Recruitment of patients who are stable and euthymic, face long-standing recalcitrant symptoms, or have a history of long mood episodes may make sense from the perspective of statistical delineation of treatment, compared with mood effects, but certainly limits the degree to which any findings would be widely applicable