Interaction between bottom-up and top-down effects during the processing of pitch intervals in sequences of spoken and sung syllables.