Hi,
Is it possible to use FDR in stable3 for data that was processed prior to the introduction of brainmask.mgz? It seems like I can view the data just fine in stable3, or stable2 for that matter (though not stable 4), but FDR gives results that seem wrong. FDR sets the threshold to -log10(p)=4.279537 [p=.00005] no matter what the starting pvalue is.
Thanks, Rob
what do you mean by "staring p value"? Do you mean the FDR itself?
Robert Levy wrote:
Hi,
Is it possible to use FDR in stable3 for data that was processed prior to the introduction of brainmask.mgz? It seems like I can view the data just fine in stable3, or stable2 for that matter (though not stable 4), but FDR gives results that seem wrong. FDR sets the threshold to -log10(p)=4.279537 [p=.00005] no matter what the starting pvalue is. Thanks, Rob
Hi,
Yes, I set the rate to various values to test it, and all resulted in the same threshold. I didn't mask to brain, because that causes an error message (it cites the lack of a brainmap.mgz as the reason, so that at least I know is due to it being processed in an older stream). So I wonder if FDR requires that new processing as well, in stable3. By the way, these are the commands I used for view the data:
setenv SUBJECTS_DIR /space/ventzl/1/users/SUBJECTS_DIR source /usr/local/freesurfer/nmr-stable3-env cd $SUBJECTS_DIR tkmedit-sess -analysis EMerror -contrast ASevASc -space tal -map sig -s BwGroupError5 -isxavg random -fthresh 1.3
Thanks, Rob
Doug Greve wrote:
what do you mean by "staring p value"? Do you mean the FDR itself?
Robert Levy wrote:
Hi,
Is it possible to use FDR in stable3 for data that was processed prior to the introduction of brainmask.mgz? It seems like I can view the data just fine in stable3, or stable2 for that matter (though not stable 4), but FDR gives results that seem wrong. FDR sets the threshold to -log10(p)=4.279537 [p=.00005] no matter what the starting pvalue is. Thanks, Rob
Try coping or linking the brain.mgz to brainmask.mgz
Robert Levy wrote:
Hi,
Yes, I set the rate to various values to test it, and all resulted in the same threshold. I didn't mask to brain, because that causes an error message (it cites the lack of a brainmap.mgz as the reason, so that at least I know is due to it being processed in an older stream). So I wonder if FDR requires that new processing as well, in stable3. By the way, these are the commands I used for view the data:
setenv SUBJECTS_DIR /space/ventzl/1/users/SUBJECTS_DIR source /usr/local/freesurfer/nmr-stable3-env cd $SUBJECTS_DIR tkmedit-sess -analysis EMerror -contrast ASevASc -space tal -map sig -s BwGroupError5 -isxavg random -fthresh 1.3
Thanks, Rob
Doug Greve wrote:
what do you mean by "staring p value"? Do you mean the FDR itself?
Robert Levy wrote:
Hi,
Is it possible to use FDR in stable3 for data that was processed prior to the introduction of brainmask.mgz? It seems like I can view the data just fine in stable3, or stable2 for that matter (though not stable 4), but FDR gives results that seem wrong. FDR sets the threshold to -log10(p)=4.279537 [p=.00005] no matter what the starting pvalue is. Thanks, Rob
Hi Doug,
I just got around to looking at this problem again and trying what you had suggested (link brain.mgz to brainmask.mgz), and the problem is, there is no brain.mgz either. I would expect to see it in $SUBJECTS_DIR/$subject/mri/brain.mgz, but instead what I see, both in individual subjects' mri directories, and in averaged brain mri directories, does not include this file. Also, the COR files in /space/ventzl/1/users/SUBJECTS_DIR/mano001/mri (for example), in the T1 volume, are dated 2003. I think the absence of brain.mgz indicates that they were generated in a stream that was prior to not only brainmask.mgz but prior to brain.mgz as well. I could not find release notes dating that far back, but it seems like a reasonable assumption. Is it correct? I can tell from the more recent release notes that COR files were used in stable2 but no longer in stable3.
Not being able to apply the Mask to Brain option is certainly caused by there not being files that would need to come from the newer recon-all stream. Would you expect this to also cause FDR to not work? Is there any way around this problem, given that there is no brain.mgz either?
