How to Get Data

The released data are hosted at multiple locations and multiple access methods are supported.


All data files are directly accessible via any web-browser. This type of access is best for a quick perusal of the data and for downloading fewer than a lot of files.

Browse Datasets


The simplest way to dowload complete data releases is to obtain data archives (tarballs) from OpenfMRI. At present, OpenfMRI hosts the following datasets:

For details on these datasets and further instructions, please visit the respective pages on

For users requiring more flexibility, rsync-based data access is provided too. Available datasets can be listed via:


When using rsync to obtain a dataset, it is recommended to use the -L/--copy-links and -a/--archive flags. For example:

rsync -aL dest

Cloud Access

While users running cloud-based analysis can also use any of the methods above, please note that all datasets are also available on Amazon's S3 cloud storage. This should be the preferred data source if you are working in their cloud (e.g. with NITRC-CE). The bucket names employ the following pattern:


Geek Mode

If you consider all of the above mediocre, then this is the option for you. git-annex offers highly flexible data access with built-in versioning and enables both tracking the complete evolution of these datasets and the obtaining of any future updates. This access method allows for complete meta-data access in combination with full or partial data download. Additional data portions can be obtained at any point in the future. Previous versions of data files are also accessible.

All data repositories for this method are publicly available on GitHub (no account is required).

  1. Obtain the meta-data repository:

    git clone phase2
  2. Enter the directory and initialize:

    cd phase2
    git annex init
  3. Download (selected) data files — repeat as necessary. Example: all BOLD time series for the movie acquisition:

    git annex get sub*/ses-movie/func/*bold.nii.gz