B2974 - ATLAS - Automated Transcription Language Analysis Software - 28/10/2017

B number: 
B2974
Principal applicant name: 
Rebecca Bright | Therapy Box Limited (United Kingdom)
Co-applicants: 
Dr Yvonne Wren
Title of project: 
ATLAS - Automated Transcription & Language Analysis Software
Proposal summary: 

ATLAS will be an app that speech and language therapists can use to screen children for developmental
language disorders. It is very important that developmental language disorders are detected because providing
the right help from speech and language therapists as early as possible is shown to make a signficant impact in
that child's potential to succeed at school and later life. We know that the benefit of providing enhanced speech
and language therapy for all children aged 6 to 10 who currently have a developmental language disorder
exceeds the cost of therapy by £741.8million.

The challenge at the moment is that speech and language therapists do not have sufficient time to complete the
screening (30 minutes) and then the transcription and analysis (90 minutes) that is needed to decide on the best
therapy plan. Our app will use automated audio and language analysis tools built into the app, along with
machine learning frameworks. Our project will look at which of these, or which combination of these tools, is
best to detect a developmental language disorder.

ATLAS will be built and trained using a very large dataset of recordings taken from the Avon Longitudinal Study
of Parents and Children so that we can compare whether the technology used is comparable to the analysis
completed on those recordings previously by speech and language therapists.

Therapy Box has an established customer base for the app. It is the market leader in speech and language
therapy apps; having had 197,500 apps downloaded from the iTunes AppStore across its own brand apps since
2011.

The team includes the technical and commercial experts at Therapy Box - who bring clinical, technical and
commercial know how to delivering this project. The team is lead by Rebecca Bright MBE who is a speech and
language therapist and since 2011 has worked to deliver specialist apps to for unmet communication needs.
Her co-founder Swapnil Gadgil has a technical and commercial focus and oversees the research and
development of the latest technologies for the company's app range and for apps for clients such as the United
Nations and universities around the world. Dr Yvonne Wren is Director of Research at Bristol Speech and
Language Therapy Research Unit. Her research has focused on speech development and disorder in children
and has included the development and trialing of software for use in speech and language therapy
intervention and extensive use of the ALSPAC dataset. Professor Steve Renals and Dr. Korin Richmond from the University of Edinburgh have a wealth of experience in applying machine learning methods to speech data

Date proposal received: 
Tuesday, 24 October, 2017
Date proposal approved: 
Saturday, 28 October, 2017
Keywords: 
Clinical research/clinical practice, Speech/language problem, Computer simulations/modelling/algorithms, Speech and language