Childhood Apraxia of Speech

Childhood Apraxia of Speech Versus Developmental Apraxia of Speech

The Committee recommends childhood apraxia of speech (CAS) as the classification term for this distinct type of childhood (pediatric) speech sound disorder. Beginning with the first word in this term, two considerations motivate replacing the widely used developmental with the word childhood. One consideration is that CAS support groups in the United States, the United Kingdom, and elsewhere have requested that developmental not be used in a classification term for this disorder. Inclusion of this word is reportedly interpreted by service delivery administrators as indicating that apraxia is a disorder that children “grow out of” and/or that can be serviced solely in an educational environment (see relevant discussion on the Apraxia-Kids listserv: www.apraxia-kids.org/talk/subscribe.html). A second rationale for the use of CAS as a cover term for this disorder, rather than alternative terms such as developmental apraxia of speech (DAS) or developmental verbal dyspraxia (DVD), is that our literature review indicated that apraxia of speech occurs in children in three clinical contexts. First, apraxia of speech has been associated causally with known neurological etiologies (e.g., intrauterine stroke, infections, trauma). Second, apraxia of speech occurs as a primary or secondary sign in children with complex neurobehavioral disorders (e.g., genetic, metabolic). Third, apraxia of speech not associated with any known neurological or complex neurobehavioral disorder occurs as an idiopathic neurogenic speech sound disorder. Use of the term apraxia of speech implies a shared core of speech and prosody features, regardless of time of onset, whether congenital or acquired, or specific etiology. Therefore, childhood apraxia of speech (CAS) is proposed as a unifying cover term for the study, assessment, and treatment of all presentations of apraxia of speech in childhood. As above, CAS is preferred over alternative terms for this disorder, including developmental apraxia of speech and developmental verbal dyspraxia, which have typically been used to refer only to the idiopathic presentation.

Apraxia Versus Dyspraxia

Rationales for the second and third words in the classification term CAS reflect empirical findings for children suspected to have this disorder. The alternative terms—apraxia of speech versus (verbal) dyspraxia—each have established traditions in international literatures. Apraxia of speech is more widely used in the United States following the Mayo Clinic traditions (Duffy, 2005), whereas verbal dyspraxia is the preferred term in many other English-speaking countries. Differentiating between these alternatives based solely on etymological distinction (i.e., total [a] vs. partial [dys] absence or lack of function) is problematic when applied to CAS. Clinical experience indicates that although a child suspected to have CAS may have very limited speech, seldom is a child completely without mastery of some speech sounds. Notwithstanding this difference, and to parallel usage for the possible acquired form of this disorder in adults (i.e., AOS), the Committee recommends use of the affix a for this classification term.

The Apraxias Versus the Dysarthrias

Several other types of apraxia and several types of dysarthria play prominent roles in the scientific foundations of CAS. Physicians and researchers recognize ideomotor and limb kinetic praxis problems that may or may not be present in persons with apraxia of speech. As discussed in this report, orofacial and limb apraxias are of particular interest as the presence of one or both in a child suspected to have CAS may provide support for the diagnosis, particularly in prelingual children. Apraxia in other systems may also play important roles in treatment. For example, the presence of limb apraxia may preclude using manual signs for functional communication. Moreover, the presence of orofacial apraxia may support the need for either more aggressive or alternative approaches to the use of phonetic placement cues in speech treatment.

Concerning dysarthria, a neuromotor disorder presumed not to involve the planning or programming deficit in apraxia (see below), some forms of these two disorders may share common speech characteristics. As discussed in later sections, a significant research challenge is to determine the diagnostic boundaries between CAS and some types of dysarthria with which it may share several speech, prosody, and voice features.

CAS Versus AOS

Although the core feature of CAS, by definition, is proposed to be similar to the core feature of AOS in adults, this relationship does not preclude the possibility of important differences in associated features. For example, Maassen (2002) noted that “a fundamental difference between [adult] AOS and [CAS]…is that in [CAS] a specific underlying speech motor impairment has an impact on the development of higher phonological and linguistic processing levels” (p. 263). Despite a much larger and well-developed literature in AOS, including many chapter-length discussions of alternative theoretical frameworks, the Committee elected not to include reviews of theory and research on acquired apraxia of speech in this report. This decision was motivated by the view that the scientific foundations of CAS should be based on research directly concerned with this and related childhood speech sound disorders. However, as discussed in several places in this document, the Committee has attempted to anticipate likely parallels between acquired apraxia of speech in adults and CAS, a task made more difficult by differences in terminology used to describe them. Treatment guidelines for acquired apraxia of speech have recently been proposed by the Academy of Neurological Communicative Disorders and Sciences (Wambaugh, Duffy, McNeil, Robin, & Rogers, 2006a, 2006b).

The Core Problem

As required of any proposed disorder classification, definitions of CAS have three elements that may be given in any order: description of the core problem, attribution of its cause or etiology, and listing of one or more diagnostic signs or markers. Definitions of CAS, such as the one above, invariably include the proposed core problem, whereas the other two elements may or may not be addressed. One of the major differences among alternative definitions of CAS is whether the core problem is proposed to include input processing as well as production, and if so, whether auditory, sensory, and prosodic aspects of perception may prefigure in the deficit. An example of a framework that might implicate the latter is the speech motor control model in development by Guenther and colleagues (e.g., Guenther, 2006; Guenther & Perkell, 2004). Whereas some of the definitions of CAS reviewed by the Committee view the core problem as one of planning and programming the spatiotemporal properties of movement sequences underlying speech sound production, others propose that the deficit extends to representational-level segmental and/or suprasegmental units in both input processing and production.

Etiology

Definitions of CAS have universally ascribed its origin to neurologic deficits, with alternative viewpoints differing with respect to specific neuroanatomic sites and circuits. There is also clear agreement that whatever the neural substrates of CAS, they differ from those underlying the several types of dysarthria. The definition of CAS proposed for this report is also clearly consistent with this neurogenic perspective.

Signs and Markers

In addition to the core problem and etiology, the third element in the proposed definition of CAS and those reviewed by the Committee is the inclusion of the key diagnostic features of the disorder. Three such features are included in the present definition, with discussion of other candidate features reviewed in subsequent sections of this report. The three features in the present definition of CAS represent a consensus conclusion based on our evaluations of the clinical research and our evaluation of comments from reviewers of preliminary drafts of this report. A major conclusion of this report is that there presently is no one validated list of diagnostic features of CAS that differentiates this disorder from other types of childhood speech sound disorders, including those apparently due to phonological-level deficits or neuromuscular disorder (dysarthria).

Birth-to-Three Legislation

One potential source of the apparent increased diagnostic prevalence of CAS in the past one to two decades is the impact of legislative changes during this period. Since the passage of early intervention statutes, particularly the Individuals with Disabilities Education Improvement Act of 2004 (IDEA '04, Part C), speech-language pathologists are asked to evaluate and identify communicative disorders as early as possible in infants and toddlers. A major problem in classifying young prelingual children (i.e., children with severe delays in the onset of speech) is that a diagnosis of CAS must be based on variables other than speech itself. As discussed later in this report, findings claiming that behaviors such as difficulty in feeding or excessive drooling are pathognomonic (positive signs) of CAS are tentative at best. For children suspected to have CAS who do have at least a moderate inventory of speech sounds, their communication profiles can be similar to those of children with other speech-language disorders or neurobehavioral disorders (Davis, Jakielski, & Marquardt, 1998; Davis & Velleman, 2000). Thus, although we use the term CAS for children who are the focus of the research reviewed in this document, it should be understood that the lack of a gold standard for differential diagnosis requires that all such classificatory labels be considered provisional.

Increased Information

Increased information on a disorder may both reflect and contribute to increased prevalence. For CAS, the past decade has seen dramatic increases in both. Interest in CAS is readily apparent when reviewing the increased number of research symposia (e.g., Shriberg & Campbell, 2003), clinical workshops, and parent support groups on CAS. Although there have been no formal accounts describing the history of this clear trend, it appears to parallel similar development in other disorders. From an academic perspective, information about CAS has traditionally been embedded within undergraduate and graduate courses in speech disorders in children or, more typically, in motor speech disorders in children and adults. However, for many speech-language pathologists, applied information on this topic is typically learned in workshops presented by persons with varying research and clinical backgrounds and/or experience with children suspected to have CAS. The Committee's anecdotal observations are that such workshops are currently among the most widely advertised opportunities for continuing education credits.

The major source of readily available information on CAS is the Internet, including its numerous Web sites and electronic discussion forums that include information on this topic. As with other unregulated medical and health-related information sources, the accuracy and usefulness of information presented on the Internet varies substantially. Some sites available internationally provide excellent information, including detailed guidelines for caregivers seeking service delivery options.

Reimbursement Issues

Speech-language pathologists must be knowledgeable about reimbursement alternatives and insurance guidelines. Because insurance companies frequently require that a child have a medical diagnosis to approve coverage, there may be increased use of CAS as a diagnostic classification for a severe childhood speech sound disorder. However, insurance claims for children with this diagnosis may sometimes be denied due to the continuing controversial status of CAS as a clinical entity and its increased prevalence in diagnostic coding.

Lack of Diagnostic Guidelines

Clearly the major source of overdiagnosis of CAS is the inconsistent and conflicting behavioral features purported to be diagnostic signs of CAS (Shriberg, Campbell, et al., 2003; Shriberg & McSweeny, 2002). In addition to children who may be misdiagnosed as false positives (persons said to have a disorder who do not), diagnostic guidelines also may result in false negatives (persons said not to have a disorder who do). Later discussion addresses this fundamental issue.

