[Freesurfer] 20% changes
Hi, We found 20% differences in temporal lobe brain volume and ca. 5% difference in the hippocampus volume between two data sets of a young normal volunteer scanned 2 days apart. We use the one-step 19 hours recon-all -all procedure and directly sum select values taken from the lh.aparc.stats, rh.aparc.stats, and wmparc.stats. The values summed were based on information given from https://mail.nmr.mgh.harvard.edu/pipermail/freesurfer/2007-April/005000.html My question is, are these observations normal for Freesurfer? Any comments are greatly appreciated. Thanks! -- Pom & Jessica ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer 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 http://www.partners.org/complianceline . 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.
Re: [Freesurfer] 20% changes
Hi Martin, Thanks for getting back to me. Actually I have two young controls who have 20% differences between two scans a day apart. Visually, there are not abnormalities for the surfaces and hippocampus in both subjects i.e. on tkmedit, the color of the hippocampus area (yellow) looks alright to me. I don't see any motion artifacts either. Jessica On Wed, Jul 27, 2011 at 12:15 PM, Martin Reuter wrote: > Hi Jessica, > > I don't think this is normal. Is this a single subject? Of course in a > single subject lots of stuff can go wrong. Are the surfaces correct? And > the hippo label? Are there motion artifacts in the image etc. > > Anyway, you should process this with the longitudinal stream: > http://freesurfer.net/fswiki/LongitudinalProcessing > which should increase repeatability. > > Best, Martin > > > On Wed, 2011-07-27 at 09:01 -0700, Jessica Liu wrote: > > Hi, > > > > We found 20% differences in temporal lobe brain volume and ca. 5% > > difference in the hippocampus volume between two data sets of a young > > normal volunteer scanned 2 days apart. > > > > We use the one-step 19 hours recon-all -all procedure and directly sum > > select values taken from the lh.aparc.stats, rh.aparc.stats, and > > wmparc.stats. The values summed were based on information given from > > > https://mail.nmr.mgh.harvard.edu/pipermail/freesurfer/2007-April/005000.html > > > > My question is, are these observations normal for Freesurfer? Any > > comments are greatly appreciated. Thanks! > > -- > > Pom & Jessica > > > > > > > > ___ > > Freesurfer mailing list > > Freesurfer@nmr.mgh.harvard.edu > > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer > > > > 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 > http://www.partners.org/complianceline . 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 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 http://www.partners.org/complianceline . 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.
Re: [Freesurfer] 20% changes
Hi Bruce, Well for one of the subjects, the total temporal lobe volume from the first scan was 149937 mm3, and the second scan's total temporal lobe volume was 183121 mm3. These numbers were found by summing the volumes from the entorhinal, fusiform, parahippocampal, temporal pole, transverse temporal, and the inferior, middle, and superior temporal sections found in the lh.aparc.stats, rh.aparc.stats, and wmparc.stats for the respective scans. Hippocampus volumes were found by summing the right and left parts found in aseg.stats. Jessica On Wed, Jul 27, 2011 at 5:37 PM, Bruce Fischl wrote: > how are you computing the 20% difference? We definitely don't see this, and > Martin has done extensive testing of repeatability. > Bruce > > > On Wed, 27 Jul 2011, Jessica Liu wrote: > > Hi Martin, >> >> Thanks for getting back to me. Actually I have two young controls who >> have 20% differences between two scans a day apart. Visually, there >> are not abnormalities for the surfaces and hippocampus in both subjects >> i.e. on tkmedit, the color of the hippocampus area (yellow) looks >> alright to me. I don't see any motion artifacts either. >> >> Jessica >> >> On Wed, Jul 27, 2011 at 12:15 PM, Martin Reuter < >> mreu...@nmr.mgh.harvard.edu> wrote: >> Hi Jessica, >> >> I don't think this is normal. Is this a single subject? Of course in >> a >> single subject lots of stuff can go wrong. Are the surfaces correct? >> And >> the hippo label? Are there motion artifacts in the image etc. >> >> Anyway, you should process this with the longitudinal stream: >> >> http://freesurfer.net/fswiki/**LongitudinalProcessing<http://freesurfer.net/fswiki/LongitudinalProcessing> >> which should increase repeatability. >> >> Best, Martin >> >> >> On Wed, 2011-07-27 at 09:01 -0700, Jessica Liu wrote: >> > Hi, >> > >> > We found 20% differences in temporal lobe brain volume and ca. 5% >> > difference in the hippocampus volume between two data sets of a young >> > normal volunteer scanned 2 days apart. >> > >> > We use the one-step 19 hours recon-all -all procedure and directly sum >> > select values taken from the lh.aparc.stats, rh.aparc.stats, and >> > wmparc.stats. The values summed were based on information given from >> > https://mail.nmr.mgh.harvard.**edu/pipermail/freesurfer/2007-** >> April/005000.html<https://mail.nmr.mgh.harvard.edu/pipermail/freesurfer/2007-April/005000.html> >> > >> > My question is, are these observations normal for Freesurfer? Any >> > comments are greatly appreciated. Thanks! >> > -- >> > Pom & Jessica >> > >> > >> > >> > __**_ >> > Freesurfer mailing list >> > Freesurfer@nmr.mgh.harvard.edu >> > https://mail.nmr.mgh.harvard.**edu/mailman/listinfo/**freesurfer<https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer> >> >> >> >> 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 >> http://www.partners.org/**complianceline<http://www.partners.org/complianceline>. >> 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 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 http://www.partners.org/complianceline . 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.
