I'm a strong proponent of MRI's for first episode psychosis, w/w/out contrast. The risks with MRI contrast is pretty low relative to iodine, unless they have end stage renal disease(Nephrogenic Systemic Fibrosis), which is uncommon in first stage psychosis patients. ED's and internists often order a non-con CT, but that's fairly useless except for acute bleeds or blatant structural abnormalities. If you're seriously investigating, then in my opinion an MRI is the only way to go. Otherwise you've never really ruled anything out.
If you look at the few papers that're out there, such as by Freudenreich, there's no real consensus on what constitutes an adequate workup for first episode psychosis.