[Eeglablist] question about statisical analysis with ERSP results

[Xiaojun Tang]

Dear EEGlab user:
 
I'm analyzing some EEG data by EEGlab (mainly ERSP). The EEG data is
collected from 12 subjects in 3 conditions. As we know, the ERSP results are
two dimentions (freq and time). I want to knwo the difference between the 3
conditions. Is there any statistical analysis I can use?
 
The bootstrap in EEGlab seems to identify the real ERSP in one condition.
When I use bootstrap, do I need to merge all the trials across the subjects
to form one dataset?? 
 
Any suggestion will be appreciated.
 
Xiaojun Tang