USE OF WEARABLE MOTION SENSORS TO ASSESS THE BEHAVIOUR OF ALZHEIMER'S PATIENTS IN SIMULATED DOMESTIC ENVIRONMENTS

Abstract

It has been postulated that the development of a dementia-friendly environment can enhance autonomy in Alzheimer's Disease. However, there a few studies that have investigated this in a controlled way. Data from wearable sensors can provide continuous objective measures of performance of task activity. Therefore, we have created controlled simulated real-world environments, using wearable sensors to evaluate activity. Three groups of participants were studied: patients with either Posterior Cortical Atrophy or Alzheimer's Disease, and age-matched controls. Simulated controlled environments were created within the Pedestrian Accessibility and Movement Environment Laboratory (PAMELA) to assess performance of different tasks: navigation along corridors, ascending and descending stairs, navigating rooms to an open door, and reaching for objects. Conditions within the environment could be varied systematically, by altering lighting levels, providing visual cues, or adding clutter to the environment. Depending on the specific task, inertial measurement units (IMUs) were placed on the feet, head and hands to assess performance. For some data series, simultaneous recordings from an opto-electronic motion capture system were performed. Foot movement was calculated from accelerations of the foot sensors, by double integration and correcting for sensor drift by identifying zero-velocity periods. This allowed calculation of spatio-temporal parameters of gait, in addition to being able to track participants moving within the environment. Locations of hesitation can then be identified. Tracking of foot movement has also shown abnormal activity with stair climbing in some patients (Figure1). Analysis of head movement has also shown that a number of Alzheimer's participants scan the environment when performing some tasks to a greater degree compared with controls (Figure 2). Use of the controlled environment of PAMELA to simulate a number of domestic environments allowed conditions to be varied in a reproducible way, and enabled the identification of situations where environmental factors may be beneficial or otherwise. The technology used for the assessment of performance is completely transferable, requiring no specific infrastructure to perform the measurement. This would allow follow-up studies to be performed within the home, day centre or care home, in order to improve design, evaluate aids or assess risk. Examples of movement of the heel measured by IMUs for control subject (left) and participants with PCA (right) when ascending a short flight of stairs. The PCA participant shows considerable uncertainty in performing this task. Examples of head orientation for a control participant (left) and an Alzheimer's subject (right) when crossing a room from one side to an open door on the other side. Compared with the control, the Alzheimer's subject takes much longer to complete the task, and shows considerably more side-to-side movement of the head in completing the task.