Hello, Freesurfer experts:
I noticed for analyzing the results from autorecon2, the intracranial volume are controlled as covariate in some publications. Is there any corresponding covariate for cortical results from autorecon3, cortical thickness, surface, and volume?
In my current study, there are two group: one is a group with less optimal developmental environment and the other is control. I noticed that without any covariate, the r square from cortical measure and group variable was really low, only accounting for 2%. Is any standard way to analyze the cortical data?
Thanks!
Karl _________________________________________________________________ MSN十年回馈，每位用户可免费获得价值25元的卡巴斯基反病毒软件2010激活码，快来领取！ http://kaba.msn.com.cn/?k=1
Hi Karl,
you could try including average cortical thickness over the entire hemisphere as a covariate. This was Mike Harms' suggestion, and I think a good one.
cheers Bruce
On Thu, 8 Apr 2010, Liukarl wrote:
Hello, Freesurfer experts:
I noticed for analyzing the results from autorecon2, the intracranial volume are controlled as covariate in some publications. Is there any corresponding covariate for cortical results from autorecon3, cortical thickness, surface, and volume?
In my current study, there are two group: one is a group with less optimal developmental environment and the other is control. I noticed that without any covariate, the r square from cortical measure and group variable was really low, only accounting for 2%. Is any standard way to analyze the cortical data?
Thanks!
Karl _________________________________________________________________ MSNÊ®Äê»ØÀ¡£¬Ã¿Î»Óû§¿ÉÃâ·Ñ»ñµÃ¼ÛÖµ25ÔªµÄ¿¨°Í˹»ù·´²¡¶¾Èí¼þ2010¼¤»îÂ룬¿ìÀ´ÁìÈ¡£¡ http://kaba.msn.com.cn/?k=1
Hi,
you could try including average cortical thickness over the entire
hemisphere as a covariate. This was Mike Harms' suggestion, and I think a good one.
why is it a good one? for cortical thickness: i'm not so sure. my 2 cents (and i may be wrong!): - controlling for intracranial volume is done to remove differences due to overall size variability. - the relationship between thickness and size is much less strong than between volume and size (or surface and size). - suppose your groups differ in mean thickness, if you would like to include mean thickness as a covariate in a mapping study (vertex-wise comparisons) of cortical thickness you would have to show that the group-difference in mean thickness is due to an effect that's spread out over the whole cortex (an overall effect, not localized), otherwise the mean thickness covariate is not a good estimator of the overall size factor that you'd like to remove.
check Peter Kochunov's recent publications on this.
-joost
cheers
Bruce
On Thu, 8 Apr 2010, Liukarl wrote:
Hello, Freesurfer experts:
I noticed for analyzing the results from autorecon2, the intracranial volume are controlled as covariate in some publications. Is there any corresponding covariate for cortical results from autorecon3, cortical thickness, surface, and volume?
In my current study, there are two group: one is a group with less optimal developmental environment and the other is control. I noticed that without any covariate, the r square from cortical measure and group variable was really low, only accounting for 2%. Is any standard way to analyze the cortical data?
Thanks!
Karl _________________________________________________________________ MSN十年回馈，每位用户可免费获得价值25元的卡巴斯基反病毒软件2010激活码，快来领取！ http://kaba.msn.com.cn/?k=1
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.
it all depends on your hypothesis. If you think there are regionally varying differences in the thickness and want to be conservative, then overall mean is a good way to test this, and is probably a more stable measure than ICV. It's all speculation I think as no one has really quantified the effects of different normalizations a as far as I know
cheers Bruce
On Thu, 8 Apr 2010, j janssen wrote:
Hi,
you could try including average cortical thickness over the entire
hemisphere as a covariate. This was Mike Harms' suggestion, and I think a good one.
why is it a good one? for cortical thickness: i'm not so sure. my 2 cents (and i may be wrong!):
- controlling for intracranial volume is done to remove differences due to
overall size variability.
- the relationship between thickness and size is much less strong than
between volume and size (or surface and size).
- suppose your groups differ in mean thickness, if you would like to include
mean thickness as a covariate in a mapping study (vertex-wise comparisons) of cortical thickness you would have to show that the group-difference in mean thickness is due to an effect that's spread out over the whole cortex (an overall effect, not localized), otherwise the mean thickness covariate is not a good estimator of the overall size factor that you'd like to remove.
check Peter Kochunov's recent publications on this.
-joost
cheers
Bruce
On Thu, 8 Apr 2010, Liukarl wrote:
Hello, Freesurfer experts:
I noticed for analyzing the results from autorecon2, the intracranial volume are controlled as covariate in some publications. Is there any corresponding covariate for cortical results from autorecon3, cortical thickness, surface, and volume?
In my current study, there are two group: one is a group with less optimal developmental environment and the other is control. I noticed that without any covariate, the r square from cortical measure and group variable was really low, only accounting for 2%. Is any standard way to analyze the cortical data?
Thanks!
Karl _________________________________________________________________ MSNÊ®Äê»ØÀ¡£¬Ã¿Î»Óû§¿ÉÃâ·Ñ»ñµÃ¼ÛÖµ25ÔªµÄ¿¨°Í˹»ù·´²¡¶¾Èí¼þ2010¼¤»îÂ룬¿ìÀ´ÁìÈ¡£¡ http://kaba.msn.com.cn/?k=1
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.
Hi Joost, If you want to control for possible global differences in thickness, then an appropriate covariate of some sort is need. To me, a logical covariate for a thickness analysis is the mean cortical thickness, as this is more directly related to the measure of interest than something like the cube-root of volume (which other studies have used when mean cortical thickness was not available as a measure).
