Background: Validation procedures following the fitting of hearing devices are crucial for accurately assessing the effectiveness of those fittings and to ensure they meet the auditory needs of users. Traditional methods include subjective reporting and behavioural assessments of speech intelligibility. These methods have limitations in terms of consistency, reliability, and applicability, particularly when working with young children, individuals with cognitive impairments, or those unable to provide accurate feedback. The methods often fail to offer objective, quantifiable data on the ability to discriminate sounds heard via the hearing device in these populations, which is essential for effective hearing device validation. Recent investigations into the Mismatch Response (MMR) have shown promise in providing an objective biomarker for auditory discrimination. MMRs are an electrophysiological measure that can index neural responses to changes in sound, which can offer speech intelligibility information without behavioural feedback. However, the process of transitioning MMR-based testing protocols from the research setting towards a clinic-ready tool for hearing device validation remains an area of active exploration. A key challenge for the clinical setting is analysing and interpreting MMR data at the level of the individual listener. This study aims to evaluate the detection sensitivity and specificity of a protocol used for eliciting MMRs and analysing the ensuing data via Hotelling’s T² - a statistical analysis technique that can be used for determining electrophysiological response presence or absence in individuals.

Design: Electrophysiological data from three cohorts of listeners informed the present study. MMR detection sensitivity was estimated from data involving 16 normally hearing infants (3-13 months) and 17 normally hearing adults receiving stimuli to trigger an MMR, while no-sound trial data from 76 normally hearing adults was used to estimate detection specificity.

Results: The Hotelling’s T² detection algorithm demonstrated acceptable specificity, with false detection rates of 6.6%, which decreased to 4% after applying upper-confidence interval exclusionary criteria with 10,000 bootstrap samples. Combined with detection sensitivity of 93% from the MMR data, performance of detection algorithm overall was observed to be high.

Conclusion: This study demonstrates that MMR testing, with high specificity among adults, shows promise as a reliable tool for early hearing device validation. While further refinement and research is needed, MMR testing holds strong potential for clinical use.