A Note on Terms

In any discussion of speech motor control, or speech production generally, the terms variable and inconsistent are likely to arise. They are often used interchangeably and without precise definitions. This is of particular concern with respect to CAS, as some clinical investigators use inconsistency as a key classification criterion for the disorder. Some common uses of variable and inconsistent include the following:

1. differential use of a certain phoneme or sound class in different word positions (e.g., the child produces /k/ accurately in final position but substitutes [t] for /k/ in prevocalic position);

2. differential use of a certain phoneme or sound class in different word targets, even in the same word position (e.g., the child produces /m/ accurately in certain well-rehearsed words such as “mommy,” but does not produce it accurately in similar or even seemingly easier words such as “moo”);

3. differential use of a certain phoneme or sound class in multiple repetitions of the same word (e.g., the child produces “fish” once as “pish”, once as “pit”, once as “fit”, and another time as “shiff”). This may include measures of the number of different errors the child made in the word (e.g., in the example above, errors consisted of stopping of /f/, stopping of /∫/, and metathesis) or measures of the frequency at which a given error type is used (e.g., in the example above, stopping was the most consistent error type because stopping was used four times, and metathesis only once). This type of inconsistency is sometimes referred to as “token-to-token variability” (Seddoh et al., 1996).

Except where specified otherwise within in this document, inconsistency refers to differences in multiple productions of the same target word or syllable (i.e., token-to-token variability). Variability is used elsewhere, when meaning 1 or meaning 2, or more than one of the above meanings, is included within the findings being reported or when parameters other than speech production (e.g., pitch) are being discussed.

Oral-Motor Development

Beginning this review with research on typical oral-motor development, studies indicate that jaw control is established by about 15 months, before control is established for the upper and lower lips (Green, Moore, Higashikawa, & Steeve, 2000; Green, Moore, & Reilly, 2002). Motor development is slower for structures, such as the lips, that have more degrees of freedom of movement (Green et al., 2002). Tongue development is also gradual, with extrinsic tongue movements necessary for swallowing and sucking developed prior to the intrinsic tongue movements required for fine motor control (S. G. Fletcher, 1973; Kahane, 1988). Such findings are hypothesized to account for the high frequency of occurrence of infants' production of syllables that can be articulated without changes in lip or tongue configuration—including labial consonants with low and neutral vowels, coronal (alveolar and dental) consonants with high front vowels, and dorsal (velar) consonants with high back vowels (Davis & MacNeilage, 1995; MacNeilage & Davis, 1990). The high prevalence of such syllables is claimed to be associated with infants' early ability to open and close the jaw, creating the consonant-vowel alternation necessary for the syllable, with the lower lip (for labials) or the tongue (for alveolars and velars) essentially “going along for the ride.” Some clinical reports indicate that these immature patterns may persist in children suspected to have CAS (Velleman, 1994).

Through processes of differentiation and refinement, the slightly older child acquires independent control over individual articulators (lips, different portions of the tongue) and learns to produce more specialized configurations to grade movements, eventually sequencing these articulatory postures without extraneous movements (Davis & MacNeilage, 2000; Green et al., 2000). Thus, automaticity and flexibility develop over time. Both neuromotor maturation and practice are believed to underlie this developmental process, with vocal experience leading to the formation of specific neuronal pathways for finer levels of control (Green et al., 2000). Coarticulation that reflects poor temporal control or poor differentiation of structures decreases, whereas coarticulation that reflects language-specific efficiency increases, as the child becomes more adept (Nijland et al., 2002; Nijland, Maassen, van der Meulen, et al., 2003). One model of the role of perception in this process was provided by Guenther and colleagues (e.g., Guenther, 2006; Guenther & Perkell, 2004). In a following section, we will see that these developmental changes may not occur spontaneously in children suspected to have CAS.

In the present context it is especially relevant to note that mastication and deglutition (swallowing) skills are not direct precursors to speech. Motor control of feeding functions is separate from motor control for vocalization early in infancy (Moore & Ruark, 1996), as is motor control for speech breathing versus breathing at rest (Moore, Caulfield, & Green, 2001). Although “the labiomandibular movement patterns established for feeding may influence initial attempts to coordinate these structures for speech” (Green et al., 2000, p. 252), this influence is more likely to be negative than positive, as feeding patterns involve tight linking of lips with jaw in a highly rhythmic stereotyped pattern. To produce a variety of syllables within varied prosodic patterns requires the child to overcome the interdependent inflexible patterns associated with sucking. Speech requires finer levels of coordination (Green et al., 2000) but lower levels of strength than are available for other oral-motor activities (Forrest, 2002). Thus, a consensus opinion among investigators is that nonspeech oromotor therapy is not necessary or sufficient for improved speech production (see also Professional Issues: Treatment).

When children reach middle school age and even beyond, their speech production continues to be more variable, less flexible, and less accurate than adult speech. Variability is especially noted during the initial portion of speech or speech-like movements, with more feedback required for unfamiliar speech tasks (Clark, Robin, McCullagh, & Schmidt, 2001). Furthermore, as discussed in Clark et al., children's speech may be constrained by resource allocation needs, such as the need to scale back the extent of a movement in order to complete it more quickly. For example, children between the ages of 5 and 6 years are able to partially compensate for the presence of a bite block between their teeth without an increase in variability or a change in coarticulation patterns, although vowel accuracy is decreased somewhat and segment durations are increased (Nijland, Maassen, & van der Meulen, 2003). Maximum performance rates have been shown to increase with age, with changes from 3.7 same syllable repetitions of /pΛ/ per second and only 1.3 repetitions of “patty-cake” at age 2;6–2;11 [years;months] to 5.5 same syllable repetitions and 1.6 repetitions of “patty-cake” at age 6;6–6;11 (Robbins & Klee, 1987). Maximum performance rates continue to increase with maturity, with young adult same syllable repetitions typically reported at average rates between 6 and 7 per second and 5.8 to 6.9 repetitions per second of /pΛtΛkΛ/ (Baken & Orlikoff, 2000). However, Williams and Stackhouse (1998, 2000) reported that rate of speech may be a less reliable measure of motor control in preschool children than accuracy and consistency of response. Again, many of the core questions about CAS address the possibility that children suspected to have CAS have different developmental trajectories on these and other motor control parameters.

Prelinguistic Period

Speech development begins long before the first word is spoken. Development of this system occurs as a child gains motor control of the speech mechanism and learns the phonological rules for production of the ambient language or languages. Prelinguistic perceptual and vocal experiences lay the groundwork for later speech and language. For example, the frequency of a child's vocalizations at 3–6 months is correlated with several later developmental milestones, including performance on the Bayley Verbal Scale at 11–15 months and expressive vocabulary size at 27 months (Stoel-Gammon, 1992).

One of the most important motor precursors to first oral words is canonical babbling, the rhythmic production of repetitive consonant-vowel (CV) sequences with complete consonant closures and fully resonant vowels (Ejiri, 1998; Oller, 1986). The frequency of occurrence of “true” supraglottal nonglide consonants in babble is positively correlated with phonological development and even with language skills (Stoel-Gammon, 1992). Children who demonstrate consistent vocal motor schemes, or favorite babbles, tend to develop words earlier (McCune & Vihman, 1987). Children suspected to have CAS who are reported by their parents to have babbled very little or with very little phonetic diversity (Davis & Velleman, 2000) are at a linguistic disadvantage long before word production begins. The frequency and characteristics of early vocalizations also can be affected by perceptual factors such as early otitis media with effusion (Petinou, Schwartz, Mody, & Gravel, 1999; Rvachew, Slawinski, Williams, & Green, 1999), as well as by physiological and other factors (see Kent, 2000). Research suggests that the earliest stages of speech development in monolingual and bilingual speakers are highly similar regardless of language environment (Buhr, 1980; Davis & MacNeilage, 1995; Gonzalez, 1983; Kent, 1992; Locke & Pearson, 1992; MacNeilage & Davis, 1990; Oller & Eilers, 1982; Poulin-Dubois & Goodz, 2001; Thevenin, Eilers, Oller, & Lavoie, 1985; Zlatic, MacNeilage, Matyear, & Davis, 1997). Babbling includes stops, nasals, and glides at the labial and coronal places of articulation (Davis & MacNeilage, 1995; Kent & Bauer, 1985; Locke, 1983; Oller, Eilers, Urbano, & Cobo-Lewis, 1997), nonrounded vowels (Davis & MacNeilage, 1990; Kent & Bauer, 1985; Levitt & Aydelott-Utman, 1992), and simple CV and CVCV syllable shapes (Boysson-Bardies, Sagart, & Bacri, 1981; Buhr, 1980; Oller & Eilers, 1982; Vihman, Ferguson, & Elbert, 1986).

Linguistic Period

In the first linguistic stage, from 12 to 18 months, babbling decreases and word production increases. While slight differences in frequencies of sounds and word shapes are reported cross-linguistically (Boysson-Bardies & Vihman, 1991; Maneva & Genesee, 2002), the considerable cross-linguistic similarities observed in babbling also exist in first words. Children from various language environments mainly produce coronal and labial stops, nasals, and glides, and simple CV syllable shapes in their first words (Boysson-Bardies & Vihman, 1991; Eilers, Oller, & Benito-García, 1984; Gildersleeve-Neumann, 2001; Goldstein & Cintrón, 2001; Oller, Wieman, Doyle, & Ross, 1976; Teixeira & Davis, 2002; Vihman et al., 1986). In addition, limited research on English-learning infants and infants in other monolingual language environments suggests that low front, non-rounded vowels are most frequent in first words (Davis & MacNeilage, 1990; Gildersleeve-Neumann, 2001; Levitt & Aydelott-Utman, 1992; So & Dodd, 1995; Stoel-Gammon & Dunn, 1985; Teixeira & Davis, 2002). Research on sounds in the first words of simultaneous bilinguals is extremely limited; however, it appears that similar consonants (Keshavarz & Ingram, 2002) and word shapes (Kehoe & Lleo, 2003) predominate. Information is not currently available on possible cross-dialectal differences.

Children's early speech patterns include phonotactic errors such as reduplication (e.g., “wawa” for “water”), consonant harmony (e.g., “goggie” for “doggie”), and final consonant deletion (e.g., “da” for “dog”) during their first 12–18 months of word production; these error patterns typically are markedly diminished by 3 years of age in children who are typically developing, although not, as reviewed later, in children suspected to have CAS. Apparent regression, in which the child produces a word less accurately but also with less variability than before, may also occur during the first year of word production as children systematize their phonologies (Vihman & Velleman, 1989). Individual sounds may be produced variably, even within the same word, although speech production patterns (i.e., frequent phonological processes) are consistent (Demuth, 2001; Ferguson & Farwell, 1975; Taelman & Gillis, 2002; Velleman & Vihman, 2002).

