PTA3 with penalisations and metrics
The same previous analysis using non-identity metric (the inverse of the within group as above) and penalisation for time and space.
Now the same analysis for the penalisation for time and space but
once using the above (above above...) metric and once not. Nonetheless the metric is used for the subject space (i.e. a 12x12 matrix) computed using mean observations over epochs at each voxel (i.e.
each subject has got nvoxel*nepoch observations, here nepoch=10 and the length of each epoch is 10).
The brain observations are limited to grey matter to reduce the dimension and to be more sensitive.
(now the number of voxels is 576 instead of 1266)
back to front-page
summary of the decomposition (without w^(-1))
Notice the stability of analysis changing all voxels to only grey matter voxels with a noticeable decrease of subject variability (good!).
summary of the decomposition (with w^(-1))
Using the metric modified the first PT in the brain map loosing a temporo-parietal positive activation (?), which may be related only to few subjects as now the variability of subjects seem diminished (glovbally and within sex). The pattern of the time-course is showing now something new: a periodicity every other ON epoch.
Last modified: Thu Jul 5 10:25:10 BST 2001