I am trying to quantify evidence for no differences in gray matter volume between groups and would therefore like to use Bayesian estimation. 

I saw the Bayesian option in the model specification, review and estimation section in SPM for PET/SPECT. However, when I load in the specified model, it asks to do an ML estimation first. So do I need to do a classical estimation first and then use the Bayesian option? Or do I need to use the Bayesian 1st or 2nd level option in the 'estimate' option? 

Any help would be greatly appreciated. 

Thanks,

David