Innovation - Earlier diagnosis through phonetics | University of Oxford
Speech analysis trial example
Speech analysis trial example

Innovation - Earlier diagnosis through phonetics

Our voices can be changed by variety of medical disorders, including Parkinson’s Disease and strokes.

Such diseases are not only debilitating, patients often face the cost and inconvenience of repeated travel to see a specialist. Pioneering work by phonetics expert Dr Ladan Baghai-Ravary has opened the possibility of potential benefits to both patients and hospitals.

Dr Baghai-Ravary developed software which can analyse a simple phone conversation, searching for early signs of Parkinson’s. The software looks at the tone, pitch and dynamics of speech, as well as the way that different sounds are produced. It also tests cognitive function using memory and reasoning tests.

The test is a brief automated questionnaire. People can call in and receive quick feedback, encouraging them to make a medical appointment if warning signals are picked up. Patients can put off going to the doctor even if they have early symptoms of conditions like Parkinson’s. The  software offers obvious cost and time savings for patients and the NHS.

Dr Baghai-Ravary worked on the software as a Knowledge Exchange Fellow with The Oxford Research Centre in the Humanities (TORCH). TORCH promotes high-quality research at Oxford which crosses traditional disciplinary boundaries and engages with the wider world.

Created in 2013, TORCH now supports more than 20 interdisciplinary research networks and 9 major research programmes, including ‘Medical Humanities’, ‘Environmental Humanities’ and ‘Race and Resistance’. Each links Oxford’s humanities departments with researchers in the sciences, and with business and cultural organisations.

TORCH supported and collaborated with almost 400 researchers last year and facilitated research events with audiences totalling more than 15,000. TORCH’s Headline Research Series in 2017 is ‘Humanities & Identities’ highlighting and exploring research themes related to diversity and inclusivity.

Voice analysis for everyone: health monitoring by phone