Between the ages of 2 and 3 years, the speech sound system of typically developing children expands in complexity, resulting in productions of a greater variety of consonants, vowels/diphthongs, and word shapes. By this age, children in English-learning environments begin to produce the more complex sounds—velars, fricatives, affricates, and liquids—generally mastering the majority of sounds with these features by approximately 5 years of age (Stoel-Gammon & Dunn, 1985). The few studies that have examined vowel and diphthong development suggest that accurate production of all vowels and most diphthongs (but not rhotic vowels) is achieved by age 3 (Bassi, 1983; Larkins, 1983; Pollock & Berni, 2003). In Pollock and Berni's study, the average percentage of vowels correct for children between 18 and 23 months was 82%, increasing to 92% for 24- to 29-month-olds, 94% for 30- to 35-month-olds, and 97% by 36 months of age. As subsequently discussed, the picture is very different for children suspected to have CAS. For typically developing children, more complex word shapes become frequent during this early period, with many consonant clusters, final consonants, and unstressed syllables correctly produced, resulting in a large increase in accuracy and intelligibility (Stoel-Gammon & Dunn, 1985). Consonant clusters emerge by the first third of the fourth year (36–40 months), usually appearing first in final position in speakers of Mainstream American English (Kirk & Demuth, 2003). Typically developing children are reportedly 26%–50% intelligible by 2 years, 71%–80% intelligible by 3 years, and 100% intelligible by 4 years (Coplan & Gleason, 1988; Weiss, 1982). It is also after age 2 that the diverse effects of a child's ambient language become most apparent (see below; Goldstein & Washington, 2001; Johnson & Wilson, 2002; Walters, 2000).

In typical and most atypical, nonapraxic speech during this age period, earlier developing sounds tend to be substituted for later developing sounds that the child may not be able to produce as easily (e.g., stops substitute for fricatives and glides substitute for liquids). Children with nonapraxic speech sound disorders appear to be most successful at producing the correct voicing features of a segment and least successful at maintaining the correct place of articulation (Forrest & Morrisette, 1999). Although perceptual and articulatory constraints are the primary posited source of English-learning children's difficulty with affricates, fricatives, and liquids, the frequency of sounds in a particular environment also plays an important role in the age and order of phoneme mastery. Children from language environments with a greater frequency of occurrence of certain less universally common sounds (e.g., liquids, fricatives) tend to produce these sounds earlier and better, suggesting the early influence of the ambient language. For example, Russian children who are exposed to many palatalized consonants as well as non-palatalized consonants typically master the palatalized ones first (Zharkova, 2004). Other research in non-English monolingual language environments has shown ambient language effects on the greater earlier accuracy of fricatives and affricates (Pye, Ingram, & List, 1987) as well as dorsal sounds and multisyllabic words (Gildersleeve-Neumann, 2001; Teixeira & Davis, 2002). In addition, children in other language environments may produce words with different error patterns. For instance, it is common for young Finnish children to have initial consonant deletions, an atypical phonological process in English-speech acquisition (Vihman & Velleman, 2000).

Ambient language effects on speech sound acquisition are also observed in bilingual children. Bilingual children may show an effect of each language on their productions within that language, such as reported for a Hindi-English simultaneous bilingual child who used predominantly monosyllables in English and predominantly disyllables in Hindi (Bhaya Nair, 1991; Vihman & Croft, in press).

Simultaneously bilingual children produce different segments and word shapes depending on which of their two language environments they are in (Holm & Dodd, 1999; Holm, Dodd, Stow, & Pert, 1999; Johnson & Lancaster, 1998; Kehoe & Lleo, 2003; Keshavarz & Ingram, 2002). Mixing of errors has been observed in the carryover of the phonetic and phonological properties of one language to the other, resulting in greater rates of error when compared to monolingual peers (Goldstein & Cintrón, 2001). Although bilingual children may follow the general developmental path, their speech patterns might still be expected to be influenced by the phonology/ies of their native language(s). Clearly, the large, cross-linguistic literature on typical and atypical speech sound acquisition provides a rich database for comparative research on speech development in children suspected to have CAS.

Prelinguistic Period

Infants' early discrimination of prosody (see Speech Perception) is followed by production of language-specific prosodic patterns. By 6–12 months, their vocalization patterns reflect the dominant prosodic contours (e.g., falling vs. rising pitch; Whalen, Levitt, & Wang, 1991) of the ambient language.

Linguistic Period

English-speaking children use falling intonation contours first, then rising contours, to mark phrase and utterance boundaries (Tonkava-Yampolskaya, 1973). Typical English-learning children have been shown to use frequency, amplitude, and duration appropriately to mark sentential emphasis (Skinder, Strand, & Mignerey, 1999), as do children with speech delays (Shriberg, Aram, & Kwiatkowski, 1997b, 1997c). The primary period for the development of prosody occurs from approximately 5 to 8 years of age (Local, 1980; Wells, Peppe, & Goulandris, 2004). However, even typically developing children may not have adultlike comprehension and production of prosody until 10 or 12 years of age (Allen & Hawkins, 1980; Morton & Trehub, 2001). In English, later-developing prosodic functions include the production of compound words, rise-fall or fall-rise prosody on a single word to convey emotion, high rising pitch to request clarification, accent on a nonfinal word to convey emphasis (e.g., “I want a black bus”), and the comprehension of another person's use of accent to emphasize a certain part of an utterance (Wells et al., 2004).

Children's stress patterns parallel the dominant stress patterns of their languages in late babbling and early words (e.g., predominantly trochaic stress-first patterns in English; iambic stress-last patterns in French; Vihman, DePaolis, & Davis, 1998). By 2½ years of age, English learners' vowel durations differ appropriately in stressed versus unstressed syllables (Smith, 1978), and they are able to produce weak syllables in initial position (e.g., the first syllable of “giraffe”) and between two stressed syllables (e.g., the second syllable of “telephone”; Gerken, 1994; Kehoe & Stoel-Gammon, 1997). Children with speech delay but not apraxia cease to delete such weak syllables by age 6 (Velleman & Shriberg, 1999); as reviewed later, children suspected to have CAS may persist in such patterns. By age 6, typically developing children have different coarticulatory and temporal patterns depending on the syllable structure of a word. For example, Dutch-learning children have coarticulation and duration patterns that differ with the metrical structure of the word (Maassen, Nijland, & van der Meulen, 2001; Nijland, Maassen, van der Meulen, et al., 2003).

Prelinguistic Period

Between birth and 2 months of age, human beings are already able to discriminate among languages with different rhythmic patterns (Mehler et al., 1988), among words that differ by number of syllables (Bijelac-Babic, Bertoncini, & Mehler, 1993), and among different vowels (Kuhl & Miller, 1975) and different consonants (Eilers, 1977; Eilers & Minifie, 1975; Jusczyk, Murray, & Bayly, 1979; Levitt, Jusczyk, Murray, & Carden, 1988). Some of these capacities may be innate, but others are learned through perceptual experience. For example, the neonate attends longer to her own mother's voice (DeCasper & Fifer, 1980) and to her own language prosody in conversational speech (Mehler et al., 1988). Speech perception skills become more and more language-specific as the child approaches 1 year of age. By 10 months, infants display preferences for stress patterns (Jusczyk, Cutler, & Redanz, 1993; Morgan, 1996; Weissenborn, Hohle, Bartels, Herold, & Hofmann, 2002), consonants, and sequences of consonants and vowels from their own language (Gerken & Zamuner, 2004; Jusczyk, Friederici, Wessels, Svenkerud, & Jusczyk, 1993; Jusczyk, Luce, & Charles-Luce, 1994). Furthermore, at 10–12 months, babies are less able than at earlier ages to discriminate segmental contrasts that are not relevant to their own languages (Werker & Tees, 1984).

Linguistic Period

At 4 years of age, children with nonapraxic speech disorders are significantly worse than their typically speaking peers at discriminating commonly misarticulated sounds from sounds that are generally substituted for them. There is a significant difference between the two groups' ability to identify whether a sound was produced correctly versus incorrectly within a word (e.g., [tæt] vs. [kæt] for “cat”) (Rvachew, Ohberg, Grawberg, & Heyding, 2003).

Nonspeech Motor Behaviors

Nonspeech motor behaviors are primarily used to differentiate children suspected to have CAS from children with various types of dysarthria, although there is some overlap between the two motor speech disorders. Nonspeech motor signs of CAS that are most commonly proposed in the literature (some of which are also cited as signs of dysarthria) include the following: general awkwardness or clumsiness, impaired volitional oral movements, mild delays in motor development, mildly low muscle tone, abnormal orosensory perception (hyper- or hyposensitivity in the oral area), and oral apraxia (e.g., Davis et al., 1998; McCabe et al., 1998; Shriberg et al., 1997a). The nonspeech motor features typically listed for oral apraxia are impaired volitional oral movements (imitated or elicited postures or sequences such as “smile-kiss”) and groping (e.g., Davis et al., 1998; McCabe et al., 1998; Shriberg et al., 1997a). Murdoch, Attard, Ozanne, and Stokes (1995) documented weaker lingual muscles and reduced tongue endurance in children who demonstrated oral apraxia than in typically developing children; a nonapraxic, phonologically disordered control group was not included in the study. Dewey, Roy, Square-Storer, and Hayden (1988) found that limb, oral, and verbal apraxia tend to co-occur in children. They highlighted the transition difficulties (moving from one action in a sequence to the next) exhibited by children “with a specific deficit in verbal sequences of consonantvowel syllables” (p. 743) and noted that repetition of the same action was far less of a problem. They also stressed the volitional aspect of the disorder, as did Maassen, Groenen, and Crul (2003) and D. Nelson (1995). Specifically, Dewey et al. (1988) found that demonstrating the action of an object was a problem for their participants with CAS only when the children were miming the action without the object in hand. Crary and Anderson (1991) also noted that compared to children without a diagnosis of CAS, children with this diagnosis had slower rates and less accurate performance on sequences of hand and facial movements.

