Chapter 45 Developmental Language Disorders
Introduction
Developmental language disorder (DLD; also known as Specific Language Impairment – SLI) is a condition in which a child with normal intelligence and hearing fails to develop language in an age-appropriate fashion. DLD is a clinical diagnosis, based on the presence of a normal nonverbal IQ, evidence of expressive and/or receptive language at least 1.5 standard deviations below the mean for age, and absence of other specific conditions such as autism, mental retardation, metabolic or genetic disorders, or severe environmental deprivation. The prevalence of DLD varies widely in different reports with a range of 1–11 percent [Bishop, 2002a; Shriberg et al., 1999]. Box 45-1 lists normal language milestones as a baseline for assessing a child’s developing language competence.
Pathophysiology
The underlying neural dysfunction that causes DLD is not known. Studies over the past 30 years have suggested that DLD is the result of a deficit in processing rapid auditory information. Evidence for this hypothesis comes from research showing that children with DLD have difficulty discriminating both nonspeech and speech sounds that are presented rapidly or very briefly in time [Tallal et al., 1993; Tallal and Benasich, 2002]. Some investigators argue for a speech-specific processing deficit [Mody et al., 1997], while others believe that the deficit is more general and that processing of all auditory information is slowed in DLD, whether the information presented is linguistic or nonlinguistic (e.g., environmental sounds) [Cummings and Ceponiene, 2010]. Studies using event-related brain potentials (ERPs) have demonstrated clear differences between DLD and control children, with language-impaired individuals having delayed N400 responses to incongruous picture–word combinations [McArthur et al., 2009].
MRI studies have shown relatively subtle differences between the brains of DLD and controls. Some children and adults with DLD (as well as relatives of DLD probands) do not have the typical planum temporale asymmetry pattern [Jackson and Plante, 1996; Gauger et al., 1997]. The absence of the typical planum asymmetry may be the result of aberrant neurogenesis, which leads to reduced cell development in the perisylvian regions or atypical patterns of cell death. Atypical right-biased asymmetries have also been reported in the prefrontal region [Jernigan et al., 1991]. An extra sulcus in the inferior frontal gyrus was associated with a history of DLD [Clark and Plante, 1998] in a group of 41 neurologically normal adults. DLD children may have decreased white-matter volumes in a left-hemispheric network comprising the motor cortex, the dorsal premotor cortex, the ventral premotor cortex, and the planum polare in the superior temporal gyrus [Jancke et al., 2007]. Rare reports document right hemisphere abnormalities in the DLD child that are suggestive of a right hemisphere contribution to language acquisition [Plante et al., 2001]. Trauner et al. [2000], in a series of 35 children with DLD, found evidence of structural abnormalities in one-third of the children. These included ventricular enlargement (n = 5), central volume loss (3), and white-matter abnormalities (4). The findings were consistent with bilateral white-matter disruption, and suggested that connectivity between different brain regions important for language development might be disrupted in DLD.
Other structural differences in brain development have been observed in some individuals with specified DLD types. Perisylvian abnormalities of varying degrees and associated with language disorders of varying severity have been reported, particularly in verbal dyspraxia and the phonological syntactic syndromes (see below for nosology). Complete opercular agenesis has been reported in association with suprabulbar palsy (Worster–Drought syndrome). Polymicrogyria has also been reported in the perisylvian region in children with DLD. Patients with the most extensive disease have the greatest language impairments, while those with posterior parietal polymicrogyria have milder symptoms [Guerreiro et al., 2002; Nevo et al., 2001]. One form of DLD, semantic pragmatic syndrome, has been reported in patients with agenesis of the corpus callosum and with hydrocephalus, which supports a possible localization in the subcortex and its connections or a disconnection effect. Consistent with this, callosal size may be decreased in some children with DLD [Preis et al., 2000]. In the KE family (see below) the caudate nucleus and inferior frontal gyrus are reduced in size bilaterally, while the left frontal opercular region (pars triangularis and anterior insular cortex) and the putamen bilaterally have a greater volume of gray matter [Watkins et al., 2002a]. An insufficient dosage of a critical forkhead transcription factor during embryogenesis may lead to malformations of regions of the brain necessary for speech and language development [Lai et al., 2001].
