Automated System for Accurate Determination of Activities of Daily Living for Independently-Living Persons with Alzheimers Disease and Related Dementias

点击翻译按钮获取中文摘要

Project Abstract In this Phase I SBIR project, ASTER Labs will develop and evaluate an innovative system to automatically and accurately detect Activities of Daily Living (ADL) performed by persons with Alzheimer’s disease and related dementias. It will address a current need to equip caregivers and health care providers, including physicians and cognitive rehabilitation therapists, with reliable information on patients’ ongoing abilities to perform these important activities. The confirmation of these abilities will directly support the capability of a functioning independent lifestyle, while producing informed decisions on interventions and level of care based on disease progression. An estimated 5.8 million Americans in 2020 live with Alzheimer’s dementia. Nearly a third of these individuals live alone and are more likely to experience poorer health outcomes than cohabitating persons. Ongoing assessment of ADL is highly recommended for establishing diagnosis of dementia and progression of the disease over time. Existing and proposed approaches to automate this assessment in the home have ranged from cameras or vision-based sensors to beacon-based signal processing techniques. However, these approaches have been subject to limitations and critique due to privacy concerns, poor accuracy, limited coverage, and requirements of significant infrastructure alterations. Common commercial activity trackers have concentrated primarily on fitness activities, and typically rely on non-discreet wearable devices that, due to unfamiliarity, may be unacceptable to dementia patients. Clinical research has indicated interventions involving compensatory memory techniques and devices may help prevent or delay dementia onset, and preserve functional independence. Use of both manual and digital memory notebooks to help patients record when they performed certain activities have shown significant promise. However, seniors with memory impairment and dementia may risk inaccurate recollection of activities performed throughout the day, and have faced difficulty interacting with recent digital implementations of these interventions. An unmet need exists in the ability of those caring for these individuals to receive verifiable information on whether the activity was completed at the time reported by the patient, or at all. ASTER Labs’ proposed Activlog system leverages intelligent processing of WiFi, GPS, inertial, and audio sensor data from a small hardware suite concealed in a shoe insole, unnoticeable to the wearer, that uses high-precision location and multi-sensor association to accurately and continuously monitor and detect ADL. In Phase I, the prototype system will be assembled, with feasibility demonstrated by functional evaluation conducted through a focus group study with caregivers, physicians, and cognitive rehabilitation therapists of patients with dementia. Activity classification accuracy of the device will be determined in timed experiments by ASTER Labs’ engineers wearing the prototype insoles. Phase I testing will provide the success criteria for the start of the Phase II program, which will include a human study of the fully- operational system in home and independent living settings to establish the efficacy of this approach.