Speech Motor Behaviors

Motoric aspects of speech, especially repetitions of syllables (maximum repetition rate [MRR]) and productions of alternating syllables (diadochokinesis [DDK] or alternating motion rate [AMR]), are commonly used to diagnose CAS both clinically and for research participant selection. The utility of these measures has been verified in several research studies, including Davis et al. (1998), McCabe et al. (1998), Nijland et al. (2002), Thoonen, Maassen, Gabreëls, and Schreuder (1999), and Thoonen, Maassen, Wit, Gabreëls, and Schreuder (1996). Thoonen et al. (1996), for example, reported that maximum sound prolongation of vowels (e.g., producing /a/ for as long as possible) and MRRs for single syllables (e.g., /pΛtΛkΛ/ etc.) differentiated children with a diagnosis of spastic dysarthria from both children with a CAS diagnosis and those who were typically developing. Maximum sound prolongation of fricatives and maximum repetition rate of trisyllabic sequences (/pΛtΛkΛ/) differentiated children with apraxia from those who were typically developing. Thus, the differences between children with CAS and those who were typically developing were only significant for the more complex tasks (prolongations of more difficult consonant sounds; sequences of different syllables). Control groups of children with other speech sound disorders of unknown origin were not tested. Lewis et al. (2004) found significant differences between preschool and school-age children with CAS and matched children with non-CAS speech delay in their ability to repeat nonwords and multisyllabic words, with the CAS group performing more poorly. Children with CAS also had significantly lower Total Function scores on the Robbins and Klee (1987) oral-motor assessment, which includes DDK. Moreover, children with CAS had more difficulty on the Fletcher Time-by-Count test of DDK (S. G. Fletcher, 1978) at school age.

Lists of the speech behaviors proposed to characterize CAS abound in the research and clinical literatures. Frequent characteristics include some features that clearly are shared with other speech sound disorders (McCabe et al., 1998), including slow development of speech, reduced phonetic or phonemic inventories, multiple speech sound errors, reduced percentage of consonants correct, and unintelligibility. Commonly proposed characteristics (Davis et al., 1998; McCabe et al., 1998; Shriberg et al., 1997a) that are less likely to be found in children with nonapraxic speech sound disorders include reduced vowel inventory, vowel errors, inconsistency of errors, increased errors in longer or more complex syllable and word shapes (especially omissions, particularly in word-initial position), groping, unusual errors that “defy process analysis,” persistent or frequent regression (e.g., loss of words or sounds that were previously mastered), differences in performance of automatic (overlearned) versus volitional (spontaneous or elicited) activities, with volitional activities more affected, and errors in the ordering of sounds (migration and metathesis), syllables, morphemes, or even words. However, many of these features are found in children who do not fit the overall pattern of CAS (McCabe et al., 1998), leading some reviewers to question their diagnostic specificity for CAS (e.g., Macaluso-Haynes, 1978). Moreover, as discussed later, many of these posited features are not consistent with a deficit in praxis. For example, motor speech theories typically assign selection and sequencing of sounds, syllables, and words to a processing stage that precedes the planning and programming of movements needed to realize these units as manifest speech. Error patterns that are not consistent with a praxis deficit but are especially common in children suspected to have CAS need to be studied to understand whether or not they are causally related and, if they are, to identify the explanatory mechanisms.

Detailed studies of differences between children suspected to have CAS compared to those with typical development or with other subtypes of speech delay have sought to identify the diagnostic characteristics of CAS. As noted previously, all such studies have research design limitations due to the lack of certainty that the children suspected to have CAS indeed have this disorder. As just one of many examples, Maassen et al. (2001) reported that children with CAS have less predictable speech errors than children who are typically developing. These authors provided useful acoustic data documenting a lack of systematic effects of given phonetic contexts on certain sounds in the speech production of children with CAS. However, because of the absence of a control group of children with other speech sound disorders and because the only inclusionary criterion information provided was that “Clear cases of [CAS] were selected according to clinical criteria described by Hall, Jordan, and Robin (1993) and Thoonen et al. (1996)” (Maassen et al., 2001, p. 146), it is difficult to evaluate claims that variability of this type may be a unique feature of CAS.

Speech sampling methods may also be crucial to interpretation of findings. Shriberg et al. (1997b) reported that a group of children, chosen by five individual researchers as exemplars of these researchers' diagnosis of CAS, did not have any speech production errors in conversational speech that could be used to differentiate them from control children with speech delay of unknown origin. A potential constraint on these findings, as discussed more recently in Shriberg, Campbell, et al. (2003), is that these findings were based on conversational speech samples, rather than on children's responses to challenging speech production tasks designed to evoke more discriminative error patterns.

In a widely cited study of speech motor behaviors, McCabe et al. (1998) attempted to identify potential features of apraxia retrospectively (from clinic files) in a mixed group of 50 children with speech disorders, 9 of whom had been identified as having apraxia of speech by their speech-language pathologists. They described characteristics of CAS in some of the 50 children who had been classified as speech disordered (non-CAS). The characteristics most often identified in the total group were “changed level of awareness of own speech errors, problems with imitation of speech, breathing difficulties/asthma/allergies, decreased performance on DDK tasks, and presence of ‘soft’ neurological signs or minimal brain damage” (McCabe et al., 1998, p. 113). These were also the most commonly reported symptoms in the subset of children who previously had been identified by their speech-language pathologists as having apraxia. However, McCabe et al. reported “inconsistent speech performance,” vowel errors, and incorrect production of “lingual phonemes” (/l/, /r/) as best differentiating this CAS group from their other participants. Other differences that distinguished the two groups quantitatively included slow development of speech, idiosyncratic sound substitutions, and syllable omissions.

Lewis et al. (2004) compared a group of children suspected to have CAS to two other groups of children longitudinally: one group with non-CAS speech sound disorders only and one group with both speech and language disorders. The CAS group was selected based on both a diagnosis of CAS by the child's speech-language pathologist and on the child meeting at least four out of eight criteria for CAS. The CAS group differed from the speech disorder group, especially at school age, on syllable structures, sound sequencing, vowel and voicing errors, unusual types of errors, and the persistence of their error patterns. At school age, the children with CAS had more speech errors overall, more unusual errors, and more syllable sequencing errors in conversational speech than the children with both speech and language disorders (but see McNeil, Robin, & Schmidt, 1997, for an alternative interpretation of phoneme-level sequencing errors in AOS).

Ball, Bernthal, and Beukelman (2002) used a very careful procedure to identify participants with CAS, including diagnosis by a speech-language pathologist and administration of the Screening Test for Developmental Apraxia of Speech (Blakeley, 1980) and the Tasks for Assessing Motor Speech Programming Capacity (Wertz, LaPointe, & Rosenbeck, 1984). A panel of three speech-language pathologists then rated each child on a scale of 1 (not CAS) to 5 (definitely CAS) based on a list of 17 characteristics of CAS. The 36 children included in the study each had an average rating of at least 3. They also had other co-occurring language, social, and behavioral impairments. The purpose of the study was to attempt to identify more inclusive communication profiles of children with CAS. Following this identification procedure, an assessment battery of tests and measures was administered to the participants, and test results were subjected to cluster analysis to identify groups of participants who shared particular patterns of communication performance. Twelve of the participants who had been rated as having a high likelihood/severity of CAS had notable deficits in the following areas compared to participants in the other clusters: receptive language, vocabulary, MLU, percentage of consonants correct, intelligibility, and behavior.

Acoustic analyses have been used by several authors to characterize more precisely the speech production differences of children with CAS. Participants with CAS in these studies have demonstrated decreased differentiation of stop place of articulation (Sussman, Marquardt, & Doyle, 2000), decreased differentiation of vowels (Nijland et al., 2002), higher degrees of anticipatory coarticulation within syllables (Maassen et al., 2001; Nijland, Maassen, van der Meulen, et al., 2003), lack of impact of syllable boundaries or syllable shape on coarticulation (Maassen et al., 2001; Nijland, Maassen, van der Meulen, et al., 2003), lack of intersyllabic coarticulation, and variable idiosyncratic patterns (Nijland et al., 2002) that were less predictable acoustically in any given phonetic context (Maassen et al., 2001). Nijland, Maassen, van der Meulen, et al. (2003) further noted that children with CAS had higher scores than typically developing children on measures of coarticulation and vowel accuracy when a bite block was placed between their teeth. As noted previously, additional studies using control groups of children with other forms of speech delay would strengthen the claims of this carefully executed study series.

With respect to severity, McCabe et al. (1998) found that severity of speech impairment, as defined by the percentage of consonants correct, is correlated with the number of features of CAS that a child exhibits even among children without this diagnosis and that CAS may be quantified on a continuum of severity as measured in this way. Relative to prognosis, Lewis et al. (2004) reported that, at school age, participants with CAS had more persistent difficulties in repeating nonsense words and sequencing syllables than participants who had previously been diagnosed with a non-CAS speech sound disorder.

Prosodic Characteristics

A consistent finding in the literature is that individuals suspected to have CAS have atypical prosody, including a variety of types of prosodic deficits (Davis et al., 1998; McCabe et al., 1998; Shriberg et al., 1997a). Also often noted are variations in rate, including both prolonged sounds and prolonged pauses between sounds, syllables, or words, which gives the listener the impression of staccato speech (syllable segregation), with sounds, syllables, or words produced as independent entities lacking smooth transitions to other structural units (Shriberg, Green, Campbell, McSweeny, & Scheer, 2003). As in other motor speech disorders, reduced range of or variable pitch, as well as reduced range of or variable loudness, gives the listener the impression of monotone, monoloud speech, respectively. Variable nasal resonance (sometimes hyponasal, sometimes hypernasal) has also been noted in the clinical research literature. Duration, pitch, and loudness combine to form the percept of stress in English; this, too, is commonly reported to be atypical in children suspected to have CAS. In a series of studies, Shriberg et al. (1997a, 1997b, 1997c) documented excessive-equal stress (all or most syllables in a word or sentence receiving prominent stress) in approximately 50% of each of three different samples of children suspected to have CAS. They noted that younger children with CAS were also rated as more involved than children with speech delay on perceptual measures of rate and resonance. However, excessive-equal stress was the only feature that reliably distinguished any of several CAS subgroups from control groups of children with speech delay of unknown etiology. Those children who exhibited excessive-equal stress also produced more distortions of early consonant sounds than the other children, but their error types (relative proportions of substitutions vs. omissions vs. distortions) and their severity and variability levels did not differ from those of the children with speech delay. The authors speculated that the children with a diagnosis of CAS who did not demonstrate excessive-equal stress may either have been incorrectly diagnosed or were possibly exhibiting another type of CAS. In a later article, Shriberg, Campbell, et al. (2003) suggested that the presence of stress errors may change over time within an individual with CAS. Odell and Shriberg (2001) further noted that prosodic disturbances may be different in adults with acquired apraxia of speech versus children with CAS. Children with CAS in their sample had excessive-equal stress, but, in contrast to the sample of adults with acquired apraxia of speech, did not have inappropriate phrasing or rate.

