Feedback, discussions, and comments are welcome. For information on how and where to discuss this data set, please see the OpenfMRI FAQ: https://openfmri.org/faq/ item "Is there a place to discuss these datasets with the larger community?"
The data set is organized in BIDS version 1.0.0rc3 (http://bids.neuroimaging.io) format.
The participants.tsv file contains subject IDs with demographic informations as well as an inventory of the scans that are included for each subject.
The /derivaties folder contains summary information that reflects the data and its contents:
- Final_Scan_Count.pdf - Plot showing the over all scan inclusion, for quick reference.
- parameter_plots/ - Folder contains many scan parameters plotted over time. Plot symbols are color coded by imaging site. Intended to provide a general sense of protocol adherence throughout the study. Individual parameters scan be found in the scan .json sidecar file. A single file containing the combined data from all of the imaging .json sidecars if provided in parameter_plots/MR_Scan_Parameters.tsv file.
- physio_plots/ - Folder contains a plot of the physiological recording trace for the Breath Hold and Resting State functional scans. For the BHT, the instructional cue timings are represented by shaded background.
- event_plots/ - Simple plots of the function task events files. The x-axis is always time (onset), and the y-axis can be task-specific. Also intended as a quick reference or summary.
- mriqcp/ - Output of the current version (as of 27 Jan 2016) of MRIQCP (MRI Quality Control Protocol: https://github.com/poldracklab/mriqc). Included are numeric results of anatomical and functional protocols as well as single subject results plotted against group distribution.
All scan files were converted from scanner DICOM files using dcm2niix (0c9e5c8 from https://github.com/neurolabusc/dcm2niix.git). Extra DICOM metadata elements were extracted using GDCM (http://gdcm.sourceforge.net/wiki/index.php/Main_Page) and combined to form each scan's .json sidecar.
Note regarding scan and task timing: In most cases, the trigger time was provided in the task data file and has been transferred into the TaskParameter section of each scans _bold.json file. If the trigger time is available, a correction was performed to the onset times to account for trigger delay. The uncompensated onset times are included in the onset_NoTriggerAdjust column. There will be an 8 second discrepancy between the compensated and uncompensated that accounts for pre-scans (4 TRs) performed by the scanner. In the cases where the trigger time is not available, the output of (TotalScanTime - nVolsRepetitionTime) may provide an estimate of pre-scan time.
Defacing was performed using freesurfer mri_deface (https://surfer.nmr.mgh.harvard.edu/fswiki/mri_deface)
Bischoff-Grethe, Amanda et al. "A Technique for the Deidentification of Structural Brain MR Images." Human brain mapping 28.9 (2007): 892–903. PMC. Web. 27 Jan. 2016.
The larger amount of missing PAM scans is due to a task design change early in the study. It was decided that data collected before the design change would be excluded.
The Stop Signal task consisted of both a training task (no MRI) and the in-scanner fMRI task. The data from the training run is included in each subject's beh folder with the task name "stopsignaltraining".