AI in Healthcare: Eye-tracking as a Tool for Diagnosing Dementia

Abstract

Dementia is an umbrella term for the neurodegenerative process that underlies the progressive im­pairment in remembering, thinking, or making decisions that interferes with doing everyday activities. Diagnosing this disease is done using neuropsychological evaluation methods. However, these meth­ods are time consuming and can lack sensitivity, which can make the diagnosis less reliable in some individuals and makes an early diagnosis difficult. Applications of eye­-tracking for the diagnosis and possible tracking of dementia have been reviewed in order to determine to what extend eye-­tracking based tests can help resolve these issues. A closer look into the findings suggests that eye-­tracking is a valuable technique in combination with the visual paired comparison task to help diagnose im­paired memory function, which in turn can help diagnose MCI or dementia. It also has great potential when combined with the antisaccade task to measure biomarkers such as oculometrics, which has shown valuable in the diagnosis of dementia because these eye movement deficits start early in the disease, even before cognitive deficits become noticeable. These eye-­tracking based tasks thus show that eye­-tracking can help diagnose dementia in an earlier stage, however, it is inconclusive whether eye-­tracking could increase the reliability of a dementia diagnosis. Future research should thus involve a wider variety of oculometric biomarkers in order to possibly improve the reliability of eye-­tracking. In addition, more research on improving eye-tracking algorithms could enable eye-­tracking to be a more reliable method, and thus could make a dementia diagnosis using eye­-tracking more reliable.