First Release: Local CSF Pipeline

This is the first public release of the local CSF correction pipeline, a modular Python-based toolset designed to extract and regress out localized cerebrospinal fluid (CSF) signals from subcortical fMRI data. This pipeline enables more region-specific physiological noise correction, particularly in high-resolution fMRI studies.

What's Included

Core Functions

• process_roi_mask(): Resamples input ROI mask to template space (e.g., MNI).

• threshold_roi_mask(): Binarizes probabilistic masks.

• dilate_binary_roi_mask(): Dilates the binary ROI mask to create a surrounding search region for CSF extraction.

• extract_local_csf_mask(): Identifies local CSF voxels adjacent to the ROI.

• extract_local_csf_time_series(): Extracts average time series from local CSF voxels.

• add_local_csf_time_series_to_confound_file(): Appends CSF regressors to confound files.

• compute_functional_timeseries(): Computes corrected ROI time series using nuisance regression.

Please refer to the README.md for full installation instructions, usage examples, and expected input/output file structures.

Future updates will include additional code to support this workflow within a general linear model framework for task-based fMRI.