Testing rsHRF-Python

The aim is to test the rsHRF-Toolbox (python version) against the stable-matlab version. For more details, click here.

VoxelTest

The functionality of both the toolboxes are compared against each other, for a single voxel (for a swift analysis).

VoxelTest
|   parameters.dat
|   rsHRF_matlab.m
|   rsHRF_python.py
|   compare.py
|   Data
└─>  └─> <Empty>

parameters.dat

This file contains the parameters used for HRF estimation. Each parameter is specified as a numerical value on a separate line. The correspondence is as follows:

1. <skipped>
2. order
3. TR
4. passband-LHS
5. passband-RHS
6. T
7. T0
8. min_onset_search
9. max_onset_search
10. AR_lag
11. thr
12. len
13. voxel_id

The voxel_id corresponds to the particular voxel under analysis. For more information about the other parameters, please refer to the documentation of the rsHRF-toolbox.

rsHRF_matlab.m

This script executes the matlab-version of the rsHRF-toolbox.

Executing this script retrieves the HRF (with the parameters in the parameters.dat file), using all the basis-sets, and saves the retrieved time-series as a .mat file in the Data directory. The file name contains the information of the estimation-rule that was used to retrieve the HRF.

rsHRF_python.py

This script executes the python-version of the rsHRF-toolbox.

Most of the details are similar to the rsHRF_matlab.m. The retrieved time-series gets stores in a .txt file in the Data directory.

compare.py

Having run both the above scripts, compare.py compares the retrieved HRF corresponding to the python and matlab versions. It returns the Pearsons' Correlation between the time-series estimated through each estimation rule.

Note

The fMRI data has not been provided in the repository. To download, visit this link, and set the paths to input files accordingly.