MRI
Reconstructing Faces from fMRI Patterns using Deep Generative Neural Networks.

OpenNeuro Accession Number: ds001761Files: 36947Size: 86.4GB

BIDS Validation

3 Warnings Valid
FilesDownloadMetadata
Reconstructing Faces from fMRI Patterns using Deep Generative Neural Networks.
Reconstructing Faces from fMRI Patterns using Deep Generative Neural Networks.
  •   CHANGES
  •   dataset_description.json
  •   participants.tsv
  •   README
  • derivatives
  • stimuli
  • sub-01
  • sub-02
  • sub-03
  • sub-04

README

Reconstructing Faces from fMRI Patterns using Deep Generative Neural Networks

Data folder related to the paper: "Reconstructing Faces from fMRI Patterns using Deep Generative Neural Networks" by VanRullen & Reddy (2019) ### General: 4 participants participated in 8 (f)MRI sessions. Each of these 8 sessions consisted of between 11 and 14 runs.

The stimuli folder contains the stimuli that were used on each trial (see description of Events.tsv below). The stimuli folder also contains a file called ImageNames2Celeba.txt, which links each stimulus with the original files in the celeba imageset.

The derivatives folder : see the corresponding README file within the derivatives folder.

The data folder for each participant is organized as follows:

  • sub/ses/anat: contains the T1 anatomical scans (defaced) for this session. Most anat folders contain one T1.nii.gz file. However, when a subject was taken out of the scanner and re-positioned during a session, a second T1 scan was performed. As a result the anat folder for this session will contain two T1.nii.gz files. Defacing was performed using Freesurfer's mri_deface (https://surfer.nmr.mgh.harvard.edu/fswiki/mri_deface).
  • sub/ses/func/: contains the _bold.nii.gz and -events.tsv files for each experimental run.
    • Task Names for BOLD runs
      • Face task: This consists of the main task during which subjects viewed unique face images on each trial.
      • Localizer tasks: In addition, a few localizer runs were performed which consisted of blocks of faces and objects.
    • Events.tsv contains stimulus information for each run.
      • onset of each event
      • duration of each event.
      • For face tasks:
        • trial_type : "fix" (fixation trial), "train_face" (a face from the training set), "test_face" (a face from the test set), "img" (an imagery trial).
        • stimulus: path to the stimulus that was displayed. The stimuli used during the experiment are in the stimulus folder.
        • one_back: a code of 1 indicates that this was a one-back trial.
      • For localizer tasks:
        • trial_type : "fix", "face" or "object" blocks

Comments

Please sign in to contribute to the discussion.