A multidisciplinary team out of the University of Alberta, led by music professor Michael Frishkopf and in collaboration with Osmar Zaïane (Fellow & Canada CIFAR AI Chair at Amii), is researching the use of AI-powered sound therapy in ICU wards.

As described in a recent article by Folio, this machine learning-based sound system is designed to read physiological feedback (e.g. heart rate, breathing, sweat gland response) to customize calming sounds for individual patients. It aims to reduce stress and anxiety in critically ill patients, whose common experiences of delirium and sleep deprivation can lead to poorer health outcomes.

The system is designed to dynamically select, tune, and mix files from an audio library comprising natural, musical, and synthetic samples. Reinforcement learning -- a branch of machine learning that enables AI systems to learn through experience -- plays a key role in the system, guiding and optimizing individual soundscapes with the aim of helping patients to better relax.

Zaïane has been involved in early design meetings and has provided training for students in Nursing (PhD) and Computing Science (BSc), as well as a talented high school student, who has continued on the project in an evaluator capacity.

The project involves students and researchers from multiple faculties at the University of Alberta, including the faculties of arts, science, nursing, and medicine and dentistry.

Read the full article on the Folio website.