How People Living With Motor Neurone Disease Use Personalised Automatic Speech Recognition Technology To Support Social Interaction

Advanced search
Browse by:

Department | Year

UCL Theses | Latest

Deposit your research

How People Living With Motor Neurone Disease Use Personalised Automatic Speech Recognition Technology To Support Social Interaction

Cave, Richard; (2024) How People Living With Motor Neurone Disease Use Personalised Automatic Speech Recognition Technology To Support Social Interaction. Doctoral thesis (Ph.D), UCL (University College London). Green open access

[thumbnail of Cave_10198447_Thesis_id_removed.pdf]

Preview

Text
Cave_10198447_Thesis_id_removed.pdf
Download (17MB) | Preview

Abstract

Motor Neurone Disease (MND) is a progressive, ultimately fatal disease causing increasing muscular weakness. Most people living with MND (plwMND) experience acquired neurogenic motor speech difficulties (dysarthria) during the course of the condition. Although more than 80% of plwMND will become unable to communicate using natural speech, many express a desire to use speech for as long as possible, even when unintelligible to others. Project Relate, an app from Google, is argued to support people with ‘non-standard’ speech to be better understood by others. Relate builds a personalised automated speech recognition model (ASR) that may better recognise speech with dysarthria, generating real-time captioning designed to help listener understanding. This study videoed interactions between three plwMND and their communication partners over a year, to better understand the effects of personalised ASR captions and how the use of ASR may change as MND progresses. This study assessed ASR caption accuracy and how well it preserved meaning. Conversation Analysis (CA) was used to identify participants’ own organisational practices in the accomplishment of interaction in the presence of ASR captioning. Semi-structured interviews were completed and assessed using Thematic Analysis. Trouble and repair were observed as a feature of the dyad extracts. Similarities and differences in allocation of trouble responsibility, as well as gaze direction were observed when comparing Relate to other forms of augmentative and alternative communication (AAC). The expectations of all dyads were not met using Relate. Forecast caption accuracy was higher than actual and the percentage of captioned phrases that preserved meaning was frequently low. Issues were identified with app design, usability and support available. The high level of incorrect ASR captioning may reinforce perceptions of self-blame for voice change. The study provides wide-ranging recommendations, including user-centred review of app design, support, training for professionals and significant others.

Type:	Thesis (Doctoral)
Qualification:	Ph.D
Title:	How People Living With Motor Neurone Disease Use Personalised Automatic Speech Recognition Technology To Support Social Interaction
Open access status:	An open access version is available from UCL Discovery
Language:	English
Additional information:	Copyright © The Author 2024. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request.
Keywords:	Motor Neurone Disease, Automated Speech Recognition, AAC, Assistive Technology, AI
UCL classification:	UCL UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Computer Science
URI:	https://discovery.ucl.ac.uk/id/eprint/10198447