AbstractBackgroundThere is growing interest in using digital biomarkers, like recorded spontaneous speech, as indicators of Alzheimer’s disease progression. Several studies have shown that changes in connected speech and language (CSL), including speech fluency markers and linguistic content, are associated with preclinical cognitive decline. Little is known about the relationships between CSL and the lifestyle factors that may be largely dependent upon it, particularly social participation. Social engagement (SE) (e.g., in activities such as informal social participation, volunteering) has been associated with multiple health indicators in aging. Our aim was to determine if SE at baseline, or change in SE, was related to declines in CSL measures that have been linked to cognitive decline. Sensitivity analyses determined if baseline CSL predicted baseline SE and change in SE.MethodLongitudinal CSL measures (Figure 1) were automatically extracted from picture description transcripts from Wisconsin Registry for Alzheimer’s Prevention participants (N=794). SE composite scores were created from selected Florida Cognitive Activities Scale items (Table 1). Statistics included t‐tests, chi‐square tests, ANCOVA, and linear mixed‐effects models with CSL as outcomes and baseline SE, and normalized change from baseline to most recent SE, as predictors.ResultTable 2 shows speech metrics and demographics of participants stratified by low, medium and high SE. Participants reporting “high” SE were more likely to be female, had higher baseline WRAT‐3 reading scores, and had lower depression scores. Participants reporting “low” SE expressed fewer content words, unique words, and total words than those with mid‐ or high SE. Positive change in SE was associated with slower age‐related change in speech fluency over time (Table 3, Figure 2). Sensitivity analyses indicated worse baseline speech fluency was significantly associated with higher baseline SE but also positive change in SE, adjusting for age, sex WRAT‐3 and depression (Table 4).ConclusionThis study shows that spontaneous speech metrics are positively associated with self‐reported social engagement in late middle age. Crucial next steps include determining if SE moderates the effect of amyloid, tau, and neurodegeneration on measures of cognition including speech. SE is a potentially modifiable lifestyle behavior that may have positive implications for improved outcomes in communication and cognitive function.