Multifactorial models of stuttering suggest cognitive, affective, linguistic, and social factors influence the speech motor system of people who stutter and these factors may be different across people who stutter. However, little research has described the influence of contextual factors on the speech motor processes and autonomic responses of adults who stutter (AWS). Therefore, the purpose of this study was to examine the effect of linguistic, memory, and social factors on the perceptually fluent speech and affective responses of AWS.

A total of 8 AWS and 8 adults who do not stutter (AWNS) participated in this study. Each participant completed three speaking tasks that imposed either a linguistic, memory, or social demand. Autonomic data (heart rate and pulse volume), perceived anxiety, and acoustic data were collected during each speaking task. Acoustic data was analyzed for differences of mean central tendency and intra-speaker variability for phrase duration, word duration, vowel duration, voice onset time, F2 transition duration, F2 transition rate, and F2 transition extent.

Acoustic results showed that AWS were not different than AWNS on temporal and spectral measures of central tendency as well as temporal variability. However, AWS were significantly more variable in F2 transition extent than AWNS across all speaking tasks suggesting greater variability in posterior to anterior tongue advancement. Results also showed the linguistic task generally contributed to longer and more variable temporal durations when compared to the control, audience, or memory tasks. Autonomic results showed AWS were similar to AWNS in their levels of autonomic arousal and perceived anxiety across the speaking tasks. Analyses of individual participants revealed that the greatest increase in autonomic arousal or perceived anxiety during the speaking tasks did not always relate to an increase in temporal or spectral intra-speaker variability. History of stuttering and treatment for stuttering did not predict trends in intra-speaker variability. Interestingly, a negative relationship existed for AWS between heart rate and perceived anxiety during the audience task.

The findings are discussed relative to clinical implications for the field of stuttering and multifactorial models of stuttering. Directions for future research are also proposed.