Dear SPMers
in my design I have four parametric modulators (on first level).
I did second level t-tests using images that represent each specific modulator of the first level
(i.e. in the first level I put an '1' on the PM and zero on all other regressors and onsets).
When I plot results of the 4th regressor at the second level (plot-->fitted responses--> predicted-->scan or time)
it gives some very large values (in the range of -1000 or +500) for around 8 out of 55 subjects.
The rest subjects have responses ranging between -10 and 10.
Other PMs give "normal" responses, i.e. between -2 and 2.
Any ideas what could be the reason for these large numbers?
Thanks
j
