Voice-Assisted Technology as a Potential Tool for Addressing Speech and Voice Concerns Experienced by People With...

People with neurological conditions such as Parkinson's Disease are at risk of speech and voice difficulties that impact volume, clarity of speech and intelligibility. Voice-assisted technology (VAT), such as Alexa, poorly recognises speech difficulties, and this often prompts people to change their speech to enable interaction.

This study aims to identify and map available literature regarding the utilisation of VAT to address speech and voice difficulties associated with neurological conditions. We explore (1) how well VAT recognises speech difficulties or dysarthric speech, across different severities and intelligibilities, and (2) the impact of using VAT on intelligibility, clarity or volume for people with dysarthria or speech difficulties.

A review of available literature was guided by Arksey and O'Malley's scoping review framework between September 2023 and November 2023. Five electronic databases were systematically searched, yielding 840 results. Results were screened by title, abstract and full text using inclusion and exclusion criteria. Eligible papers were identified and articles using 'out of the box' (i.e., not customised) VAT, speech difficulties or dysarthria and speech outcomes were included. Relevant data were extracted using an adapted version of the Joanna Briggs Institute extraction tool, and study outcomes were narratively reviewed. We report findings using the Preferred Reporting Items for Systematic Reviews and Meta-analysis extension for scoping reviews (PRISMA-ScR).

Five papers were reviewed that included varying speech difficulties, including dysarthria, stammering and reduced intelligibility. Results yielded limited evidence regarding the role of VAT for assessment in speech and language therapy (SLT). However, literature suggests that VAT should be used cautiously as a tool for managing volume, intelligibility and clarity of speech. Indeed, implications for activity, participation and well-being were noted, although findings are mixed. Early outcomes indicate that further research is needed to validate the evidence base for using VAT in SLT. Future research should explore the views of people with Parkinson's and speech and language therapists (SaLTs) who are using VAT to understand current usage and direct future development.

This review provides tentative conclusions that VAT may contribute to improve volume, intelligibility and clarity of speech for people with dysarthria or speech difficulties, such as people with Parkinson's Disease. However, caution is advised when using VAT with dysarthric ethnic speakers, as word error rates of devices are high, and it is unclear if this is related to device limitations or disordered speech. Biofeedback from devices may contribute to the mechanism of the effect.

What is already known on this subject Voice-assisted technology poorly and inconsistently recognises dysarthric or disordered speech, often without a clear indication of whether poor recognition is a limitation of the technology itself or the result of dysarthric or disordered speech. What this paper adds to existing knowledge Voice-assisted technology may hold promise as a future SLT management tool to improve volume, clarity and intelligibility of speech, with implications for activity, participation and well-being. Feedback from devices and the enablement of home practice may contribute to the technology's mechanism of effect; however, device limitations in recognition rates, variability and transactional interactions must also be considered. Due to a lack of evidence-based intervention development, no firm recommendations can be made on the specific impact for different populations, on dosage or duration of use during interventions, or on the usage of voice-assisted technology in assessment. Additional research is needed to fully understand clinical implications. What are the potential or actual clinical implications of this work? Voice-assisted technology may be used to promote changes in voice and speech, including improved volume, intelligibility and clarity of speech and potential functional gains. However, considering the numerous variables that can influence the consistency of VAT, clinicians should undertake a thorough risk-benefit analysis and determine its appropriateness on a case-by-case basis, in light of the potential adverse outcomes, whether real or perceived. People living with Parkinson's may benefit from using VAT as part of SLT. Devices may allow unlimited practice attempts, can prompt increased self-awareness of speech difficulties and motivate continued practice. However, error rates, ethnocentric bias and potential impacts on well-being warrant caution and clinicians should use clinical judgement to carefully consider clinical use of VAT. Overall, this widely available technology has potential as an SLT tool to enable self-management and support maintenance, alongside or following therapy. Future research should seek to investigate biofeedback as a mechanism of speech change.