Hi Greg,
1) To verify PreFreeSurfer worked correctly, you can compare these files
to the templates:
${StudyFolder}/${Subject}/MNINonLinear/T1w.nii.gz
${StudyFolder}/${Subject}/MNINonLinear/T1w_restore.nii.gz
${StudyFolder}/${Subject}/MNINonLinear/T1w_restore_brain.nii.gz
${StudyFolder}/${Subject}/MNINonLinear/T2w.nii.gz
${StudyFolder}/${Subject}/MNINonLinear/T2w_restore.nii.gz
${StudyFolder}/${Subject}/MNINonLinear/T2w_restore_brain.nii.gz
A variety of images are used for fMRI processing, however if the above 6
are all okay, then everything else from PreFreeSurfer is okay.
2) We try to avoid using BET in the pipelines because it is very finicky.
Initial brain extraction is performed using a FNIRT registration to your
template data and then brining the atlas brain mask back to your
individual subject, which is more robust than BET. Final brain extraction
is performed by FreeSurfer’s more precise algorithm (though it is
“assisted” to be more robust by having an internal FreeSurfer registration
stage use the initial brain extracted image to the brain extracted
FreeSurfer template instead of the image with skull to the FreeSurfer
template with skull). I don’t recommend doing additional bias field
corrections prior to using them in the HCP Pipelines (any operation of
this sort would need to be performed identically on the T1w and T2w
images). For T1w and T2w images, the HCP Pipelines perform bias field
correction (have a look at the paper to see the methods used). Note that
the behavior of bias field correction for fMRI data is likely to change in
the future (currently it is based on the T1w and T2w computed bias field,
but whether or not to do bias field correction has become debatable in
some circumstances, and there are almost certainly a variety of better
ways to do it).
Peace,
Matt.
On 3/13/15, 8:39 AM, "Book, Gregory" <[email protected]> wrote:
>Matt,
>I finally understood how to make the template you were suggesting, I was
>making it more complicated that it needed to be. So far, using the custom
>template has salvaged a significant number of the bad analyses. Thanks!
>
>I still have a couple questions though:
>1) which image is used for the fMRI processing? T1w_acpc_dc_restore_brain
>or T1w_acpc_dc_restore_brain_1mm? The *_1mm usually have poorer skull
>stripping. Basically, for our own determination of whether the HCP
>preprocessing pipelines were successful, which image(s) should we look at?
>2) When running BET, all by itself on the MPRAGEs, we found that using
>intensity corrected images produces much better skull stripping than
>using the original images. Does the HCP preprocessing pipeline do any
>kind of intensity correction? And if we did do intensity correction on
>the MPRAGEs prior to passing them to the pipeline, would it have a
>negative effect?
>
>-Greg
>
>
>-----Original Message-----
>From: Glasser, Matthew [mailto:[email protected]]
>Sent: Tuesday, March 03, 2015 1:09 PM
>To: Book, Gregory
>Subject: Re: [HCP-Users] manual AC-PC realignment
>
>Hi Greg,
>
>If you use the files I specified from MNINonLinear, they are already
>aligned. You can reuse the existing template brain masks (and to create
>the masked templates).
>
>Peace,
>
>Matt.
>
>On 3/3/15, 10:15 AM, "Book, Gregory" <[email protected]> wrote:
>
>>Whoops, i assumed when you said template that you meant a full template.
>>Now i understand it. i assume the mean images should be realigned
>>before averaging?
>>G
>>________________________________________
>>From: Glasser, Matthew [[email protected]]
>>Sent: Tuesday, March 03, 2015 10:54 AM
>>To: Book, Gregory; [email protected]
>>Subject: Re: [HCP-Users] manual AC-PC realignment
>>
>>Hi Greg,
>>
>>I¹m a bit worried that you didn¹t just do the simple averaging of the
>>good subjects as I suggested. The reason is you will not produce the
>>same kind of templates we include with the pipelines. The brain mask
>>will not be the same and it sounds like you used a different nonlinear
>>registration algorithm, which can cause group registration drift away
>>from HCP¹s MNI templates. How about trying it the way I suggested and
>>see if the problem persists.
