mvpa2.datasets.sources.native.load_tutorial_data¶
-
mvpa2.datasets.sources.native.
load_tutorial_data
(path=None, roi='brain', add_fa=None, flavor=None)¶ Loads the block-design demo dataset from PyMVPA dataset DB.
Parameters: path : str, optional
Path to the directory with the extracted content of the tutorial data package. This is only necessary for accessing the full resolution data. The
1slice
, and25mm
flavors are shipped with PyMVPA itself, and the path argument is ignored for them. This function also honors the MVPA_LOCATION_TUTORIAL_DATA environment variable, and the respective configuration setting.roi : str or int or tuple or None, optional
Region Of Interest to be used for masking the dataset. If a string is given a corresponding mask image from the demo dataset will be used (mask_<str>.nii.gz). If an int value is given, the corresponding ROI is determined from the atlas image (mask_hoc.nii.gz). If a tuple is provided it may contain int values that a processed as explained before, but the union of a ROIs is taken to produce the final mask. If None, no masking is performed.
add_fa : dict, optional
Passed on to the dataset creator function (see fmri_dataset() for more information).
flavor: str, optional
Resolution flavor of the data to load. By default, the data is loaded in its original resolution. The PyMVPA source distribution contains a ‘25mm’ flavor that has been downsampled to a very coarse resolution and can be used for quick test execution. Likewise a
1slice
flavor is available that contents a full-resultion single-slice subset of the dataset.