Remove cardiac signal and Bandpass

AFNI Message Board
1

AFNI version info (afni -ver): 26

Hi afni experts,

I have resting state and task scan that I analyzed with afni_proc.py. But, I have not removed the cardiac signal and I did not use the bandpass option. TR = 2 sec

So, this topic relates to whether I should bandpass the data and/or whether removing the cardiac signal would remove this nuisance variable and avoid using the bandpass. I think AFNI does not recommend to use bandpass 0.01-.0.1 because of the loss of df.

-1-
Can I remove the cardiac signal after data have been processed or do I have to re-run the whole afni_proc.py? I read somewhere it's better to remove all nuisances variables together and once.

-2-
If I can remove cardiac signal after afni_proc.py, it seems I need to run 3dretroicor?
I have processed my cardiac signal with physio_calc.py and have a subj_Rest_EKG_slibase.1D file. Is it the input file to 3dretroicor for the -card flag?
And what about -cardphase -threshold -order , I'm unclear what goes here!

thanks
pb

3dretroicor \
-card subj_Rest_EKG_slibase.1D \
-cardphase ??  \
-threshold ??    \
-order ??           \
-prefix errts_subj_EKG \
errts_subj+tlrc
2

Hi PB,

To be sure, is this for a single run?

Yes, it is preferable to give this information to afni_proc.py as part of the command. The main reason is that the 8 slicewise regressors are projected out of the data very early on, before any alignment or even slice timing correction.

The other regressors in the slibase file can be regressed out just as well later on.

So if you want to regress out everything, then it would mean re-processing the data.

If you want to just re-run the final regression with the addition of the 5 RVT terms (per run), then it would not be necessary to re-do the full afni_proc.py processing. In this case it would be enough to extract those first 5 columns and pass them via -regress_extra_ortvec.

Also, physio_calc.py will very soon get a large update that will, among other things, change the outputs. For example, it will create separate outputs for slicewise regressors and volumetric regressors. And afni_proc.py is prepared to handle them as 2 separate inputs. So just be ready for that. :)

-rick