Cortical Auditory Evoked Potentials in Response to Frequency-specific Speech Tokens Measured in Infants

Abstract

Background: the introduction of universal newborn hearing screening has seen an increase in infant hearing aid fittings. Currently, clinical practice lacks an objective measure of auditory discrimination to validate hearing aid fittings for infants. Obligatory cortical auditory evoked potentials (CAEPs) have the potential to become a useful clinical tool for this purpose. Aims: 1) develop three frequency-specific speech tokens suitable for assessing auditory discrimination with aided CAEPs, and assess the effects of hearing aid processing on the speech tokens and the subjective speech-like characteristics of the speech tokens; 2) measure CAEPs in normally hearing infants to evaluate the frequency specificity of the speech tokens and, if frequency-specific, establish normative data. Methods: the phonemes a/, /ʃ/ and /s/ were digitally manipulated to create three speech tokens representing 1000, 4000 and 6000 Hz, respectively. The effects of hearing aid processing on the speech tokens were assessed using two different hearing aids, and the speech tokens were subjectively assessed by six adults. CAEPs were measured in nine normally hearing infants in response to the speech tokens, and three narrow band noise (NBN) stimuli with centre frequencies corresponding to the speech tokens, as a reference for frequency specificity. Stimuli were presented in the sound field for clinical relevance. Results: the speech tokens were accepted as speech by the hearing aids, but their rise times were altered by hearing aid processing. The /ʃ/ speech token was perceived as most speechlike, while the /a/ speech token was perceived as least speech-like. All speech tokens and NBN stimuli evoked repeatable CAEPs in all participants tested with the stimuli. Frequency specificity of the speech tokens could not be confirmed. Conclusion: presently the speech tokens have little clinical utility. The speech tokens should be reassessed for frequency specificity, and future research must also focus on understanding the effects of hearing aid processing on the speech tokens, and the subsequent effects on aided CAEPs must be better understood prior to clinical use.