COMPutational Assessment of Stroke Survivors (COMPASS)
Project Description
Background & Significance Stroke affects 152,000 UK citizens every year. Over 50% of stroke survivors have cognitive impairment. Currently 850,000 people live with dementia in the UK and stroke is one of the biggest risk factors. National guidelines promote early cognitive testing on all people who have had a stroke. However, current pen-and-paper based tests are not always appropriate for stroke survivors who often have motor, visual or language difficulties. In addition, longitudinal follow-up is required to detect emerging cognitive impairment. Finally, assessments typically take place in hospital settings, are costly and often inconvenient for patients.
Aim & Objectives This project aims to create an easy-to-use cognitive assessment tool specifically designed for the needs of stroke survivors. It will be based on our stratification tool COCOA (COmputational COgnitive Assessment), developed for detecting early signs of dementia. The tool uses automatic analysis of conversations that patients have with an on-screen digital doctor. This includes verbal fluency and a picture description task. The conversation and responses to questions are analysed for signs of cognitive decline using speech recognition and machine learning classification. We have demonstrated high stratification accuracy in distinguishing between healthy controls, people with mild cognitive impairment (MCI) and people with Alzheimer’s disease. Feedback suggests that patients find the system straightforward and pleasant to use.