Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Are the Time-reversed Syllables Perceptible？Unraveling the question through MEG responses
Poster Session C, Friday, October 25, 4:30 - 6:00 pm, Great Hall 3 and 4
Yuting Meng1, Hua Fan1, Hongwei Bian1, Ziyi Zhang1, Ling Liu1; 1Beijing Language and Culture University
Time-reverse speech/ backward speech refers to a temporal reversal of natural speech signals, characterized by acoustic-phonetic distortions. As a non-speech tool, time-reverse speech is widely used in research on early stages of speech perception (acoustic-phonological processing). However, the neural mechanisms underlying the perception of reverse speech, especially at the syllabic level, remain unclear. Therefore, we employed magnetoencephalography (MEG) to investigate the dynamic perception of consonants, vowels, and lexical tones in time-reversed syllables. In the experiment 1, 27 participants (mean age 23) completed a speech perception discrimination task, in which they were asked to determine whether two consecutively presented forward/ backward syllables were identical. The forward syllables consisted of eight Mandarin Chinese syllables, each 250ms long, constructed from vowels /i/ and /u/, consonants /b/ and /d/, and tones 1 and 4. The backward syllables were generated by temporally reversing the forward syllables. MEG data were analyzed using multivariate pattern analysis (MVPA) to understand the decoding of different phonemes in the brain at various time points. The results indicated that all three phonemes of the forward syllables could be decoded within 100-500ms after sound onset, whereas consonants in the backward syllables either had very low decoding accuracy or could not be decoded at all. Comparatively, it was found that for the syllable_1, backward consonant decoding became more difficult between 280-330ms, while backward lexical tone decoding became easier between 100-220ms. Additionally, we observed bilateral brain involvement in processing for both forward and backward conditions. Given that children with autism may pay different attention to non-speech sounds compared to other children, it is possible that autistic children perceive time-reversed syllables differently. To account for different age groups and special populations, we designed Experiment 2. The experiment 2 employed an oddball paradigm with 20 participants (mean age 23). The same speech materials were used as in Experiment 1. For the forward condition, the syllable bi1 was the standard stimulus, while the other three syllables (di1 with consonant variation, bu1 with vowel variation, and bi4 with lexical tone variation) served as deviant stimuli. Results showed that in the forward syllable condition, vowel variation elicited both an early N200 effect and a late P300 effect, while lexical tone variation only elicited an N200 effect, with a smaller effect size compared to vowel variation. Consonant variation elicited only a P300 effect, smaller than that elicited by vowel variation. In the backward syllable condition, all deviant conditions elicited mismatch negativity (MMN) effects, with the largest MMN effect elicited by lexical tone variation, followed by vowel variation, and the smallest by consonant variation. In summary, both experiments indicate that at the early stage of speech perception, the brain shows differences in timing and perceptual sensitivity to different phonemes.
Topic Areas: Speech Perception, Phonology