confound2_custom for estimates of task-driven activity

[Figsandcitrus]

I want to regress-out a task estimate as a custom confound. ( https://xcpengine.readthedocs.io/modules/confound.html#confound2-custom )

I believe I should change the following in my dsn file: confound2_custom[cxt]=

Is there a way to add .1D estimates of task-driven activity files in cohort file? Currently I don't see any additional file in the custom2 folder when I do the following.

Cohort file example

id0, id1,img,confound2_custom
sub-01,run-01,/path/to/sub-01.nii.gz,/path/to/sub-01.1D

Design File This is what I assumed I should do: Paste your entire design (.dsn) file between the triple backticks

confound2_custom[cxt]= 1

Runtime Information running on Singularity tried both locally and on HCP

Additional context Further explanation for this part of the documentation would help https://xcpengine.readthedocs.io/modules/confound.html#confound2-custom

[a3sha2]

You can add it in cohort file, that is good.

This is not necessary anymore confound2_custom[cxt]= 1, just leave it empty

To verify that the custom time series are added to nuisance regressors or not, check this file *_modelParameterCount.txt in confound2 to see number of nuisance regressors. If you are regressing 36p (36 motion parameters) , the modelParameterCount will be 36+N where N is number of column of custom regressors you added.

[Figsandcitrus]

Thanks. Yes it seems it changed it to 37 from 36. There is little effect on the regress timecourse and the fcon module results, so I guess I have to review how I make my 1D file for the task, I created it in SPM12. Does it have to be in FEAT? appreciate if you can update the documentation and answer this question.

I can close this issue it was a very helpful answer!

[a3sha2]

It is not necessary to be in FEAT format but the time series has to be convolved with hemodynamic response function (HRF). Thank you, documentation will be updated!