>>
>>Peace,
>>
>>Matt.
>>
>>From: <Book>, Gregory
>><[email protected]<mailto:[email protected]>>
>>Date: Tuesday, March 3, 2015 at 8:38 AM
>>To: Matt Glasser
>><[email protected]<mailto:[email protected]>>,
>>"[email protected]<mailto:[email protected]>"
>><[email protected]<mailto:[email protected]>>
>>Subject: RE: [HCP-Users] manual AC-PC realignment
>>
>>Hi Matt,
>>I¹ve created the templates using 42 of our good subjects. The templates
>>look like nice. I made them using the following methods, where N is 1
>>and
>>2:
>>
>>TNwTemplate: ANTs, end resolution same as input: 0.8mm
>>TNwTemplateBrain: FSL¹s BET, 0.8mm
>>TNwTemplate2mm: flirt¹s applyisoxfm to output 2.0mm images
>>TemplateMask: ended up using the T2 mask output from the BET, because
>>the BET was a little better for the T2
>>Template2mmMask: used flirt to resample the 0.8mm mask to 2.0mm
>>
>>So far, its improving some of the analyses. I have a couple questions
>>about the resulting images and what goes in to the fMRI preprocessing
>>pipeline after the structural, so I can determine by visually
>>inspecting images whether they are valid to continue on to the fMRI
>>pipeline. The following thumbnails are from the same subject 1) Which
>>of these images is used in the fMRI pipeline 2) why are they different?
>>
>>T1w_acpc_dc_restore_brain:
>>[http://ado2:8080/preview.php?image=/mount/home/pipeline/onrc/data/pipe
>>lin
>>e/S1869KAR/1/HCPMPRAGE/analysis/shots/T1w_acpc_dc_restore_brain.png]
>>
>>T1w_acpc_dc_restore_brain_1mm:
>>[http://ado2:8080/preview.php?image=/mount/home/pipeline/onrc/data/pipe
>>lin
>>e/S1869KAR/1/HCPMPRAGE/analysis/shots/T1w_acpc_dc_restore_brain_1mm.png
>>]
>>
>>Based on the first image, I¹d say this subject is fine. But based on
>>the second image, I¹d say it¹s not fine. This kind of discrepancy
>>happened with the MNI 0.7mm, 0.8mm, and our own templates.
>>Thanks!
>>-Greg
>>
>>From: Glasser, Matthew [mailto:[email protected]]
>>Sent: Tuesday, February 24, 2015 9:50 AM
>>To: Book, Gregory;
>>[email protected]<mailto:[email protected]>
>>Subject: Re: [HCP-Users] manual AC-PC realignment
>>
>>The masks would be the same, so regenerating six template files. To
>>make the non masked templates you average the files I mentioned below.
>>To make the masked (_brain) templates, apply the brain masks to the
>>appropriate files.
>>
>>Peace,.
>>
>>Matt.
>>
>>From: <Book>, Gregory
>><[email protected]<mailto:[email protected]>>
>>Date: Tuesday, February 24, 2015 at 7:32 AM
>>To: Matt Glasser
>><[email protected]<mailto:[email protected]>>,
>>"[email protected]<mailto:[email protected]>"
>><[email protected]<mailto:[email protected]>>
>>Subject: RE: [HCP-Users] manual AC-PC realignment
>>
>>Thanks. I tried the 0.8mm templates instead of the 0.7mm, so it would
>>match our data, but it didn¹t really help. So if I were to use my own
>>templates, I would need to make all 8 of the templates specified below?
