Hello SPMers,
I am running a multi-subject conditions x subjects interaction using
proportional scaling to 50. Although the analysis seems to run without
problems, listing the "files and factors" used in the analysis shows that
the global scaling factors are incorrect - either in the range of 0.1 or
100 instead of all ranging around 50. Are there restrictions on the use
of proportional scaling of which I am not aware that are causing the scale
factor problem? This has not been a problem in other analyses using,
e.g., AnCova by subject, which I understand does not use the global
scaling factors.
Another related problem is that preprocessed image files which previously
have shown correct scale factors (e.g. intensity Y = 0.0011641 X) are now
showing incorrect scale factors (e.g. intensity Y = 1 X). These files
have been used in various analyses, but should not have been rewritten or
altered in the process. Does anyone have some insight as to how the scale
factors could have been changed and how to prevent such alteration of the
data in the future? Any suggestions would be appreciated.
Brandall Y. Suyenobu
UCLA / CURE Neuroenteric Disease Program
Veterans Administration Greater LA Heathcare System
(310) 478-3711 ext 40580
[log in to unmask]
|