Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysis https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysis
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to contrast the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the rest of the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
yes, that is correct
On 5/24/17 10:38 AM, Taylor Hilton wrote:
Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysis
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to contrast the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the rest of the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Okay then we’re having an issue somewhere along the line. Using the cache.th40.neg.y.ocn.dat file to make a csv file, we pulled out the individual values and compared across groups using a linear regression in R and didn’t find a group difference for each cluster. Any thoughts as to why that might be? This is assuming of course that the file goes top to bottom in ascending order, and that the columns are each cluster from left to right.
- Taylor
________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve greve@nmr.mgh.harvard.edu Sent: Wednesday, May 24, 2017 10:47:24 AM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Contrast matrix format
yes, that is correct
On 5/24/17 10:38 AM, Taylor Hilton wrote: Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysishttps://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsurfer.nmr.mgh.harvard.edu%2Ffswiki%2FFsTutorial%2FGroupAnalysis&data=01%7C01%7CBTH22%40pitt.edu%7Ce59dd79b995c400da33c08d4a2b3d3bb%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=F%2BYoNdQhEqF0iWAtpPJ8rbx6sRcEcJQgLwTm7wGUEa0%3D&reserved=0
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to contrast the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the rest of the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurferhttps://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mgh.harvard.edu%2Fmailman%2Flistinfo%2Ffreesurfer&data=01%7C01%7CBTH22%40pitt.edu%7Ce59dd79b995c400da33c08d4a2b3d3bb%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=TlpipUOHlI9wBEcBdMHORvLFpGzlav4t4ulwWjmXFDY%3D&reserved=0
After some digging I found this thread where someone was having a similar issue, so I can give this matlab code you posted a try.
http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg32951.html
________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve greve@nmr.mgh.harvard.edu Sent: Wednesday, May 24, 2017 10:47:24 AM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Contrast matrix format
yes, that is correct
On 5/24/17 10:38 AM, Taylor Hilton wrote: Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysishttps://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsurfer.nmr.mgh.harvard.edu%2Ffswiki%2FFsTutorial%2FGroupAnalysis&data=01%7C01%7CBTH22%40pitt.edu%7Ce59dd79b995c400da33c08d4a2b3d3bb%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=F%2BYoNdQhEqF0iWAtpPJ8rbx6sRcEcJQgLwTm7wGUEa0%3D&reserved=0
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to contrast the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the rest of the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurferhttps://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mgh.harvard.edu%2Fmailman%2Flistinfo%2Ffreesurfer&data=01%7C01%7CBTH22%40pitt.edu%7Ce59dd79b995c400da33c08d4a2b3d3bb%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=TlpipUOHlI9wBEcBdMHORvLFpGzlav4t4ulwWjmXFDY%3D&reserved=0
Running that code it turns up these values:
pvalues =
1.0e-04 *
0.2084 0.1849
While the p-values are 0.1340 and 0.02200 respectively in the summary file, why the difference?
________________________________ From: Hilton, Benjamin Taylor Sent: Thursday, May 25, 2017 2:43:35 PM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Contrast matrix format
After some digging I found this thread where someone was having a similar issue, so I can give this matlab code you posted a try.
http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg32951.html
________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve greve@nmr.mgh.harvard.edu Sent: Wednesday, May 24, 2017 10:47:24 AM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Contrast matrix format
yes, that is correct
On 5/24/17 10:38 AM, Taylor Hilton wrote: Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysishttps://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsurfer.nmr.mgh.harvard.edu%2Ffswiki%2FFsTutorial%2FGroupAnalysis&data=01%7C01%7CBTH22%40pitt.edu%7Ce59dd79b995c400da33c08d4a2b3d3bb%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=F%2BYoNdQhEqF0iWAtpPJ8rbx6sRcEcJQgLwTm7wGUEa0%3D&reserved=0
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to contrast the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the rest of the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurferhttps://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mgh.harvard.edu%2Fmailman%2Flistinfo%2Ffreesurfer&data=01%7C01%7CBTH22%40pitt.edu%7Ce59dd79b995c400da33c08d4a2b3d3bb%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=TlpipUOHlI9wBEcBdMHORvLFpGzlav4t4ulwWjmXFDY%3D&reserved=0
Not sure what you did. Sound like you found the max vox from the cluster, then extracted the values from each subject and ran those values through the glm to compare p-values against those in the summary file. Right? Now, which values from the summary file are you comparing against? The cluster-wise p-value is the p-value of the cluster, not of a voxel in the cluster. The cluster p-values is the chance that you see a cluster of that size by chance.
