n_subs overwritten in model init

[lucywowen]

When passing a list of brain objects, n_subs has to be defined.

[jeremymanning]

the way we had previously defined this is:

• each brain object and/or nifti object counts as 1 subject
• each model objects count as Model.n_subj subjects
• each numerator/denominator/corrmat counts as n_subj (i.e. the specified number of subjects) subjects
• the final model object should sum over the numbers of subjects for all objects it's being initialized by

[jeremymanning]

(and i think this is how it should work, if it doesn't currently work this way)