Prosodic processing in people who do and do not stutter: Evidence from pause perception

Stuttering is a neurodevelopmental disorder characterised by frequent repetitions, prolongations, and pauses. Although stuttering is commonly viewed as a speech production disorder, recent evidence suggests that speech perception may also be affected, e.g. people who stutter (PWS) have been shown to process auditorily presented words and sentences differently to people who do not stutter (PWDS). According to the motor theory of speech perception, speech production mechanisms may in part determine the way speech is perceived. Since PWS’ speech production is impaired, it is possible that their perception of speech could be too. Recently, PWS have been found to have disrupted rhythm perception which has been attributed to aberrant functioning in neural timing networks. Given that speech prosody is comprised of speech rhythm and intonation then it might be expected that PWS exhibit prosody processing abnormalities. The aim of this project was to investigate whether PWS processed prosodic information differently to PWDS. To this end, prosodically expected grammatical pauses located at clause boundaries and prosodically unexpected ungrammatical pauses located within syntactic phrases were presented to a group of PWS and an age-matched group of PWDS while their electroencephalogram (EEG) was recorded. Since existing data on neurophysiological correlates of pause perception was scarce, the first part of the project (Experiment 1) was dedicated to identifying neural correlates of grammatical and ungrammatical pause perception in a non-stuttering population. 40 participants listened to sentences that contained grammatical and ungrammatical pauses and watched an unrelated video clip. Event-related analysis of the EEG showed two time windows, -100-100 ms and 100-230 ms relative to pause onset, where the event-related potential (ERP) amplitude difference between grammatical and ungrammatical pauses was statistically significant. In the early time window, ungrammatical pauses were associated with a significantly larger positivity than grammatical pauses. This was unexpected as processing of phrase boundaries is typically marked with a large slow potential, the closure positive shift (CPS). In the late time window, ungrammatical pauses elicited an enhanced N1 response which was interpreted as a marker of prosodic violation processing. The time windows and regions of statistically significant difference were used to define time and regions of interest (TOI/ROIs) for Experiment 2. Experiment 2 compared ERPs elicited by grammatical and ungrammatical pauses in PWS and PWDS. Participants were 15 PWS and 15 age- and sexmatched PWDS. The experimental design was identical to that of Experiment 1. A 2x2 mixed factorial ANOVA test was performed on the mean ERP amplitude for each TOI using pauses (grammatical, ungrammatical) and fluency (stuttering, non-stuttering) as factors. The results showed that there was no significant difference in ERPs between stuttering and non-stuttering subjects in the early time interval. In the late time interval, a significant interaction between pauses and fluency was found. Stuttering participants produced a significantly reduced N1 response to ungrammatical pauses than non-stutterers. Responses to grammatical pauses were similar across both groups. This finding was interpreted as a diminished ability (at least at the neural level) in PWS to discriminate between expected and unexpected prosodic events during early stages of speech processing and attributed to partial failure to successfully predict prosodic elements in speech. Further studies are needed to determine the behavioural significance of these findings.