It is true that a difference in mean thickness doesn't necessarily imply that the difference in thickness is truly equally spread out over the entire cortex. But an analogous comment could be applied to volume analyses -- a difference in total brain volume between groups doesn't necessarily imply that differences in volume are present to an equal degree in all structures. Yet, using a volume covariate of some sort is a standard part of analyses of regional volumes. As a practical matter, I think it is fair to say that you are not going to find differences in mean thickness or total volume between groups unless a "substantial" portion of the brain is affected.
cheers, -Mike H.
On Thu, 2010-04-08 at 10:39 +0200, j janssen wrote:
Hi,
you could try including average cortical thickness over the entire hemisphere as a covariate. This was Mike Harms' suggestion, and I think a good one.why is it a good one? for cortical thickness: i'm not so sure. my 2 cents (and i may be wrong!):
- controlling for intracranial volume is done to remove differences
due to overall size variability.
- the relationship between thickness and size is much less strong than
between volume and size (or surface and size).
- suppose your groups differ in mean thickness, if you would like to
include mean thickness as a covariate in a mapping study (vertex-wise comparisons) of cortical thickness you would have to show that the group-difference in mean thickness is due to an effect that's spread out over the whole cortex (an overall effect, not localized), otherwise the mean thickness covariate is not a good estimator of the overall size factor that you'd like to remove.
check Peter Kochunov's recent publications on this.
-joost
cheers Bruce On Thu, 8 Apr 2010, Liukarl wrote: Hello, Freesurfer experts: I noticed for analyzing the results from autorecon2, the intracranial volume are controlled as covariate in some publications. Is there any corresponding covariate for cortical results from autorecon3, cortical thickness, surface, and volume? In my current study, there are two group: one is a group with less optimal developmental environment and the other is control. I noticed that without any covariate, the r square from cortical measure and group variable was really low, only accounting for 2%. Is any standard way to analyze the cortical data? Thanks! Karl _________________________________________________________________ MSN十年回馈，每位用户可免费获得价值25元的卡巴斯基反病毒软件2010激活码，快来领取！ http://kaba.msn.com.cn/?k=1 _______________________________________________ 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.
Hello,
Thanks for all the suggestion. Just a few follow-up questions:
1 How should the average thickness be calculated? (The average thickness measure was defined as a ratio between the sum of product of thickness and surface area measures in each individual ROI and sum of surface measure each individual ROI (over 30 ROIs from autorecon3) in the hemisphere?)
2 Since autorecon3 also generates volume and surface measure for each cortical ROI, any suggestion for covariate for cortical volume and surface?
3 I noticed to calculate the average thickness, the left and right hemisphere is separated out. But for analyzing results from autorecon2, such as left putamen volume, the ICV is used as a covariate. Is there is a separate left and right ICV available?
Thanks
Karl
From: mharms@conte.wustl.edu To: joost.janssen76@gmail.com Date: Thu, 8 Apr 2010 07:26:33 -0500 CC: schnack@gmail.com; freesurfer@nmr.mgh.harvard.edu; rememberer@hotmail.com Subject: Re: [Freesurfer] Covariate for results of cortical parcellation
Hi Joost, If you want to control for possible global differences in thickness, then an appropriate covariate of some sort is need. To me, a logical covariate for a thickness analysis is the mean cortical thickness, as this is more directly related to the measure of interest than something like the cube-root of volume (which other studies have used when mean cortical thickness was not available as a measure).
It is true that a difference in mean thickness doesn't necessarily imply that the difference in thickness is truly equally spread out over the entire cortex. But an analogous comment could be applied to volume analyses -- a difference in total brain volume between groups doesn't necessarily imply that differences in volume are present to an equal degree in all structures. Yet, using a volume covariate of some sort is a standard part of analyses of regional volumes. As a practical matter, I think it is fair to say that you are not going to find differences in mean thickness or total volume between groups unless a "substantial" portion of the brain is affected.
cheers, -Mike H.
On Thu, 2010-04-08 at 10:39 +0200, j janssen wrote:
Hi,
you could try including average cortical thickness over the entire hemisphere as a covariate. This was Mike Harms' suggestion, and I think a good one.
why is it a good one? for cortical thickness: i'm not so sure. my 2 cents (and i may be wrong!):
- controlling for intracranial volume is done to remove differences
due to overall size variability.
- the relationship between thickness and size is much less strong than
between volume and size (or surface and size).
- suppose your groups differ in mean thickness, if you would like to
include mean thickness as a covariate in a mapping study (vertex-wise comparisons) of cortical thickness you would have to show that the group-difference in mean thickness is due to an effect that's spread out over the whole cortex (an overall effect, not localized), otherwise the mean thickness covariate is not a good estimator of the overall size factor that you'd like to remove.
check Peter Kochunov's recent publications on this.
-joost
cheers Bruce
On Thu, 8 Apr 2010, Liukarl wrote:
Hello, Freesurfer experts:
I noticed for analyzing the results from autorecon2, the intracranial volume are controlled as covariate in some publications. Is there any corresponding covariate for cortical results from autorecon3, cortical thickness, surface, and volume?
In my current study, there are two group: one is a group with less optimal developmental environment and the other is control. I noticed that without any covariate, the r square from cortical measure and group variable was really low, only accounting for 2%. Is any standard way to analyze the cortical data?
Thanks!
Karl
MSN十年回馈，每位用户可免费获得价值25元的卡巴斯基反病毒软件2010激活码，快来领取！ http://kaba.msn.com.cn/?k=1
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.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_________________________________________________________________ 一张照片的自白――Windows Live照片的可爱视频介绍 http://windowslivesky.spaces.live.com/blog/cns!5892B6048E2498BD!889.entry
freesurfer@nmr.mgh.harvard.edu
-
Bruce Fischl -
j janssen -
Liukarl -
Michael Harms