Re: [Freesurfer] 20% changes
Hi Martin, When I did longitudinal analyses in the past, the output was very strange even though I think I did everything correctly. If it's not too much trouble for you, could you please take a look at my data and tell me what I've done wrong? Since the files are so large, how would I send over my data? Also, I'm using is a head coil 1.5T single-channel GE scanner. The sequence I'm using is 3D SPGR with TR = 14 msec, TE = 6.2 msec, TI = 450.0, FOV 24 cm, data acquisition matrix = 256x192, and 124 slices with 1.6 mm thickness. Your help is truly appreciated! Jessica On Thu, Jul 28, 2011 at 7:37 AM, Martin Reuter wrote: > Hi Jessica, > > I'd recommend to run the -base and the -long on these and see what > happens. If you then still see these differences, maybe we can take a > look at your data to figure out what causes this. It could be that the > WM segmentation fails somewhere. Anyway, the longitudinal stream should > help fix these things. > > Are these mprages or multi echo mprages? 3T? what coil? > > Best, Martin > > On Wed, 2011-07-27 at 18:01 -0700, Jessica Liu wrote: > > Hi Bruce, > > > > Well for one of the subjects, the total temporal lobe volume from the > > first scan was 149937 mm3, and the second scan's total temporal lobe > > volume was 183121 mm3. These numbers were found by summing the > > volumes from the entorhinal, fusiform, parahippocampal, temporal pole, > > transverse temporal, and the inferior, middle, and superior temporal > > sections found in the lh.aparc.stats, rh.aparc.stats, and wmparc.stats > > for the respective scans. Hippocampus volumes were found by summing > > the right and left parts found in aseg.stats. > > > > Jessica > > > > On Wed, Jul 27, 2011 at 5:37 PM, Bruce Fischl > > wrote: > > how are you computing the 20% difference? We definitely don't > > see this, and Martin has done extensive testing of > > repeatability. > > Bruce > > > > > > > > On Wed, 27 Jul 2011, Jessica Liu wrote: > > > > Hi Martin, > > > > Thanks for getting back to me. Actually I have two > > young controls who have 20% differences between two > > scans a day apart. Visually, there > > are not abnormalities for the surfaces and hippocampus > > in both subjects i.e. on tkmedit, the color of the > > hippocampus area (yellow) looks > > alright to me. I don't see any motion artifacts > > either. > > > > Jessica > > > > On Wed, Jul 27, 2011 at 12:15 PM, Martin Reuter > > wrote: > > Hi Jessica, > > > > I don't think this is normal. Is this a single > > subject? Of course in a > > single subject lots of stuff can go wrong. Are > > the surfaces correct? And > > the hippo label? Are there motion artifacts in > > the image etc. > > > > Anyway, you should process this with the > > longitudinal stream: > > > > http://freesurfer.net/fswiki/LongitudinalProcessing > > which should increase repeatability. > > > > Best, Martin > > > > > > On Wed, 2011-07-27 at 09:01 -0700, Jessica Liu wrote: > > > Hi, > > > > > > We found 20% differences in temporal lobe brain > > volume and ca. 5% > > > difference in the hippocampus volume between two > > data sets of a young > > > normal volunteer scanned 2 days apart. > > > > > > We use the one-step 19 hours recon-all -all > > procedure and directly sum > > > select values taken from the lh.aparc.stats, > > rh.aparc.stats, and > > > wmparc.stats. The values summed were based on > > information given from > > > > > > https://mail.nmr.mgh.harvard.edu/pipermail/freesurfer/2007-April/005000.html > > > > > > My question is, are these observations normal for > > Freesurfer? Any > > > comments are greatly appreciated
[Freesurfer] Difference in hippocampal volume between scanners
Hi everyone! I'm currently a summer student working at a lab which uses Freesurfer to measure hippocampal volumes. I noticed that scanning the same patient first on a GE 3T (Multi-channel head coil) and then a GE 1.5T (single-channel head coil) 24 hours later resulted in a 30% increase in the left hippocampus volume and a 18% increase for the right hippocampus based on the values in aseg.stats. Can anyone please comment on this and explain the large discrepancies? Thanks! Jessica Liu ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer 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 http://www.partners.org/complianceline . 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.