Velleman and Shriberg (1999) completed metrical analyses of the lexical stress patterns of children with CAS who had inappropriate stress, children with CAS who did not have inappropriate stress, and children with speech delays of unknown origin. They found that the pattern of stress errors of the children with CAS did not differ substantially from the error pattern of younger, typically developing children, suggesting that either the children with CAS were misdiagnosed or that such errors reflect prosodic delay rather than disorder. That is, participants with CAS who had inappropriate stress tended to either omit or overstress weak (unstressed) syllables, especially in the initial position of words, as do typically developing 2-year-old children. However, whereas the children with speech delay ceased to make such errors after the age of 6, lexical stress errors of this type had persisted into adolescence in the participants suspected to have CAS with inappropriate stress.

Stress differences in CAS have also been examined using acoustic analyses. Munson, Bjorum, and Windsor (2003) reported that the vowel durations, fundamental frequencies, vowel intensities and f0 peak timing of stressed syllables produced by children with CAS were appropriate despite the fact that the children were perceived as producing inappropriate stress patterns. Skinder et al. (1999) also found that children with CAS marked stress in the same ways as children who were typically developing, although there was more variability within the CAS group. Again, listeners had judged the children with CAS as less accurate in their stress production than the typically developing children, but the acoustic measures used did not identify the source of these perceptions. Skinder et al. suggested that listeners were confused or distracted from attending to prosodic details by the higher number of segmental errors produced by the children with CAS. Shriberg, Campbell, et al. (2003), in contrast, reported that ratios based on acoustic measures of stressed versus unstressed syllables differed in children with CAS who perceptually were noted to produce excessive-equal stress compared to control children with speech delay. The acoustic differences were quantitative rather than qualitative. Thus, it may not be that children with CAS have uniquely different stress patterns. Rather, it may be their inability to fully contrast stressed versus unstressed syllables that leads to the impression of inappropriate stress patterns. Note that these and associated stress findings are consistent with findings reviewed earlier indicating that the phonetic distinctiveness of vowels/diphthongs and consonants is reduced in the speech of children suspected to have CAS.

Speech Perception Characteristics

A few studies have addressed the hypothesis that children suspected to have CAS have deficits in auditory perception, auditory discrimination, and/or auditory memory. Bridgeman and Snowling (1988) reported that compared to control children, children with CAS have more difficulty discriminating sound sequences in nonsense words. Groenen and Maassen (1996) found that children with CAS did not have difficulty identifying the place of articulation of a consonant but did have difficulty discriminating consonants with subtle acoustic differences associated with place of articulation. Furthermore, deficits in place discrimination were found to be correlated with deficits in accurate production of place. Maassen et al. (2003) also reported that compared to children with typically developing speech, children with CAS had poorer identification as well as poorer discrimination of vowels. Given the likelihood of phoneme-specific relationships between production and perception in children with other speech sound disorders (Rvachew, Rafaat, & Martin, 1999), an important research goal for theories of CAS is to determine if the speech perception deficits described above are replicable and whether they are unique to children with CAS.

Language Characteristics

There is general agreement in reviews of the literature that children suspected to have CAS typically also have significant language deficits (e.g., Crary 1984, 1993; Ozanne, 1995; Velleman & Strand, 1994). As with the perceptual findings reviewed in the previous section, a research challenge is to determine how such constraints are associated with the praxis deficit in planning and programming that defines CAS. One possibility is that language impairments are a consequence of having any type of disorder affecting neurological development (Robin, 1992). In response to another possiblity—that all expressive language deficits in children with CAS are due to their speech involvements—Ekelman and Aram (1983) documented language errors in a group of children with CAS that were clearly not due to the children's phonological deficits. Their participants made incorrect choices of pronouns and verbs. They also failed to invert auxiliary (helping) and copula (be) verbs in questions. More recently, Lewis et al. (2004) found that language impairments were more significant and persistent in children with CAS than in children with non-CAS speech sound disorders. The authors concluded that language symptoms are “a key aspect of the disorder” (p. 131) based on the following observations: (1) gains in articulation did not eliminate language deficits (e.g., morphological omissions of plural, possessive, third person singular, and past tense markers are not simply due to an inability to produce final consonant clusters); (2) receptive as well as expressive language deficits were noted, although expressive language consistently lagged behind receptive language; and (3) there was a strong family history of language impairment in the families of the children with CAS (Lewis et al., 2004).

Language symptoms that might differentiate children suspected to have CAS from children with SLI have been implied in the literature. For example, Velleman and Strand (1994) modeled CAS as a disorder of hierarchical organization, which suggests that language errors should take the form of part-whole and sequencing difficulties. Lewis et al. (2004) failed to find language differences between children with CAS and children with a combined language and non-CAS speech sound disorder on standardized tests at the preschool level. As indicated above, they did identify more persistent receptive and expressive language difficulties among the children in the CAS group at school age; analyses of the children's spontaneous spoken and/or written language would have strengthened this claim. Overall, the literature remains inconclusive on whether there are differences in the language profiles of children with CAS versus children with SLI or with combined language and nonapraxic speech sound disorder.

Metalinguistic/Literacy Characteristics

Children with any sort of speech production deficit are at higher risk for difficulty with phonological awareness, which itself is a “critical element of literacy development” (Justice & Schuele, 2004, p. 378). Although CAS researchers frequently cite literacy and other academic difficulties as a characteristic of the disorder, few studies have explored this topic and some have been limited by the lack of a speech delayed comparison group. For example, Marion, Sussman, and Marquardt (1993) demonstrated that children with CAS have more difficulty perceiving and producing rhymes than do children with typically developing speech. Marquardt et al. (2002) similarly showed that children with CAS score lower than typically developing children on metaphonological (phonological awareness) tasks, such as tapping to count the syllables in a word and using blocks to represent the structure of a word (e.g., using black blocks to represent the consonants and white blocks to represent the vowels in the word blue [i.e., black black white]). Given that a history of speech delay puts a child at increased risk for phonological awareness deficits, it will be important to cross-validate such interesting findings with control groups who have speech sound disorders other than CAS.

Lewis et al. (2004), as cited previously, found that children with CAS had deficits in word attack, word identification, and spelling in comparison to children with speech disorders only. Their participants with CAS also scored significantly lower on tasks requiring them to spell unpredictable words, compared to scores from children with a combination of language and nonapraxic speech disorders.

Finally, it is useful to note that children suspected to have CAS have sometimes been described as having increased self-awareness of their own speech production limitations (McCabe et al., 1998; Velleman & Strand, 1994). This is a special type of metalinguistic awareness (the ability to reflect consciously about or comment on linguistic elements, structures, or processes) that, to date, has not been addressed in controlled research.

CAS research to date has almost exclusively focused on English-speaking participants in several countries, with the exception of several cohorts of Dutch children with CAS studied by Maassen and colleagues (Maassen et al., 2001, 2003; Nijland, Maassen, & van der Meulen, 2003; Thoonen et al., 1999). Although not exploring cross-linguistic similarities or differences between individuals with CAS who speak Dutch or English, these investigators have used CAS criteria from studies of English-speaking participants. Dutch and English are similar in phonetic and phonotactic properties and it appears that features of CAS may be similar in the two language environments. The Committee did not identify any studies that have compared aspects of CAS in individuals speaking different dialects of English or speaking languages that differ markedly from English in phonetic, phonemic, and phonotactic properties. Cross-linguistic studies of CAS could provide greater understanding of the effects of language and culture on its short- and long-term expression.

Suprasegmental Perspectives

There appears to be widespread agreement that syllables and prosody are affected in more profound, distinctive ways in CAS than are other aspects of speech or phonology. Some researchers have hypothesized that deficits in the syllabic framework of speech result in prosodic symptoms (Davis et al., 1998; Maassen, 2002; Marquardt et al., 2002; Nijland, Maassen, van der Meulen, et al., 2003). Others have proposed the reverse: that fundamental prosodic deficits affect syllable and segment production (Boutsen & Christman, 2002; Odell & Shriberg, 2001). Other researchers have emphasized the critical roles of timing (e.g., Shriberg, Green, et al., 2003) and sequencing deficits (e.g., Thoonen et al., 1996) as core features underlying many of the other segmental and suprasegmental characteristics of CAS.

Sensorimotor Perspectives

Several theoretical frameworks for CAS posit that core deficits are in the relationship between perception or sensory processing and some aspect of motor processing. Maassen (2002), for example, proposed that deficient sensorimotor learning leads to weak prelinguistic articulatory-auditory mappings, which in turn fail to support full phoneme-specific mappings. He noted that “higher-level knowledge…must be acquired by the child via the problematic speech production and perception skills” (p. 265). Maassen suggested that unlike typically developing children, children with CAS seem to process real words more similarly to the way they process nonsense words. Maassen speculated that such processing renders their linguistic systems (e.g., lexical representations) less able to support online language processing tasks. Barry (1995a), Boutsen and Christman (2002), and Odell and Shriberg (2001) focused on the related issue of online self-monitoring and feedback systems. These investigators proposed that children with CAS may have weak sensorimotor feedback loops or decreased ability to respond to such feedback. Thus, children with CAS may be unable to either benefit immediately from feedback in order to self-correct or to appropriately grade actions, or may be unable to use this feedback to alter incomplete representations or motor plans for future retrieval. Such a sensorimotor deficit could also underlie proposed difficulties in automating motor programs, such that each word production must be planned anew (Barry, 1995a; Nijland et al., 2002; Nijland, Maassen, van der Meulen, et al., 2003).

