This page was created to keep a running list of
notes specific to this dataset. To add notes to this page, just open this
page with Netscape Composer and add your notes. This page was created 10/6/03.
Last modified:
GENERAL
Though there are 90 scans,
there are only 45 subjects. Scans L272 and L842 are the same scan (same
subject, same time) so was excluded from the study.
8/03
DENSITY
Reminder: Time 2 was registered
to Time 1 (timelocked). Therefore the binary gray images used for density
are non timelocked for Time 1 and timelocked for Time 2 but ALL images
are considered to be timelocked.
THICKNESS
The order in which the transformations
are carried out in pipeline is important to get the segmented images in
sulcal space to look like the binary gray images.
Scans L649 and L779 were found to have incomplete left ucfs AFTER all the analysis had been done. The cause was that the system crashed before the ucf was finished cause it to have the default value of "0.000000" in the 4th dimension. The subject ucfs were then fixed by rerunning the softmean script and ALL analyses were redone. Very minimal change in the results occured.
DX files: Because the ucfs
for the left hemisphere were remade the dx files are a bit confusingly
located in two places. All correct left hem dx files are found in /loni/edevel/LEONARD_THICKNESS/STAT_MAPS.
For the right hem, the one_samp ttest as well as the annualized and avg_thickness
maps are found in /loni/edevel/LEONARD_THICKNESS/STAT_MAPS/SOFTMEAN_two_bad_subjects.
The right hem data in "two_bad_subjects" are CORRECT and not BAD (it was
only the left hem that was bad, these files were removed to avoid confusion).