The parametric bootstrap joint (pbj) methods are designed for voxel-wise and cluster-extent inference and include semi-PBJ (sPBJ) inference that is robust to variance misspecification using an estimating equations approach. The package is designed for neuroimaging data and allows for input and output of neuroimaging data.
The pbj package is in alpha stage as many features are still being added.
- A graphical user interface
- Semiparametric longitudinal multidimensional inference (F-test in repeated measurements models)
- Semiparametric confidence set inference
- Effect size based inference
- Nonparametric bootstrap-based inference
You can install the released version of pbj from GitHub:
devtools::install_github("simonvandekar/pbj")