The body of my email was lost because of the attachments: here it is again for those who are curious. Sorry about that, I didn't know a text file attachment would do that.
Not sure if this bug has been mentioned before, but there is a bug with using aparcstats2table for the aparc.a2005s parcellation map. If some of the subjects in the group have missing values of a certain type, aparcstats2table fails to designate a 0 for them. So for instance, it's often the case that a subject has no lh_unknown_volume in their aparc.a2005s.stats. Certain other, seemingly random areas will fail to print out as well. This will lead to shifts in the dataset, so that every volume after the missing value for a subject will be incorrect, and will need to be shifted for the blank. The only way I have found to reliably check that the data is valid is to generate seperate stats tables for each subject.
Anyway, this is a serious problem for people who collect and analyse data with aparcstats2table using the aparc.a2005s parcellation map. Attached are two examples illustrating the problem. To make it clearer, paste the contents of both into excel, and look at the column mismatch; some of the subjects' data don't make it all the way to the end, and a shift in the values is clear.
Victor Laluz : Imaging Coordinator : UCSF Memory and Aging Center 415-476-1680 : 415-476-1816 (fax)
Victor,
We are aware of this bug. Unfortunately, there is no 'quick fix', which is why it has not been repaired yet. But yes, I agree its an ugly one which creates havoc with automated scripts. Hopefully a fix will appear in the next release.
Nick
On Thu, 2008-10-23 at 13:19 -0700, Laluz, Victor wrote:
The body of my email was lost because of the attachments: here it is again for those who are curious. Sorry about that, I didn’t know a text file attachment would do that.
Not sure if this bug has been mentioned before, but there is a bug with using aparcstats2table for the aparc.a2005s parcellation map. If some of the subjects in the group have missing values of a certain type, aparcstats2table fails to designate a 0 for them. So for instance, it’s often the case that a subject has no lh_unknown_volume in their aparc.a2005s.stats. Certain other, seemingly random areas will fail to print out as well. This will lead to shifts in the dataset, so that every volume after the missing value for a subject will be incorrect, and will need to be shifted for the blank. The only way I have found to reliably check that the data is valid is to generate seperate stats tables for each subject.
Anyway, this is a serious problem for people who collect and analyse data with aparcstats2table using the aparc.a2005s parcellation map. Attached are two examples illustrating the problem. To make it clearer, paste the contents of both into excel, and look at the column mismatch; some of the subjects’ data don’t make it all the way to the end, and a shift in the values is clear.
Victor Laluz : Imaging Coordinator : UCSF Memory and Aging Center 415-476-1680 : 415-476-1816 (fax)
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Laluz, Victor -
Nick Schmansky