The goal of this tool is to offer a simple interface for downloading, storing, and getting access to neuroimaging datasets. We want to:
- Decrease the amount of time spent by data scientists in accessing new datasets
- Decrease the difficulty of scientists in sharing their data with the world
- Increase the visibility of available data
The types of data we wish to expose include:
- MRI / fMRI / rsMRI / dMRI
- EEG / ERP / MEG / ECoG
Not all data sources have been implemented. Our list of known data sources can be found here: https://github.com/nidata/nidata/wiki/Data-sources
Current data sources will be available via the website, when implemented: http://nidata.github.io/
nidata
is tested in Python 2.6, 2.7, and 3.4. The only package-level dependency is pip.
Individual datasets may have package dependencies for downloads or examples. If so, nidata
attempts to install them via pip. These packages include:
- nilearn - Machine learning for neuroimaging, contains generic download tools and logic for accessing fMRI datasets
- nibabel - Tools for accessing many formats of MRI data
sudo pip install git+https://github.com/nidata/nidata
To run an example,
python nidata/multimodal/hcp/example1.py
To download data,
from nidata.multimodal import HcpDataset
HcpDataset().fetch(n_subjects=1)