Global access to digits-in-noise hearing tests

This project aims to develop a procedure to create digits-in-noise tests in all languages fully automatically for different target populations overcoming the current limitations. The digits-in-noise test, invented and developed in the early 2000s by us, has proven to be a very successful test internationally to screen for hearing loss, but also for use in the clinic. This project helps to develop specific self-administered home versions of the digits-in-noise test for cochlear implant users as well.

The World Health Assembly (WHA) of the WHO adopted a resolution in 2017 which calls on hearing healthcare stakeholders to invest in improved access to health services and to develop, implement and monitor screening programs for early identification of hearing loss. This project has the potential to make hearing screening affordable and accessible in low and middle-income countries, and to improve the care pathway of cochlear implant users substantially, leading to home care and self-management.

The DIN test has a limitation because it uses language-specific speech signals (the digits). To allow comparability and ensure uniform use, each language version needs to be developed separately. This is an expensive and time-consuming process. In the first part of the project it will be explored if the recording and selection of the digits by a text-to-speech system (TTS) can be replaced which can generate digits in almost all languages. The generated digits should sound natural and should have the desired acoustical properties. Next, automatic speech recognition (ASR) systems will be evaluated. The system should be able to detect the same intelligibility differences between digits as for human listeners. For the specific tests for cochlear implant users, the digits will first be modified through cochlear implant processor simulation software.

The ultimate aim of the project is to develop a system that produces a digits-in-noise test for hearing screening or for CI users fully automatic in only one click, for any language.