On the early neural perceptual integrality of tones and vowels

•Findings.-Cantonese vowels and tones are perceptually integrated.-The perceptual integration of Cantonese vowels and tones occurs at auditory and phonological levels.•Implications.-The perceptual integration of vowels and tones is consistent with the TTRACE model, but contrary to the modified TRACE model.-We propose the TTRACE+ model in which vowels and tones are viewed as the same phonological unit.

The current study adopted the MMN additivity approach to examine the pre-attentive perceptual integration of vowels and tones. Twenty Cantonese listeners participated in the ERP experiment. Using the passive oddball paradigm, we elicited tone-MMN, vowel-MMN and double-MMN in the speech condition; and fundamental frequency-MMN, formant frequency-MMN and double-MMN in the non-speech condition. In both conditions, the double-MMNs were significantly smaller in amplitude than the sum of single feature MMNs. Morphological comparisons showed no significant difference in the latency and topographic patterns between vowel-MMN and tone-MMN, and marginal significant differences between formant frequency-MMN and fundamental frequency-MMN. Collectively, results reflect the perceptual integration of tones and vowels at the phonological level, and partial integration of fundamental frequency and formant frequency at the auditory level.