Activity and Behavioral Recognition Using Sensing Technology in Persons with Parkinson’s Disease or Dementia: An Umbrella Review of the Literature

BACKGROUND With a progressively aging global population, the prevalence of Parkinson’s Disease and dementia will increase, thus multiplying healthcare burden worldwide. Sensing technology can complement current measures used for symptom management and monitoring. The aim of this umbrella review is to provide future researchers with a synthesis of current methodologies and metrics for use of sensing technologies for the management and monitoring of activities and behavioral symptoms in older adults with neurodegenerative disease. This is of key importance when considering the rapid obsolescence of and potential for future implementation of these technologies into real-world healthcare settings. METHODS Seven medical and technical databases were searched for systematic reviews (2018-2024) which met inclusion/exclusion criteria. Articles were screened independently using Rayyan. PRISMA guidelines, the Cochrane Handbook for Systematic Reviews, and the Johanna Briggs Institute Critical Appraisal Checklist for Systematic Reviews were utilized for assessment of bias, quality, and research synthesis. A narrative synthesis combines the study findings. RESULTS After screening 1458 articles, 9 systematic reviews were eligible for inclusion, synthesizing 402 primary studies. This umbrella review reveals that use of sensing technologies for observation and management of activities and behavioral symptoms is promising, however diversly applied, heterogenous in methods used, and currently challenging to apply within clinical settings. CONCLUSIONS Human Activity and Behavioral Recognition requires true interdisciplinary collaborations between engineering, data science, and healthcare domains. Standardization of metrics, ethical AI development, and a culture of research-friendly technology and support are the next crucial developments needed for this rising field.