Cleaned movement regressors for HCP-YA 2025 release?

[Dan Lurie]

Hello,

I am wondering if the revised HCP-YA pipeline produces a file containing cleaned movement regressors (i.e. equivalent to Movement_Regressors_hp0_clean.txt from the Aging/Development releases), and if so, if there is any plan to release these.

I understand that HCP (via Glasser et al. 2019) no longer recommend aggressive removal of movement regressors in addition to sICA, but it would be useful to have these files nonetheless. 

If these files do not exist, or if there is no plan to release them, it would be great to get some guidance on how to generate these myself. I've done something similar before on the S1200 release data to get a FIX-denoised global mean signal (i.e. use the FIX mixing matrix and component signal/noise labels to remove noise variance), but I'm not entirely clear how the tICA components would fit in to this.

Thanks,

Dan

[Glasser, Matthew]

That file will exist, but will not have the effects of temporal ICA cleanup applied.  You would need to do the following:

 

1. Demean, detrend, and concatenate the movement regressors.
2. Identify the tICA timeseries for the relevant fMRI data: ${StudyFolder}/${Subject}/MNINonLinear/fsaverage_LR32k/${Subject}.${fMRIConcat}_d${tICAdim}_WF${WDs}_WR_tICA_MSMAll_ts.32k_fs_LR.sdseries.nii.
3. Get the list of noise components for the relevant data (I guess from us as I don’t know that we released that).
4. Regress the noise tICA components out of the concatenated movement regressors.
5. Unconcatinate the movement regressors.

 

This is all completely not recommended though, and we did not do this for people because we didn’t see a good reason to do it after tICA.  The only reason we made the other files available was that without temporal ICA cleanup, Mike Harms found some global signal related to moving in and out of the head coil that was encoded by movement regressors.  With temporal ICA cleanup, that will be addressed.

 

If you want a measure of when the subject was moving, you would use the unmodified movement regressors.  The effects of head motion are removed by a combination of spatial and temporal ICA clean up.  The additional effects of movement regression beyond that is to remove neural signal (Glasser et al., 2019 Neuroimage), which is why we do not do this anymore.

 

I would think about this all a different way.  With sICA and tICA cleanup the artifacts are removed.  If there are neural signal components that you don’t want in the data, you can regress them out.  It would be important to justify that decision in your methods section.  Alternatively, you might benefit from using a multi-variate approach to study different components separately.

 

Matt.

--
You received this message because you are subscribed to the Google Groups "HCP-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hcp-users+...@humanconnectome.org.
To view this discussion visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/c87d0b06-6c90-4882-bd85-a2d57b3e3cefn%40humanconnectome.org.

 

---

The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.