Hello Gang!

This might be a silly question, but I was intrigued when reading about BayesianGroupAna.py. Especially since you emphasise that the standard approach can be very over-penalizing. I just don’t know what part of our traditional analysis pipeline that could be replaced by this (for both resting state and task fmri). I don’t seem to see that you input any brain data into this function?

The example command I found was this:

```
BayesianGroupAna.py -dataTable my_roi_data.txt \
-prefix dock_of_the_bayes \
-y zscore -x some_x other_x \
-chains 4 -iterations 1000 \
-plot -more_plots rhat violin
```

Where the table could look like this

```
Subj ROI some_y some_x other_x
S001 roi1 0.12 0.056 0.356
S001 roi2 0.65 0.232 0.231
S002 roi1 0.14 0.456 0.856
S002 roi2 0.64 0.432 0.431
...
```

Excuse my ignorance, but are the values in these ROIs (some_y, some_x) average beta-coefficients (or R-scores form seed based resting state) for a specific task regressor? And this makes the group analysis part (e.g. 3dMVM) obsolete? Or how do you compare these ROI-values between two groups of subjects? If this is the case, can you get the same information from this as MVM when it comes to group interactions and covariets and such?

Really sorry for a basic question but this function sound really good and something we would want to explore! I’m pretty tired of all these p-values =).

Thanks in advance!