Hi everyone,
I’m having a little trouble understanding the exact meaning of the 20% default threshold used in the freeview -tv option and dmri_pathstats --pthr. I’ve seen multiple threads where Anastasia said that it is 20% of the maximum value in the probability distribution—which corresponds to the maximum number of sample paths intersecting a single voxel—but the default thresholds I’ve seen set by the -tv option are generally lower than that. For example, I have one subject in whom the right CST has a maximum value in its path.pd.nii.gz file of 300. I would expect based on what I’ve read in the mail archives that the threshold would be set at 60, but when I open the merged file with the -tv option the default threshold for the right CST is 35.
My current best guess is that the default threshold is set to produce a volume that has a probability sum equal to 20% of the sum of the pre-threshold volume. (That is, the sum of intensities of all voxels in the path.pd.nii.gz file should be 5 times the sum of the voxels above the default threshold.) Is that accurate?
Thanks for all you’ve done to develop this tool. It’s brilliant and I want to understand as much of it as I can.
-Dillan
Hi Dillan - Thank you for your support!
Because the maximum value can sometimes be an outlier, we use the values of the 99th percentile instead. In the absence of an outlier this would be very close to the maximum.
Best,
a.y
On Tue, 26 Jul 2016, Newbold, Dillan wrote:
Hi everyone,
I’m having a little trouble understanding the exact meaning of the 20% default threshold used in the freeview -tv option and dmri_pathstats --pthr. I’ve seen multiple threads where Anastasia said that it is 20% of the maximum value in the probability distribution—which corresponds to the maximum number of sample paths intersecting a single voxel—but the default thresholds I’ve seen set by the -tv option are generally lower than that. For example, I have one subject in whom the right CST has a maximum value in its path.pd.nii.gz file of 300. I would expect based on what I’ve read in the mail archives that the threshold would be set at 60, but when I open the merged file with the -tv option the default threshold for the right CST is 35.
My current best guess is that the default threshold is set to produce a volume that has a probability sum equal to 20% of the sum of the pre-threshold volume. (That is, the sum of intensities of all voxels in the path.pd.nii.gz file should be 5 times the sum of the voxels above the default threshold.) Is that accurate?
Thanks for all you’ve done to develop this tool. It’s brilliant and I want to understand as much of it as I can.
-Dillan
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Thanks for clarifying. Is the same true for the --pthr option of dmri_pathstats as well? i.e., --pthr is specifying the portion of the 99th percentile, not the strict maximum?
thx, -- Michael Harms, Ph.D.
----------------------------------------------------------- Conte Center for the Neuroscience of Mental Disorders Washington University School of Medicine Department of Psychiatry, Box 8134 660 South Euclid Ave.Tel: 314-747-6173 St. Louis, MO 63110Email: mharms@wustl.edu
On 7/26/16, 11:40 PM, "freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Anastasia Yendiki" <freesurfer-bounces@nmr.mgh.harvard.edu on behalf of ayendiki@nmr.mgh.harvard.edu> wrote:
Hi Dillan - Thank you for your support!
Because the maximum value can sometimes be an outlier, we use the values of the 99th percentile instead. In the absence of an outlier this would be very close to the maximum.
Best,
a.y
On Tue, 26 Jul 2016, Newbold, Dillan wrote:
Hi everyone,
I’m having a little trouble understanding the exact meaning of the 20% default threshold used in the freeview -tv option and dmri_pathstats --pthr. I’ve seen multiple threads where Anastasia said that it is 20% of the maximum value in the probability distribution—which corresponds to the maximum number of sample paths intersecting a single voxel—but the default thresholds I’ve seen set by the -tv option are generally lower than that. For example, I have one subject in whom the right CST has a maximum value in its path.pd.nii.gz file of 300. I would expect based on what I’ve read in the mail archives that the threshold would be set at 60, but when I open the merged file with the -tv option the default threshold for the right CST is 35.
My current best guess is that the default threshold is set to produce a volume that has a probability sum equal to 20% of the sum of the pre-threshold volume. (That is, the sum of intensities of all voxels in the path.pd.nii.gz file should be 5 times the sum of the voxels above the default threshold.) Is that accurate?
Thanks for all you’ve done to develop this tool. It’s brilliant and I want to understand as much of it as I can.
-Dillan
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
________________________________ The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.
Yes, same thing!
On Wed, 27 Jul 2016, Harms, Michael wrote:
Thanks for clarifying. Is the same true for the --pthr option of dmri_pathstats as well? i.e., --pthr is specifying the portion of the 99th percentile, not the strict maximum?
thx,
Michael Harms, Ph.D.
Conte Center for the Neuroscience of Mental Disorders Washington University School of Medicine Department of Psychiatry, Box 8134 660 South Euclid Ave.Tel: 314-747-6173 St. Louis, MO 63110Email: mharms@wustl.edu
On 7/26/16, 11:40 PM, "freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Anastasia Yendiki" <freesurfer-bounces@nmr.mgh.harvard.edu on behalf of ayendiki@nmr.mgh.harvard.edu> wrote:
Hi Dillan - Thank you for your support!
Because the maximum value can sometimes be an outlier, we use the values of the 99th percentile instead. In the absence of an outlier this would be very close to the maximum.
Best,
a.y
On Tue, 26 Jul 2016, Newbold, Dillan wrote:
Hi everyone,
I’m having a little trouble understanding the exact meaning of the 20% default threshold used in the freeview -tv option and dmri_pathstats --pthr. I’ve seen multiple threads where Anastasia said that it is 20% of the maximum value in the probability distribution—which corresponds to the maximum number of sample paths intersecting a single voxel—but the default thresholds I’ve seen set by the -tv option are generally lower than that. For example, I have one subject in whom the right CST has a maximum value in its path.pd.nii.gz file of 300. I would expect based on what I’ve read in the mail archives that the threshold would be set at 60, but when I open the merged file with the -tv option the default threshold for the right CST is 35.
My current best guess is that the default threshold is set to produce a volume that has a probability sum equal to 20% of the sum of the pre-threshold volume. (That is, the sum of intensities of all voxels in the path.pd.nii.gz file should be 5 times the sum of the voxels above the default threshold.) Is that accurate?
Thanks for all you’ve done to develop this tool. It’s brilliant and I want to understand as much of it as I can.
-Dillan
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Anastasia Yendiki -
Harms, Michael -
Newbold, Dillan