cortex.mapper.get_mapper

cortex.mapper.get_mapper(subject, xfmname, type='nearest', recache=False, **kwargs)

Get a Mapper object that projects data between volume and vertex (surface) space for the given subject and transform, loading a cached version from disk if one already exists.

Parameters:

subjectstr

Subject identifier. Must exist in the pycortex database.

xfmnamestr

Transform name. Must exist in the pycortex database.

typestr, optional

Mapping method to use for projecting between volume and surface space. One of ‘nearest’, ‘trilinear’, ‘gaussian’, ‘lanczos’, ‘const_patch_nn’, ‘const_patch_trilin’, ‘const_patch_lanczos’, ‘line_nearest’, ‘line_trilinear’, ‘line_lanczos’. Default is ‘nearest’.

recachebool, optional

If True, ignore any cached mapper file on disk and recompute it. Default is False.

**kwargs

Extra keyword arguments are passed to the chosen mapper’s sampling function, and which ones matter depends on type:

sigmafloat (default 1)

Standard deviation of the gaussian kernel. Only used when type=’gaussian’.

windowint (default 3)

Kernel window size, in voxels. Used when type=’gaussian’ or type=’lanczos’.

nptsint (default 64)

Number of random sample points per patch/line segment. Used by the ‘const_patch_*’ and ‘line_*’ types.

mpbool (default True)

Whether to parallelize sampling across processes. Used by the ‘const_patch_*’ types.

Returns:

Mapper

A Mapper subclass instance (e.g. PointNN, LineTrilin, …) that can project data between volume and vertex space for this subject/transform.