On 05/25/2017 03:06 PM, Hilton, Benjamin Taylor wrote:
Running that code it turns up these values:
pvalues =
1.0e-04 *
0.2084 0.1849While the p-values are 0.1340 and 0.02200 respectively in the summary file, why the difference?
*From:* Hilton, Benjamin Taylor *Sent:* Thursday, May 25, 2017 2:43:35 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
After some digging I found this thread where someone was having a similar issue, so I can give this matlab code you posted a try.
http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg32951.html
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve greve@nmr.mgh.harvard.edu *Sent:* Wednesday, May 24, 2017 10:47:24 AM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
yes, that is correct
On 5/24/17 10:38 AM, Taylor Hilton wrote:
Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be* https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysis https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsurfer.nmr.mgh.harvard.edu%2Ffswiki%2FFsTutorial%2FGroupAnalysis&data=01%7C01%7CBTH22%40pitt.edu%7Ce59dd79b995c400da33c08d4a2b3d3bb%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=F%2BYoNdQhEqF0iWAtpPJ8rbx6sRcEcJQgLwTm7wGUEa0%3D&reserved=0
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to contrast the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the rest of the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be* https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
I was trying to follow along with the suggestion you provided someone else here: http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg32951.html
I'm not sure how relevant that solution is to my problem though, and I'm pulling values from the "cache.th40.neg.y.ocn.dat" file which is output by the mri_glmfit-sim method. I believe this has the mean cluster value for each subject and each cluster. I am comparing the p-values to the cluster wise p-values.
________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu Sent: Thursday, May 25, 2017 6:06:30 PM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Contrast matrix format {Disarmed}
Not sure what you did. Sound like you found the max vox from the cluster, then extracted the values from each subject and ran those values through the glm to compare p-values against those in the summary file. Right? Now, which values from the summary file are you comparing against? The cluster-wise p-value is the p-value of the cluster, not of a voxel in the cluster. The cluster p-values is the chance that you see a cluster of that size by chance.
On 05/25/2017 03:06 PM, Hilton, Benjamin Taylor wrote:
Running that code it turns up these values:
pvalues =
1.0e-04 *
0.2084 0.1849While the p-values are 0.1340 and 0.02200 respectively in the summary file, why the difference?
*From:* Hilton, Benjamin Taylor *Sent:* Thursday, May 25, 2017 2:43:35 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
After some digging I found this thread where someone was having a similar issue, so I can give this matlab code you posted a try.
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-arc...
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve greve@nmr.mgh.harvard.edu *Sent:* Wednesday, May 24, 2017 10:47:24 AM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
yes, that is correct
On 5/24/17 10:38 AM, Taylor Hilton wrote:
Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be* https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsurfer.nmr.... https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsurfer.nmr.mgh.harvard.edu%2Ffswiki%2FFsTutorial%2FGroupAnalysis&data=01%7C01%7CBTH22%40pitt.edu%7Ce59dd79b995c400da33c08d4a2b3d3bb%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=F%2BYoNdQhEqF0iWAtpPJ8rbx6sRcEcJQgLwTm7wGUEa0%3D&reserved=0
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to contrast the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the rest of the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be* https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg... https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2... Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners... . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
It does not matter whether you used the maxvox or the mean of the cluster. The two p-values are measuring fundamentally different phenomena. The cluster pvalue is measuring the likelihood of seeing a cluster of that size by chance. The p-value you have computed is the likelihood of seeing a mean value in the cluster greater than 0.
On 05/26/2017 12:43 PM, Hilton, Benjamin Taylor wrote:
I was trying to follow along with the suggestion you provided someone else here: http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg32951.html
I'm not sure how relevant that solution is to my problem though, and I'm pulling values from the "cache.th40.neg.y.ocn.dat" file which is output by the mri_glmfit-sim method. I believe this has the mean cluster value for each subject and each cluster. I am comparing the p-values to the cluster wise p-values.
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu *Sent:* Thursday, May 25, 2017 6:06:30 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format {Disarmed} Not sure what you did. Sound like you found the max vox from the cluster, then extracted the values from each subject and ran those values through the glm to compare p-values against those in the summary file. Right? Now, which values from the summary file are you comparing against? The cluster-wise p-value is the p-value of the cluster, not of a voxel in the cluster. The cluster p-values is the chance that you see a cluster of that size by chance.