Metabolic imaging suggests abnormalities in the left temporal region and may vary by DLD subtype. Some children with DLD may be right hemisphere language-dominant [Bernat and Altman, 2003]. Whitehouse and Bishop [2008] compared language organization in four groups of adults – those with persisting SLI, those with a history of SLI, those on the autistic spectrum (ASD) with language impairment, and matched controls – using functional transcranial Doppler ultrasonography (fTCD), which assesses blood flow through the middle cerebral arteries. The participants were asked to generate words starting with a given letter silently and then later were required to verbalize these words. All of the participants in the SLI-history group and the majority of participants in the ASD (81.8 percent) and typical (90.9 percent) groups had greater activation in the left compared to the right middle cerebral arteries, while the majority of participants in the persistent SLI group had bilateral (27 percent) or right hemisphere (55 percent) language function. The investigators suggest that atypical language dominance may be a marker of persisting SLI. All 17 of the DLD children studied by Im et al. [2007] had grossly normal magnetic resonance imaging (MRI); however, 87.5 percent had decreased metabolic activity on positron emission tomography (PET) studies, most frequently in the thalamus, but also in both frontal, temporal, and right parietal areas, and significantly increased metabolism in both occipital areas as compared to a control group. Children with SLI showed significantly lower cerebral blood flow (CBF) values in the right parietal region and in subcortical regions compared to an attention-deficit hyperactivity disorder (ADHD) group. In addition, the DLD group had symmetric CBF distributions in the left and right temporal regions, whereas the ADHD group showed the usual asymmetry with left-sided hemispheric predominance in the temporal regions. The findings provide further evidence of anomalous neurodevelopment with deviant hemispheric lateralization as an important factor in the etiology of SLI. They also point to the role of subcortical structures in language impairment in childhood [Ors et al., 2005].
Although DLD is often referred to as “specific” language impairment, individuals with this disorder often have other accompanying problems, including abnormalities in gross and fine motor function [Trauner et al., 2000; Visscher et al., 2007]. These findings are indicative of more global neurological dysfunction and may reflect the fact that language disorders do not occur in isolation, but are merely the most prominent symptom of a more widespread neural network malfunction. Stuttering is a disorder involving the rhythm and fluency of speech production, as opposed to language. The stutterer knows what s/he wishes to say but cannot get the words out without significant dysfluency and hesitation. The neurobiological basis of stuttering has been the subject of a number of recent studies [Brown et al., 2005; Watkins et al., 2008]. During speech production, regardless of fluency or auditory feedback, stutterers showed overactivity relative to controls in the anterior insula, cerebellum, and midbrain bilaterally, and underactivity in the ventral premotor, rolandic opercular and sensorimotor cortex bilaterally, and Heschl’s gyrus on the left. These results are consistent with a recent meta-analysis of functional imaging studies in developmental stuttering. Overactivity occurred in the midbrain, at the level of the substantia nigra, and extended to the pedunculopontine nucleus, red nucleus, and subthalamic nucleus. This overactivity is consistent with suggestions in previous studies of abnormal function of the basal ganglia or excessive dopamine in stutterers. Underactivity of the cortical motor and premotor areas was associated with articulation and speech production. Analysis of the diffusion data revealed that the integrity of the white matter underlying the underactive areas in ventral premotor cortex was reduced in people who stutter. The white matter tracts in this area, via connections with posterior superior temporal and inferior parietal cortex, provide a substrate for the integration of articulatory planning and sensory feedback, and via connections with primary motor cortex, a substrate for execution of articulatory movements. These data would lead to the conclusion that stuttering is a disorder related primarily to disruption in the cortical and subcortical neural systems supporting the selection, initiation, and execution of motor sequences necessary for fluent speech production [Watkins et al., 2002b]. A recent meta-analysis of these studies identified three “neural signatures” of stuttering: people who stutter show more activity than fluent-speaking controls in the cerebellar vermis and in the right anterior insular cortex, with an “absence” of activity in the auditory cortices in the superior temporal lobe [Brown et al., 2005]. These abnormal levels of activity were observed during speech production, regardless of the presence or absence of stuttered speech during scan acquisition. Surprisingly, the meta-analysis did not reveal abnormal levels of activity in the basal ganglia circuitry, despite early imaging work on small samples showing abnormal metabolism.
Factors Associated with DLD
As with many neurodevelopmental disorders, there is a higher incidence of DLD in males. One recent study showed a ratio of 1.66:1 males to females in a group of children with communication disorders [McLeod and McKinnon, 2007]. The cause for the gender differences is not known. A number of biological and environmental risk factors for DLD have been identified. Box 45-2 lists a number of disorders associated with language impairment. Low birth weight and prematurity, as well as prenatal exposure to drugs (e.g., cocaine) and to cigarettes, adversely affect language development [Lewis et al., 2007]. Although frequent episodes of otitis media have been suggested as causing language impairment, there is little evidence from controlled studies to indicate a causal relationship. Intermittent hearing loss may interfere with language development in at-risk children, but is not likely to cause long-term language issues in otherwise normally developing children. Other environmental factors that may potentially adversely affect language development have been studied. For example, early and excessive TV exposure has been associated with language delay. Children who started watching television before 12 months of age and watched television for more than 2 hours per day were approximately six times more likely to have language delays than children without such early TV exposure [Chonchaiya et al., 2008]. Such associations, however, do not indicate causation.
