Relating speech perception in noise to temporal-processing auditory capacities in childhood: effects of typical development and of sensori-neural hearing impairment

Temporal cues (e.g., amplitude modulation, AM) play a crucial role in speech intelligibility for adults. This study aims to characterize the development of sensory (AM encoding) and non-sensory mechanisms (i.e., processing efficiency) constraining auditory temporal processing and their relationship with speech in noise perception (SIN) for normal-hearing (NH) and hearing-impaired (HI) children.
Eighty-one NH children from three groups of age ranging from 6 to 10 years participated in the study, and pilot data were collected from eight 5-to-7-year-old children with sensorineural mild-to-moderate HI.
Three 3I-3AFC adaptive tasks were designed. The first task assessed AM sensitivity using pure tone carriers and 3 modulation rates (4, 8, 32 Hz). The second task assessed AM masking by comparing AM detection thresholds at 8 Hz modulation using 3 carriers varying in their inherent AM fluctuations: tones, narrowband noises with small inherent AM fluctuations and noises with larger fluctuations. The third task assessed temporal integration, the effect of increasing the number of AM cycles (between 2 and 8 cycles) on AM detection using tones modulated at 4 Hz. Finally, identification thresholds were measured in speech-shaped noise using fricative consonants (/f/-/v/-/ʃ/-/ʒ/-/s/-/z/).
For the NH children, overall the youngest group (6-7-years) had poorer AM detection thresholds in all tasks. However, their thresholds were affected by AM rate, large carrier fluctuations and number of AM cycles in the same way as older children, suggesting that sensory mechanisms may be mature by 6 years. Nevertheless, only 6-7-year-olds did not show any difference between tone and noise with small fluctuations, reflecting perhaps a change in the magnitude of internal noise with age for AM processing, and lower processing efficiency. Regarding SIN, regression analyses indicated that better AM detection with 8 cycles predicted better SIN thresholds (R2adj=20.9%). Moreover, the slope of the temporal-modulation transfer function (an estimate of temporal acuity) predicted an additional 3.5% of the variance.
Preliminary results with HI children show good AM detection abilities and similar effects of rate, carrier fluctuations and AM cycles on AM detection, but not surprisingly, worse SIN thresholds than NH children. Regarding the relationship between AM tasks and SIN, the same trend as NH was observed. These results suggest that SIN perception in childhood is related to temporal processing, and better AM processing might contribute to better SIN perception for children with HI. Computational modelling will help to better disentangle the relationship between sensory and non-sensory processing of AM in SIN during childhood.