Date Published: April 26, 2019
Publisher: Public Library of Science
Author(s): Mehdi Bakhtiar, Caicai Zhang, So Sze Ki, Simone Sulpizio.
There have been controversial debates across multiple disciplines regarding the underlying mechanism of developmental stuttering. Stuttering is often related to issues in the speech production system; however, the presence and extent of a speech perception deficit is less clear. This study aimed to investigate the speech perception of children who stutter (CWS) using the categorical perception paradigm to examine their ability to categorize different acoustic variations of speech sounds into the same or different phonemic categories. In this study, 15 CWS and 16 children who do not stutter (CWNS) completed identification and discrimination tasks involving acoustic variations of Cantonese speech sounds in three stimulus contexts: consonants (voice onset times, VOTs), lexical tones, and vowels. The results showed similar categorical perception performance in boundary position and width in the identification task and similar d’ scores in the discrimination task between the CWS and CWNS groups. However, the reaction times (RTs) were slower in the CWS group compared with the CWNS group in both tasks. Moreover, the CWS group had slower RTs in identifying stimuli located across categorical boundaries compared with stimuli located away from categorical boundaries. Overall, the data implied that the phoneme representation evaluated in speech perception might be intact in CWS as revealed by similar patterns in categorical perception as those in CWNS. However, the CWS group had slower processing speeds during categorical perception, which may indicate an insufficiency in accessing the phonemic representations in a timely manner, especially when the acoustic stimuli were ambiguous.
Stuttering is a speech fluency deficit that is mainly manifested by involuntary repetition and prolongation of sounds and syllables, as well as momentary blocks in speech production, which can induce negative impacts on the social, occupational, and academic life of people who stutter (PWS) . Earlier epidemiological studies have reported that 70% of children who stutter (CWS) naturally outgrow their stuttering , whereas more recent studies have suggested a higher rate of natural recovery, close to 90%, and the remaining 10% will experience persistent stuttering across their life span . The underlying mechanisms of persistent stuttering have been debated based on many theories, including speech motor control  and psycholinguistic theories .
The purpose of this study was to identify any speech perception differences between Cantonese-speaking CWS and CWNS by examining the categorical perception of three types of speech sound distinctions (i.e., VOTs, lexical tones, and vowels). The results showed that there were no significant differences between the CWS and CWNS groups in terms of the boundary position and boundary width across the three stimulus continua in the identification task. Regarding the discrimination task, although the overall patterns indicated that the subjects’ perceptions were categorical (i.e., higher d’ scores for the between-category pairs compared with the within-category pairs), again, no significant difference was found between the CWS and CWNS groups with regard to the benefit of between-category discrimination relative to within-category discrimination (Fig 6). These results suggest similar capacities in categorical perception between the CWS and CWNS groups and refute a robust deficit in the organization of phonemic representation measured in speech perception among the CWS. In contrast, a former study  that used a similar paradigm found that the phonemic boundaries were located at a longer VOT continuum for AWS compared with TFS. Although our study found a similar pattern (i.e., a higher phonemic boundary position and width in the VOT continuum for the stuttering group), the difference was not statistically significant. One important difference between the two studies is the target age group, as the former study examined AWS and the current study focused on CWS. As argued by others, there are some differences between adults and children in terms of their attention capacities toward auditory stimuli and the time course of auditory perception development  that might have played a role in the different findings in the two studies.
Overall, the findings from this study may not fully support the presence of a robust deficit in the organization of phonemic representation evaluated in speech perception (i.e., categorical perception) in CWS; however, it supports the presence of slower perceptual speech processing and difficulty in accessing phonemic representation in a timely manner, especially when the acoustic stimuli were more ambiguous. Given the fact that the behavioral measures mainly accounted for the endpoints of the cognitive processes, one future direction could involve the use of ERPs, which offer a high temporal resolution in examining the online cognitive processes of speech perception in CWS.