Top Lang Disorders Vol. 31, No. 2, pp. 112–127 Copyright c 2011 Wolters Kluwer Health | Lippincott Williams & Wilkins Subtyping Children With Speech Sound Disorders by Endophenotypes

Barbara A. Lewis, Allison A. Avrich, Lisa A. Freebairn, H. Gerry Taylor, Sudha K. Iyengar, and Catherine M. Stein

Purpose: The present study examined associations of 5 endophenotypes (i.e., measurable skills that are closely associated with speech sound disorders and are useful in detecting genetic in- fluences on speech sound production), oral motor skills, phonological memory, phonological awareness, vocabulary, and speeded naming, with 3 clinical criteria for classifying speech sound disorders: severity of speech sound disorders, our previously reported clinical subtypes (speech sound disorders alone, speech sound disorders with impairment, and childhood apraxia of speech), and the comorbid condition of disorders. Participants and Method: Chil- dren with speech sound disorders and their siblings were assessed at early childhood (ages 4– 7 years) on measures of the 5 endophenotypes. Severity of speech sound disorders was deter- mined using the z score for Percent Consonants Correct—Revised (developed by Shriberg, Austin, Lewis, McSweeny, & Wilson, 1997). Analyses of variance were employed to determine how these endophenotypes differed among the clinical subtypes of speech sound disorders. Results and Conclusions: Phonological memory was related to all 3 clinical classifications of speech sound disorders. Our previous subtypes of speech sound disorders and comorbid conditions of language impairment and reading disorder were associated with phonological awareness, while severity of speech sound disorders was weakly associated with this endophenotype. Vocabulary was as- sociated with mild versus moderate speech sound disorders, as well as comorbid conditions of language impairment and reading disorder. These 3 endophenotypes proved useful in differentiat- ing subtypes of speech sound disorders and in validating current clinical classifications of speech sound disorders. Key words: endophenotypes, oral motor skills, phonological awareness, phonological memory, severity, speech sound disorders, speeded naming, subtypes, vocabulary HILDREN with early speech sound dis- Author Affiliation: Departments of Psychological C orders (SSD) of unknown etiology com- Sciences (Dr Lewis and Ms Freebairn), Epidemiology prise a heterogenous group with high preva- and Biostatistics (Ms Avrich and Drs Iyengar and lence rates, especially during the preschool Stein), and Pediatrics (Dr Taylor), Case Western Reserve University, Cleveland, Ohio. years (Campbell et al., 2003). Subgrouping This research was supported by the National Institutes children into clinically validated subtypes of of Health, National Institute on Deafness and Other SSD allows for more precision in the diag- Communication Disorders, Grant DC00528, awarded nosis, treatment, and prognosis of these dis- to Barbara A. Lewis and a Multidisciplinary Clinical Re- search Development Programs Grant (KL22RR024990) orders. However, to date, there is no sin- from the National Center for Research Resources gle classification system of SSD that is em- awarded to Catherine M. Stein. ployed uniformly by SLPs (speech-language The authors thank Dr Lawrence D. Shriberg, the pathologists). Children with SSD may be speech/language pathologists who assisted them in re- cruiting participants, and the families who generously grouped differently depending on whether agreed to participate. or not the subgroups are based on hypoth- Corresponding Author: Barbara A. Lewis, PhD, esized etiology (Shriberg et al., 2005), psy- Department of Psychological Sciences, Case Western Reserve University, 11635 Euclid Avenue, Room 330, cholinguistic models (Stackhouse & Wells, Cleveland, OH 44106 (e-mail: [email protected]). 1997), or behavioral symptoms of SSD (e.g., DOI: 10.1097/TLD.0b013e318217b5dd phonological processes; Dodd, 2005). These

