AI algorithm development for wearable bio-monitoring device to help people who stammer

"About 1.3% of the population stammers (Reilly et al., 2009). This condition has distressing symptoms that deeply affect social and professional interactions and can define the life choices of those people who stammer (PWS): 60% of PWS suffers from social anxiety disorder (SAD) (McAllister et al., 2017). Even if it is currently not possible to cure stammering, there exist a series of practices that, when mastered, help PWS to gain fluency levels similar to that of a non-stammerer. In addition to mainstream therapist led approaches, several companies such as the Starfish, DelFerro or McGuire Program, currently run courses worldwide to help people to overcome their stammer via these learned techniques.

The largest challenge to a PWS is turning these practices into habits, as this requires patience and dedication. We believe that technology can help in this endeavour: with the support of Innovate UK, we are developing a prototype of the first wearable device, RespiraBelt, that tracks breathing and speech patterns, providing real-time feedback and performance history to help PWS turn fluency promoting techniques into habits. Respira will also provide a platform to share the user's performance data with the speech and language therapist (SLT), who will supervise and customize the therapy. Finally, a discretionary pool of anonymous data will be shared with the scientific community to further stammering research. Respira acts as virtual coach for PWS; as a therapeutic tool for SLTs; and a scientific instrument. We have patented the innovative technology behind the RespiraBelt as it is the first of its kind, and in December 2017 our idea was awarded the Terence Barry Grant Award from the Stammer Trust. Detailed information about Respira can be found at www.respira.io.

The current project is aimed at developing the algorithms that will unequivocally detect and evaluate when the user is practising habits that will improve their fluency. For that purpose, AI and machine learning methods will be used. This project is conducted in consortium with the Stammer Trust, which enables a trustful relationship with the PWS taking part in our research, as well as provides assurances that the best interests of our future clients are at the core of our project."