Thanks much, Rob
[vantz:mri] (nmr-dev-env) echo $SUBJECTS_DIR/
Doug Greve wrote:
Try coping or linking the brain.mgz to brainmask.mgz
Robert Levy wrote:
Hi,
Yes, I set the rate to various values to test it, and all resulted in the same threshold. I didn't mask to brain, because that causes an error message (it cites the lack of a brainmap.mgz as the reason, so that at least I know is due to it being processed in an older stream). So I wonder if FDR requires that new processing as well, in stable3. By the way, these are the commands I used for view the data:
setenv SUBJECTS_DIR /space/ventzl/1/users/SUBJECTS_DIR source /usr/local/freesurfer/nmr-stable3-env cd $SUBJECTS_DIR tkmedit-sess -analysis EMerror -contrast ASevASc -space tal -map sig -s BwGroupError5 -isxavg random -fthresh 1.3
Thanks, Rob
Doug Greve wrote:
what do you mean by "staring p value"? Do you mean the FDR itself?
Robert Levy wrote:
Hi,
Is it possible to use FDR in stable3 for data that was processed prior to the introduction of brainmask.mgz? It seems like I can view the data just fine in stable3, or stable2 for that matter (though not stable 4), but FDR gives results that seem wrong. FDR sets the threshold to -log10(p)=4.279537 [p=.00005] no matter what the starting pvalue is. Thanks, Rob
You can get brain.mgz by converting the COR files in /space/ventzl/1/users/SUBJECTS_DIR/mano001/mri/brain to mgz using mri_convert --in_type cor --out_type mgz
Robert Levy wrote:
Hi Doug,
I just got around to looking at this problem again and trying what you had suggested (link brain.mgz to brainmask.mgz), and the problem is, there is no brain.mgz either. I would expect to see it in $SUBJECTS_DIR/$subject/mri/brain.mgz, but instead what I see, both in individual subjects' mri directories, and in averaged brain mri directories, does not include this file. Also, the COR files in /space/ventzl/1/users/SUBJECTS_DIR/mano001/mri (for example), in the T1 volume, are dated 2003. I think the absence of brain.mgz indicates that they were generated in a stream that was prior to not only brainmask.mgz but prior to brain.mgz as well. I could not find release notes dating that far back, but it seems like a reasonable assumption. Is it correct? I can tell from the more recent release notes that COR files were used in stable2 but no longer in stable3. Not being able to apply the Mask to Brain option is certainly caused by there not being files that would need to come from the newer recon-all stream. Would you expect this to also cause FDR to not work? Is there any way around this problem, given that there is no brain.mgz either?
Thanks much, Rob
[vantz:mri] (nmr-dev-env) echo $SUBJECTS_DIR/
Doug Greve wrote:
Try coping or linking the brain.mgz to brainmask.mgz
Robert Levy wrote:
Hi,
Yes, I set the rate to various values to test it, and all resulted in the same threshold. I didn't mask to brain, because that causes an error message (it cites the lack of a brainmap.mgz as the reason, so that at least I know is due to it being processed in an older stream). So I wonder if FDR requires that new processing as well, in stable3. By the way, these are the commands I used for view the data:
setenv SUBJECTS_DIR /space/ventzl/1/users/SUBJECTS_DIR source /usr/local/freesurfer/nmr-stable3-env cd $SUBJECTS_DIR tkmedit-sess -analysis EMerror -contrast ASevASc -space tal -map sig -s BwGroupError5 -isxavg random -fthresh 1.3
Thanks, Rob
Doug Greve wrote:
what do you mean by "staring p value"? Do you mean the FDR itself?
Robert Levy wrote:
Hi,
Is it possible to use FDR in stable3 for data that was processed prior to the introduction of brainmask.mgz? It seems like I can view the data just fine in stable3, or stable2 for that matter (though not stable 4), but FDR gives results that seem wrong. FDR sets the threshold to -log10(p)=4.279537 [p=.00005] no matter what the starting pvalue is. Thanks, Rob
freesurfer@nmr.mgh.harvard.edu
-
Doug Greve -
Michael Ho -
Robert Levy