Seminars in Hearing Research

Date: Thursday, April 10, 2025

Location: Nelson 1215

Time: Noon - 1:00 pm

Speaker: Ananthanarayan (Ravi) Krishnan, Professor, SLHS 

Title: Neural representation of temporal fine structure and not envelope periodicity determines perceived pitch of Mandarin lexical tones.

Abstract:  Pitch is a robust perceptual attribute that plays an important role in the perception of speech, music, and in the segregation of concurrent sounds. Complex sounds like speech have a slow varying temporal envelope (ENV) and a rapidly varying temporal fine structure (TFS). There is considerable interest in determining the relative contributions of ENV and TFS cues in speech, music, and lexical tone perception. Previous perceptual studies, manipulating temporal ENV and TFS cues, have shown that the temporal envelope cue is sufficient for recognition of English speech. In contrast for melodies, and dynamic pitch contours (as in Mandarin lexical tones) it is the TFS cue, rather than ENV cue that is dominant in perception. While different experimental approaches have been used to evaluate the relative contributions of ENV and TFS cues, we focus here on using auditory Chimeras where the envelope from one complex sound is paired with the TFS of another. Specifically, two Mandarin Chinese monosyllables [T2 (rising) and T4 (falling)] were divided into 4, 8, or 16 frequency bands and the fine structure and envelope of one tone pattern were exchanged with those of another tone pattern of the same monosyllable. Based on previous perceptual studies we hypothesize that the perceived pitch (and therefore the tone categorization) of our stimuli will correspond to the neural representation of the TFS (as reflected in the FFR) and not the envelope periodicity (as reflected in the EFR). FFRs were recorded to the unprocessed T2, and T4; ET2TFST4 (4, 8, 16 bands); ET4TFST2 (4, 8,16 bands) in 15 native speakers of Mandarin, and English. Both the Envelope following Response (EFR) and the neural phase-locked activity to the TFS (FFR) were extracted. Both a spectral (spectrograms and instant spectra) domain and a time domain (Autocorrelogram and Autocorrelation function) analysis was performed on all data. Results were consistent with our hypothesis that for both groups the neural phase locking to TFS corresponded with the perceived pitch of both lexical tones and for all chimeric stimuli. Responses were more robust for the Chinese listeners in general, and particularly for the TFS representation. These results provide strong electrophysiologic evidence that lexical tone categorization is largely dependent on the neural representation of TFS. Implications of these results will be discussed.

 ---

The working schedule is available here: https://purdue.edu/TPAN/hearing/shrp_schedule

The titles and abstracts of the talks will be added here: https://purdue.edu/TPAN/hearing/shrp_abstracts