Neural correlates of segmental and tonal information in speech perception

Abstract

The Chinese language provides an optimal window for investigating both segmental and suprasegmental units. The aim of this cross‐linguistic fMRI study is to elucidate neural mechanisms involved in extraction of Chinese consonants, rhymes, and tones from syllable pairs that are distinguished by only one phonetic feature (minimal) vs. those that are distinguished by two or more phonetic features (non‐minimal). Triplets of Chinese monosyllables were constructed for three tasks comparing consonants, rhymes, and tones. Each triplet consisted of two target syllables with an intervening distracter. Ten Chinese and English subjects were asked to selectively attend to targeted sub‐syllabic components and make same‐different judgments. Direct between‐group comparisons in both minimal and non‐minimal pairs reveal increased activation for the Chinese group in predominantly left‐sided frontal, parietal, and temporal regions. Within‐group comparisons of non‐minimal and minimal pairs show that frontal and parietal activity varies for each sub‐syllabic component. In the frontal lobe, the Chinese group shows bilateral activation of the anterior middle frontal gyrus (MFG) for rhymes and tones only. Within‐group comparisons of consonants, rhymes, and tones show that rhymes induce greater activation in the left posterior MFG for the Chinese group when compared to consonants and tones in non‐minimal pairs. These findings collectively support the notion of a widely distributed cortical network underlying different aspects of phonological processing. This neural network is sensitive to the phonological structure of a listener's native language.