112 Subtyping Using Endophenotypes 113 classification systems have provided insight process analyses or classification of errors as into the nature of SSD and have resulted in omissions, substitutions, or distortions, imply different diagnostic categories and treatment an assumption that children with similar error approaches. However, comorbidity of sub- types share a common etiology and/or may types (e.g., a combined articulation and benefit from similar treatment approaches. phonological disorder), as well as different Children with a large proportion of omis- combinations of related disorders such as lan- sion errors, for example, have been consid- guage impairment (LI) and reading disorders ered to have a linguistically based disorder (RD), result in overlapping diagnostic cate- rather than a motor-based disorder and have gories, complicating understanding of causal been viewed as qualitatively different from relationships and choices of treatment ap- children with primarily substitution or dis- proaches. tortion errors (Panagos, 1974). Preston and Some past studies have subtyped children Edwards (2010) found that children with atyp- with SSD based on the clinical features of ical speech sound errors had poorer phono- the disorders, such as speech sound errors logical awareness skills and lower receptive and comorbid LI. Other studies have exam- vocabularies than children with more typical ined associations of cognitive–linguistic skills speech sound errors. or endophenotypes with SSD. Endopheno- Other classification systems have grouped types are objectively measurable cognitive, children with SSD on the basis of a combi- linguistic, or neuropsychological parameters nation of factors that include history, sever- that are closely associated with a specific be- ity, and associated conditions, as well as havioral trait and are useful in detecting ge- speech sound error characteristics (Dodd, netic influences on the behavioral phenotype 2005; Shriberg, 1994; Shriberg & Kwiatowski, (Gottesman & Gould, 2003; Inoue & Lupski, 1988; Shriberg, Kwiatkowski, Best, Hengst, 2003), for example, phoneme awareness has & Terselic-Weber, 1986). Such systems have been associated with RD. Few studies, how- been useful clinically in deriving diagnostic ever, have investigated the extent to which profiles of children with SSD. For example, endophenotypes are related to classifications a system proposed by Dodd (2005) classi- based on the clinical features of SSD. As en- fies SSD into five subtypes: articulation dis- dophenotypes relate more directly to under- order, delayed phonological acquisition, con- lying genetic influences (Gottesman & Gould, sistent deviant disorder, inconsistent deviant 2003), differential association of these skills disorder, and other (including dysfluency, with clinical features of SSD may help to elu- , and apraxia of speech). This sys- cidate the sources of important clinical varia- tem does not consider the etiology of the SSD, tions. and, therefore, children within a subtype may have diverse hypothesized causal factors that SUBTYPING SSD ON THE BASIS OF may influence prognosis and treatment. SPEECH ERRORS Another framework for subtyping children with SSD that does consider etiology, pro- Previous attempts to subtype children with posed by Shriberg and colleagues (Shriberg, SSD have relied on various classifications of 1993; Shriberg et al., 2005; Shriberg & features of the child’s speech. Traditionally, Kwiatkowski, 1982, 1994; Shriberg, 2010), clinical severity ratings of the disorder as mild, is the Classification Sys- moderate, or severe have depended on the tem (SDCS). This classification system pro- SLP’s judgment of the child’s intelligibility poses seven subtypes of SSD: — in conversational speech or the number of genetic, speech delay—otitis media with effu- speech sound errors that the child exhibits sion, speech delay—apraxia, speech delay— on a standardized articulation test. Analyses dysarthria, speech delay—developmental of speech error types, such as phonological psychosocial involvement, and two categories 114 TOPICS IN LANGUAGE DISORDERS/APRIL–JUNE 2011 of speech errors limited to distortions of dren with isolated SSD. This suggests that co- speech sounds. Children with speech delay— morbid conditions such as LI may not be use- apraxia, also called childhood apraxia of ful in forming subgroups of children with SSD speech (CAS), are of particular interest be- at early childhood. cause of the severity of the disorder, the prolonged course of treatment, and the dis- tinct speech characteristics (see the American ASSOCIATIONS OF ENDOPHENOTYPES Speech-Language-Hearing Association [ASHA] WITH SSD position statement report for a detailed de- scription; ASHA, 2007). In our previous stud- Recently, genetic studies of SSD have exam- ies, we have demonstrated that children with ined endophenotypes as the presumed her- CAS represent a distinct subtype of SSD itable components that reflect genetic influ- with a long-term impact on school-age aca- ences leading to SSD. Endophenotypes are demic skills (Lewis, Freebairn, Hansen, Tay- thought to influence clinical phenotypes such lor, Iyengar, & Shriberg, 2004; Lewis, Free- as SSD, LI, and RD (Bishop & Snowling, bairn, Hansen, Iyengar, & Taylor, 2004). 