>>
>>From: Glasser, Matthew [mailto:[email protected]]
>>Sent: Thursday, February 12, 2015 11:32 AM
>>To: Book, Gregory;
>>[email protected]<mailto:[email protected]>
>>Subject: Re: [HCP-Users] manual AC-PC realignment
>>
>>Well the first thing to try would be to use the 0.8mm templates instead
>>of the 0.7mm templates (note that this will generally be more robust
>>and your outputs will be sampled at the acquired resolution of 0.8mm
>>rather than being up sampled to 0.7mm). If that doesn¹t work, you¹ll
>>want to specify these variables:
>>
>> # Templates
>>T1wTemplate="${HCPPIPEDIR_Templates}/MNI152_T1_0.7mm.nii.gz" #Hires T1w
>>MNI template
>>T1wTemplateBrain="${HCPPIPEDIR_Templates}/MNI152_T1_0.7mm_brain.nii.gz"
>>#Hires brain extracted MNI template
>>T1wTemplate2mm="${HCPPIPEDIR_Templates}/MNI152_T1_2mm.nii.gz" #Lowres
>>T1w MNI template
>>T2wTemplate="${HCPPIPEDIR_Templates}/MNI152_T2_0.7mm.nii.gz" #Hires T2w
>>MNI Template
>>T2wTemplateBrain="${HCPPIPEDIR_Templates}/MNI152_T2_0.7mm_brain.nii.gz"
>>#Hires T2w brain extracted MNI Template
>>T2wTemplate2mm="${HCPPIPEDIR_Templates}/MNI152_T2_2mm.nii.gz" #Lowres
>>T2w MNI Template
>>TemplateMask="${HCPPIPEDIR_Templates}/MNI152_T1_0.7mm_brain_mask.nii.gz"
>>#Hires MNI brain mask template
>>Template2mmMask="${HCPPIPEDIR_Templates}/MNI152_T1_2mm_brain_mask_dil.nii
>>.
>>gz" #Lowres MNI brain mask template
>>
>>You can leave the masks alone (and use them to create the _brain
>>templates from the unmasked data). Those files are bias field
>>corrected, T1w and T2w files. The ones with .2. Are 2mm voxel spacing
>>instead of whatever resolution you acquired your data with.
>>
>>Peace,
>>
>>Matt.
>>
>>From: <Book>, Gregory
>><[email protected]<mailto:[email protected]>>
>>Date: Thursday, February 12, 2015 at 10:21 AM
>>To: Matt Glasser
>><[email protected]<mailto:[email protected]>>,
>>"[email protected]<mailto:[email protected]>"
>><[email protected]<mailto:[email protected]>>
>>Subject: RE: [HCP-Users] manual AC-PC realignment
>>
>>Thanks, so we would specify them in here, in the PreFS script? The
>>ŒT1wTemplate¹ variable would be the mean T1w_restore template image,
>>and the ŒT1wTemplateBrain¹ variable would be the T2w_restore.2 image?
>>
>> #Templates
>> T1wTemplate="${HCPPIPEDIR_Templates}/MNI152_T1_0.7mm.nii.gz" #Hires
>>T1w MNI template
>> T1wTemplateBrain="${HCPPIPEDIR_Templates}/MNI152_T1_0.7mm_brain.nii.gz"
>>#Hires brain extracted MNI template
>> T2wTemplate="${HCPPIPEDIR_Templates}/MNI152_T2_0.7mm.nii.gz" #Hires
>>T2w MNI Template
>> T2wTemplateBrain="${HCPPIPEDIR_Templates}/MNI152_T2_0.7mm_brain.nii.gz"
>>#Hires T2w brain extracted MNI Template
>>
>>As a side question, what are the T1w_restore.nii.gz,
>>T2w_restore.nii.gz, T1w_restore.2.nii.gz, T2w_restore.2.nii.gz files?
>>
>>
>>From: Glasser, Matthew [mailto:[email protected]]
>>Sent: Thursday, February 12, 2015 11:09 AM
>>To: Book, Gregory;
>>[email protected]<mailto:[email protected]>
>>Subject: Re: [HCP-Users] manual AC-PC realignment
>>
>>You would specify this in PreFreeSurfer. Where you put the templates
>>and what you call them is up to you, but the obvious choice is
>>${GitRepo}/global/templates/${YourStudyTemplates}.nii.gz
>>
>>Peace,
>>
>>Matt.
