Electroacoustic and Behavioural Assessment of Hearing Aids and PSAPs

Direct-to-consumer (DTC) hearing devices include the unregulated Personal Sound Amplification Products (PSAPs) and the yet-to-be regulated Over-the-Counter hearing aids (OTC HAs) (American Academy of Audiology, 2018).  Evidence from the few studies comparing the performance of PSAPs and conventional HAs reveals that: (a) audiologist-driven fitting and verification approaches lead to better outcome measures; (b) some, but not all, PSAPs produce statistically similar performance as conventional HAs; and (c) some PSAPs may produce unsafe output sound pressure levels.  It is therefore important to assess the performance characteristics of PSAPs (and OTC HAs) before they are used by the listener.

Standards for electroacoustic characterization of HAs and PSAPs do exist (ANSI S3.22 and the recent ANSI/CTA-2051 respectively).  While the tests specified in these standards allow measurement of important device characteristics such as total harmonic distortion, self-generated noise levels, and maximum acoustic output, they are not predictive of behavioural speech intelligibility and sound quality outcome measures.   This talk will present results from an ongoing study at the National Centre for Audiology on electroacoustic and behavioural assessment of PSAPs and HAs.  In particular, this talk will: (a) present speech intelligibility and sound quality data collected from hearing impaired listeners for low, mid, high-end PSAPs and HAs, and (b) evaluate the correlation between the subjective data and the objective speech intelligibility and quality predictors employing computational auditory models.

Learning Objectives:

  1. Discuss the performance differences among different HA and DTC devices, as assessed through subjective speech quality and intelligibility experiments.
  2. Describe the standardized electroacoustic tests for characterizing HA and DTC device performance, including those specified in the recent ANSI/CTA-2051 (2017).
  3. Gain insights into the prediction of subjective speech quality and intelligibility data through computational modeling.