Supporting diagnostic decision for early detection of a neurodegenerative disease on a behavioural altered pattern basis

This paper describes the various aspects and steps of a multidisciplinary project called BEDMOND which aims at the development of an ICT-based system for the early detection of Alzheimer's disease (AD) and other neurodegenerative diseases on the basis of data assessment with health professional criteria. BEDMOND project addresses a system that supports the decision making process for the doctor, automating the information process related, first, to the recognition and modelling of the daily activity performed by the elder while being at home and, then, to the interpretation of deviations and behavioural changes detected. The system is a future-market extension for current tele-assistance technology and service enhanced with smart-home environment. BEDMOND considers that daily activity under two main scopes: first through the elder's daily routine -considered as a sequence of activities of daily living (ADL) - and then concretely through some specific behaviours highly concerned with mild cognitive impairment (MCI) typical symptoms (oversights and forgetfulness -medication intake, appointments, etc.-, disorientation -spatial and temporal-, loss of interest and isolation, etc.). First the requirements of such a behaviour pattern based assistant are discussed, then the system architecture deployed is introduced; next step deepens into the applied reasoning layers for the situation recognition, interpretation and data representation layers. Finally, next steps and some conclusions are also depicted.