2004; Castellanos & Tannock, 2002; Fisher & DeFries, 2002; Pennington, 1997). The en- SUBTYPING BY COMORBID LI dophenotype is hypothesized to involve fewer genes than the clinical phenotype, sim- One factor that appears to distinguish plifying the genetic analysis (Gottesman & subgroups of SSD is the presence or ab- Gould, 2003). Endophenotypes, such as oral sence of comorbid LI. Shriberg, Tomblin, and motor skills, phonological awareness, phono- McSweeny (1999) reported 11% to 15% co- logical memory, speed of processing, and vo- morbidity of speech delay with LI at 6 years cabulary, are postulated to influence SSD. of age, with considerably higher rates of co- Several studies have demonstrated that chil- morbidity for preschool children with speech dren with speech and language disorders delay (40%–60%; Shriberg & Kwiatkowski, present with poorer motor skills than nor- 1994). A growing body of literature suggests mally developing children (Bishop, 2002; that children with LI in addition to the SSD Bishop & Adams, 1990; Bishop & Edmundson, have poorer outcomes than children with SSD 1987; Hill, 2001). Bishop (2002) employed a in isolation (Aram & Hall, 1989; Bishop & tapping task and a peg movement task that Adams, 1990; King, Jones, & Laskey, 1982; were timed. She found that the tapping task Lewis, Freebairn, & Taylor, 2000; Nathan, was related to accuracy Stackhouse, Goulandris, & Snowling, 2004; and the peg movement task was related to Shriberg & Austin, 1998; Young et al., 2002). nonword repetition. Hill (2001) reviewed 28 In our own studies, children with SSD ac- studies assessing children with LI employing companied by LI performed more poorly than motor tasks such as peg moving, bead thread- children with isolated SSD at school age on ing, and finger tapping and found associations measures of , reading decoding, read- of poor motor skills to LI. Timed tasks such as ing comprehension, and written language, as those employed by Bishop (2002) may be con- well as on language measures and phono- founded with slowed speed of processing in logical awareness. Eighteen percent with iso- general. Studies specific to oral motor skills lated SSD and 75% with combined SSD and LI have demonstrated that children with primar- had reading problems at school age (Lewis, ily articulation disorders perform more poorly Freebairn, & Taylor, 2000; Lewis, O’Donnell, on diadochokinetic tasks than their normal Freebairn, & Taylor, 2002). Despite the peers (Bernthal, Bankson, & Flipson, 2009). poorer academic outcomes, children with Other factors such as tongue strength, fine combined SSD and LI did not present with tongue control, and stability may also play a more severe SSD at early childhood than chil- role. Subtyping Using Endophenotypes 115

Phonological awareness, also an endophe- 30% of the variance in phonological aware- notype associated with SSD, is assessed by ness in preschool and young school-age chil- tasks that require children to manipulate dren (Bishop & Adams, 1990; Elbro, Borstrom, phonemes in spoken words, such as by & Peterson, 1998; Rvachew & Grawberg, phoneme deletion or reversal. Deficits in 2006). Metsala (1999) hypothesized that chil- phonological awareness have been identi- dren who know more words have phonologi- fied as one of the strongest earliest predic- cal representations that are more adult-like in tors of RD (e.g., Raitano, Pennington, Tu- their features and organization. nick, Boada, & Shriberg, 2004). Studies also have demonstrated that children with SSD, PURPOSE OF THE STUDY both with and without comorbid LI, per- form more poorly than controls on phono- The preferred practice patterns for the pro- logical awareness measures (Colledge, et al., fession of speech language pathology (ASHA, 2002; Kovas et al., 2005; Nathan, Stackhouse, 2004) indicate that clinicians should docu- Goulandris & Snowling, 2004; Preston & Ed- ment the type and severity of the SSD and wards, 2010). Thus, phonological awareness its associated conditions. Yet, as reported by is a useful endophenotype for SSD as well Flipsen, Hammer, and Yost (2005), there is as RD. little