New brain imaging study links childhood speech disfluencies to neural development patterns

A groundbreaking study of five-year-olds reveals that speech disfluencies, often viewed as developmental norms, may also serve as indicators of how speech and language networks evolve in the brain, offering new insights into childhood fluency and stuttering disorders.

Speech disfluencies in young children are often a normal part of development, but a new study suggests they may also help reveal how speech and language networks mature in the brain. The research focuses on five-year-olds and treats disfluency as a continuous trait rather than placing children into fixed categories, a shift that reflects growing evidence that fluency varies from day to day and across situations. According to the American Speech-Language-Hearing Association, ordinary hesitations, repetitions and filler words such as “uh” and “um” are common, but more disruptive patterns can signal stuttering or another fluency disorder.

The study builds on a long line of imaging work in children and adults who stutter, much of which has pointed to differences in frontal, temporal and subcortical regions involved in speech and motor control. Prior research using voxel-based morphometry and FreeSurfer has repeatedly implicated the left inferior frontal gyrus, superior temporal areas, the putamen, caudate nucleus and premotor cortex, although findings have varied in direction and strength. Surface-based studies have also reported reduced cortical thickness in left frontal and motor regions, reinforcing the view that fluency depends on a distributed speech-motor system.

The authors say the new work is intended to probe that system from a different angle. For stuttering-like disfluencies, they draw on neurocomputational models such as DIVA and GODIVA, which describe how cortical and basal ganglia circuits support speech initiation and sequencing. For other disfluencies, such as revisions and filled pauses, they point to language-planning systems linked to the dual-stream model of speech processing, which separates motor-related pathways from those involved in lexical and semantic formulation. In practical terms, that means the study is looking not only at whether children stutter, but at how often they produce different kinds of interruptions in everyday speech.

Five-year-olds were chosen because the age marks an important stage in language development and in the course of childhood stuttering. Many children who begin to stutter do so between ages two and four, and the likelihood of new onset falls sharply after about age five, according to longstanding clinical research and guidance from the National Institute on Deafness and Other Communication Disorders. The authors also note that the preschool years are a period of rapid brain change, with cortical thickness and grey matter still developing in a regionally uneven way. Against that backdrop, the study uses whole-brain analyses of grey matter volume alongside measures of cortical thickness and surface area to see whether different kinds of disfluency map on to different neural signatures.