Hi folks-
I would like to be run iterative cycles of a simple group analysis (paired
or unpaired t-tests) using a leave-one-out strategy (ie a jackknife
approach). The goal of this is to specify the images in an automated way
(leaving one out in turn), run the group analysis, extract peak voxel
activity or activity in a sphere around the peak voxel, and spit out the
values for each of the N most significant peak voxels for each of the
original images....all in an automated command-line fashion. This would then
be used to feed into a classifier (ie as a feature selection approach).
Does anybody have any SPM5 scripts that would do any or all of this
iterative model setup, estimation, results assessment and time course
extraction?
much appreciated!!
Amit
|