If you are viewing this file through github, please see our readthedocs page for links to render properly.
| docs | |
|---|---|
| tests |    |
| binder | 
|
| package | 
|
NiBetaSeries is BIDS-compatible application that calculates betaseries correlations. In brief, a beta coefficient (i.e. parameter estimate) is calculated for each trial (or event) resulting in a series of betas that can be used to correlate regions of interest with each other.
There are potential insights hidden in your task fMRI data. Rest fMRI enjoys a multitude of toolboxes which can be applied to task fMRI with some effort, but there are not many toolboxes that focus on making betaseries. Betaseries can then be used for correlations/classifications and a multitude of other analyses. While a couple alternatives exist (pybetaseries and BASCO), NiBetaSeries is the only application to interface with BIDS organized data with the goal of providing a commandline application experience like fmriprep.
Currently NiBetaSeries returns symmetric r-z transformed correlation matrices, with an entry for each parcel defined in the atlas. Soon, NiBetaSeries will also return the betaseries images themselves, so you can take them and flexibly apply another analysis method.
Note
The betas (i.e. parameter estimates) are generated using Least Squares Separate. Please read the :ref:`betaseries` page for more background information. There are plans to support Least Squares All in future iterations.
NiBetaSeries takes BIDS and preprocessed data as input that satisfy the BIDS deriviatives specification. In practical terms, NiBetaSeries uses the output of fmriprep, a great BIDS-compatible preprocessing tool. NiBetaSeries requires the input and the atlas to already be in the same space (e.g. MNI space). For more details, see :ref:`usage` and the tutorial (:ref:`sphx_glr_auto_examples_plot_run_nibetaseries.py`)
This is a very young project that still needs some tender loving care to grow. That's where you fit in! If you would like to contribute, please read our :ref:`code_of_conduct` and contributing page (:ref:`contributing`).
This project heavily leverages nipype, nilearn, pybids, and nistats for development. Please check out their pages and support the developers.