Early Dementia Detection through Conversations to Virtual Personal Assistant

Early detection and routine follow up of dementia are important because it can slow down the progress of the disease. The most common way to detect dementia is based on cognitive tests. The tests are usually done in the clinical setup with the help of a psycho-metrically trained examiner. Revised Hasegawa’sDementia Scale (HDS-R) is one of the prominent screening tests for dementia. We propose a method for early dementia detection by using a Virtual Personal Assistant (VPA) on a computer that has a natural language user interface, such asAmazon Echo, Apple Siri, Google Home, Microsoft Cortana, Soft bank Pepper, Sharp RoBoHon, etc. In our proposal, we consider HDS-R as a guideline to examine dementia. A VPA extracts the necessary features from the verbal and interactive response of the patient to compute the level of dementia.Such implicit checking is physically and mentally much comfortable for old people. We believe the proposed method will be able to contribute future society.