Research Aims to Improve Speech Perception Testing for Spanish Speaking Children

Tuesday, June 27, 2017

​Speech perception testing is a critically important tool for assessing children's auditory skills, determining candidacy for sensory devices, and guiding language intervention. A large number of children in the US are raised in Spanish-speaking households (>15%), but speech perception testing is typically performed in English or omitted altogether, due to a lack of appropriate test materials and a shortage of Spanish-speaking audiologists. Moreover, speech perception testing is typically performed in quiet or in relatively steady state noise despite the prevalence of complex and time varying sounds in children's real-world listening environments. The proposed work will develop an easy-to-administer clinical test of speech perception, allowing audiologists (regardless of their language proficiency) to appropriately evaluate Spanish- and English-speaking children in both noise and two-talker backgrounds.

Specific Aims

•  Aim 1: Refine test materials and streamline procedures
• Aim 2: Establish reliability and validity across listener age and hearing status
• Aim 3: Evaluate procedures in a clinical setting

IMPACT: The research team assembled to carry out the proposed work includes expertise in basic hearing science, auditory development, speech perception, clinical research, bilingualism, biostatistics, language assessment, and audiology, ensuring the final produce will be a clinical instrument that is both rigorous and practical. The end result is expected to improve our ability to evaluate and thus treat hearing loss in children.

Speech perception testing is one of the most important clinical tools we have for evaluating children's communication abilities. The current failure to comprehensively evaluate speech perception in bilingual children is a significant public health problem that restricts our ability to differentially diagnose auditory disorders, monitor performance over time, and evaluate the effectiveness of interventions in this population. Assessment and intervention resources for pediatric audiologists have fallen behind the rapidly changing demographics in the US. We are under both ethical and legal obligation to address this service gap: the Individuals With Disabilities Education Act specifies that children be assessed in their native language whenever feasible (US Department of Education, 1997).

There is growing consensus that speech recognition performance in a complex speech masker is a better predictor of children's functional listening skills than performance in steady noise or multi-talker babble. Speech perception tests are typically administered in quiet, in relatively steady-state noise, or in multi-talker babble. Informational masking is widely believed to reflect a failure of higher-order auditory processes such as sound source segregation and selective attention; thus, inclusion of a complex speech masker provides an opportunity to tap into the perceptual and cognitive processes that underlie speech understanding in real-world environments containing complex and time-varying sounds (e.g., Hillock-Dunn et al. 2014).

Innovation

The proposed work will yield a scientifically rigorous and clinically efficient test of children's English and Spanish speech perception. The development of this speech perception tool takes into account the fact that most audiologists in the US are not fluent in Spanish.  The proposed clinical tool offers an innovative solution for audiologists with limited Spanish proficiency. Specifically, the audiologist administering the test does not need to speak the test language to obtain a valid score. The development of the present tool avoids these problems by using recorded materials for task instruction and stimulus familiarization, and asking children to respond by pointing to the picture associated with the target word. This general approach has been used with some success in the past to evaluate bilingual children's speech perception in quiet (e.g., Comstock and Martin, 1984); the present work is novel in that it evaluatesmasked speech perception with English and Spanish stimuli that are rigorously controlled and balanced across languages

The proposed test will allow pediatric audiologists, for the first time, to characterize speech-in-noise performance as well as functional hearing in a two-talker masker. Completion of this work will yield a speech perception tool used to measure performance in the presence of speech-shaped noise or two-talker speech.

Our approach of adapting masker level and using a closed-set response extends the range of ages and auditory abilities of children than can be tested using the same clinical test. Several features of the proposed speech perception measure facilitate its use with children spanning a wide range of ages and auditory skills. Closed-set word recognition reduces the impact of working memory and linguistic familiarity, which can impact sentence recognition (e.g., memory or linguistic familiarity; Klem et al., 2015). Limiting the set of alternative responses makes the task relatively easy (Miller et al., 1951), such that all but the most severely impaired children can perform the task in quiet, whereas adapting the masker level prevents ceiling effects in the best performers. The forced-choice, picture-pointing response eliminates scoring errors associated with speech production problems or non-native accents (Ross and Lerman, 1970;Rimikis et al., 2013), and this type of task has been used successfully for evaluating monolingual English-speaking children as young as 4 years of age.