Project Outline
Understanding the perception and embodied experience of sound in dementia is pivotal to designing future support systems that facilitate meaningful community engagement. Advances in machine learning and signal processing techniques offer opportunities to optimise acoustic environments for people living with dementia across different social and musical settings.
By personalising auditory experiences and working directly with organisations that support people living with dementia in Sheffield (Age UK Sheffield, JoJanGles music focus group), this research directly targets people more at risk of isolation in a noisier future. Ultimately, this project goes towards ensuring future sound environments are actively engaging and supporting social connection for vulnerable populations.
Further project information
The increasingly noisy environments in the future, intersecting with rising incidences of dementia across the world presents a crucial challenge. A key part of living independently with dementia for longer is the ability to interact with others in public and community spaces. What we see currently throughout the dementia journey is withdrawal from these social spaces, which can often be a result of sensory overwhelm. Sounds intended to induce pleasurable experiences, or to specifically calm and soothe, can instead add to panic, stress and confusion.
Engaging with multisensory experiences and particularly music is accepted as a widely valued experience in dementia, and one that can promote identity, belonging, and general wellbeing. Yet how music and sound are processed in dementia in potentially noisy social spaces is rarely examined, with the bulk of research assuming a hearing typical and cognitively intact population.
The proposed project looks to i) examine sound and music processing within cognitive impairment and ii) to develop technology solutions to reduce sensory overwhelm in social spaces, preserving the ability to interact with sound or music-based programmes. Examining the behavioural responses of people living with dementia to various musical and aural stimuli will identify the acoustic characteristics which are easier or more engaging for an individual.
The project will then investigate whether signal processing can adapt the material to better match the individual’s threshold for comfort. This could include personalised predictive overload models which train user-specific machine learning models to evaluate ambient acoustic metrics (e.g. spatial clutter, reverberation, psychoacoustic sharpness) and predict or anticipate stressful auditory situations.
Another approach could involve content-preserving hearable scene processing, i.e. developing real-time acoustic processing for hearables that selectively softens and restructures distressing background soundscapes to create a calmer audio mix without suppressing critical speech, safety cues, or desired musical elements.
Subject Areas
- Dementia
- Music
- Sound
- Machine Learning
- Computer Science
- Technology co-design
Required technical skills/academic qualifications/experience
A minimum 2.1 or higher in an undergraduate degree or a relevant master’s degree is required. Applicants with experience or interests in psychology, signal processing or computer science are encouraged to apply.
We welcome candidates keen to develop skills across disciplines, and relevant experience need not come only from formal degree study.
We particularly encourage applications from women and other groups under‑represented in research. All appointments will be made on merit.
Mode of attendance
Successful applicants are expected to register as full-time students. However, eligible UK home fees applicants may request to be considered for part-time mode of attendance by the supervisory team.
Supervisors
Project Partners
This project will be hosted at the University of Sheffield with support from Age UK Sheffield
