Processing and analyzing tone-learning fMRI data. WIP - KRS 2022.10

Processing pipeline

Dicom conversion: ./dicom_conversion/

1. Peek at the dicom .tsv file using initialize_dicoms_heudiconv.sh
2. Create heuristic.py based on your MRI sequences
3. Convert dicoms to .nii using convert_dicoms_heudiconv.sh

MRI preprocessing: ./fmriprep/

1. Preprocess anatomical and functional MRI with run_fmriprep.sh

(Note: this runs using a Singularity image, so may need to create that first)

Behavioral data conversion: ./behav_conversion/

1. Run convert_behav_to_bids.py to get psychopy outputs into BIDS-compatible format

Univariate analysis: ./univariate/

1. Run univariate_analysis.py
2. Run group_level.ipynb for group-level GLM and output maps/figures

Multivariate analysis: ./multivariate/

1. Create trial-specific beta estimates with modeling_firstlevel_singleevent_LSS.py

(Note: depending on the stimulus set, this will yield different results than modeling_first_level_stimulus_perrun_LSS.py. For our 16-stimulus set, we repeat each sound 3 times per run, so these outputs would be different. For the 40-stimulus set, each sound is only used once per run, so the estimates would be the same [although the output names would be different].)

2. Create grey matter mask for searchlight using make_gm_mask.py (WIP)
3. Run whole-brain searchlight with multivariate_searchlight.py
4. Run region-based decoding with confusion_matrix_plots.py
5. (Work-in-progress) Group-level searchlight decoder statistics with group_level_searchlight_WIP.ipynb

Representational similarity analysis: ./rsa/

1. Run whole-brain RSA searchlight
2. Run region-based RSA