Difference Between rsfMRI_REST and fMRI_CONCAT_ALL

260 views
Skip to first unread message

Gabi Champion

unread,
Sep 8, 2022, 4:49:37 PM9/8/22
to HCP-Users
Hi experts,

I was hoping someone could explain the difference between the rsfMRI_REST files in the rsfMRI Preprocessed Recommended package and the fMRI_CONCAT_ALL files in the rsfMRI Preprocessed Extended package.

I'm using the HCP-aging data and planning to do a functional connectivity analyses using regions of specific resting-state networks. There are several files in each package and I need help figuring out which files to start with.

I also wanted to see what you all thought about using the CONN toolbox. My initial plan was to use the CONN toolbox for my analysis, but in another conversation I saw an expert recommended not to use the CONN toolbox. Could someone give more insight to this please?

Thanks so much for your help and advice in advance!

Best,
Gabi Champion

Graduate Student
Cognitive and Affective Neuroscience 
Georgia State University

Glasser, Matt

unread,
Sep 8, 2022, 5:42:40 PM9/8/22
to hcp-...@humanconnectome.org

rfMRI_REST are just the concatenated resting state runs.  fMRI_CONCAT_ALL are task and resting state runs all concatenated together. 

 

I don’t know too much about the CONN toolbox; however, we would recommend using this file for any functional connectivity analysis:

 

${StudyFolder}/${Subject}/MNINonLinear/Results/rfMRI_REST/rfMRI_REST_Atlas_MSMAll_hp0_clean.dtseries.nii. 

 

You can parcellate it with wb_command -cifti-parcellate using a functionally relevant parcellation: https://balsa.wustl.edu/file/87B9N

 

And then compute area to area functional connectivity with wb_command -cifti-correlation.


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 on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/d99cc0ea-0e88-4516-920b-012ac75b3716n%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.

Gabi Champion

unread,
Sep 19, 2022, 11:24:19 AM9/19/22
to HCP-Users, glas...@wustl.edu
Thank you Matt!

I have another question I hope you can answer. Could you tell me where to find the rsfMRI_REST1_AP_hp0_clean.nii.gz files for HCP aging? It seems that those files exist for other releases, but I can't seem to find these files for the aging dataset. I'd like the cleaned files post FIXICA before they are put into CIFTI format. Do these files exist?

Thanks,
Gabi 

Glasser, Matt

unread,
Sep 19, 2022, 11:47:33 AM9/19/22
to hcp-...@humanconnectome.org, Elam, Jennifer

I see that file locally.  Perhaps Jenn can help you find it on the NDA. 

 

In general, doing cross-subject analyses of the cerebral cortex using volume data are not publishable, particularly when better aligned data are available and recommended. 

Elam, Jennifer

unread,
Sep 19, 2022, 11:57:28 AM9/19/22
to HCP-Users, Glasser, Matt
Hi Gabi,
rsfMRI_REST1_AP_Atlas_hp0_clean.dtseries.nii.gz is in fmriresults01/HCA{SubID}_V1_MR/MNINonLinear/Results/rfMRI_REST1_AP/ in the data downloaded from NDA.

File names and directory structure are listed in this Appendix to LS Release 2.0: https://www.humanconnectome.org/storage/app/media/documentation/LS2.0/LS_2.0_Release_Appendix_1.pdf
Lifespan 2.0 Reference Manual – Appendix 1 | Connectome Coordination Facility Page 5 Introduction This document lists all file names, directories, and subdirectories obtained when downloading available
Best,
Jenn
From: Gabi Champion <gabich...@gmail.com>
Sent: Monday, September 19, 2022 10:24 AM

To: HCP-Users <hcp-...@humanconnectome.org>
Cc: Glasser, Matt <glas...@wustl.edu>
Subject: Re: [hcp-users] Difference Between rsfMRI_REST and fMRI_CONCAT_ALL
 

* External Email - Caution *

Gabi Champion

unread,
Sep 19, 2022, 2:06:23 PM9/19/22
to HCP-Users, el...@wustl.edu, glas...@wustl.edu
Thank you Jenn!

Just to clarify, are these .dtseries files in CIFTI format? 