On 05/25/2017 03:06 PM, Hilton, Benjamin Taylor wrote:
Running that code it turns up these values:
pvalues =
1.0e-04 *
0.2084 0.1849While the p-values are 0.1340 and 0.02200 respectively in the summary file, why the difference?
*From:* Hilton, Benjamin Taylor *Sent:* Thursday, May 25, 2017 2:43:35 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
After some digging I found this thread where someone was having a similar issue, so I can give this matlab code you posted a try.
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-arc...
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve greve@nmr.mgh.harvard.edu *Sent:* Wednesday, May 24, 2017 10:47:24 AM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
yes, that is correct
On 5/24/17 10:38 AM, Taylor Hilton wrote:
Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be*
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsurfer.nmr....
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to contrast the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the rest of the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be*
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg... https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2... Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners... . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Okay but if we run a simple regression analysis on the mean values between groups for the clusters which mri_glmfit-sim found significant between groups, why isn't that significant? Are we leaving a covariate out?
________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu Sent: Friday, May 26, 2017 12:50:07 PM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Contrast matrix format {Disarmed}
It does not matter whether you used the maxvox or the mean of the cluster. The two p-values are measuring fundamentally different phenomena. The cluster pvalue is measuring the likelihood of seeing a cluster of that size by chance. The p-value you have computed is the likelihood of seeing a mean value in the cluster greater than 0.
On 05/26/2017 12:43 PM, Hilton, Benjamin Taylor wrote:
I was trying to follow along with the suggestion you provided someone else here: https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-arc...
I'm not sure how relevant that solution is to my problem though, and I'm pulling values from the "cache.th40.neg.y.ocn.dat" file which is output by the mri_glmfit-sim method. I believe this has the mean cluster value for each subject and each cluster. I am comparing the p-values to the cluster wise p-values.
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu *Sent:* Thursday, May 25, 2017 6:06:30 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format {Disarmed} Not sure what you did. Sound like you found the max vox from the cluster, then extracted the values from each subject and ran those values through the glm to compare p-values against those in the summary file. Right? Now, which values from the summary file are you comparing against? The cluster-wise p-value is the p-value of the cluster, not of a voxel in the cluster. The cluster p-values is the chance that you see a cluster of that size by chance.
On 05/25/2017 03:06 PM, Hilton, Benjamin Taylor wrote:
Running that code it turns up these values:
pvalues =
1.0e-04 *
0.2084 0.1849While the p-values are 0.1340 and 0.02200 respectively in the summary file, why the difference?
*From:* Hilton, Benjamin Taylor *Sent:* Thursday, May 25, 2017 2:43:35 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
After some digging I found this thread where someone was having a similar issue, so I can give this matlab code you posted a try.
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-arc...
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve greve@nmr.mgh.harvard.edu *Sent:* Wednesday, May 24, 2017 10:47:24 AM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
yes, that is correct
On 5/24/17 10:38 AM, Taylor Hilton wrote:
Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be*
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsurfer.nmr....
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to contrast the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the rest of the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be*
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg... https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2... Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners... . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg... https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2... Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners... . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Aren't your p-values sig? They are both in the range of .0001
On 05/26/2017 02:23 PM, Hilton, Benjamin Taylor wrote:
Okay but if we run a simple regression analysis on the mean values between groups for the clusters which mri_glmfit-sim found significant between groups, why isn't that significant? Are we leaving a covariate out?
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu *Sent:* Friday, May 26, 2017 12:50:07 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format {Disarmed} It does not matter whether you used the maxvox or the mean of the cluster. The two p-values are measuring fundamentally different phenomena. The cluster pvalue is measuring the likelihood of seeing a cluster of that size by chance. The p-value you have computed is the likelihood of seeing a mean value in the cluster greater than 0.
On 05/26/2017 12:43 PM, Hilton, Benjamin Taylor wrote:
I was trying to follow along with the suggestion you provided someone else here:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-arc...
I'm not sure how relevant that solution is to my problem though, and I'm pulling values from the "cache.th40.neg.y.ocn.dat" file which is output by the mri_glmfit-sim method. I believe this has the mean cluster value for each subject and each cluster. I am comparing the p-values to the cluster wise p-values.
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu *Sent:* Thursday, May 25, 2017 6:06:30 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format {Disarmed} Not sure what you did. Sound like you found the max vox from the cluster, then extracted the values from each subject and ran those values through the glm to compare p-values against those in the summary file. Right? Now, which values from the summary file are you comparing against? The cluster-wise p-value is the p-value of the cluster, not of a voxel in the cluster. The cluster p-values is the chance that you see a cluster of that size by chance.
