HCP pipelines in Philips 3T Achieva

[Miguel Ángel Rivas Fernández]

Dear all,

I conducted the structural HCP preprocessing pipeline with our MRI data acquired with a 3T Philips Achieva scanner (32 channel head coil). The outputs are very nice, however I have two important questions:

1. We have not a T2w SPACE sequence, can I use the HCP preprocessing pipelines using MRI data acquired with a Philips 3T scanner using a T2w different sequence than those included in the HCP protocol? In particular, we have:

• Magnetization Prepared Rapid Gradient-Echo (MPRAGE) 3D T1 weighted image: repetition time/echo time =7.45 ms / 3.40 ms, flip angle = 8◦; 180 slices, voxel size = 1 ×1 x 1 mm, matrix size = 240 ×240 mm.

• 3D T2 weighted image: repetition time/echo time =2500 ms / 291.48 ms, flip angle = 90◦; 225 slices, voxel size = 0.35 × 0.35 x 0.35 mm, matrix size = 640 ×640 mm.

2. We are interested in obtain myelin maps using the T1w/T2w ratio. Taking into account that one step of the HCP preprocessing pipeline involves a coregistration between the T1w image and the T2w image, there would be some problem in estimating these maps from two images acquired with different resolution (T1w: 1mm vs T2w: 0.35mm).

Any help would be very appreciated!!

Thank you very much in advance.

Best regards,

[Glasser, Matt]

As I said on the FreeSurfer list, I’d like to know more details on how you ended up in this situation before commenting further (e.g., are you really scanning living people at 3T at 0.35mm isotropic uninterpolated resolution?).

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/9a0dbccc-4a56-4b56-9792-0f7332945f6en%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.

[Harms, Michael]

 

Maybe you could post the protocol (.txt) files that you used on the Achieva for both the T1w and T2w scans?

 

-- 

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

 

From: "Glasser, Matt" <glas...@wustl.edu>
Reply-To: "hcp-...@humanconnectome.org" <hcp-...@humanconnectome.org>
Date: Tuesday, November 15, 2022 at 11:22 AM
To: "hcp-...@humanconnectome.org" <hcp-...@humanconnectome.org>
Subject: Re: [hcp-users] HCP pipelines in Philips 3T Achieva

 

* External Email - Caution *

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/611DFC50-46DB-4F08-817D-A23767483816%40wustl.edu.

[Miguel Ángel Rivas Fernández]

Dear Matt,

Apologies for cross-posting. We did not notice that you replied us in the Freesurfer support list because we did not receive any email. Now, we read your answer in the Freesurfer mailing archive website, thanks a lot for your clear explanation. I was wrong in my previous message posted here, I confused the reconstruction voxel size with the acquisition voxel size. The acquisition voxel size of our T2w images is 0.7 mm. I understand the limitations of having a different spatial resolution in the T1w image and the T2w image, it would be necessary to check that the white and pial surfaces are locate in the right place. I have some questions on which I would appreciate your help: 

1. Bias field correction: I have seen that one of the final steps in the PreFreeSurferPipeline is to conduct a bias field correction. In Glasser & Van Essen, 2011 this method has been tested with T1w MPRAGE and T2w images acquired at Siemens (T2w SPACE sequence) and GE (T2w XETA sequence). Could we use it with images acquired on a Philips scanner?.

2. Quality control: after the FreeSurferPipeline and before the PostFreeSurferPipeline, we would like to check and correct (when necessary) pial and white matter surface errors using the Freeview toolkit and then run again the recon-all command with the required flags (i.e. autorecon2-cp, autorecon2-wm, autorecon3, etc...). Can we apply this quality control as we always do when working with FreeSurfer?. In other words, in the particular case of the HCP pipelines, it is necessary to include some additional flags when running the final recon-all before continue to the PostFreeSurferPipeline?

3. Files and software to conduct statistical analyses: after the PostFreeSurferPipeline we get the myelin maps. Which files do I need to perform the statistical analyses?  Also, which software can I use to conduct statistical analyses with files obtained with the HCP-pipelines, FreeSurfer for example? I have previously analyzed data in a .gii format using the CAT12 toolbox but I don't know if it is appropriate for this data.

Thank you very much in advance for your help.

Best regards,

[Glasser, Matt]

I would treat this as 0.7mm T2w data and downsample it to that.  You can do this by choosing the 0.7mm templates in the HCP Pipelines.

 

1. It should work similarly.  See Glasser et al., 2022 Neuroimage for caveats.
2. Yes, you can do hand edits of FreeSurfer in the HCP Pipelines. Erin Reid (CCed) can provide more details if desired. 
3. I would use PALM for statistical analysis.  Definitely read Glasser et al., 2022 Neuroimage to understand some of the problems with T1w/T2w myelin mapping analysis across subjects and how we have resolved them in the HCP.

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/7493b927-3863-46a2-8e20-fc4658ba2e9fn%40humanconnectome.org.

[Miguel Ángel Rivas Fernández]

Thank you very much for recommending this excellent paper Matt!. This correction will undoubtedly allow statistical comparisons of the relative T1w/T2w differences between individuals and groups. If I have understood correctly, this correction will be implemented in the future version of the HCP protocols.

1. Do you have an estimation of when this new version will be released?

2. Since bias field correction is conducted in the PreSurferPipeline, it will be necessary to reprocess the data again when the new version is available?.

3. While this new version is not available, we would like to do a pilot study with a group of subjects simply in order to familiarising ourselves with the procedure to be followed in PALM to conduct statistical analyses. Thus, although we are aware of the problems described in Glasser et al., 2022, what file would we select to create the 4D.nii file that we include in PALM as input, perhaps the SubID.SmoothedMyelinMap_BC.164k_fs_LR.dscalar.nii?

Thank you very much for your help.

[Glasser, Matt]

1. We are working on getting this pipeline into the HCP Pipelines repo now. 
2. No, this is a separate process only for the myelin maps that runs after existing pipelines.
3. I wouldn’t use smoothed data.  _BC removes the interesting between subject differences.

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/9ac05c24-2705-4d13-814a-4979ae9a1bc8n%40humanconnectome.org.