Deficits in the preprogramming, programming, and execution of speech motor events (Klapp, 1995, 2003) have each been proposed as a core deficit in CAS. Unfortunately, explicit definitions for each of these processes are, themselves, a source of debate in associated literatures. Most theoretical proposals place the source of the speech production difficulties in CAS further “upstream” than the actual execution of the motor plan. Marquardt, Jacks, and Davis (2004), for example, attributed high inconsistency levels in children with CAS to “lack of neural instantiation of phonemic representations” (p. 142) and unstable motor programs for word targets. They noted that increases in accuracy are associated with increases in stability (i.e., decreases in inconsistency), presumably reflecting more specific, stable motor plans for words. A further common theme in all such discussions is a deficit in integration or coordination across different levels proposed to be relevant to speech production (and, in some cases, speech perception as well). Such levels include syllabic, phonemic, or motor representations; motor plans and/or programs; and neuromotor group networks. Thus, multiple levels of speech motor processing, and the relationships among them, have been implicated in processing perspectives on CAS.

In prior decades, discussion of the core deficit(s) in CAS was often framed as a debate between linguistic/psycholinguistic perspectives versus motor perspectives. Currently, this opposition is more appropriately described as a debate between motor + linguistic versus motor-only views. As discussed previously, the primary question is how to reconcile the linguistic behaviors that have been associated with CAS in the research literature—differences in speech perception, phonological awareness, phonological patterns, and expressive language—with the core problem of praxis from which this disorder takes its name. In widely cited papers on AOS, McNeil and colleagues have argued on formal grounds that such deficits cannot be accommodated as core features of AOS (McNeil, 1997; McNeil et al., 1997). Rather, if present, they likely reflect secondary consequences of apraxia of speech. Thus, from a theoretical perspective, the lines are drawn fairly sharply. Following McNeil's rationale, if research validates deficits in speech processes that precede planning/programming of movement sequences for speech, reconsideration would have to be given to the appropriateness of the term apraxia for this clinical entity.

Descriptive-Linguistic Findings

Despite the wide-ranging, cross-disciplinary impact of research on the KE family, there are few published descriptions of the segmental and supra-segmental error profiles of affected individuals. Hurst, Baraitser, Auger, Graham, and Norell (1990), the first paper by the U.K. (London and Oxford) researchers, included brief case summary paragraphs for 6 family members. These reports primarily described the speakers' apparent impairment in the organization of manual movements for signing and speech movements, with the clinical speech profile of affected family members interpreted by the researchers as consistent with CAS (“developmental verbal dyspraxia”; p. 352).

The Hurst et al. (1990) report was followed by a series of papers by Canadian researchers (primarily at McGill University) providing descriptive-linguistic analyses of selected affected KE family members. Using a variety of corpora, they interpreted their findings to suggest that the core deficit in these individuals was in their grammatical morphology (Gopnik, 1990a, 1990b; Gopnik & Crago, 1991; Matthews, 1994). More relevant for the present focus, Fee (1995) provided a perceptually based comprehensive description of the consonant errors of 8 affected family members sampled at two points in time. She reported that, even as adults, these speakers had deletion and substitution errors, especially on final consonants and consonant clusters. Goad (1998) provided a thorough analysis of the grammatical impairment in plurals in 5 affected adult family members, focusing on alternative theoretical explanations in prosodic versus morphological domains. Also, in a study assessing knowledge of lexical stress rules, Piggott and Kessler Robb (1999) reported that affected family members had considerably more incorrect and variable judgments of what constitutes appropriate lexical stress than unaffected family members. Although the investigators in this group did not use the term apraxia or dyspraxia, they reported that affected family members produced polysyllabic words that had “prominent pauses separating them” and that were “evenly stressed” (p. 61). Thus, prosodic impairment has been described as a key feature shared among affected family members.

Genetic Findings

As indicated in Table 1, the molecular genetic findings for the KE family began with the report by Fisher, Vargha-Khadem, Watkins, Monaco, and Pembrey (1998), which identified a region on chromosome 7 in affected family members that was subsequently narrowed to a susceptibility locus (i.e., a region of increased risk) on 7q31 termed SPCH1. This finding was the bridge between the earlier descriptive-linguistic characteristics summarized above and later identification of the FOXP2 gene within the SPCH1 region. Two studies from the London/Oxford research group (Lai et al., 2000; Lai, Fisher, Hurst, Vargha-Khadem, & Monaco, 2001) provided information on this transcription gene, FOXP2 (see Table 1). As a transcription gene, the protein products of FOXP2 influence the function of other genes, which in turn reportedly may affect both linguistic and sensorimotor aspects of speech and language acquisition. The U.K. group is conducting two large-scale projects to identify all genes “downstream” of FOXP2 to determine how products of these genes may contribute to speech-language acquisition and disorder (cf. Marcus & Fisher, 2003).

A number of studies (e.g., Liégeois et al., 2001; MacDermot et al., 2005; Tyson, McGillivary, Chijiwa, & Rajcan-Separovic, 2004; Zeesman et al., 2006) have supported the association of FOXP2 with apraxia of speech, as well as with a variety of other deficits first observed in affected members of the KE family. In addition, more recent research related to FOXP2 and the KE family has also suggested possible mechanisms by which CAS and dysarthria may co-occur. Morgan, Liégeois, Vogel, Connelly, and Vargha-Khadem (2005) used FMRI and electropalatography (EPG) to study 5 affected members of the KE family and 5 sex-, age-, and handedness-matched controls. In addition to brain abnormalities reported previously in the motor cortex, the EPG data were reportedly consistent with speech sound distortions, with excessive variability in lingual-palatal contacts. Morgan and colleagues suggested that the FOXP2 mutation in the KE family has disrupted the development and function of the brain regions involved in both planning and execution of speech movements (i.e., the latter process consistent with dysarthria). Finally, Shriberg et al. (2006) described a mother and a 19-year-old daughter who from early ages were treated for apraxia of speech associated with a balanced 7;13 translocation affecting FOXP2. Detailed speech and prosody analyses indicated that the mother's and daughter's speech profiles are consistent with both apraxia of speech and spastic dysarthria.

As indicated in the introduction to this section, findings from the KE family have prompted widespread interdisciplinary interest in the FOX family of genes. Examples at the time this report was in preparation include studies tracing the evolutionary history of FOXP2 (e.g., Enard et al., 2002) and studies describing transcription processes and other molecular features of FOXP2 and the larger family of FOX genes (e.g., Bruce & Margolis, 2002; Ferland, Cherry, Preware, Morrisey, & Walsh, 2003; Takahashi, Liu, Hirokawa, & Takahasi, 2003; Tamura, Morikawa, Iwanishi, Hisaoka, & Senba, 2003; B. Wang, Lin, Li, & Tucker, 2003; Zhang, Webb, & Podlaha, 2002). A major finding for definitional issues in CAS is that, as described in the studies cited immediately above, both FOXP1 and FOXP2 genes are expressed (switched on) widely in the brain. Importantly, these locations include many of the primary neuroanatomic sites that subserve speech-language development and processing. Additional work focusing on these genes and their counterparts in animals has suggested the potential to develop animal models for speech-language disorders; for example, Teramtizu, Kudo, London, Geschwind, and White (2004) in songbirds and Shu et al. (2005) in mice.

A number of studies in the emerging discipline of linguistic genetics (genetic studies of verbal traits and disorders) have sought to determine whether deficits in FOXP2 are linked to other neurobehavioral disorders, including language impairment, dyslexia, and autism. As indicated previously, with the exception of a finding in autism (Gong et al., 2004), FOXP2 deficits have not been found in children with other neurobehavioral disorders (see OMIM for current findings).

Neuropsychological Findings

Following their initial rejoinders to the Canadian group's interpretation of the communicative deficit in affected KE members (P. Fletcher, 1990; Vargha-Khadem & Passingham, 1990), the U.K. research group (i.e., Hurst et al., 1990) published four descriptive papers. The papers described findings from an extensive battery of neuropsychological and other measures given to family members and a number of control groups, including information on affected individuals' articulatory and prosodic involvements. As shown in Table 1, Vargha-Khadem, Watkins, Alcock, Fletcher, and Passingham (1995) reported that the orofacial apraxia used as the phenotype for affected family members was accompanied by an array of deficits in other verbal and nonverbal domains, involving both comprehension and production. In particular, they summarized their alternative descriptive-explanatory perspective on the KE family as suggesting “a broad phenotype which transcends impaired generation of syntactical rules and includes a striking articulatory impairment as well as defects in intellectual, linguistic, and orofacial praxic functions generally” (Vargha-Khadem et al., 1995, p. 930).

Among other findings in two subsequent papers, Alcock and colleagues (Alcock, Passingham, Watkins, & Vargha-Khadem, 2000a, 2000b) reported that affected family members' deficits involve both comprehension and production of rhythms, as assessed using verbal and nonverbal (i.e., tapping) modalities. These authors speculated that a core problem in timing may underlie the performance deficits of affected family members on the diverse comprehension and production tasks included in the protocol.

The fourth paper (Watkins, Dronkers, & Vargha-Khadem, 2002) provided extensive neuropsychological information on cognitive-linguistic involvements in affected family members, who were compared to a sample of adults with aphasia. These findings are of particular interest for issues addressing similarities and differences in acquired adult AOS and CAS, as well as suggesting neostriatal (basal ganglion) involvement in praxic deficits in movement sequencing and procedural learning. If such findings are replicated in studies of other individuals with FOXP2 or other genetic deficiencies, they will have important implications for assessment and treatment of CAS.

Neuroimaging Findings

As sampled in Table 1, a series of neuroimaging findings in affected members of the KE family has provided unprecedented information on neuroanatomic structures and circuits associated with this subtype of orofacial apraxia and apraxia of speech. Findings have possible implications for the developmental neurobiology of CAS symptoms and, more generally, speech-language acquisition.