On 05/25/2017 03:06 PM, Hilton, Benjamin Taylor wrote:
Running that code it turns up these values:
pvalues =
1.0e-04 *
0.2084 0.1849While the p-values are 0.1340 and 0.02200 respectively in the summary file, why the difference?
*From:* Hilton, Benjamin Taylor *Sent:* Thursday, May 25, 2017 2:43:35 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
After some digging I found this thread where someone was having a similar issue, so I can give this matlab code you posted a try.
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-arc...
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve greve@nmr.mgh.harvard.edu *Sent:* Wednesday, May 24, 2017 10:47:24 AM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
yes, that is correct
On 5/24/17 10:38 AM, Taylor Hilton wrote:
Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be*
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsurfer.nmr....
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to contrast the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the rest of the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be*
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop:
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg...
https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2...
Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners...
. If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg... https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2... Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners... . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Sorry I don't think I've been very clear, got off on a tangent about the program before that sort of sidestepped the original question.
We're running a linear regression in R with the individual values from the cache.th40.neg.y.ocn.dat file, with values of one cluster as the dependent variable and our study groups as the independent variable.
The R format being lm(cluster ~ groups).
This analysis is showing p values well above .05 (.233 and 0.887, for clusters 1 and 2 respectively, of the left hemisphere), even though the clusters are supposed to be significantly different between our groups according to the output of mri_glmfit-sim . I'd like to know if there's some covariate left out of the analysis, or am I looking at the wrong individual values perhaps? We are hoping to get the value of the cortical thickness for each individual subject in a specific cluster.
________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu Sent: Friday, May 26, 2017 5:38:35 PM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Contrast matrix format {Disarmed}
Aren't your p-values sig? They are both in the range of .0001
On 05/26/2017 02:23 PM, Hilton, Benjamin Taylor wrote:
Okay but if we run a simple regression analysis on the mean values between groups for the clusters which mri_glmfit-sim found significant between groups, why isn't that significant? Are we leaving a covariate out?
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu *Sent:* Friday, May 26, 2017 12:50:07 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format {Disarmed} It does not matter whether you used the maxvox or the mean of the cluster. The two p-values are measuring fundamentally different phenomena. The cluster pvalue is measuring the likelihood of seeing a cluster of that size by chance. The p-value you have computed is the likelihood of seeing a mean value in the cluster greater than 0.
On 05/26/2017 12:43 PM, Hilton, Benjamin Taylor wrote:
I was trying to follow along with the suggestion you provided someone else here:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-arc...
I'm not sure how relevant that solution is to my problem though, and I'm pulling values from the "cache.th40.neg.y.ocn.dat" file which is output by the mri_glmfit-sim method. I believe this has the mean cluster value for each subject and each cluster. I am comparing the p-values to the cluster wise p-values.
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu *Sent:* Thursday, May 25, 2017 6:06:30 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format {Disarmed} Not sure what you did. Sound like you found the max vox from the cluster, then extracted the values from each subject and ran those values through the glm to compare p-values against those in the summary file. Right? Now, which values from the summary file are you comparing against? The cluster-wise p-value is the p-value of the cluster, not of a voxel in the cluster. The cluster p-values is the chance that you see a cluster of that size by chance.
On 05/25/2017 03:06 PM, Hilton, Benjamin Taylor wrote:
Running that code it turns up these values:
pvalues =
1.0e-04 *
0.2084 0.1849While the p-values are 0.1340 and 0.02200 respectively in the summary file, why the difference?
*From:* Hilton, Benjamin Taylor *Sent:* Thursday, May 25, 2017 2:43:35 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
After some digging I found this thread where someone was having a similar issue, so I can give this matlab code you posted a try.
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-arc...
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve greve@nmr.mgh.harvard.edu *Sent:* Wednesday, May 24, 2017 10:47:24 AM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
yes, that is correct
On 5/24/17 10:38 AM, Taylor Hilton wrote:
Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be*
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsurfer.nmr....
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to contrast the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the rest of the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be*
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop:
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg...
https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2...
Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners...
. If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg... https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2... Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners... . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg... https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2... Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners... . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
What p-values are you comparing to? You should run the analysis with mri_glmfit.m Are you sure that the design matrix is exactly the same in R? The design matrix in Xg.dat
On 05/31/2017 02:10 PM, Hilton, Benjamin Taylor wrote:
Sorry I don't think I've been very clear, got off on a tangent about the program before that sort of sidestepped the original question.
