- Python 97.8%
- Shell 2.2%
| .python-version | ||
| analyze.py | ||
| pyproject.toml | ||
| README.md | ||
| run.sh | ||
| stats.py | ||
| uv.lock | ||
MRI hours
The goal of this little project is to answer the question: "How many fMRI hours are there on OpenNeuro?"
which was asked by pmin on Neurostars:
Total amount of fMRI hours on OpenNeuro?.
The approach was suggested by effigies in the same thread:
fetch a kilobyte from each BOLD file with datalad-fuse,
read the NIfTI header with nibabel, and calculate TR * nvols.
This relies on PyBIDS to list BOLD files and get corresponding TR. Datasets for which a BIDSLayout object can not be initialized are skipped (reasons can vary, and can also be related to modalities other than BOLD). Files for which an S3 URL is not available are also skipped. For this reason, the total time reported will be slightly lower than what could be accessed if each dataset got debugged individually, but including files accessible through standard means should be a reasonable concession.