ENGAGE-DEM : a model of engagement of people with dementia

摘要

Dementia is a neurodegenerative disease that affects cognition, producing a reduction in thinking, problem-solving, and mnemonic abilities, functioning, preventing affected people to care for themselves and carry out activities of daily living (e.g., self-feeding, dress oneself), and psychosocial well-being, causing the appearance of disorders of thought content, mood, and behavior (e.g., depression, apathy, anxiety). The reduction of cognition, the disorientation in space and time, and the inability to complete basic tasks and function independently are the major causes of institutionalization in dementia. Care facilities are extremely efficient in meeting the physical and environmental needs of persons with dementia (e.g., food, self-care, drugs). However, they often fail in addressing mental and social needs. Several studies show that people with dementia living in institutionalized contexts spend most of their time inactive and isolated. Just as many studies demonstrate that engagement in playful activities is crucial to ensure quality of life and psychosocial well-being in dementia. On the one hand, the centrality of the medical aspects of dementia is due to the legacy of the biomedical approach to care. On the other hand, it is caused by the objective difficulty of understanding the needs of somebody who struggles to communicate. As a matter of fact, we are in great need of models enabling us to make meaning of how people with dementia express their psychological states. According to the literature, engagement is the psychological state of proactive involvement with an object (e.g., a game, an interactive system) or an agent (e.g., a person, a social robot) that has a positive affective nuance. In healthy adults, engagement can be measured on three different levels, according to three distinct response systems: experiential/subjective (i.e., self-reports), behavioral/expressive (i.e. overt behavior), and peripheral-physiological (i.e., human bodily reactions). In adults with dementia, the experiential/subjective level is rarely accessible due to cognitive impairment, the behavioral/expressive level might be blunted by motivational disorders like apathy and depression, and the peripheralphysiological level might p rovide insightful r esults, but is often overlooked. As the three response systems are not always equally accessible, it becomes crucial to combine them to achieve the most exhaustive possible measurement of engagement in people with dementia. The objective of this dissertation is twofold. First, it aims at exploring new techniques to assess engagement in dementia with the help of unobtrusive physiological sensors and systematic behavior observation. Second, it focuses on the development of a model of engagement of people with dementia that could formalize the relationships among these assessment techniques and outline their relative meaning in the economy of the overall engagement state. In order to pursue these two goals, the doctoral research was organized in three studies. First, we carried out an extensive ethnographic study to understand people with dementia in their context of living and get acquainted with the activities proposed by nursing homes. Second, we conducted an exploratory study to investigate the reactivity of people with dementia to an experimental setting and deploy a sensible research protocol for data collection. Third, we performed an experimental study and collected a database of multimodal data (e.g., video recordings, electrodermal activity signals, accelerometer signals) while people with dementia were involved in two types of activities: a game-based cognitive stimulation (i.e., jigsaw puzzles, shape puzzles, and a match with dominoes) and a robot-based free play (with the dinosaur robot Pleo). As a first result, we came up with three techniques to measure different aspects of engagement in people with dementia: electrodermal activity (EDA), the Ethographic and Laban-Inspired Coding