HCP-A rfMRI Preprocessed Legacy Volume Questions

[Matthew Amandola]

Hello,

I had a quick question about the HCP-Aging rfMRI Preprocessed Legacy Volume images. 

The documentation says that these images have had both GenericfMRIVolumeProcessingPipeline and hcp_fix_multi_run applied to them. 

As far as I understand, GenericfMRIVolumeProcessingPipeline applies to masking/MNI registration, and hcp_fix_multi_run applies ICA-based denoising to the data.

If I understand correctly, this means that if I wanted to use the volumetric data, since hcp_fix_multi_run has been applied to these images, that I wouldn't need to apply denoising regressors, such as white matter/CSF/detrending, or typical motion scrubbing parameters to the model. Would someone mind telling me if this is the case?

Thank you!

- Matt

[Tim Coalson]

hcp_fix_multi_run indeed takes the place of what motion regressors are intended to do.  The data also gets detrended during the process.  White matter and CSF regression have sometimes been done with masks that don't fully exclude gray matter partial voluming, and as such have ended up accidentally doing something very similar to global signal regression, with no obvious benefit to our data when the masks are properly conservative, so we generally don't recommend those.  For the purpose of removing respiration effects from the data without affecting the neural signal, we implemented temporal ICA, which is not part of hcp_fix_multi_run.

However, we recommend using surface-based processing for group analysis of cortical data, specifically the cifti-format outputs.  The reasoning for this is described in our paper here (the short version is: human cortical folding is very different across subjects in most regions, so volume registration can't fully align cortex with cortex across subjects, which is why volume-based group results are so blurry):

https://www.pnas.org/doi/10.1073/pnas.1801582115

Tim

--
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 on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/b0d5c955-c4f9-41ff-ba5d-2eeb5073a98an%40humanconnectome.org.

[Matthew Amandola]

Thanks so much for the information! If I am understanding correctly, then, the rfMRI_REST_hp0_clean.nii.gz images are preprocessed fully, and regressors/temporal filtering/scrubbing/etc do not need to be applied to these images if I wanted to do some network calculations?

This was mostly just to run some quick volumetric-based tests and comparisons to other volumetric data we have, not for the final analysis. For the final analysis we will definitely use surface-based processing.

Thank you!

- Matt

You received this message because you are subscribed to a topic in the Google Groups "HCP-Users" group.
To unsubscribe from this topic, visit https://groups.google.com/a/humanconnectome.org/d/topic/hcp-users/mJ__GJ4ULA4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to hcp-users+...@humanconnectome.org.
To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CANm3QU6avNnbnMco%3D7T_cpD4vHpYmwN6BizpqH3%2BfoSTd6ZCCQ%40mail.gmail.com.

--

Matt Amandola (he/him)

PhD Candidate in Integrative Neuroscience

Dept. of Psychology

Stony Brook University

[Tim Coalson]

Yes, you can use them directly, motion effects and intensity drift have been removed.

Tim

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CAPGh%3Dq9E-onXRQeTrokx94KeV0GOH-b9iUdRkuKmZxxSR8304g%40mail.gmail.com.

[Matthew Amandola]

Thanks so much - one last quick question. I've downloaded all of the rfMRI legacy volumes, and I noticed that many of the subjects don't have a full rfMRI_REST directory (though they have the constituent *_AP/PA/1/2 directories). Also, within the rfMRI_REST directories, some of them only have the rfMRI_REST_hp0_clean_vn.nii.gz files, rather than the rfMRI_REST_hp0_clean.nii.gz files - about ~250 of them had the rfMRI_REST_hp0_clean.nii.gz files. Was this an issue with my download, or are there subjects that do not have the full legacy volume preprocessing protocol done?

Further, just checking that the *_vn files are just the outputs from variance normalization from the ICA, and are not suitable for use in analysis. Thanks for your patience.

- Matt

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CANm3QU7EydThAGeUEeBreeKexhrPHqdDHj97DNBaX9-tZq_9nw%40mail.gmail.com.

[Elam, Jennifer]

Hi Matthew,

We didn't intentionally partially process some subjects and do intend for all subjects in the HCA rfMRI Legacy Volume package have all the files listed in the Files and Directory structure Appendix for that package.  It's possible that you downloaded the package from NDA before it was fully made in NDA and showed as ready to download on your Data Packages page. I just made the package and the NDA Download Manager lists it as having 74,458 files and being 2174 Gb.

Can you list the subjects that are missing files in your package?

