Introduction
The VoCS project is a network of 19 doctoral candidates across Europe that span a wide and interdisciplinary set of research themes within voice, speech, hearing, perception, cognition, and technology. Eriksholm is involved in activities related to hearing loss within those areas.
Recent research shows that people with hearing loss struggle with communication and turn-taking. However, why this happens is not clear. The doctoral project at Eriksholm aims to understand why turn-taking is challenging for people with hearing loss and identifying the relevant cues. In addition, we also aim to understand what other consequences problems with these cues can cause for hearing.
The VoCS project has been funded by the European Union under GA 101168998.
Aims
People with hearing loss often struggle with turn-taking in conversation, and this project aims to understand why. Previous research indicates that they make less use of fine-grained pitch cues – an important component of prosody that helps listeners anticipate when a speaker is about to finish. Building on this finding, the project will examine which vocal cues, including pitch patterns, signal upcoming turn completions and whether these cues support smoother conversational timing.
The project will compare people with normal hearing and those with hearing impairment to understand how they differ in perceiving and producing these cues, based on how precisely they time their turns in naturalistic interaction. It will also investigate how conversation partners adapt to each other’s speaking style, providing a fuller picture of how individuals with hearing loss use available cues during real-time communication. Ultimately, the project aims to deepen our understanding of how hearing loss shapes turn-taking and conversational flow.
Methodology
The project will analyze existing conversational datasets involving participants with and without hearing loss (e.g. the AMEND project). Signal processing methods will be applied to extract detailed acoustic cues, while AI and linguistics-based techniques will derive higher-level linguistic features. Machine learning and causal inference models will then be used to uncover patterns that link these cues to turn-taking behavior. Finally, knowledge about auditory processing and hearing loss will be incorporated to interpret the resulting models and understand how impairment shapes the perception and production of conversational cues.
