Clinical Phenotypes of Feedback and Feedforward Control in Adults who Stutter - Stuttering affects approximately 80 million people worldwide and, for many, significantly hinders communication. However, treatment outcomes are variable, which is due in large part to significant behavioral and neurological heterogeneity. Current theories of motor control posit that stuttering develops due to weakened feedforward (FF) commands (i.e., automated adjustments in anticipation of errors) and/or aberrant feedback (FB) control (i.e., monitoring and correcting for errors). While these models agree on the general role of FF and FB systems, they differ in how much each system is affected across individuals who stutter. Furthermore, recent evidence suggests that there may be distinct subgroups of adults who stutter (AWS) characterized by different patterns of motor control. The proposed project will test the hypotheses that distinct stuttering behaviors are associated with behavioral and neural correlates of FB and FF control (Aim 1), and that distinct clinical phenotypes of FB and FF control can be identified in AWS (Aim 2). For Aim 1, we will conduct three linear mixed effects models, with rates of (1) repetitions, (2) prolongations, and (3) blocks as the outcome variables, both reflexive and adaptive responses as the predictors, and participant as a random intercept. We will also conduct three LASSO regressions with rates of (1) repetitions, (2) prolongations, and (3) blocks as the outcome variables and all ROI-to-ROI connections from the FB and FF control networks of the DIVA/GODIVA models as predictors. We expect repetition and prolongation rates to be associated with FB- based behavioral and neural mechanisms (i.e., reflexive response and functional connectivity in the FB control network). In contrast, we expect block rates to be associated with FF-based behavioral and neural mechanisms (i.e., adaptive response and functional connectivity in the FF control network). For Aim 2, we will use K-means clustering to group individuals based on patterns across stuttering behaviors, perturbation responses, and functional connectivity metrics. We will then characterize each cluster by input-variable means and test whether behavioral and neural measures covary with stuttering as predicted by FB- and FF-dominant phenotypes. We expect to find a FB-dominant cluster characterized by elevated repetition and prolongation rates, heightened reflexive responses to auditory perturbation, and stronger functional connectivity within the FB control network—relative to other clusters. Conversely, we expect to find a FF-dominant cluster characterized by increased block rates, reduced adaptive responses, and weaker connectivity within the FF control network—relative to other clusters. Results will have an important scientific and clinical impact by (1) strengthening our mechanistic understanding of stuttering and (2) deepening our clinical insights into factors that may influence therapy outcomes. This project falls under the 2023-2027 NIDCD Strategic Plan’s Theme 3 to “promote a precision medicine approach to prevention, diagnosis, and treatment”, as it investigates subtypes of a source of heterogeneity in AWS that may inform individualized treatments.
