LifeBio-ALZ: AI driven digital biomarker engine leveraging natural conversation to widely scale accessibility for early detection and assessment of Alzheimers disease progression

点击翻译按钮获取中文摘要

Alzheimer’s Disease (AD) is one of the most common forms of dementia to occur in elderly populations, affecting over 30 million individuals worldwide. As the U.S. elderly population continues to increase, AD incidence rises as well, as there is no neuroprotective therapy or cure. Common symptoms include memory loss, cognitive impairment, disorientation, and psychiatric issues. Traditionally, diagnosis is achieved through a combination of clinical criteria such as neurological examination, mental status tests & brain imaging. However, these strategies are challenging for detection of early AD or patients with mild symptoms, specifically during the mild cognitive impairment (MCI) stage. Mental status tests & subjective journals, kept by patients or caregivers, can track AD progression, but have low sensitivity and reliability. The most strongly established biomarkers for AD, including amyloid beta, tau protein, & phosphorylated tau, are all obtained thru CSF requiring invasive lumbar puncture. The LifeBio-ALZ technology will provide a convenient and accessible, yet comprehensive digital biomarker and analytics suite to detect & assess Alzheimer’s progression. The platform will integrate a suite of assessment domains all seamlessly captured through a single, patient-centric app that engages users in natural video chat conversation via smart digital assistant. During brief, but regular sessions, an individual answers questions following a smart sequence to evaluate awareness, engagement, cognition, reaction time, speech patterns, & emotional state. The platform will record audio/video during the conversation. Type and timing of assessments, as well as specific questions will be adaptively modulated based on AD stage, personal demographics and previous analytics to minimize user burden while still providing rich data for algorithms. Quantitative features across multiple domains will be extracted from digital speech and eye movements, and then used as inputs to an AI engine to detect and assess Alzheimer’s’ disease progression. Data will be aggregated in secure cloud storage with clinician access to dashboard visualization tools. Phase I will demonstrate core feasibility. Development will build on a strong tech foundation of an existing LifeBio platform to increase likelihood of success. Currently, LifeBio is deployed in several formats including web, phone, & mobile apps to record life histories of people reaching advanced age or facing life-threatening illnesses or memory loss. Natural language processing tools parse information into life stories shared by family or used by staff to personalize engagement in care facilities. While the existing tech provides a base, significant enhancements will be executed in Phase I. More specifically, Phase I tasks will first update platform architecture to integrate novel data domains, build on smart sequenced multidimensional questions, and enhance patient workflow interfaces. Once the enhanced app passes all technical verification testing, it will be deployed in a field data collection and usability study with wide ranging AD patient demographics and stages. Finally, collected data will be used to build and validate an AI engine for detection and assessment of Alzheimer’s progression.