Dear Alfonso and CONN users,

I have used CONN v.16 to perform standard seed-based rs-FC analyses, using the default preprocessing pipeline for volume-based analyses on MNI-space.

A reviewer has recently asked us the following questions regarding the ART-based scrubbing: which steps of ART were implemented (e.g., spike, slice, volume correction)? what % of the data was interpolated? were there any participants that were discarded entirely, based on what threshold?

If I understand correctly, ART as implemented in CONN is only used to identify outliers (with thresholds for subject motion and a global-signal Z value), and then, these identified time-points are used as a first-level nuisance covariate in order to exclude them from subsequent analysis. They are not removed from the original time-series, nor are they interpolated. 

Am I missing something? is my description correct, and are there any other steps of the ART procedure which are being used in the default preprocessing pipeline? 

Thank you!
Zohar

Zohar,

My understanding is the same as yours, but I’d like to hear others weigh in on this. I’ve found that there can be misunderstanding when the term “scrubbing” is used, hence I avoid it and just describe what CONN is doing with the ART nuisance regressors.

Warm regards,
Jeff

Dear Jeff,
Thank you very much for your helpful response.

If anyone here has any additional insights on this issue, please inform me as I want to be completely sure before answering the Reviewers...

Thank you very much in advance,
Zohar