Recognition of emotional prosody and verbal components of spoken language: an fMRI study