>>
>>From: <Book>, Gregory
>><[email protected]<mailto:[email protected]>>
>>Date: Thursday, February 12, 2015 at 8:56 AM
>>To: Matt Glasser
>><[email protected]<mailto:[email protected]>>,
>>"[email protected]<mailto:[email protected]>"
>><[email protected]<mailto:[email protected]>>
>>Subject: RE: [HCP-Users] manual AC-PC realignment
>>
>>After I create the new templates, where would I specify the new
>>template (either in the HCP scripts, or what path to place the
>>templates) when I rerun the subjects through the pipeline?
>>
>>From: Glasser, Matthew [mailto:[email protected]]
>>Sent: Wednesday, February 11, 2015 10:43 AM
>>To: Book, Gregory;
>>[email protected]<mailto:[email protected]>
>>Subject: Re: [HCP-Users] manual AC-PC realignment
>>
>>Are these being acquired axially? Are you using the 0.8mm templates?
>>If so, I think it is the different FOV that is causing the issue. I
>>recommend taking the subjects that did work and averaging these files:
>>
>>${StudyFolder}/${Subject}/MNINonLinear/T1w_restore.nii.gz
>>${StudyFolder}/${Subject}/MNINonLinear/T2w_restore.nii.gz
>>${StudyFolder}/${Subject}/MNINonLinear/T1w_restore.2.nii.gz
>>${StudyFolder}/${Subject}/MNINonLinear/T2w_restore.2.nii.gz
>>
>>Then you can use these averages in place of the current pipeline
>>templates (so now you¹ll have a study specific template that has the
>>same FOV as all your images). It might be best for you to rerun all
>>subjects with the same template just for consistency.
>>
>>Peace,
>>
>>Matt.
>>
>>From: <Book>, Gregory
>><[email protected]<mailto:[email protected]>>
>>Date: Wednesday, February 11, 2015 at 9:20 AM
>>To: Matt Glasser
>><[email protected]<mailto:[email protected]>>,
>>"[email protected]<mailto:[email protected]>"
>><[email protected]<mailto:[email protected]>>
>>Subject: RE: [HCP-Users] manual AC-PC realignment
>>
>>We run 5 MPRAGEs (*tfl3d1_16) and pass those into the HCP pipeline.
>>0.8x0.8x0.8mm voxel size. TR=2200, TE=2.88ms, flip angle 13deg. The
>>slice box does not include the neck, it cuts off just below the brain
>>stem.
>>We are aware that the quality of our images is not great. Even with 5
>>averages, it¹s not as good as the stock Siemens T1w sequence. But
>>unfortunately, we¹ve been collecting the MPRAGEs for a while now, so we
>>can¹t change the sequence mid study.
>>Just want to see if there is any manual intervention we can do that may
>>save some data.
>>-G
>>
>>From: Glasser, Matthew [mailto:[email protected]]
>>Sent: Tuesday, February 10, 2015 5:36 PM
>>To: Book, Gregory;
>>[email protected]<mailto:[email protected]>
>>Subject: Re: [HCP-Users] manual AC-PC realignment
>>
>>Can you post some details about your subjects and T1w acquestion
>>protocol. Am I correct in noting that the inferior region below the
>>brain is not being imaged?
>>
>>Peace,
>>
>>Matt.
>>
>>From: <Book>, Gregory
>><[email protected]<mailto:[email protected]>>
>>Date: Tuesday, February 10, 2015 at 4:27 PM
>>To: Matt Glasser
>><[email protected]<mailto:[email protected]>>,
>>"[email protected]<mailto:[email protected]>"
>><[email protected]<mailto:[email protected]>>
>>Subject: RE: [HCP-Users] manual AC-PC realignment
>>
>>Failure for us also means that subsequent steps don¹t work. I should
>>also include the brain extraction in the possible failures, but for any
>>that the brain extraction failed, the ac/pc alignment is usually not
>>correct going into it.
