link the correct stimulus appertures to the respective bold files by creating events.tsv

For running prfprepare you need to provide the following folder structure:

<BASE_DIR>/BIDS has to contain: -) the preprocessed data in fsnative space (for analysis_space=fsnative) or T1w space (for analysis_type=volume) and the corresponding freesurfer segmenataion (fMRIPrep output in BIDS/derivatives/fmriprep/analysis-XX; this can be set when calling the fMRIPrep)

-) the log file from vistadisp (in BIDS/derivatives/sourcedata/vistadisplog/sub-XX/ses-XX) and the stimulus images (in BIDS/derivatives/sourcedata/stimuli)

-) unprocessed subject data in BIDS-compatible format (e.g. from HeuDiConv) for the BIDS layout

<BASE_DIR>/BIDS/

We provide an example rundocker.sh script with all necessary bindings and an example_config.json file in this repository. example_config.json, * mandatory

Subject and sessions can be defined as "all" (all that are found in BIDS folder), "001" for single or "[001,002,003]" for a list. Same works for rois. Set the analysisSpace either to "volume" for 3D analysis or "fsnative" for surface analysis.