Masked speech recognition and reading in children (Miller et al., 2018)
datasetposted on 2018-03-05, 19:11 authored by Gabrielle Miller, Barbara Lewis, Penelope Benchek, Emily Buss, Lauren Calandruccio
Purpose: The relationship between reading (decoding) skills, phonological processing abilities, and masked speech recognition in typically developing children was explored. This experiment was designed to evaluate the relationship between phonological processing and decoding abilities and 2 aspects of masked speech recognition in typically developing children: (a) the ability to benefit from temporal and spectral modulations within a noise masker and (b) the masking exerted by a speech masker.
Method: Forty-two typically developing 3rd- and 4th-grade children with normal hearing, ranging in age from 8;10 to 10;6 years (mean age = 9;2 years, SD = 0.5 months), completed sentence recognition testing in 4 different maskers: steady-state noise, temporally modulated noise, spectrally modulated noise, and two-talker speech. Children also underwent assessment of phonological processing abilities and assessments of single-word decoding. As a comparison group, 15 adults with normal hearing also completed speech-in-noise testing.
Results: Speech recognition thresholds varied between approximately 3 and 7 dB across children, depending on the masker condition. Compared to adults, performance in the 2-talker masker was relatively consistent across children. Furthermore, decreasing the signal-to-noise ratio had a more precipitously deleterious effect on children’s speech recognition in the 2-talker masker than was observed for adults. For children, individual differences in speech recognition threshold were not predicted by phonological awareness or decoding ability in any masker condition.
Conclusions: No relationship was found between phonological awareness and/or decoding ability and a child’s ability to benefit from spectral or temporal modulations. In addition, phonological awareness and/or decoding ability was not related to speech recognition in a 2-talker masker. Last, these data suggest that the between-listeners variability often observed in 2-talker maskers for adults may be smaller for children. The reasons for this child–adult difference need to be further explored.
Supplemental Material S1. Phonological processing/reading measures: children (ordering of data is consistent across Supplemental Materials S2 and S4).
Supplemental Material S2. Masker condition signal recognition thresholds (SRTs, dB SNR) in children (ordering of data is consistent across Supplemental Materials S1 and S4).
Supplemental Material S3. Masker condition signal recognition thresholds (SRTs, dB SNR) in adults (ordering of data is consistent with Supplemental Material S5).
Supplemental Material S4. Masker condition signal recognition slopes in children (ordering of data is consistent across Supplemental Materials S1 and S2).
Supplemental Material S5. Masker condition signal recognition slopes in adults (ordering of data is consistent with Supplemental Material S3).
Supplemental Material S6. Least-squares (LS) means estimates of masker condition signal recognition thresholds (SRTs) in adults and children.
Supplemental Material S7. Least-squares (LS) means estimates of masker condition signal recognition slopes in adults and children.
Supplemental Material S8. Two-talker masker speech recognition threshold (SRT) regression results: children.
Supplemental Material S9. Spectral modulation benefit regression results: children.
Supplemental Material S10. Amplitude modulation benefit regression results: children.
Supplemental Material S11. Steady-state noise slope regression results: children.
Miller, G., Lewis, B., Benchek, P., Buss, E., & Calandruccio, L. (2018). Masked speech recognition and reading ability in school-age children: Is there a relationship? Journal of Speech, Language, and Hearing Research, 61, 776–788. https://doi.org/10.1044/2017_JSLHR-H-17-0279
Funding was provided by the ACES + ADVANCED Opportunity Grant from the Office of the Deputy Provost at Case Western Reserve University awarded to Lauren Calandruccio. This project was completed in partial fulfillment of Gabrielle Miller’s PhD degree program.
speechchildrenschool-agereadingspeech recognitionmasked speechdecoding skillsphonological processing skillsmasked speech recognitiontypically developingtemporal modulationspectral modulationnoise maskerspeech maskersteady-state noisespeech-in-noisespeech recognition thresholdsignal-to-noise ratioLinguistic Processes (incl. Speech Production and Comprehension)Language