Best,
Gabi 

Elam, Jennifer

unread,
Sep 19, 2022, 2:57:37 PM9/19/22
to Gabi Champion, HCP-Users, Glasser, Matt
Yes, all *.dtseries.nii files are CIFTI (NIFTI-2).
From: Gabi Champion <gabich...@gmail.com>
Sent: Monday, September 19, 2022 1:06 PM
To: HCP-Users <hcp-...@humanconnectome.org>
Cc: Elam, Jennifer <el...@wustl.edu>; Glasser, Matt <glas...@wustl.edu>

Gabi Champion

unread,
Sep 20, 2022, 10:36:19 AM9/20/22
to HCP-Users, el...@wustl.edu, glas...@wustl.edu
Sorry for the continued questions, I'm just making sure I have clarification on everything. Can you tell me where in the preprocessing steps do the files get put into CIFTI format? 

It looks like in previous releases putting the data into CIFTI format happened after ICA was run, but for the aging data it looks like this is reversed. For example, this CONN script takes in connectome data that is not in CIFTI format (https://github.com/alfnie/conn/blob/master/conn_batch_humanconnectomeproject.m).

The file you pointed me to is what I need for my analysis, but I need it not in nifti format. If this file is only in CIFTI format, is the connectome wb the only method to convert these files to nifti?

Thank you,
Gabi 

Coalson, Timothy Scott (S&T-Student)

unread,
Sep 20, 2022, 6:32:13 PM9/20/22
to HCP-Users, el...@wustl.edu, glas...@wustl.edu
I am fairly sure that the released HCP-YA data computed ICAFIX using the cifti data, the CONN toolbox is not part of the HCP (and you don't need to follow the mistakes of others).  The functional data gets put into cifti format almost immediately after alignment (see the fMRIVolume and fMRISurface pipelines).  We may have also provided volume data cleaned using the ICAFIX computed components computed from the cifti data, but we have always recommended using the cifti-format outputs, because they have better cross-subject alignment and exclude locations like skull and white matter that don't generate functional activation signal.  You can simply do indexwise math between 91k grayordinate files and get valid results that are better localized than the "equivalent" volume pipeline, see our paper:


wb_command is probably the most capable tool for converting cifti files, and can in fact map the data back to a volume file with some effort (which can be valid when dealing with only a single subject, but for group analysis of human cortex, is a bad idea).  It can also separate the surface data into gifti files, which can be imported into freesurfer or other surface-based tools.  For tools written for volume files that don't apply spatial steps like smoothing or gradient, it can even convert to something that is formatted like a volume file, but actually contains the cifti data values as-is.  You can also read the cifti files into matlab or python (or R, with some limitations on which subtypes are supported) and do whatever you want.

What type of analysis do you want to run on this data, and what restricts that analysis such that it can't be implemented on accurately aligned surface-based data?

Tim

From: Gabi Champion <gabich...@gmail.com>
Sent: Tuesday, September 20, 2022 9:36 AM
To: HCP-Users <hcp-...@humanconnectome.org>
Cc: el...@wustl.edu <el...@wustl.edu>; glas...@wustl.edu <glas...@wustl.edu>

Subject: Re: [hcp-users] Difference Between rsfMRI_REST and fMRI_CONCAT_ALL
 
WARNING: This message has originated from an External Source. This may be a phishing expedition that can result in unauthorized access to our IT System. Please use proper judgment and caution when opening attachments, clicking links, or responding to this email.

Harms, Michael

unread,
Sep 20, 2022, 10:32:11 PM9/20/22
to hcp-...@humanconnectome.org, Elam, Jennifer, Glasser, Matt

 

It’s the other way around.  The ICA and FIX signal/noise classification is done on the volume data, and then the cleanup is applied to the CIFTI data.

 

-- 

Michael Harms, Ph.D.

-----------------------------------------------------------

Associate Professor of Psychiatry

Washington University School of Medicine

Department of Psychiatry, Box 8134

660 South Euclid Ave.                        Tel: 314-747-6173

St. Louis, MO  63110                          Email: mha...@wustl.edu

Reply all
Reply to author
Forward
0 new messages