We're running a linear regression in R with the individual values from the cache.th40.neg.y.ocn.dat file, with values of one cluster as the dependent variable and our study groups as the independent variable.
The R format being lm(cluster ~ groups).
This analysis is showing p values well above .05 (.233 and 0.887, for clusters 1 and 2 respectively, of the left hemisphere), even though the clusters are supposed to be significantly different between our groups according to the output of mri_glmfit-sim . I'd like to know if there's some covariate left out of the analysis, or am I looking at the wrong individual values perhaps? We are hoping to get the value of the cortical thickness for each individual subject in a specific cluster.
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu *Sent:* Friday, May 26, 2017 5:38:35 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format {Disarmed} Aren't your p-values sig? They are both in the range of .0001
On 05/26/2017 02:23 PM, Hilton, Benjamin Taylor wrote:
Okay but if we run a simple regression analysis on the mean values between groups for the clusters which mri_glmfit-sim found significant between groups, why isn't that significant? Are we leaving a covariate out?
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu *Sent:* Friday, May 26, 2017 12:50:07 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format {Disarmed} It does not matter whether you used the maxvox or the mean of the cluster. The two p-values are measuring fundamentally different phenomena. The cluster pvalue is measuring the likelihood of seeing a cluster of that size by chance. The p-value you have computed is the likelihood of seeing a mean value in the cluster greater than 0.
On 05/26/2017 12:43 PM, Hilton, Benjamin Taylor wrote:
I was trying to follow along with the suggestion you provided someone else here:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-arc...
I'm not sure how relevant that solution is to my problem though, and I'm pulling values from the "cache.th40.neg.y.ocn.dat" file which is output by the mri_glmfit-sim method. I believe this has the mean cluster value for each subject and each cluster. I am comparing the p-values to the cluster wise p-values.
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu *Sent:* Thursday, May 25, 2017 6:06:30 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format {Disarmed} Not sure what you did. Sound like you found the max vox from the cluster, then extracted the values from each subject and ran those values through the glm to compare p-values against those in the
summary
file. Right? Now, which values from the summary file are you comparing against? The cluster-wise p-value is the p-value of the cluster,
not of
a voxel in the cluster. The cluster p-values is the chance that
you see
a cluster of that size by chance.
On 05/25/2017 03:06 PM, Hilton, Benjamin Taylor wrote:
Running that code it turns up these values:
pvalues =
1.0e-04 *
0.2084 0.1849While the p-values are 0.1340 and 0.02200 respectively in the
summary
file, why the difference?
*From:* Hilton, Benjamin Taylor *Sent:* Thursday, May 25, 2017 2:43:35 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
After some digging I found this thread where someone was having a similar issue, so I can give this matlab code you posted a try.
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-arc...
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve greve@nmr.mgh.harvard.edu *Sent:* Wednesday, May 24, 2017 10:47:24 AM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
yes, that is correct
On 5/24/17 10:38 AM, Taylor Hilton wrote:
Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be*
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsurfer.nmr....
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to
contrast
the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the
rest of
the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be*
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop:
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg...
https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2...
Outgoing:
ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners...
. If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop:
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg...
https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2...
Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners...
. If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg... https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2... Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners... . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
I'm comparing to the cluster-wise p-value in the cluster summary file from mri_glmfit-sim. We are running mri_glmfit and then using that output for mri_glmfit-sim to determine significant clusters. It should be the same matrix though just to be certain I just ran mri_glmfit with just the subjects in the relevant groups (contrasted groups only) and it turns up only one cluster in the same location as the original first cluster but slightly smaller. I did this in case the other groups weren't being handled as expected.
The linear regression in R makes the matrix for you given two groups. I've verified that the same groups are given for each participant between the linear regression and mri_glmfit. The contrast matrix looks correct, it gives a +1 for when the participant is in the group and a +0 when not with two columns for the two groups.
________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu Sent: Wednesday, May 31, 2017 3:12:43 PM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Contrast matrix format {Disarmed}
What p-values are you comparing to? You should run the analysis with mri_glmfit.m Are you sure that the design matrix is exactly the same in R? The design matrix in Xg.dat
On 05/31/2017 02:10 PM, Hilton, Benjamin Taylor wrote:
Sorry I don't think I've been very clear, got off on a tangent about the program before that sort of sidestepped the original question.
