Using digital voice analyis to detect MCI/Dementia

  • Research type

    Research Study

  • Full title

    Assessing the diagnostic accuracy of a three one-minute tests of verbal fluency, using artificial intelligence, in detecting mild cognitive impairment and early dementia.

  • IRAS ID

    312213

  • Contact name

    Simon Vann Jones

  • Contact email

    s.vannjones@nhs.net

  • Sponsor organisation

    Cornwall Partnership NHS Foundation Trust

  • Duration of Study in the UK

    2 years, 0 months, 1 days

  • Research summary

    Primary aim: To assess the accuracy of a battery of three one-minute tests of verbal fluency delivered on a digital device (e.g. smartphone, laptop) compared to both the standard pen-and-paper test used in memory clinic (ACE III) and clinical diagnosis. The test involves three tasks lasting a maximum of 1-minute each: 1. Name as many animals as possible in one minute; 2. Image description – describe a complex image; 3. Open question – answer an open question. The first of these is already included in the paper-and-pen test so the only difference to the routine clinical pathway is that this part of the test will be recorded on a digital device and the second and third tasks will follow immediately, representing the only (minimal) deviation from normal clinical practice.

    Performance on these tasks will subsequently be analysed by artificial intelligence to identify potentially pathological elements of voice and language which have been shown to accurately correlate with cognitive problems. This software has shown a high degree of diagnostic accuracy (90%) in clinical settings in Spanish speaking populations. This will be the first study in an English-speaking clinical population.

  • REC name

    London - London Bridge Research Ethics Committee

  • REC reference

    22/PR/0963

  • Date of REC Opinion

    3 Aug 2022

  • REC opinion

    Favourable Opinion