In the first neuroimaging study of the KE family, Vargha-Khadem et al. (1998) reported that affected KE family members have diverse, bilateral neuroanatomic differences from unaffected members, primarily involving the neostriatum and associated neural circuits. Although the FOXP2 gene was identified 2 years later as the gene deficit transmitted to affected family members, the authors' speculations about alternative genetic origins of these findings (e.g., see the following excerpts) remain relevant:

Our data suggest that development of the neural mechanisms mediating the acquisition of fine oromotor coordination (both vocal and nonvocal) and of speech and language are interdependent, such that abnormality in the one will be associated with abnormality in the other…

…a central abnormality affecting speech production could have a cascading effect resulting in intellectual defects…

At this stage, we cannot discount the alternative possibility that the different components of the phenotypic profile are the consequence of abnormalities in several different neural networks resulting from disruption of either a single gene or even several contiguous genes (Vargha-Khadem et al., 1998, p. 12700)

As summarized in Table 1, the study by Watkins, Vargha-Khadem, et al. (2002) reported significant differences in white matter volumes bilaterally in affected compared to nonaffected KE family members and controls, with affected family members having both larger and smaller volumes at different neuroanatomic sites. These morphometric data underscore the complexity of the pathophysiology of CAS in these family members. Using different neuroimaging methods, Belton, Salmond, Watkins, Vargha-Khadem, and Gadian (2003) provided additional neuroanatomic findings, again supporting bilateral involvements and morphological differences in areas that subserve both motor and language processing. Finally, Liégeois, Baldeweg, Connelly, Gadian, and Vargha-Khadem (2003) used both functional neural imaging methods and verbal processing measures to attempt to relate structural findings to behavioral findings in the affected individuals. The extended discussion of neural and neurocognitive findings in this paper provides a promising research agenda for studies in process on the genetic substrates of speech-language challenges.

Autism

Limb apraxias, oral apraxia, and apraxia of speech have been frequently reported for children with autism or a pervasive developmental disorder (e.g., Boyar et al., 2001; Page & Boucher, 1998; Rogers, Bennetto, McEvoy, & Pennington, 1996; Seal & Bonvillian, 1997). Well-controlled studies are needed to test the hypothesis that apraxia of speech is more prevalent in autism than as occurs idiopathically in the general population. At the time this report was in preparation, several studies in process were studying this question using contemporary inclusionary/exclusionary criteria for both autism and CAS.

Epilepsy

CAS has been noted as comorbid with or a sequela of several forms of epilepsy, including benign rolandic epilepsy and autosomal dominant rolandic epilepsy, the latter of which is a rare form associated with more severe and long-term communicative disorders. Scheffer et al. (1995; see Table 2) provided interesting research hypotheses on the diagnostic significance of comorbid epilepsy and apraxia, again underscoring the value of studying apraxia in the context of well-characterized neurological and complex neurobehavioral disorders.

Fragile X Syndrome

Fragile X syndrome is a genetically transmitted complex neurobehavioral disorder in which speech and prosody deficits are associated with reduced intelligibility (Roberts, Hennon, & Anderson, 2003). Reports indicate that some of these deficits overlap with diagnostic criteria for CAS, but the measures used to assess the nature of speech and prosody involvement have typically not been well developed. At the time this report was in preparation at least one research study in process was attempting to replicate the Spinelli, Rocha, Giacheti, and Ricbieri-Costa (1995) findings (see Table 2) of apraxia of speech in 40% of a small sample of children with fragile X syndrome.

Galactosemia

Some form of CAS reportedly also occurs in 40%–60% of children with one of the several genetic forms of the metabolic disorder, galactosemia (Elsas, Langley, Paulk, Hjelm, & Dembure, 1995; Hansen et al., 1996; C. D. Nelson, Waggoner, Donnell, Tuerck, & Buist, 1991; D. Nelson, 1995; Robertson, Singh, Guerrero, Hundley, & Elsas, 2000; Webb, Singh, Kennedy, & Elsas, 2003). At the time this report was in preparation, a study of CAS in galactosemia was in process using a relatively large sample of children with this disorder.

Rett Syndrome

Limb and speech apraxia are reportedly part of the sequence of neurological dysfunctions that characterize the degenerative course of expression of Rett syndrome. Because the apraxic disorder is so profound that children at this stage essentially do not speak (Bashina, Simashkova, Grachev, & Gorbachevskaya, 2002; Schanen et al., 2004), it is difficult to study speech apraxia in individuals with this neurobehavioral disorder. Genetic studies indicate that the molecular regions involved in Rett syndrome include susceptibility genes for a number of disorders reported to involve speech-language deficits (N. J. Wang, Liu, Parokonny, & Schanen, 2004).

Chromosome Translocations Involving Deletions and Duplications

One of the most active and productive areas of genetic research in complex neurobehavioral disorders involves the identification of persons with translocations that affect speech processing. The case study reported in Weistuch and Schiff-Meyers (1996) (see Table 2) illustrates the potential for CAS research in chromosomal translocations. Recall that it was a child with a translocation involving a breakpoint in chromosome 7 that helped the U.K. investigators identify the SPCH1 susceptibility region for the apraxic disorder found in the KE family. Somerville et al. (2005) reported a child with chromosome duplications affecting genes at 7q11.23 (the Williams-Beuren syndrome microdeletion locus) who has “severe delay in expressive speech.” Kriek et al. (2006) also described a child with a duplication in the same region who reportedly also has significant speech delay (cf. Tassabehji & Donnai, 2006). These two papers have prompted a large-scale study now in process seeking to determine if duplications of this locus are present in children who reportedly have CAS. Lichtenbelt et al. (2005) described a child and 4 other reported cases with a rare supernumerary ring chromosome on 7q. All 5 cases have severely delayed expressive speech. Finally, Shriberg, Jakielski, Patel, and El Shanti (Shriberg, 2006) described 3 siblings with an unbalanced 4q;16q translocation whose speech and prosody profiles also are consistent with CAS and with dysarthria.

Linguistic Approaches

Powell (1996) described a case study of a child who had been diagnosed with CAS and oral apraxia. Previous treatment at two different facilities, in which the child was typically seen for two 30-minute sessions per week, had yielded little progress. Treatment had reportedly appeared to focus on production of early developing sounds and those that were emerging in the child's phonetic inventory, as well as on the use of AAC. Powell initiated intensive treatment (four 1-hour sessions per week for a 3-month period in the summer) that included a significant focus on increasing the child's stimulability for sounds not appearing in his speech. The rationale for increasing stimulability, characterized as a phonologic approach, was based on findings from research (Powell, 1993) indicating that for children with speech sound disorders, “stimulable sounds are likely to be added to the speaker's phonetic inventory whereas non-stimulable sounds will continue to be excluded” (Powell, 1996, p. 319). This goal was supplemented by other components more typical of a traditional articulation approach (Bernthal & Bankson, 2004), including stabilization of inconsistently used sounds in words and generalizations of known sounds to the conversational level. Over the 3-month period, the child's productive repertoire went from 11 to 17 phones, a 55% increase. The author suggested that these findings indicate the potential value of targeting stimulability in children with CAS. However, the overall lack of control, characteristic of a case study (see Table 4, Level III), provides only a low level of evidence for the findings. For example, the multiple components included in the rich intervention protocol prohibit clear assignment of the source of the findings to the stimulability activities.

Motor Programming Approaches

Motor programming approaches, which may also be termed articulatory or phonetic approaches, include integral stimulation (Strand & Debertine, 2000; Strand & Skinder, 1999) as well as a number of commercially available intervention programs (e.g., Dauer, Irwin, & Schippits, 1996; Kaufman, 1995; Kirkpatrick, Stohr, & Kimbrough, 1990; Strode & Chamberlain, 1993; Williams & Stephens, 2004). Of these examples, efficacy research has been reported only for the integral stimulation approach (Strand & Debertine, 2000; Strand & Skinder, 1999). Integral stimulation is a modification of a treatment approach developed for adults with apraxia by Rosenbek, Lemme, Ahern, Harris, and Wertz (1973). It incorporates principles of motor learning described by earlier authors in the field (e.g., Rosenbek, Hansen, Baughman, & Lemme, 1974) and also by researchers from outside of the field of speech-language pathology. Notable in the latter category is the research by Schmidt and colleagues (e.g., Schmidt, 1991; Schmidt & Bjork, 1992). Manipulation of parameters that affect motor learning, such as frequency and nature of practice opportunities and knowledge of results and performance, are fundamental elements of the integral stimulation approach. Strand and Debertine completed a multiple-baseline-across-behaviors single subject design to examine the efficacy of integral stimulation over 33 sessions (30 minutes, four times per week) for a girl (age 5;9) with low comprehensibility (10%–20%) and CAS. Speech production gains for a small number of functional one- and two-word phrases (e.g., “Hi, Dad”, “Not now!”, “No!”) were observed when probe data for these treated items were compared against baseline and control measures.

The generalizability of Strand and Debertine's (2000) findings is limited by a lack of replication across subjects and because single subject experimental designs are not recognized within the SIGN hierarchy shown in Table 4. Nonetheless, such designs are thought to demonstrate a high degree of experimental control, especially for heterogeneous and rare participant populations for which randomized control trials may prove unfeasible or even ill-advised (Ylvisaker et al., 2002).

Combined Linguistic-Motor Programming Approaches

For the time period reported at publication of this review, approximately the past 12 years, the only published treatment study that can be readily classified using Hall's (2000b) category of combined linguistic-motor programming treatment is the research described in Bahr et al. (1999). This exploratory study described an inclusion classroom staffed by a speech-language pathologist and elementary school teacher as the context for treatment of 4 children diagnosed with suspected CAS (referred to as “oral motor impairments”; Bahr et al., 1999, p. 25), as well as 5 children with speech sound disorders. It may be described as a combined linguistic-motor programming approach because it was based on the work of Velleman and Strand (1994), who, as described in Bahr et al., proposed that CAS represents “an underlying impairment of the ability to generate hierarchical plans or sequences of behaviors—whether they be motor or linguistic or both” (p. 21). Speech production served as a focus within a regular kindergarten curriculum in which 5 children with typically developing speech also participated. Classroom activities ranged from those focusing on oral-motor and oral sensory experiences (e.g., light touch and brushing of the face and articulators) to more phonological aspects of speech production focusing on phoneme practice. The latter activities made use of tactile cues as well as descriptive phrases to define important sound characteristics, such as those used in Metaphon (Dean & Howell, 1986). Sounds were practiced in varying phonetic contexts and, over time, in longer and more complex syllabic structures. Prosody also served as a treatment focus. If children were unable to complete activities within the group context, they received individual treatment on an as-needed basis. The authors' impression was that children made positive progress in speech production and intelligibility, but no data were provided to support those observations.

