Poster E5, Saturday, August 18, 3:00 – 4:45 pm, Room 2000AB
Resting state and task-related neural oscillations in adults who stutter and controls implicate deficits in sensorimotor integration
Andrew Bowers1, Dan Hudock2, Lisa Bowers1, Heather Ramsdell-Hudock2;1University of Arkansas, 2Idaho State University
Introduction: Developmental stuttering is a speech fluency disorder affecting speech-motor timing that often remits in childhood and persists in 1% of adults. Nonword repetition performance is one cognitive-linguistic factor that predicts recovery or persistence and is associated with lower performance at high-loads in adults who stutter (AWS), but it is unclear how the mechanisms underlying stuttering might be related to nonword repetition performance. Recent work suggests that cyclic, oscillatory timing in the dorsal stream is related to nonword repetition load performance, suggesting a potential link between timing for speech production and phonological working memory (PWM). The aim of the current proposal was to determine whether intrinsic differences in oscillatory power in the resting state of AWS were different from those of fluent speakers and related to power and oscillatory timing during a syllable repetition task. Methods: Data was collected from 20 AWS and 20 controls matched for sex, age, and education level. Electroencephalography (EEG) was recorded from 129 channels in a resting-state, eyes-open condition and in delayed 2 and 4 syllable bilabial repetition conditions presented in random order. One channel was used to record myographic artifact. Following analysis methodology used in previous studies, independent component analysis (ICA) was used to identify the same independent component (IC) network in the resting state and task. Spectral power (1-65Hz) using a short fast-fourier transform was used to quantify spectral power in the resting state and in time-periods of interest in the task along event-related spectral perturbations (ERSPs) were computed relative to a 2-second interstimulus interval. Permutation statistics were used to test for the condition and group differences with an FDR correction at p<.05. Pearson correlations were used to quantify relationships between baseline and task-related spectral power. Preliminary results: Preliminary data analysis of 7 AWS and 7 controls implicated an IC frontal, temporal, and temporal-parietal network active during the encoding and maintenance task phases (i.e., delta/theta/alpha/beta/gamma ERSPs) that was also associated with lower resting state power in AWS relative to controls (1-65 Hz). A second frontal-parietal network was active primarily during repetition and was associated with higher power in AWS relative to controls (1-65Hz). Dipole estimates for both IC networks were consistent with the dorsal stream, somatosensory association, and prefrontal regions. ERSPs showed group differences in timing of the alpha, beta, and gamma bands (p<.05 uncorrected). A right-hemisphere frontal IC cluster showed gamma suppression in AWS not present in any of the controls. Baseline log spectral power was significantly correlated (p<.05) with task-related power in the encoding, maintenance, and execution phases of the task in both groups and across both identified networks. There were no group differences in behavioral performance. Conclusions: Although data analysis is ongoing, preliminary findings from 7 AWS and 7 controls suggest that intrinsic resting state differences in power are correlated with power changes in the task and are related to timing differences in ERSPs. Findings will be discussed relative to a dorsal stream sensorimotor integration deficit in AWS on a PWM task.
Topic Area: Phonology and Phonological Working Memory