Prevalence Estimates

Prevalence estimates of speech sound disorders are essential in conducting behavioral and molecular genetic studies. They are used to calculate an individual's risk of having a disorder as well as to test different genetic models of transmission. Prevalence rates for speech disorders may vary based on the age and sex of the individual, the type of disorder, and the comorbid conditions associated with it. In an epidemiological sample, Shriberg and Austin (1998) reported the prevalence of speech delay in 6-year-old children to be 3.8%. Speech delay was approximately 1.5 times more prevalent in boys than in girls. Shriberg and Austin (1998) also found that children with speech involvement have a two to three times greater risk for expressive language problems than for receptive language problems. Estimates of the comorbidity of receptive language disorders with speech disorders ranged from 6% to 21%, based on whether receptive language was assessed by vocabulary, grammar, or both. Similarly, estimates of comorbidity of expressive language disorders with speech disorders ranged from 38% to 62%, depending on the methods used to assess expressive skills.

Phenotype Definitions

Phenotype definitions (i.e., the behavior that is under study) are also crucial for genetic studies of speech disorders. Phenotype definitions may be broadly or narrowly defined, according to the hypothesis to be tested. A broad phenotype may include language as well as speech disorders and sometimes related language learning difficulties such as reading and spelling disorders (Tallal, Ross, and Curtiss, 1989). An individual exhibiting a single disorder or a combination of disorders is considered to be affected. Such a broad phenotype may test a general verbal trait deficit hypothesis which holds that there is a common underlying genetic and cognitive basis for speech and language disorders that is expressed differently in individual family members (i.e., variable expression). An alternative explanation is that each disorder has a unique underlying genetic and cognitive basis. Some investigators have narrowly defined the phenotype as a specific speech disorder, such as phonology (Lewis, Ekelman, and Aram, 1989). Even if the proband (i.e., the child with a disorder from whom other family members are identified) is selected by a well-defined criterion, nuclear family members often present a varied spectrum of disorders. Some studies, while narrowly defining the phenotype for the proband, have used a broad phenotype definition for family members. Since older siblings and parents often do not demonstrate speech sound errors in their conversational speech, researchers have relied on historical reports, rather than direct observations of the speech disorder. Figure 1 shows a typical family pedigree of a child with a speech disorder.

Figure 1..  

A typical family pedigree of a child with a speech disorder. The arrow indicates the proband child. Male family members are represented by squares and female family members are represented by circles. Individuals who are affected with speech disorders are shaded in black. Other disorders are coded as follows: Read = reading disorder, Spell = spelling disorder, Lang = language disorder, LD = learning disability, Apraxia = apraxia of speech, Artic = articulation disorder.

Narrow phenotype definitions may examine subtypes of phonology disorders with postulated distinct genetic bases. One schema for the subtyping of phonology disorders may be based on whether or not the phonology disorder is accompanied by more pervasive language disorders (Lewis and Freebairn, 1997). Children with isolated phonology disorders experience fewer academic difficulties than children with phonology disorders accompanied by other language disorders (Aram and Hall, 1989). Shriberg et al. (1997) propose at least two forms of speech sound disorders of unknown origin: those with speech delay and those with questionable residual errors. These two subtypes may have different genetic or environmental causes.

Sex as a Risk Factor

A robust finding in studies of familial speech and language disorders has been a higher prevalence of disorders in males than in females, ranging from a 2:1 to a 3:1 ratio (Neils and Aram, 1986; Tallal et al., 1989; Tomblin, 1989; Lewis, 1992). Explanations for this increased prevalence in males include referral bias (Shaywitz et al., 1990), immunoreactive theories (Robinson, 1991), differences in rates and patterns of neurological maturation (Plante, 1996), variation in cognitive phenotypes (Bishop, North, and Donlan, 1995), and differences in genetic transmission of the disorders. An X-linked mode of transmission of speech and language disorders has not been supported by pedigree studies (Lewis, 1992; Beitchman, Hood, and Inglis, 1992). However, a sex-specific threshold hypothesis that proposes that girls have a higher threshold for expression of the disorder, and therefore require a higher genetic loading (more risk genes) before the disorder is expressed, has been supported (Tomblin, 1989; Lewis, 1992; Beitchman, Hood, and Inglis, 1992). Consistent with this hypothesis, a higher percentage of affected relatives are reported for female (38%) than for male probands (26%). Differing sex ratios may be found for various subtypes of phonology disorders (Shriberg and Austin, 1998). Hall, Jordan, and Robin (1993) reported a 3:1 male to female ratio for developmental apraxia of speech. Similarly, boys with phonology disorders were found to have a higher rate of comorbid language disorders than girls with phonology disorders (Shriberg and Austin, 1998). Lewis et al. (1999) found that probands with phonology disorders alone demonstrated a more equal sex ratio (59% male and 41% female) than probands with phonology disorders with language disorders (71% male and 29% female).