Re: [Freesurfer] Difference in hippocampal volume between scanners
Hi Dr. Fischl, Thanks for your comments. The aim of my study is to analyze the brain volume in AD subjects. Most of the study was scanned on the 1.5T (single-channel head coil) with a few subjects rescanned on the 3T (12-channel head coil), so we would like to compare the brain volume of the same subject on both scanners. In addition, the protocols on the two scanners didn't match exactly due to the limitations of our over 15-years-old 1.5T scanner (i.e. non-isotropic voxel size). From your comment, I gather that the 1.5T data is rather imprecise/unreliable. Since I'm a summer student, do you recommend that I should concentrate on analyzing the 3T data only, and hopefully all of the subjects can be rescanned on the 3T? Thanks! Jessica Liu 2011/6/23 Bruce Fischl > Hi Jessica, > > the 1.5t scan looks kind of washed out and low contrast, but of course it's > hard to tell from an image. The 3T looks much crisper. Were the sequence > parameters matched? I'm not sure what you're trying to show with this. The > 3T multi-channel scan will have a much, much higher SNR than the 1.5T single > channel, so the results will not at all be comparable. > > > Bruce > > > On Thu, 23 Jun 2011, Jessica Liu wrote: > > Looking at the scans (attached, 3T on the left and 1.5T on the right, >> yellow >> hippocampus), I think you've made a good point. More comments are much >> appreciated, thanks! >> >> Jessica Liu >> >> 2011/6/23 Pedro Paulo de Magalhães Oliveira Junior >> This is not normal. Maybe you should check the aseg volume for >> improper segmentation. >> >> >> On Thu, Jun 23, 2011 at 19:27, Jessica Liu >> wrote: >> Hi everyone! >> >> I'm currently a summer student working at a lab which uses >> Freesurfer to measure hippocampal volumes. I noticed that >> scanning the same patient first on a GE 3T (Multi-channel head >> coil) and then a GE 1.5T (single-channel head coil) 24 hours >> later resulted in a 30% increase in the left hippocampus volume >> and a 18% increase for the right hippocampus based on the values >> in aseg.stats. Can anyone please comment on this and explain >> the large discrepancies? Thanks! >> >> Jessica Liu >> >> __**_ >> Freesurfer mailing list >> Freesurfer@nmr.mgh.harvard.edu >> https://mail.nmr.mgh.harvard.**edu/mailman/listinfo/**freesurfer<https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer> >> >> >> 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 >> http://www.partners.org/**complianceline<http://www.partners.org/complianceline>. >> 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 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 http://www.partners.org/complianceline . 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.