agreement on ratings of severity of SSD Phonological memory refers to coding even by experienced SLPs, especially in the information phonologically for temporary middle of the scale range. Past studies have storage in working or short-term memory not examined possible endophenotypic dif- (Gathercole & Baddeley, 1990; Torgesen & ferences between clinical subtypes. Discov- Wagner, 1992). It is usually assessed by ask- ery of such differences would shed light on ing children to repeat nonwords. Deficits in the deficits underlying clinical variations in phonological memory are thought to impair SSD and help guide further studies of underly- an individual’s ability to learn both spoken ing genetic mechanisms. To address this gap and written new words. Poor phonological in knowledge, this study examined associa- memory may result in weak phonological rep- tions of endophenotypes of oral motor skills, resentations that underlie SSD and RD (Adams phonological memory, phonological aware- & Gathercole, 1995). ness, vocabulary, and speeded naming with Speeded naming has been implicated as an clinical classifications on the basis of sever- endophenotype in studies that have shown ity and comorbidity. These classifications in- that children with LI and/or RD have a cluded (1) severity of SSD (moderate SSD, generalized slowing of information process- mild-moderate SSD, mild SSD, recovered SSD, ing when compared with control children and no SSD), (2) comorbidity of SSD with LI (Windsor & Hwang, 1999). Slower speed of and RD (SSD alone, SSD with RD, SSD with LI, processing may also impact the establishment SSD with RD, and LI), and (3) our previously of phonological representations for words as defined subtypes of SSD (SSD alone, SSD with the phonological information held in memory LI, and CAS). may decay before the representation is estab- lished (Montgomery, Magimairaj, & Finney, METHOD 2010). Vocabulary knowledge is an endopheno- Participants type that plays a critical role in reading acqui- Participants were 237 children who were sition and skill (Wise, Sevcik, Morris, Lovett, enrolled in a large longitudinal genetic study & Wolf, 2007). The semantic content of a of SSD, including both the probands with SSD word has been shown to influence phono- and their nonreferred, typically developing logical skills as well. For example, vocabu- siblings. Probands were identified as having lary has been shown to account for 25% to SSD (n = 168) at 4 to 7 years of age (n = 79 116 TOPICS IN LANGUAGE DISORDERS/APRIL–JUNE 2011 participants aged 5 years and younger) from Procedures the clinical caseloads of SLPs working at com- The participants were tested individually munity speech and hearing centers or in pri- in two sessions to reduce potential effects vate practice in the greater Cleveland area. All of fatigue on test results. Testing was car- participants demonstrated the following: (1) ried out in a speech research laboratory at normal hearing acuity as defined by passing a Case Western Reserve University or, at the pure tone audiometric screening test at 25 dB parent’s request, in a quiet and adequately HL ISO for 500, 1000, 2000, and 4000 Hz bi- lit room in the family’s home. Speech pro- laterally and fewer than six reported episodes ductions were recorded on high-quality au- of otitis media before the age of 3 years as diocassette. Responses were recorded initially reported by the parent; (2) absence of a his- online using broad phonetic transcription. tory of neurological disorders or developmen- Speech sound samples were later transcribed. tal delays other than speech and language All tests were presented in a counter balanced as reported by the parent; and (3) normal manner, with the more time-consuming mea- intelligence defined as a performance intel- sures alternating with less lengthy measures. ligence quotient (PIQ) of 80 or above on the Parents completed a developmental question- Wechsler Intelligence Test Scale for Children– naire and structured interview on their chil- Third Edition (WISC-III; Wechsler, 1991) or dren to document comorbid conditions of LI, the Wechsler Preschool and Primary Scales RD, and a diagnosis of CAS. of Intelligence–Revised (WPPSI-R; Wechsler, 1989). Measures of endophenotypes The severity of SSD of the probands was de- Oral motor skills fined in terms of a conversational speech sam- ple of at least 50 utterances, obtained using Diadochokinetic rates of single and multi- technical