The trap of protopathic bias in neuropsychiatric research