We're running a linear regression in R with the individual values from the cache.th40.neg.y.ocn.dat file, with values of one cluster as the dependent variable and our study groups as the independent variable.
The R format being lm(cluster ~ groups).
This analysis is showing p values well above .05 (.233 and 0.887, for clusters 1 and 2 respectively, of the left hemisphere), even though the clusters are supposed to be significantly different between our groups according to the output of mri_glmfit-sim . I'd like to know if there's some covariate left out of the analysis, or am I looking at the wrong individual values perhaps? We are hoping to get the value of the cortical thickness for each individual subject in a specific cluster.
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu *Sent:* Friday, May 26, 2017 5:38:35 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format {Disarmed} Aren't your p-values sig? They are both in the range of .0001
On 05/26/2017 02:23 PM, Hilton, Benjamin Taylor wrote:
Okay but if we run a simple regression analysis on the mean values between groups for the clusters which mri_glmfit-sim found significant between groups, why isn't that significant? Are we leaving a covariate out?
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu *Sent:* Friday, May 26, 2017 12:50:07 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format {Disarmed} It does not matter whether you used the maxvox or the mean of the cluster. The two p-values are measuring fundamentally different phenomena. The cluster pvalue is measuring the likelihood of seeing a cluster of that size by chance. The p-value you have computed is the likelihood of seeing a mean value in the cluster greater than 0.
On 05/26/2017 12:43 PM, Hilton, Benjamin Taylor wrote:
I was trying to follow along with the suggestion you provided someone else here:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-arc...
I'm not sure how relevant that solution is to my problem though, and I'm pulling values from the "cache.th40.neg.y.ocn.dat" file which is output by the mri_glmfit-sim method. I believe this has the mean cluster value for each subject and each cluster. I am comparing the p-values to the cluster wise p-values.
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N Greve greve@nmr.mgh.harvard.edu *Sent:* Thursday, May 25, 2017 6:06:30 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format {Disarmed} Not sure what you did. Sound like you found the max vox from the cluster, then extracted the values from each subject and ran those values through the glm to compare p-values against those in the
summary
file. Right? Now, which values from the summary file are you comparing against? The cluster-wise p-value is the p-value of the cluster,
not of
a voxel in the cluster. The cluster p-values is the chance that
you see
a cluster of that size by chance.
On 05/25/2017 03:06 PM, Hilton, Benjamin Taylor wrote:
Running that code it turns up these values:
pvalues =
1.0e-04 *
0.2084 0.1849While the p-values are 0.1340 and 0.02200 respectively in the
summary
file, why the difference?
*From:* Hilton, Benjamin Taylor *Sent:* Thursday, May 25, 2017 2:43:35 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
After some digging I found this thread where someone was having a similar issue, so I can give this matlab code you posted a try.
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.mail-arc...
*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve greve@nmr.mgh.harvard.edu *Sent:* Wednesday, May 24, 2017 10:47:24 AM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] Contrast matrix format
yes, that is correct
On 5/24/17 10:38 AM, Taylor Hilton wrote:
Hi all,
I’m trying to perform a group analysis according to the tutorial on this page: *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be*
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsurfer.nmr....
For my own analysis, I have 4 total groups, and I’m trying to do a contrast between two of those groups.
My matrix file is formatted ‘0 1 -1 0’, since I’m trying to
contrast
the 2nd and 3rd groups.
My fsgd file is formatted as the following: GroupDescriptorFile 1 Title PYSCan Class 1 Class 2 Class 3 Class 4 Input s70032 2 Input s70038 1 …
Wanted to verify that this is correct since there doesn’t appear to be a group difference when comparing the clusters against the
rest of
the data.
Any help is very appreciated!
Sincerely, B. Taylor Hilton Lab Data Coordinator University of Pittsburgh
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu *MailScanner has detected a possible fraud attempt from "na01.safelinks.protection.outlook.com" claiming to be*
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop:
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg...
https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2...
Outgoing:
ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners...
. If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop:
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg...
https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2...
Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners...
. If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg... https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2... Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners... . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgate.nmr.mg... https://na01.safelinks.protection.outlook.com/?url=www.nmr.mgh.harvard.edu%2... Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.nmr.mg...
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.partners... . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
freesurfer@nmr.mgh.harvard.edu
-
Douglas Greve -
Douglas N Greve
-
Hilton, Benjamin Taylor -
Taylor Hilton