A near-infrared spectroscopy (NIRS) study of lexical pitch accent processing in Japanese speakers
Background and aims: The aim of this study is to investigate the neural mechanisms underlying the processing of Japanese lexical pitch accent. The left and right cerebral hemispheres work together but differently for auditory language processing in human adults. The left side activates dominantly for processing most of the linguistic aspects of speech, including phonemic, lexical and syntactic analyses (Purves, 2001; Zatorre et al., 1992). On the other hand, the right dominant or nondominant activations are observed during processing of prosodic cues such as intonational pitch characterized by changing fundamental frequencies (Imaizumi et al., 1998; Zatorre et al., 1992). Although prosody is usually associated with the melodic features of spoken language, lexical level prosody such as Japanese pitch accent can be used to distinguish word meanings. For example, a pair of homophones with two syllables is distinguished by the pitch accent pattern which is either high-low (HL) or low-high (LH) such as a'me (HL: "rain") vs. ame' (LH: "candy"). A question then arises with regard to how the lexical pitch accent is processed, especially in terms of the functional lateralization. Methods: We employed 44-channel near-infrared spectroscopy (NIRS), which can noninvasively measure relative changes in the concentration of hemoglobin (Hb) in the localized brain tissue. The subjects were healthy right-handed Japanese speakers (n=18, nine male and nine female, 20-22 years). We used a'me and ame' tokens (14 pairs), ka'me (turtle) and kame' (ceramic jar) tokens (14 pairs), and bisyllabic Japanese words differing by pitch accent pattern (HL vs. LH) (14pairs). In addition, pure tone stimuli were created by extracting fundamental frequencies from the a'me and ame' words. Four conditions were performed in a block design paradigm. In the ame condition, the baseline block (20 s or 25 s) contained only a'me or ame' which was repeated approximately every 1.25 s. The test block (10 s) contained both a'me and ame' presented in a pseudo-random order with the equal probabilities. The kame condition was included to compare the phonemic change (/a/ to /ka/) to the lexical pitch-accent. In this condition, a'me or ame' was presented during the baseline, but ame and kame were presented in the test block. Similarly, in the variable words and the pure tone conditions, the baseline block contained either HL or LH pattern stimuli, the test block consisted of both pitch pattern stimuli. In order to assess cerebral lateralization, a laterality index, LI = (L ? R) / (L + R), was calculated from the peaks of the Oxy-Hb responses in the left (L) and the right (R) temporal areas in each condition. Results and conclusions: The results showed that the LI for the pure tone condition was significantly smaller than those for other three conditions, indicating that the lexical pitch accent is processed with more leftward shift compared with the processing of non-linguistic pure tone stimuli. Our results suggest that unlike more global prosody, which has been reported to be processed by the right hemisphere, lexical pitch accent is processed quot;linguisticallyquot; by Japanese native speakers.
Sato, Y; Mazuka, R; Sogabe, Y
Volume / Issue
International Standard Serial Number (ISSN)