Digital detection of social isolation and loneliness markers of risk for Alzheimer's disease

Project Summary/Abstract Social isolation and loneliness are associated with increased risk for cognitive decline and Alzheimer’s disease (AD) in older adults. This is a pressing public health concern given worldwide increases in social disconnectedness. Yet, research on the effect of social disconnection, especially social isolation, on risk for AD is hindered by reliance on retrospective self-report measures of social relationships and behaviors. Moreover, potentially modifiable social cognition mechanisms (e.g., apathy, defeatist social appraisals, biased threat perception) that may differentially contribute to isolation and loneliness are poorly understood. Integrated digital technology measurement approaches using ecological momentary assessment (EMA), which involves multiple daily smartphone surveys about social behavior and experiences, and passive social sensing, including GPS location and quantification of social interactions using smartphone sensors, could provide more precise and reliable probes for detection of social disconnection related to risk for AD in CN older adults, and could also reveal novel modifiable social cognition treatment targets to mitigate risk. Measurement problems, such as incomplete and inconsistent coverage of daily social behavior and experiences, have hampered observational and interventional research. Our inter-disciplinary research group has led development and validation of EMA, mobile social cognitive testing, and scalable passive sensing (GPS and voice sensing) measures, and social network analyses, to more precisely quantify social dynamics in daily life. We have also translated our real-time EMA data into interventions that reduce social cognitive biases that influence day-to-day social disconnection (e.g., social threats, defeatist attitudes). For the first time integrating these tools, we propose to investigate associations between real-time maladaptive social cognitive biases, social isolation, loneliness and AD risk biomarkers in 128 cognitively normal (CN) older adults divided into high (N=64) and low (N=64) risk based on CSF P-tau181, A42 and subtle cognitive decline (SCD) markers. We propose to administer in-lab standard measures, as well as EMA, GPS and social interaction digital detection measures, of social isolation, loneliness and social cognitive biases. We propose to compare high- and low-risk CN groups on EMA (primary outcome), passive sensing and in-lab measures, and will also examine relationships between digital social relationship measures, in-lab measures, and biomarkers. The goals of the project are to show that EMA and passive social sensing measures (1) can differentiate high- and low-risk CN groups; (2) are associated with known Aβ and P- tau biomarkers; and (3) are associated with social cognition biases that can be modified using treatments like in-person and digital cognitive-behavioral therapy. The immense data and digital products of this study would be available for future research probing real-world social processes in older adults.