Best,

Jenn

Jennifer Elam, Ph.D.
Scientific Outreach, Human Connectome Project
Washington University School of Medicine
Department of Neuroscience, Box 8108
660 South Euclid Avenue
St. Louis, MO 63110
314-362-9387
el...@wustl.edu
www.humanconnectome.org

---

From: 'Matthew Amandola' via HCP-Users <hcp-...@humanconnectome.org>
Sent: Friday, June 21, 2024 10:18 AM
To: hcp-...@humanconnectome.org <hcp-...@humanconnectome.org>
Subject: Re: [hcp-users] HCP-A rfMRI Preprocessed Legacy Volume Questions

 

* External Email - Caution *

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CAPGh%3Dq-uhU2iEJ-Pd%2BP9UpYXiG-8pr180QkC8gwwcDisNkG%3DuQ%40mail.gmail.com.

 

---

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.

[Matthew Amandola]

Sure - I've attached two CSV files.

subjects_without_rfMRI_REST_dir.csv is the list of subjects that did not have the rfMRI_REST directory, but had the constituent AP1/AP2/PA1/PA2 directories.

subjects_with_rfMRI_REST_but_only_vn.csv is the list of subjects who had the rfMRI_REST directory, but only contained the *clean_vn.nii.gz file, and no *clean.nii.gz file.

My data package looks the same as yours, so it may be an issue with my download. I will try to download it again and get back to you.

- Matt

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CH3PR02MB1030945E506A0DCE45A98DEA4CCC92%40CH3PR02MB10309.namprd02.prod.outlook.com.

[Elam, Jennifer]

Spot checking of your subject lists I don't see the missing files and directories. Hopefully your new download will have all the files.

Jenn

Jennifer Elam, Ph.D.
Scientific Outreach, Human Connectome Project
Washington University School of Medicine
Department of Neuroscience, Box 8108
660 South Euclid Avenue
St. Louis, MO 63110
314-362-9387
el...@wustl.edu
www.humanconnectome.org

---

From: 'Matthew Amandola' via HCP-Users <hcp-...@humanconnectome.org>

Sent: Friday, June 21, 2024 11:34 AM

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CAPGh%3Dq9X0Qe-tDh5aJXBRXkxDd%3DNqLbfo3yTc0649g%3D2sEeJfg%40mail.gmail.com.

[Elam, Jennifer]

To be more clear, the subjects that I checked had the expected rfMRI_REST directory and the rfMRI_REST_hp0_clean.nii.gz file (i.e. nothing was missing).

Jenn

Jennifer Elam, Ph.D.
Scientific Outreach, Human Connectome Project
Washington University School of Medicine
Department of Neuroscience, Box 8108
660 South Euclid Avenue
St. Louis, MO 63110
314-362-9387
el...@wustl.edu
www.humanconnectome.org

---

From: 'Elam, Jennifer' via HCP-Users <hcp-...@humanconnectome.org>
Sent: Friday, June 21, 2024 12:25 PM

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CH3PR02MB103091DBC1982F5431AF463E0CCC92%40CH3PR02MB10309.namprd02.prod.outlook.com.

[Tim Coalson]

Yes, _vn is just the standard deviation map that variance normalization used.

Tim

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CAPGh%3Dq-uhU2iEJ-Pd%2BP9UpYXiG-8pr180QkC8gwwcDisNkG%3DuQ%40mail.gmail.com.

[Glasser, Matthew]

At some point there is going to be a release with temporal ICA cleaned data and that would not need further denoising for structured artifacts.

Matt.

 

From: Tim Coalson <tim.c...@gmail.com>
Reply-To: "hcp-...@humanconnectome.org" <hcp-...@humanconnectome.org>
Date: Friday, June 21, 2024 at 3:35 PM
To: "hcp-...@humanconnectome.org" <hcp-...@humanconnectome.org>
Subject: Re: [hcp-users] HCP-A rfMRI Preprocessed Legacy Volume Questions

 

 

Yes, _vn is just the standard deviation map that variance normalization used.

 

Tim

 

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CANm3QU4iYM1YKG9f%2BSVGzP8bARjtBtyMVUC2xKW%2BvLnVUStjpQ%40mail.gmail.com.

[Matthew Amandola]

Hi Matt,

Would you mind clarifying what you mean by structured artifacts?

Thank you!

- Matt

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/161DEB2D-06D9-4AB5-A0C0-934434C58B1F%40wustl.edu.

[Tim Coalson]

Respiration effects are an artifact that tICA is designed to remove.  Motion-induced intensity changes in the EPI data is an artifact that spatial ICA (multirun fix) is designed to remove.  "Structured" basically just means something that ICA can model (as opposed to random noise).

Tim

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CAPGh%3Dq_keQ0YbFBr%3D31nMoLwNdYdMAWLVT13b9K5Rbe8aE_DWg%40mail.gmail.com.