Linguistic fine-tuning in the spectra of the frequency-following responses to Mandarin tones

Abstract

Our understanding of the meaningful linguistic information encoded in neural signals such as FFR is still rather limited. Previous FFR tone studies mostly have focused on F0, and it is unclear how the other rich phonetic cues are encoded in the spectra. In this study, we analyzed the spectra of the frequency following responses to four Mandarin lexical tones from 13 native Mandarin listeners and 22 native English listeners. The stimuli consisted of four 250 ms synthetic syllable /yi/ with Mandarin four lexical tones minimally distinguished by F0 contours. In general, the harmonic structure of the FFR signals for native Mandarin listeners is much better defined than for non-native listeners, and the advantage is especially clear for higher-frequency harmonics. This result suggests that native listeners can hear the pitch contours more clearly and precisely not only because of a clearer f0 representation but also because of better-defined overtones. Moreover, for native listeners, harmonic representation varies by tonal targets. For a higher-pitched tone (Tone 1), H2 is the most prominent harmonic; for a lower-pitched tone (e.g., Tone 3 and Tone 2), H1 is the most prominent harmonic. These results suggest that pitch tracking strategies may vary by tonal targets.