[Are variations of structural neuro-anatomy promising endophenotype candidates in bipolar disorder?].