Thank you Matt and Jenn!

We'd heard about the ICA-FIX pipeline and are very excited to give it a
test run!

Best,
Michelle

On Tue, Jan 30, 2018 at 3:27 PM, Glasser, Matthew <[email protected]>
wrote:

> Does that subject have NIFTIs and CIFTIs for that scan?
>
> As for ICA AROMA, I would recommend instead running the multi-run ICA+FIX
> pipeline on the task fMRI scans.  This should preform a bit better than ICA
> AROMA, as we got comparable classifier performance to ICA+FIX in the
> resting state scans when combining across task runs (and multi-run ICA+FIX
> clearly improves task fMRI statistics where single run ICA+FIX was more
> equivocal).  See:
>
> https://www.biorxiv.org/content/early/2017/12/13/193862 Figure 5 and page
> 3 of the main supplemental information.
>
> Script is here:
>
> https://github.com/Washington-University/Pipelines/tree/master/ICAFIX
>
> We hope to run this on all the HCP task fMRI data at some point, but there
> are many priorities and few resources…
>
> Peace,
>
> Matt.
>
> From: <[email protected]> on behalf of Michelle Chiu <
> [email protected]>
> Date: Tuesday, January 30, 2018 at 2:07 PM
> To: "[email protected]" <[email protected]>
> Subject: [HCP-Users] missing movement_regressors.txt file
>
> I'm trying to run ICA_AROMA on subjects in the HCP through AWS but I'm
> getting the following error back informing me there is no such file or
> directory (for a few subject numbers) for the following:
>
> /s3/hcp/126426/MNINonLinear/Results/tfMRI_GAMBLING_RL/
> Movement_Regressors.txt
>
> I see on Connectomedb that there is indeed 3T data for the "missing"
> subjects, although these subject folders appear to be missing in the s3/hcp
> folder I've mounted. Please advise, thanks!
>
>
> _______________________________________________
> HCP-Users mailing list
> [email protected]
> http://lists.humanconnectome.org/mailman/listinfo/hcp-users
>

_______________________________________________
HCP-Users mailing list
[email protected]
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Reply via email to