Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Age-Related Changes in Resting-State Functional Connectivity and Executive Function Predict Specific Types of Speech Disfluency
Poster Session C, Saturday, September 13, 11:00 am - 12:30 pm, Field House
Megan Nakamura1, Kathryn Walters1, Haoyun Zhang2, Michele Diaz1; 1Pennsylvania State University, 2University of Macau
Speech disfluencies become more frequent with age, yet the mechanisms underlying this change are not well understood. One possible mechanism is age-related decline in executive function (EF), which has been linked to increased difficulties in maintaining fluent speech. Simultaneously, aging is also associated with intrinsic large-scale functional reorganization of brain networks such as the Default Mode Network (DMN) and Multiple Demand (MD) network which play important roles in EF. Moreover, prior work using Resting-State Functional Connectivity (RSFC) to capture these intrinsic changes has linked aging with reduced functional segregation—reflected in weaker within-network coherence and stronger between-network connectivity—which may impact the specialization of systems supporting language production (Zhang & Diaz, 2023). Building on this work, the present study (n= 271) examined a) whether RSFC within and between the Language, DMN, and MD networks predicts speech disfluency patterns and b) whether EF (measured via Stroop interference) and/or RSFC mediates the relationship between age and speech disfluency. Importantly, we distinguished disfluency types (e.g., repetitions, revisions, filled/unfilled pauses, and prolongations) to identify which are most sensitive to cognitive and neural factors. Behaviorally, aging was associated with more repetitions, revisions, and unfilled pauses, but was negatively associated with filled pauses. Age also significantly predicted worse Stroop performance, consistent with age-related declines in EF. In terms of RSFC, increasing age predicted significant declines in within-network connectivity and segregation across all three networks. Increasing age also predicted decreases in between-network RSFC between the Language and DMN. Notably, greater RSFC between the Language and DMN was associated with fewer revisions (b = –1.26, p = .0095) and fewer unfilled pauses (b = –3.26, p = .0405), while higher DMN segregation was associated with fewer total disfluencies (b = –3.19, p = .042) and repetitions (b = –0.572, p = .0029). Two significant Age × Connectivity interactions emerged. First, age moderated the relationship between DMN within-network connectivity and filled pauses (p = .016) such that among older adults, greater within-DMN connectivity predicted more filled pauses, suggesting that increased coherence in this network may not support fluent speech in older age. Second, age moderated the relationship between DMN segregation and repetitions (p = .0098) such that in younger and middle-aged adults, greater DMN segregation was associated with fewer repetitions, but this effect diminished in older adults. Lastly, causal mediation analyses revealed that Stroop performance partially mediated the effect of age on filled pauses (indirect effect: B = –0.0057, p = .048), suggesting that age-related declines in EF may contribute to increased use of filled pauses. Together, these findings suggest that age-related changes in both EF and RSFC within and between brain networks related to language contribute to specific patterns of speech disfluency in adulthood. Moreover, the distinct effects across disfluency types highlight that not all disfluencies are functionally equivalent—certain types may be more tightly linked to the breakdown of domain-general cognitive control or network modularity with age.
Topic Areas: Language Production, Control, Selection, and Executive Processes