Re: [Freesurfer] Difference in hippocampal volume between scanners
Alright, thanks for all of your input! Jessica 2011/6/24 Bruce Fischl > Hi Jessica > > yes, probably. With so many differences (sequence, field strength, coil) it > will be really hard to get anything out of a combined dataset. > > sorry > Bruce > > On Fri, 24 Jun 2011, Jessica Liu wrote: > > Hi Dr. Fischl, >> >> Thanks for your comments. The aim of my study is to analyze the brain >> volume in AD subjects. Most of the study was scanned on the 1.5T >> (single-channel head coil) with a few subjects rescanned on the 3T >> (12-channel head coil), so we would like to compare the brain volume of >> the >> same subject on both scanners. In addition, the protocols on the two >> scanners didn't match exactly due to the limitations of our over >> 15-years-old 1.5T scanner (i.e. non-isotropic voxel size). From your >> comment, I gather that the 1.5T data is rather imprecise/unreliable. >> Since >> I'm a summer student, do you recommend that I should concentrate on >> analyzing the 3T data only, and hopefully all of the subjects can be >> rescanned on the 3T? Thanks! >> >> Jessica Liu >> >> 2011/6/23 Bruce Fischl >> Hi Jessica, >> >> the 1.5t scan looks kind of washed out and low contrast, but of >> course it's hard to tell from an image. The 3T looks much >> crisper. Were the sequence parameters matched? I'm not sure what >> you're trying to show with this. The 3T multi-channel scan will >> have a much, much higher SNR than the 1.5T single channel, so >> the results will not at all be comparable. >> >> Bruce >> >> >> On Thu, 23 Jun 2011, Jessica Liu wrote: >> >> Looking at the scans (attached, 3T on the left and 1.5T on >> the right, yellow >> hippocampus), I think you've made a good point. More >> comments are much >> appreciated, thanks! >> >> Jessica Liu >> >> 2011/6/23 Pedro Paulo de Magalhães Oliveira Junior >> >> This is not normal. Maybe you should check the aseg >> volume for >> improper segmentation. >> >> >> On Thu, Jun 23, 2011 at 19:27, Jessica Liu >> >> wrote: >> Hi everyone! >> >> I'm currently a summer student working at a lab which uses >> Freesurfer to measure hippocampal volumes. I noticed that >> scanning the same patient first on a GE 3T (Multi-channel >> head >> coil) and then a GE 1.5T (single-channel head coil) 24 >> hours >> later resulted in a 30% increase in the left hippocampus >> volume >> and a 18% increase for the right hippocampus based on the >> values >> in aseg.stats. Can anyone please comment on this and >> explain >> the large discrepancies? Thanks! >> >> Jessica Liu >> >> __**_ >> Freesurfer mailing list >> Freesurfer@nmr.mgh.harvard.edu >> >> https://mail.nmr.mgh.harvard.**edu/mailman/listinfo/**freesurfer<https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer> >> >> >> 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 >> >> http://www.partners.org/**complianceline<http://www.partners.org/complianceline>. >> 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 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 http://www.partners.org/complianceline . 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] recommended voxel size
Hi all, I know that freesurfer recommends an isotropic voxel size, but I found that a voxel size of 1.2x1.2x1.6 at 1.5T actually gives a better resolution for my data. Would the different voxel size significantly affect the outcome of the volumetric segmentation? Thanks! Jessica ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer 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 http://www.partners.org/complianceline . 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] longitudinal base creation error
Hi all, I'm trying to analyze longitudinal data with for a subject with six time points. However, when I try to create a within-subject template using all six time points, I get the following error message on TP 3: Resolution: 0 Iteration(single-prec): 1MRIalloc(256, 256, 256): could not allocate 262144 bytes for 140th slice Cannot allocate memory Linux HMRI-HP.SME 2.6.35.6-45.fc14.i686 #1 SMP Mon Oct 18 23:56:17 UTC 2010 i686 i686 i386 GNU/Linux recon-all -s MKA_base exited with ERRORS at Wed Jul 13 14:55:32 PDT 2011 After trying again using only five of the six time points, this occurs on TP 4: Resolution: 0 Iteration(single-prec): 1 -- diff. to prev. transform: 0.114929 Iteration(single-prec): 2terminate called after throwing an instance of 'std::bad_alloc' what(): std::bad_alloc Abort (core dumped) Linux HMRI-HP.SME 2.6.35.6-45.fc14.i686 #1 SMP Mon Oct 18 23:56:17 UTC 2010 i686 i686 i386 GNU/Linux recon-all exited with ERRORS at Wed Jul 13 11:52:31 PDT 2011 Is there any way to avoid these errors while using all of the time points? Is it even necessary to use all six time points when creating the base (it's over the course of ~1.5 years), or could I just use two or three? Thanks! Jessica ___ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer 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 http://www.partners.org/complianceline . 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.
