A reliable, accurate, and clinic-friendly objective test of speech sound detection and discrimination in sleeping infants

One of the most basic measures of hearing device validation and outcome is the ability of the aided person to successfully use their device for verbal communication. For infants, an additional outcome measure is whether the device is facilitating spoken language development. Unfortunately, there is currently no clinically-adopted or reliable objective measure for speech understanding or for language development in individual infants, and we must rely upon behavioural observations or tasks that require the infant to first reach a sufficient developmental stage. This inability to verify whether an early-fitted hearing device will optimally facilitate language development is a potential cause of sub-optimal or delayed early intervention, which can result in permanent language disability, and reduced quality of life. This paper outlines the development of a new objective and reliable method to measure speech detection and speech discrimination in sleeping infants using a technology - functional near-infrared spectroscopy (fNIRS) – that is not yet available in audiology clinics. The culmination of research efforts to optimise the test methods and to devise novel analysis algorithms to reliably and automatically detect brain responses in sleeping infants is a test and analysis method with clinically acceptable accuracy (95% specificity and greater than 78-100% sensitivity). We have shown that the methods described here are now mature enough to bring into clinics for the benefit of infants and their audiological managers.