Remove non-steady-state volumes before rapidtide

[tsalo]

What would you like to see added in fMRIPost-Rapidtide?

Rapidtide has a parameter (numtozero) that will drop initial volumes from the BOLD file. My understanding is that this is meant for removing non-steady-state volumes. If so, I'd like to do that before running rapidtide.

Do you have any interest in helping implement the feature?

Yes

Additional information / screenshots

No response

[bbfrederick]

There's actually a separate option that's somewhat more focused on that, --simcalcrange, which I added while we were working on the connectivity inflation paper. It restricts which timepoints are used in estimating the voxelwise delays and signal strengths. One of the things we saw in the FC inflation paper was that for HCP resting state data, things really didn't settle down in the noise (for frequencies above the sLFO band) until a few hundred points into the dataset. Also, because of inflation, the SNR for calculating delay will be higher near the end of the data than the beginning, so it make sense to focus on those points:

Using --simcalcrange 200 -1 (use TR's starting at 200 and going to the end of the dataset) you can actually get a better estimate of the delay. However, you can still estimate the sLFO over the ENTIRE timecourse. That way you can make voxelwise regressors that remove sLFO without having to remove points (IIRC there have been multiple debates in the fMRIPrep discussion forums about trimming data or not, and the conclusion has alway been not to do it.

As a rule of thumb, you can probably safely set the starting TR to 1/4 of the total and still get a delay estimate which is as good or better than using the entire time course.