Computer Vision to Measure Finger Tapping in Parkinson's Disease

  • Research type

    Research Study

  • Full title

    Computer vision to measure finger tapping in Parkinson’s Disease (An exploratory study)

  • IRAS ID

    256116

  • Contact name

    Stefan Williams

  • Contact email

    stefanwilliams2@nhs.net

  • Sponsor organisation

    Leeds Teaching Hospitals NHS Trust

  • Duration of Study in the UK

    1 years, 8 months, 0 days

  • Research summary

    The core clinical feature of Parkinson's disease is a slowness of movement (called 'bradykinesia'). Clinicians currently make a visual judgement of this slowness by asking the patient to tap finger and thumb together and then observing for a typical impairment of rate, rhythm or size of the movement. Diagnosis and follow-up assessment rest heavily on this. However, the evidence suggests such visual assessment is inaccurate, and so our aim is to understand whether video from an ordinary camera, such as that in a smartphone, could be used to automatically and more reliably detect and measure this core clinical sign of Parkinson's disease (bradykinesia). Our initial findings from a previous exploratory study have suggested a method by which to do this, but we wish to collect normative data from a large group of control participants, and also collect some data using a 'gold standard' hand movement detector (infrared camera) to compare with our method.

    The proposed study involves two data collection methods to record anonymous video of finger tapping (and basic medical information) from patients. One is letter invitation to a specific research clinic (some with infrared camera), the other is the offer of brief participation at the end of a routine neurology clinic appointment in Leeds Teaching Hospitals NHS Trust. In addition, we plan to collect brief anonymous video of hand movement from a large number of control participants by asking colleagues to collect videos, and a brief control participant questionnaire, returned in a sealed envelope and labelled only by number. The resulting videos will be processed by computer to analyse hand movement through pixel changes, then testing for associations / correlations with diagnosis, blinded clinical ratings, and infrared camera data, to understand the degree to which the computer processing can detect and measure bradykinesia during finger tapping.

  • REC name

    North of Scotland Research Ethics Committee 2

  • REC reference

    19/NS/0011

  • Date of REC Opinion

    4 Feb 2019

  • REC opinion

    Further Information Favourable Opinion