I am attempting to create the necessary regressors for a gPPI analysis following the AFNI6 training. In the scripts it mentions to use the basis function used in the main functional analysis. In our analysis we used a TENT function. In the example it was checking for either BLOCK or GAM. Is it valid to use other functions such as TENT instead?
To be sure, how much time did you model with the TENTs? It seems likely that either GAM or BLOCK would be fine at this point.
- rick
I modeled 14 seconds (7 TENTs with 2.0s TR).
If that is TENT(0,14,7), it is actually 6 intervals, with a TR of 2.333.
In any case, either basis function should be okay.
- rick
Thank you, Rick!
On the subject of model functions there is also the balloon haemodynamic model of Buxton, Wong & Frank which is implemented at https://github.com/afni/rmafni/blob/master/src/balloon.c; this resembles a gamma function but carries the advantage of a basis in physiological modelling. (Rick & al., I wonder whether this option might be useful to mention in the “Suggested Choices for Basis Functions” at https://afni.nimh.nih.gov/pub/dist/doc/misc/Decon/2007_0504_basis_funcs.html ?)