Box 45-2 Disorders Known to be Associated with Language Delay/Impairment
Language impairment is seen in association with specific neurological and genetic disorders [Nass and Frank, 2010]. For example, perisylvian polymicrogyria (or congenital bilateral perisylvian syndrome) is a disorder of defective neuronal migration that has a spectrum of neurological impairments that include severe epilepsy and cognitive impairment. In some children with this condition, language impairment is the most prominent feature [Brandao-Almeida et al., 2008]. Language impairment has been described in children with neurofibromatosis type 1, although those individuals often have other cognitive deficits and learning disabilities as well. Language impairment is also prominent in a number of chromosomal disorders, including Down, Klinefelter’s, and fragile X syndromes. Epileptic encephalopathies, particularly Landau–Kleffner syndrome (LKS), may present with language impairment as an isolated or primary symptom [Kleffner and Landau, 2009]. Children with LKS typically have greater receptive than expressive language dysfunction, although they may become completely aphasic and mute in the course of the disease process. Rolandic epilepsy, often considered to be “benign”, may be complicated by DLD and learning disabilities [Lillywhite et al., 2009].
Genetics
Heritability rates for DLDs run as high as 0.5 [Byrne et al., 2009], but they are very variable and are affected by the criteria used to diagnose DLD. The median incidence rate for language difficulties in the families of children with language impairment runs as high as 35 percent, compared with a median incidence rate of 11 percent in control families [Stromswold, 1998]. Increased monozygotic versus dizygotic twin concordance rates indicate that heredity, not just shared environment, is responsible for familial clustering [Bartlett et al., 2002; Stromswold, 2001]. Heritability is substantially higher if DLD is identified based on referral to speech and language pathology services [Bishop and Hayiou-Thomas, 2008] than if it is identified by language test scores. Childhood language disorders that are demonstrated in population screening are likely to have different phenotypes and different etiologies than clinically referred cases. It is also possible, however, that familial cases are more likely to come to the attention of specialists since the parents are more attuned to the problem and recognize it earlier in subsequent children.
Studies using genome-wide scanning have implicated a number of gene loci, but the same loci have not been found in a reproducible fashion [SLI Consortium, 2002; Grigorenko, 2009]. One exceptional family stands out. In the three-generation KE family, half the members are affected with a severe speech and language disorder that is transmitted as an autosomal-dominant monogenic trait involving the FOXP2 forkhead-domain gene [Watkins et al., 2002b]. Notably, however, a recent screening of 270 4-year-olds with DLD was negative for the FOXP2 mutation [Meaburn et al., 2002]. It is unlikely that there will prove to be specific genes whose function would be restricted to forming the genetic basis for speaking and language acquisition. It is much more likely that there are many genes that contribute to a variety of functions, and that these genes form networks that are recruited in the process of forming language-related representations during language acquisition [Grigorenko, 2009]. The definition of the phenotype forming the basis for specific genetic studies is crucial. To date, in addition to the KE family phenotype, nonword repetition has proven the most useful phenotype in molecular genetic research [SLI Consortium, 2004; Vernes et al., 2008]. The issue of pleiotropy, or the impact of the same genes on multiple phenotypes, has also been discussed in the literature on DLD, given the substantive overlap in regions of linkage for a variety of developmental disorders, such as speech and sound disorders (SSD) and developmental dyslexia [Miscimarra et al., 2007; Stein et al., 2004], and SLI and autism [Tager-Flusberg and Joseph, 2003]. Once again, whether these are true examples of pleiotropy or outcomes of the imprecision of phenotype definitions is yet to be determined [Grigorenko, 2009].
Diagnosis
Box 45-3 lists warning signs that suggest a DLD during the first 3 years. Language delay can be diagnosed very early, whether the delay is primarily expressive only or mixed receptive and expressive. Before the age of 2 years, however, delay may not always equal disorder. Research on late-talking toddlers reveals a lack of homogeneity within the population of children with a vocabulary delay at 2 years of age. In a recent study, only about 40 percent of children retained the diagnosis of DLD at ages 3 and 4 years [Dale et al., 2003]. This is particularly true if the early language delay was primarily expressive. Children with receptive language impairments are more likely to have a persistent DLD. The pattern of receptive language development is highly predictable during the elementary school years. In a study of 184 children age-assessed at three time points – 7 years, 8 years, and 11 years of age – receptive language disorder was associated with declining rates of language growth over time [Law et al., 2008]. Thus, concern for poor language prognosis should be heightened when receptive language deficits are identified.
Box 45-3 Warning Signs of a Developmental Language Disorder
