Benefits of music training in mandarin-speaking pediatric cochlear implant users

Abstract

Purpose: The aims of this study were to assess young (5- to 10-year-old) Mandarin-speaking cochlear implant (CI) users' musical pitch perception and to assess the benefits of computer-based home training on performance.

Method: Melodic contour identification (MCI) was used to assess musical pitch perception in 14 Mandarin-speaking pediatric CI users; the instrument timbre and the contour length were varied as experimental parameters. Six subjects received subsequent MCI training on their home computer in which auditory and visual feedback were provided.

Results: MCI performance was generally poor (grand mean=33.3% correct) and highly variable, with scores ranging from 9.3% to 98.1% correct; there was no significant effect of instrument timbre or contour length on performance (p>.05). After 4 weeks of training, performance sharply improved. Follow-up measures that were conducted 8 weeks after training was stopped showed no significant decline in MCI performance. For the 6 trained subjects, there was a significant effect of contour length for the training and follow-up measures.

Conclusion: These preliminary data suggest that although baseline MCI performance initially may be poor, training may greatly improve Mandarin-speaking pediatric CI users' melodic pitch perception.

Figures

Figure 1.

Individual and mean (n = 14) melodic contour identification (MCI) performance with the 5 TONE, 3 TONE, and 5 PIANO stimuli. The asterisks indicate subjects who participated in the subsequent MCI training. Error bars indicate 1 standard error. The dashed line shows chance-level performance (11.1% correct).

Figure 2.

Mean baseline, training, and follow-up MCI performance for the 5 TONE (black circles), 3 TONE (grey triangles), and 5 PIANO stimuli (white squares). Baseline performance (Test Week 0) is shown at left. Training performance (Test Weeks 4–10) is shown in the gray shaded area. Follow-up performance (Test Weeks 11–18) is shown at right. Error bars indicate 1 standard error. The dashed line shows chance-level performance (11.1% correct).