[TD‐P‐023]: FEASIBILITY OF AUTOMATED VOICE‐BASED COGNITIVE ASSESSMENT ON A CONSUMER VOICE PLATFORM

activity. Methods: 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. Results: 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). Conclusions:Use of the controlled environ-