>>
>>For example, here is the original T1w:
>>[http://ado2/preview.php?image=/mount/home/pipeline/onrc/data/pipeline/
>>S18 69KAR/1/HCPMPRAGE/analysis/shots/T1w.png]
>>
>>T1w_acpc.nii.gz:
>>[http://ado2/preview.php?image=/mount/home/pipeline/onrc/data/pipeline/
>>S18 69KAR/1/HCPMPRAGE/analysis/shots/T1w_acpc.png]
>>
>>And the T1w_acpc_brain.nii.gz
>>[http://ado2/preview.php?image=/mount/home/pipeline/onrc/data/pipeline/
>>S18 69KAR/1/HCPMPRAGE/analysis/shots/T1w_acpc_brain.png]
>>
>>From: Glasser, Matthew [mailto:[email protected]]
>>Sent: Tuesday, February 10, 2015 4:54 PM
>>To: Book, Gregory;
>>[email protected]<mailto:[email protected]>
>>Subject: Re: [HCP-Users] manual AC-PC realignment
>>
>>How do you define failure and what kind of subjects are these? This is
>>only intended to be a rough alignment that gets the rest of the
>>pipeline into a ³funnel² by which the alignment continues to be
>>refined. For us, failure of the AC/PC registration means that
>>subsequent steps cannot work properly.
>>
>>Peace,
>>
>>Matt.
>>
>>From: <Book>, Gregory
>><[email protected]<mailto:[email protected]>>
>>Date: Tuesday, February 10, 2015 at 3:51 PM
>>To: "[email protected]<mailto:[email protected]>"
>><[email protected]<mailto:[email protected]>>
>>Subject: [HCP-Users] manual AC-PC realignment
>>
>>We¹ve noticed a small, but not insignificant, percentage of our
>>structural analyses appear to be failing on the AC-PC alignment step.
>>I¹m wondering if there is a intervention we can do to manually get the
>>structural into a better alignment prior to the ac/pc alignment step?
>>-Greg
>>
>>_________________________________________________
>>Gregory Book
>>Senior Technology Manager
>>Olin Neuropsychiatry Research Center, Institute of Living, Hartford
>>Hospital
>>200 Retreat Avenue
>>Hartford, CT 06106
>>Tel: 860-545-7267 Fax: 860-545-7797
>>[email protected]<mailto:[email protected]>
>>http://nidb.sourceforge.net
>>
>>
>>This e-mail message, including any attachments, is for the sole use of
>>the intended recipient(s) and may contain confidential and privileged
>>information. Any unauthorized review, use, disclosure, or distribution
>>is prohibited. If you are not the intended recipient, or an employee or
>>agent responsible for delivering the message to the intended recipient,
>>please contact the sender by reply e-mail and destroy all copies of the
>>original message, including any attachments.
>>
>>_______________________________________________
>>HCP-Users mailing list
>>[email protected]<mailto:[email protected]>
>>http://lists.humanconnectome.org/mailman/listinfo/hcp-users
>>
>>________________________________
>>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.
>>
>>________________________________
>>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.
>>
>>________________________________
>>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.
>>
>>________________________________
>>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.
>>
>>________________________________
>>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.
>>
>>________________________________
>>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.
>>
>>________________________________
>>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.
>
>
>________________________________
>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.
>
>This e-mail message, including any attachments, is for the sole use of
>the intended recipient(s) and may contain confidential and privileged
>information. Any unauthorized review, use, disclosure, or distribution is
>prohibited. If you are not the intended recipient, or an employee or
>agent responsible for delivering the message to the intended recipient,
>please contact the sender by reply e-mail and destroy all copies of the
>original message, including any attachments.
________________________________
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.
_______________________________________________
HCP-Users mailing list
[email protected]
http://lists.humanconnectome.org/mailman/listinfo/hcp-users
