In this study we carried out a comprehensive evaluation of speech abilities in adolescents born preterm and examined the relationship to conventional MRI findings and DWI investigations of the primary motor pathway. We have shown that although there is no evidence of clinically significant developmental dysarthria, dyspraxia, or a speech-sound disorder in this population, specific difficulties in speech and non-speech oromotor control are common. Furthermore, with DWI, we identified changes along the CST/CBT that were predominantly left-lateralized in the impaired group.
Studies of preterm individuals conducted in early childhood have identified speech deficits, most commonly in speech-sound processing and production.3-6,28
Although this study was not prospective, our data suggest that these problems may resolve by adolescence as no significant speech-sound production or phonological awareness deficits could be demonstrated, despite one-third of the sample having received speech and language therapy. In particular, the incidence of lisps and phonological errors was not higher than expected in the general population29
and deficits in phonological awareness were entirely consistent with IQ scores.
In contrast, 31% of those assessed showed problems in oromotor and speech-motor control, including difficulties in the precision of individual and combined movements of the lips, jaw, face, and tongue. This is the first study to examine speech and oromotor abilities in adolescence, and our findings are in line with reports in younger children of very low birth weight.10
Further, we have shown that although neurologic deficits were significantly more common in this group, the presence of cerebral palsy or minor neurologic findings alone could not predict FOC difficulties (4 of 11 cases had no detectable neurologic deficit).
It is important to consider these findings in the context of the limitations of this study. The sample assessed included a large proportion of individuals who had a positive cUS finding at birth, including 10 with major brain injury. A higher incidence of oromotor problems might therefore be expected. Also, because of the retrospective nature of the study, and limited information about the nature of the initial problems and the amount of speech and language therapy received, the effect of interventions could not be assessed. This should be addressed in future prospective studies.30
The possible influence of subtle hearing problems also cannot be excluded, because a detailed audiological examination was not performed. However, there was no evidence for alterations in speech production consistent with a hearing impairment (eg, sibilant distortion), and the impaired group did not perform differently from others on any of the auditory screening tests administered, suggesting this would not have influenced the results.
Despite these limitations, it is encouraging that even in the more severely affected individuals, none were diagnosed with a persistent speech disorder at follow-up. Because this is the most comprehensive study of speech-outcome in this population to date, these findings suggest that speech problems are not a major area for concern in adolescents born prematurely.
CST damage can account for most of the motor deficits in children with cerebral palsy and neuroimaging of the internal capsule plays an important role in predicting motor outcome in such infants.31
The link between more subtle motor problems, which are common in preterm children,7
and imaging abnormalities is less well established.32
In this study, children with speech-motor and oromotor deficits were more likely to show brain injury on MRI at follow-up, but this was not a significant predictor for focal oromotor impairment in the regression analysis. In contrast, DWI measures of diffusion anisotropy in the left primary motor pathway contributed significantly to this prediction (in addition to the presence of neurologic impairment).
The sensitivity of DWI to detect primary periventricular white matter lesions and secondary changes further downstream in the PLIC has been demonstrated previously.33
In this sample, such FA reductions may therefore reflect Wallerian degeneration23
caused by periventricular white matter injury to the CST/CBT during the neonatal period.34,35
This is supported by evidence showing that the main cause for FA reduction in the PLIC in preterm children is caused by an increase in radial diffusivity,36
which is in keeping with the diffusion characteristics of secondary degeneration,23
and periventricular white matter abnormalities were one of the most common findings on MRI in this group.
DWI of the genu, PLIC, and midbrain therefore may provide a sensitive measure of the degree and laterality of damage to the primary motor pathway not easily visible on cUS or conventional MRI. This conclusion is supported by closer examination of the laterality of major lesions detected on cUS at birth: in the impaired group 3 of 5 children with asymmetrical injury had more severe lesions in the left hemisphere, while the opposite was true (right larger than left) in the 3 asymmetrical cases in the unimpaired group.
These results are not surprising considering the dominant role of the left hemisphere in speech production and articulation37
and the recent evidence showing that dysarthria is more common and severe in adults with left rather than right cerebral hemisphere lesions.38
However, from a developmental perspective, persistent speech disorders are generally associated with bilateral injury.39,40
Our findings are compatible with these data because none of the participants fulfilled the criteria for a clinically significant speech disorder.
In classical descriptions, the CST is said to descend in the PLIC, with representations of the face, hand, and foot arranged from anterior to posterior.41
Corticobulbar fibers concerned with speech articulation (representing the face, tongue, and larynx) therefore should lie anteriorly, close to the genu. Therefore we might expect deficits of speech and oromotor control to be most strongly associated with internal capsule abnormalities in this region, but this specificity was not detected. We suspect that this may reflect more generalized damage to the primary motor projections in the left hemisphere, affecting fibers passing through all regions of the posterior limb—including those of the hand and foot. DWI tracking of the individual fiber pathways was not performed here; however, we did show that individuals with oromotor problems were also slower at the hand-motor task, particularly with their right hand (see reference 42
). A more detailed examination of neurologic dysfunction may have revealed further evidence for lateralized motor deficits in this group. Although it is conceivable that difficulties in fine hand-motor control might cause real-world difficulties in these children, it is reassuring that residual impairments in FOC do not affect normal speech production in adolescence.
We have shown that everyday speech is generally preserved in adolescents born preterm, even in children with positive MRI findings, neurologic impairment, or a history of speech and language problems. Furthermore, our data suggest that speech abilities may improve with age, in contrast with evidence that intellectual function may decline with time in this population.43,44
This could be caused by the timely provision of speech and language therapy or the natural plasticity and reorganization of the developing brain. Our identification of a neural correlate for focal oromotor problems in adolescence raises the possibility of using neonatal imaging to identify individuals at greatest risk of the development of speech problems in early childhood.