Familial Aggregation

Familial aggregation refers to the percentage of family members demonstrating a disorder. Familial aggregation or family resemblance may be due to heredity, shared family environment, or both. Research has supported the conclusion that speech and language disorders aggregate within families (Neils and Aram, 1986; Tallal, Ross, and Curtiss, 1989; Tomblin, 1989; Gopnik and Crago, 1991; Lewis, 1992; Felsenfeld, McGue, and Broen, 1995; Lahey and Edwards, 1995; Spitz et al., 1997; Rice, Haney, and Wexler, 1998). Reports indicate that 23%–40% of first-degree family members of individuals with speech and language disorders are affected. Differences in reported rates of affected family members again may be attributed to differences in definitional criteria for probands and family members.

Two studies specifically examined familial aggregation of phonology disorders (Lewis, Ekelman, and Aram, 1989; Felsenfeld, McGue, and Broen, 1995). Both studies reported 33% of first-degree family members (nuclear family members) to have had speech-language difficulties. Brothers were most frequently affected.

Behavioral and Molecular Genetic Studies

The twin study paradigm has been employed to identify genetic and environmental contributions to speech and language disorders. Twin studies compare the similarity (concordance) of identical or monozygotic twins to fraternal or dizygotic twins. If monozygotic twins are more concordant than dizygotic twins, a genetic basis is implied. To date, a twin study specifically examining speech sound disorders has not been conducted. Rather, twin studies have employed a broad phenotype definition that includes both speech and language disorders. Twin studies of speech and language disorders (Lewis and Thompson, 1992; Bishop, North, and Donlan, 1995; Tomblin and Buckwalter, 1998) have consistently reported higher concordance rates for monozygotic than for dizygotic twin pairs, confirming a genetic contribution to these disorders. Concordance rates for monozygotic twins range from .70 (Bishop, North, and Donlan, 1995) to .86 (Lewis and Thompson, 1992) and .96 (Tomblin and Buckwalter, 1998). Concordance rates reported for dizygotic twin pairs are as follows: .46 (Bishop, North, and Donlan, 1995), .48 (Lewis and Thompson, 1992), and .69 (Tomblin and Buckwalter, 1998). A large twin study (3000 pairs of twins) suggested that genetic factors may exert more influence at the lower extreme of language abilities, whereas environmental factors may influence normal language abilities more (Dale et al., 1998). These studies, while supporting a genetic contribution to speech and language skills, also indicate a moderate environmental influence. Environmental factors working with genetics may determine speech and language impairment in an individual.

Consistent with these findings, an adoption study by Felsenfeld and Plomin (1997) demonstrated that a history of speech and language disorders in the biological parent best predicted whether or not a child was affected. This relationship was not found when the family history of the adoptive parents was considered. As with twin studies, a broad phenotype definition that encompassed both speech and language disorders was employed. Adoption studies have not been conducted for speech sound disorders alone.

Segregation analyses examine the mode of transmission of the disorder within a family. Segregation analyses have confirmed familial aggregation of speech and language disorders and supported both a major locus model and a polygenic model of transmission of the disorder (Lewis, Cox, and Byard, 1993). The failure to define a definitive mode of transmission may be due to genetic heterogeneity (i.e., more than a single underlying genetic basis).

Only a single study to date has reported a molecular genetic analysis of a family with apraxia of speech and other language impairments. Genetic studies of a single large pedigree, known as the K.E. family, revealed linkage to a region of chromosome 7 (Vargha-Khadem et al., 1995; Fisher et al., 1998). Subsequently, the FOXP2 gene that is postulated to result in the development of abnormal neural structures for speech and language was identified (Lai et al., 2001). Neuroimaging of family members indicated abnormalities in regions of the frontal lobe and associated motor systems. This was the first study that provided direct evidence for a genetic basis for a speech sound disorder associated with a neurological abnormality. It was the initial step in the application of molecular genetic techniques to the study of speech and language disorders. Further studies are needed to determine if FOXP2 is found in other families with speech disorders.

See also speech disorders in children: descriptive linguistic approaches; developmental apraxia of speech; phonological errors, residual.