AnatSearchlight

This repository includes code to run anatomically informed searchlight for functional MRI data using python.

The SearchlightVolume class defines a searchlight sphere in 3D space, restricted to a specific ROI. For example, you can use it for a search light for each cerebellar voxel, using only voxels from the cerebellum.

The SearchlightSurface class defines a searchlight for a cortical hemisphere. One searchlight for each vertex is computed.

Running a searchlight analysis involves two main steps:

• define: defining the searchlights. The precomputed searchlight can be saved as a .h5 file.
• run: runs the searchlight analysis for arbitrary input data and mvpa-function. The results can be saved as nifti or cifti files.

Dependencies

The code is written in python 3.9 and uses the following libraries:

• numpy
• nibabel

We are also using the surface class written by Nick Oosterhof, which was originally part of the PyMVPA project.

Documentation and Examples

For detailed usage instructions, code examples, and API documentation:

AnatSearchlight Documentation on ReadTheDocs