Approaches That Include Specific Sensory and Gestural Cueing Techniques

As with the class of approaches above, there have been no treatment methods during the review period that have focused on specific sensory and gestural cueing techniques, although this approach has been described as a component of intervention in many less recent papers (e.g., Bashir, Graham-Jones, & Bostwick, 1984; Chappell, 1974; Chumpelik, 1984; Hayden & Square, 1994; Klick, 1985). However, one recent paper that described what may be considered a sensory and gestural cueing approach consists of a literature review and case description concerned with the use of instrumentation to support children's speech production treatment (Gibbon, Stewart, Hardcastle, & Crampin, 1999). The authors introduced the use of electropalatography for children described as having persistent speech sound disorders, a category that can include children suspected to have CAS (although CAS was not considered as a diagnosis for the child described). Electropalatography is used to obtain detailed assessment of tongue movement and provides visual feedback regarding tongue contact with an artificial palate.

Summary

Assessment and diagnosis of CAS are the responsibility of the speech-language pathologist with specialized knowledge, training, and skills in this area. The symptoms of CAS change over time and may be influenced by development in other behavioral domains. Although no single differential diagnostic marker with high sensitivity and specificity has been documented to date, there is some consensus among clinical researchers on three segmental and supra-segmental features observed in children suspected to have CAS.

Intensity

Given the need for repetitive planning, programming, and production practice in motor speech disorders, clinical sources stress the need for intensive and individualized treatment of apraxia, especially for children with very little functional communication. There is emerging research support for the need to provide three to five individual sessions per week for children with apraxia as compared to the traditional, less intensive, one to two sessions per week (Hall et al., 1993; Skinder-Meredith, 2001; Strand & Skinder, 1999). Ideally, this should be done in as naturalistic an environment as possible to facilitate carry-over and generalization of skills. Although home practice is critical for optimal progress, it cannot take the place of individual treatment provided by a speech-language pathologist who has expertise in motor speech skill facilitation. For the diverse backgrounds of children seen for early intervention, including their stages of psychological/emotional development, the Committee sees value in endorsing a treatment plan for optimum progress based on provision of intensive therapy.

Individual differences among children will also underlie rationale for changing the form, content, and intensity of treatment throughout the course of intervention. If toddler and preschool-age children are seen for early intervention that targets their speechmotor deficits, the frequency of treatment may be able to be reduced over time. As long as the primary goal is to improve the motoric aspects of the child's speech production (i.e., more time for motor practice), individual therapy should be the preferred approach regardless of age. For children whose severity of involvement has decreased and whose treatment goals have begun to move toward language and pragmatic skills enhancement, a combination of both individual and small group therapy may also be optimal for some children, providing that a treatment focus is maintained on speech production.

For children with apraxia who also require other therapeutic services (e.g., occupational therapy, physical therapy), care must be taken to vary therapy activities to avoid fatigue. Collaborative decision-making is critical in such cases, where creative use of alternatives, such as co-treatment, should be considered (Davis & Velleman, 2000; Velleman & Strand, 1994).

Length of Treatment Sessions

In view of the Committee's information indicating that children are being enrolled for treatment of CAS at increasingly younger ages, careful consideration should be given to the length of the therapy session. If repetitive practice of speech-motor patterns is targeted in a therapy session, many children in the younger age ranges can remain engaged for only a maximum of 30 minutes per session. There are certainly those children for whom “adjustment” time is necessary prior to the introduction of more intensive treatment activities. Some service providers are allotted a certain number of minutes or hours per week of therapy time per child. Given the option between two 1-hour sessions and four 30-minute sessions, many clinical researchers strongly recommend the latter (e.g., “more sessions—less time per session”; Skinder-Meredith, 2001).

Treatment Strategies

The treatment literature in CAS indicates that the operating principles and strategies overlap those recommended for children with other speech sound disorders. Overall, the principles of motor learning theory and intensity of speech-motor practice appear to be the most often emphasized in an optimal treatment program. These recommendations include the need for distributed practice, in which speech-motor practice is carried out across a variety of activities, settings, and situations, and includes several exemplars per pattern (e.g., Strand & Skinder, 1999). Recall from the discussion above that speech requires more flexibility, less stereotyped rhythmicity, finer levels of coordination, and lower levels of strength than other nonspeech oral motor activities such as chewing, blowing, and the like. A systematic review addressing this topic is currently underway by an ASHA committee through its National Center for Evidence-Based Practice. Until the committee report is available, the consensus opinion is nonspeech oro-motor therapy is neither necessary nor sufficient for improved speech production. Another often-cited recommendation is to take advantage of other areas of strength for children with CAS by utilizing a multisensory approach to treatment. The use of sign language, pictures, AAC systems, visual prompts, and touch cues have been described as being extremely effective for children with CAS, providing functional communication while at the same time supporting and enhancing verbal speech production. Another important element for optimal progress and carry-over is to involve as many important people in the child's life as possible, in a culturally appropriate manner, in understanding and completing therapy goals outside the treatment setting.

Funding Treatment

Although reimbursement and funding issues are a concern for all childhood speech sound disorders, insurance funding issues in CAS have become a topic of considerable interest. As reviewed, children with CAS are likely to require intensive services over a long period of time. In a study on treatment outcomes from one large facility, Campbell (1999) reported that children with the diagnosis of CAS needed 81% more individual therapy sessions than children described as having a phonological disorder in order to achieve the same functional outcome. Of even more interest to funders are Campbell's findings indicating that the average cost of achieving the same functional outcome for a child with CAS was $11,325, compared to $2,000 for a child with a phonological disorder. This study reported that in addition to time needed for treatment of children with CAS, professionals needed additional time to identify the appropriate diagnostic codes for CAS, write reports, educate funders, and assist caregivers with advocacy needs in order to pursue reimbursement. Web sites such as that of the Childhood Apraxia of Speech Association of North America (http://www.apraxia-kids.org) have developed useful materials to help caregivers and others with the complex of resources to aid families and therapists in securing insurance funding.

Comorbid Conditions and Allocation of Resources

Recent research has continued to validate co-morbid deficits that accompany the severe speech production constraints that characterize children with CAS. As reviewed, Lewis and her colleagues in a follow-up study reported that children with CAS are likely to have deficits in both expressive and receptive language as well as in academic areas such as reading, spelling, and written expression (Lewis et al., 2004). Lewis et al. proposed that speech-language pathologists should consider offering phonological awareness and preliteracy training as part of intervention for children with CAS at risk for language learning challenges.

For some children with CAS, therapy approaches that focus exclusively on oral output are inadequate, requiring augmentative and alternative modes of communication (Cumley & Swanson, 1999). AAC systems, of all types, require time for speech-language pathologists, other school personnel, and the child to learn and use in order to expand communication opportunities. Thus, in addition to their needs for intensive, individual speech and/or AAC treatment to maximize effective communication, children with CAS will also require therapeutic time to address a number of other issues—all of which contribute to these children becoming functional communicators who can be comfortable with and learn in the school environment. This level of need places demands on the speech-language pathologist, regardless of practice setting. It also requires that both private speech therapy providers and school-based clinicians work closely with one another and with families. Collaboration is required in order to optimally integrate the child's needs into the intervention time available and to assemble the best possible program of services for affected children.

Summary

Although the specific forms of treatment may change over time, the Committee recommends that children with CAS receive intensive services, especially in the earlier stages of intervention. The rationale for this recommendation is based on the assumption that the child's potential for normalization of speech and prosody may be substantially reduced if not addressed during early periods of growth and development. There are treatment constraints (e.g., limited funding, limited staff availability) in certain settings that make it challenging to secure intensive, individual therapy. Resources need to be made available to insurance companies, school districts, and specialized programs to provide children with CAS the best opportunity to develop functional communication. Sociodemographic issues should be addressed to ensure that all children with CAS receive the type and intensity of services needed to treat this complex motor speech disorder.

The Committee also underscores the responsibility of ASHA and its membership to educate allied health care professionals on current perspectives in CAS so that timely referrals are made and appropriate therapeutic services are supported. This requires education at both local and national levels. More generally, speech-language pathologists must be adequately trained in areas such as differential diagnosis of childhood motor speech disorders, motor learning theory, cueing strategy usage, and other intervention techniques that clinical researchers have reported as effective. Such knowledge and skills training is the responsibility of academic training programs. New forms of partnerships must emerge among clinicians and across therapeutic settings to create intervention programs that maximize resources and address the multifaceted deficits presented by this clinical population. Finally, trends in the treatment literature indicate that professional education and collaboration are needed to enhance the resources and opportunities for children with apraxia of speech.

Basic Research Needs

1. Speech motor control and neurolinguistic studies using contemporary methods in such disciplines as neurophysiology, neurochemistry, neural imaging, kinematics, and acoustics to describe the pathophysiology of CAS.

2. Molecular genetic studies using contemporary genomic and bioinformatic resources to provide an eventual account of the developmental neurobiology of CAS.

3. Epidemiological studies of CAS to delineate the gender-specific risk for this disorder in children reared in different countries, languages, races, ethnicities, and cultures.

Applied Research Needs

1. Cross-linguistic longitudinal studies to identify the core behavioral features of CAS and to develop clinically efficient diagnostic protocols for valid and reliable assessment of children at prelinguistic and later stages of CAS.

2. Studies to develop treatment programs that are appropriate for children of all ages and backgrounds with idiopathic CAS, as well as multidisciplinary studies to develop treatment programs for children with apraxia of speech occurring as the sequela of neurological deficits and within complex neurobehavioral disorders.

3. Randomized control trials and smaller-scale studies to test the efficacy of alternative treatment programs for children of all ages, types, and severities of expression of CAS, with findings enabling the development of guidelines for best practices.