<![CDATA[Social isolation is a significant and modifiable risk factor for accelerated cognitive decline and dementia in aging populations. Traditional methods for detecting cognitive changes, such as clinical screenings, are often infrequent and fail to capture the subtle, early behavioral shifts that precede a formal diagnosis. This study aimed to develop and validate an artificial intelligence model designed for the early detection of cognitive decline by passively monitoring behavioral and vocal biomarkers of social isolation in older adults living independently. A 24-month, prospective longitudinal study was conducted with a cohort of 200 community-dwelling adults aged 70 and older. A suite of unobtrusive in-home sensors was used to passively collect data on movement patterns, social communication (frequency and duration of conversations), and computer/phone usage. The AI-powered system identified individuals who would later show clinically significant cognitive decline with an accuracy of 91% and a lead time of approximately 7 months before formal assessment. The model successfully distinguished between simple loneliness and the specific behavioral patterns of social withdrawal associated with cognitive impairment. AI-powered passive monitoring systems are a highly effective and ecologically valid tool for the pre-clinical detection of cognitive decline linked to social isolation.]]>
Copyrights © 2025
