Abstract: Linguistic pitch information is used to signal different aspects of spoken language such as lexical tone and intonation. An important issue that is particularly relevant to the processing of linguistic pitch information concerns language lateralization in the human brain. Two competing hypotheses based on the data from brain imaging studies were proposed to state the neural mechanisms underlying human pitch perception. However, the low temporal resolution related to neuroimaging techniques might not explicitly detect the brain activities occur at a relatively small or a determined time frame.
We used the mismatch negativity (MMN) and a source estimation technique (LORETA) to probe the brain activities underlying the early pre-attentive processing of Mandarin intonation. The target was Chinese character [lai2] which was pronounced in a declarative intonation and an interrogative intonation respectively. A passive oddball paradigm was applied to present declarative and interrogative intonation contrast for the syllable context condition, sentence context condition and their corresponding hummed versions. Thirteen Chinese college students were recruited to participate in the experiment and the participants were instructed to watch a silent movie and ignore the sounds from the headphone during the course of experiment.
The results showed that no MMN was elicited for the contrast of declarative intonation and interrogative intonation in the syllable intonation and sentence intonation conditions. However, the MMN was evoked when the linguistic information was removed from the two conditions above, and no significant differences of the MMN amplitudes existed between the hummed syllable and sentence intonation conditions. Source localization of MMNs for the hummed conditions showed the right hemispheric dominance.
In conclusion, these findings suggest that the pre-attentive cortical processing for Mandarin intonation can be modulated by the temporal window of linguistic pitch patterns manifested at syllable level and sentence level. Our data provide evidence for the acoustic hypothesis of speech processing.

Key words: Mandarin intonation, pre-attentive processing, MMN, ERP, LORETA