Some issues about scrubbing

Submitted by yaosir on

Dear Chao-Gan,

I used scrubbing method to reduce the effects of head motion following reviewer's suggestion, but I have some questions about it, as follows:

1. Does this method directly delete 'bad' time points?

2. Is there any file containing the information about the remaining time points for each subjects after scrubbing? I would like to easily check which subject to be excluded.

3. The default FD used in the DPABI is 0.5, but it is 0.2 in Power et al. (2012). I have tried FD<0.2, but a great number of participants have to be excluded due to this criterion. I was wondering if FD<0.2 is too stringent and could I try more liberal criterion such as FD < 0.5 or something else?

4. I have analyzed both ALFF and rsFC measures. Since the scrubbing was done at last when preprocessing, is it meaning the ALFF data were not scrubbed?

5. I do not like scrubbing because it led to a great loss of data (both time points and participants). I have excluded participants whose head motions exceeded 3mm or 3 degree, and used friston 24 model to further reduce the impact of head motion for remaining participants. Is it possible to persure the reviewer that I have already done enough to reduce the covariates of head motion.

Thank you in advance!

Vincent

YAN Chao-Gan

Tue, 07/21/2015 - 15:00

Hi Vincent,
1. Scrubbing regression is spike regression, somehow close to delete the time points. The send scrubbing option is delete if you choose "delete".

2. Please check HeadMotion.tsv

3. I think 0.2 is too stringent. You can use it as supplementary analysis.

4. If you select the second scrubbing, ALFF is not based on scrubbed data as it depends on temporal continuity.

5. I think some reviewers will accept. I may take scrubbing as supplementary analysis if needed.

Best,

Chao-Gan