Processing of Acoustic Cues in Lexical-Tone Identification by Pediatric Cochlear-Implant Recipients

Abstract

Purpose: The objective was to investigate acoustic cue processing in lexical-tone recognition by pediatric cochlear-implant (CI) recipients who are native Mandarin speakers.

Method: Lexical-tone recognition was assessed in pediatric CI recipients and listeners with normal hearing (NH) in 2 tasks. In Task 1, participants identified naturally uttered words that were contrastive in lexical tones. For Task 2, a disyllabic word (yanjing) was manipulated orthogonally, varying in fundamental-frequency (F0) contours and duration patterns. Participants identified each token with the second syllable jing pronounced with Tone 1 (a high level tone) as eyes or with Tone 4 (a high falling tone) as eyeglasses.

Results: CI participants' recognition accuracy was significantly lower than NH listeners' in Task 1. In Task 2, CI participants' reliance on F0 contours was significantly less than that of NH listeners; their reliance on duration patterns, however, was significantly higher than that of NH listeners. Both CI and NH listeners' performance in Task 1 was significantly correlated with their reliance on F0 contours in Task 2.

Conclusion: For pediatric CI recipients, lexical-tone recognition using naturally uttered words is primarily related to their reliance on F0 contours, although duration patterns may be used as an additional cue.

Figure 1.

Examples of amplitude waveforms, spectrograms, and pitch contours (yellow lines overlaid on the spectrograms) of the word ba spoken by a female talker (upper panels) and a male talker (lower panels) using the four lexical tones (left to right panels). F0 = fundamental frequency.

Figure 2.

The systematic manipulation in each acoustic dimension, using Praat Version 4.3 (Boersma & Weenink, 2004). One disyllabic word, yanjing—in which yan was fixed with Tone 3 and jing was varied in fundamental frequency (F0) and duration properties to represent lexical tone contrasts for Tone 1 (meaning eyes) and Tone 4 (meaning eyeglasses)—was resynthesized to generate two, eight, and six steps of variations, respectively, in (a) F0 height, (b) F0 contour, and (c) duration ratio. The entire stimulus set comprised 96 tokens (1 disyllabic stimuli × 2 F0 heights × 8 F0 variations × 6 duration ratios). As a reference, the naturally uttered tokens are most closely represented with the F0 variation values of 0 and −1.0 octave (i.e., corresponding with Tone 1 and Tone 4, respectively), and with the duration ratio values of 0.8–1.2 and 0.6–0.8 (i.e., corresponding with Tone 1 and Tone 4, respectively). For detailed acoustic depictions of lexical tones in Mandarin Chinese, please refer to Fon et al. (2004), Fon and Chiang (1999), and Tao et al. (2015).

Figure 3.

Identification accuracy of individual participants with (a) cochlear implants and (b) normal hearing for individual lexical tones.

Figure 4.

The proportion of yanjing stimuli identified as Tone 1, plotted as a function of the changes in acoustic properties. (a) Variation in fundamental frequency; (b) duration ratio. CI = cochlear implant; NH = normal hearing.

Figure 5.

(a) Overall Task 1 identification accuracy of individual participants with cochlear implants (CI) and normal hearing (NH) for lexical tones as a function of their coefficients for variation in fundamental frequency (F0) in Task 2. The data points are shown as squares and circles for individual CI and NH participants, respectively. The solid and dotted lines represent the regression line best fitted to the CI and NH groups, respectively. (b) The same data (excluding one data point; see text), with the abscissa showing log-transformed absolute values of the coefficients of F0 variation in Task 2.

Figure 6.

(a) Average tone recognition (% correct score) in Task 1, plotted against age at implantation for individual participants with cochlear implants (CI). The solid line shows the regression. (b) Coefficient of log duration ratio in Task 2, plotted against CI participants' age at implantation. The brown squares and solid line represent the correlation obtained without including the single outlier. (c) Coefficient of variation in fundamental frequency (F0) in Task 2, plotted against chronological age of individual participants with normal hearing.