Using HCP longitudinal pipelines with independent structural and fMRI sessions

[Deniz Vatansever]

Dear HCP Experts,

As I have been struggling with this for a few weeks now, I would appreciate some guidance on how best to handle a longitudinal dataset within the HCP pipelines.

We have multiple structural (T1w/T2w) sessions collected longitudinally, and a larger number of task fMRI sessions acquired at separate time points that do not coincide exactly with the structural sessions. I ran PreFreeSurfer, FreeSurfer, and PostFreeSurfer on each structural session, followed by the longitudinal FreeSurfer and PostFreeSurfer pipelines to generate a long.base template.

However, I cannot directly use the longitudinal version of the fMRIVolume pipeline because it expects session-specific T1wCross2Long transforms, implicitly assuming that each fMRI session is paired with one of the structural sessions used to build the longitudinal template. While I could associate each fMRI session with the structural session that is nearest in time, this felt a bit suboptimal (though I might be wrong).

Instead, I attempted to run the fMRIVolume/Surface pipelines cross-sectionally for each fMRI session, while reusing the longitudinal base template outputs so that all data are ultimately aligned to the same subject-specific template. In doing so, I encountered several issues:

- Missing eye.dat (seems to be the same across sessions, so can copy),

- Missing MNINonLinear/BiasField (this may not be a problem as I am using SEBased bias correction for fMRI),

- Missing lh.white_deformed / rh.white_deformed, which I understand to be the MNI-registered white surfaces required for ribbon masking and surface processing.

My main questions are:

- Does this general strategy (cross-sectional fMRI pipelines anchored to a longitudinal base template) make sense within the HCP framework?

- Is there a recommended way to generate the required *_deformed surfaces (e.g., from the longitudinal PostFreeSurfer outputs), or another workaround and/or any other issues I should be watching out for?

- Or is it preferable to explicitly associate each fMRI session with a specific structural session and use the standard longitudinal fMRIVolume pipeline as intended?

Any advice or best-practice suggestions would be greatly appreciated.

Many thanks in advance,

Deniz

[Glasser, Matthew]

I would associate fMRI with the nearest structural timepoint.

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/ddd1664b-8c98-432a-a9e6-2604bd8f5308n%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.