nimare.extract.fetch_neurostore

fetch_neurostore(version='latest', data_dir=None, overwrite=False, return_type='studyset', target='mni152_2mm', url=None, timeout=30, verify_checksum=True)[source]

Download a NeuroStore studyset release.

Added in version 0.22.0.

NeuroStore releases are the maintained source of coordinate data for NiMARE, and should be preferred over the frozen Neurosynth and NeuroQuery snapshots downloaded by fetch_neurosynth() and fetch_neuroquery(). Releases are published as archives of parquet tables at https://neurostore.org/api/neurostore-studyset-releases/ and load directly into the columnar tables a Studyset stores.

Parameters:

  • version (str, optional) – Release to download. "latest" (the default) takes the most recent dated release, "nightly" takes the rolling nightly build, and any other value is matched against the published version strings (e.g. "2026-09"). Call fetch_neurostore_releases() to see what is available.

  • data_dir (pathlib.Path or str, optional) – Path where data should be downloaded. By default, files are downloaded in the home directory. A subfolder, named neurostore, will be created in data_dir, with one subfolder per downloaded release.

  • overwrite (bool, optional) – Whether to re-download and re-extract a release that is already on disk. Default is False.

  • return_type ({"studyset", "files"}, optional) – Type of object to return. The default, "studyset", returns a Studyset; "files" returns the path to the extracted release directory without loading it.

  • target (str, optional) – Target template space for the returned Studyset. Default is "mni152_2mm". Ignored when return_type="files".

  • url (str or None, optional) – Release index to query. Default is NEUROSTORE_RELEASES_URL.

  • timeout (int or float, optional) – Timeout, in seconds, for each request. Default is 30.

  • verify_checksum (bool, optional) – Whether to check the downloaded archive against the checksum published with the release. Default is True.

Returns:

The downloaded studyset, or the path to the extracted release directory when return_type="files".

Return type:

Studyset or str

Notes

Reading the parquet tables requires pyarrow, which NiMARE declares as the parquet extra (pip install nimare[parquet]). return_type="files" works without it.

Examples

Download the most recent release and run a meta-analysis on part of it:

studyset = fetch_neurostore()
print(studyset)

See also

fetch_neurostore_releases

List the releases available for download.

nimare.io.fetch_neurostore_studyset

Download a single studyset from NeuroStore.

Examples using nimare.extract.fetch_neurostore

Neurosynth and NeuroQuery

Neurosynth and NeuroQuery

Download NeuroStore coordinates

Download NeuroStore coordinates

Machine learning in NiMARE

Machine learning in NiMARE