and interlocutor procedures for free syllables on The Oral and Speech Motor Con- speech sampling, as described by Shriberg trol Protocol (Robbins & Klee, 1987) or the and Kwiatkowski (1982). The Percentage of Fletcher-Time-by-Count technique (Fletcher, Consonants Correct-Revised (PCCR) metric 1977) were used to assess oral motor skills. (Shriberg et al., 1997) was calculated. The The same syllables were employed in both PCCR is a well-respected and widely used tests to assure that the tasks were compara- measure, both clinically and in research, to ble; z scores were obtained. estimate the severity of speech impairment. It Phonological awareness is simple to calculate and has demonstrated va- lidity and reliability (Shriberg, Austin, Lewis, Tests of phonological awareness included McSweeny, & Wilson, 1997; Shriberg & the Elision and Sound Blending Subtests of the Kwiatkowski, 1982). It is similar to the Per- Comprehensive Test of Phonological Process- cent Consonant Correct (PCC) metric, with ing (CTOPP; Wagner, Torgesen, & Rashotte, the exception that speech sound distortions 1999). Standard scores were used in data anal- are not counted as errors. Z scores were cre- ysis. The first available measurement for this ated on the basis of Shriberg’s work to ac- variable was used for analysis. If children were count for age (Shriberg et al., 1997). On the younger than 5 years at their first visit, the basis of these z scores of the PCCR, five groups CTOPP assessment from their second visit was were created: zPCCR less than −2.0 (moder- used in analyses; z scores were obtained. ate SSD), zPCCR less than −1.0 but greater than −2.0 (mild-moderate SSD), zPCCR less Phonological short-term memory than 0 but greater than −1.0 (mild SSD), Phonological short-term memory was as- zPCCR greater than 0 but reported history of sessed by the Digit Span subtest of the WISC therapy for SSD (recovered SSD), and zPCCR III, the Sentence Imitation Subtest of the greater than 0 with no history of SSD (no SSD). Test of –Primary 3rd Subtyping Using Endophenotypes 117

Edition (TOLDP:3; Newcomer & Hammill, ual measures for age and age squared. Our pre- 1997), and the repetition of multisyllabic real vious work has shown that, although perfor- (Catts, 1986) and nonsense words (Catts, mance on these measures improves with age, 1986). The rationale for using nonsense words it does not do so in a straight line (Stein et al, to assess short-term phonological memory is 2011). Adjustment for age and age squared ap- that because the word is novel, the partici- proximates this relationship so that all scores pant does not have it coded or stored in his are on the same scale. A domain or factor lexicon. Again, if children were younger than score was derived for each endophenotype’s 5 years at their first visit, the Digit Span sub- z scores for the measures comprising that fac- test from their second visit was used for analy- tor. The measures included were the diado- sis. Raw scores were age standardized accord- chokinetic rates of single and multisyllables ing to the age of the child at assessment and on either The Oral and Speech Motor Con- the age-standardized scores were used in data trol Protocol (Robbins & Klee, 1987) or the analysis. Fletcher-Time-by-Count (Fletcher, 1977) for Vocabulary the oral motor factor, the CTOPP subtests Elision and Sound Blending for the phono- The Expressive One Word Picture Vocab- logical awareness factor, the real-word and ulary Test, Revised (EOWPVT-R; Gardner, nonsense-word repetition tasks and the 1990) and the Peabody Picture Vocabulary TOLD-P-3 Sentence Imitation for the phono- Test, 3rd Edition (PPVT-III; Dunn & Dunn, logical memory factor, the PPVT and EOW- 1997) assessed receptive and expressive vo- PVT for the vocabulary factor, and the Rapid cabulary skills. Standard scores were used in Naming of Colors for the speeded naming fac- data analyses. tor. The assignment of tests to domains was Speeded naming based on results from our previous factor anal- Rapid Naming of Colors (Denckla & Rudel, yses of the test battery (Lewis et al., 2006; 1976) assessed speeded naming, a measure of Stein et al., 2004; Sices, Taylor, Freebairn, speed of processing. Rapid naming requires Hansen, & Lewis, 2007). As shown in Table efficient retrieval of phonological informa- 1, significant correlations (p < .05) were ob- tion. z scores derived from the normative data served among the phonological memory fac- by Denckla and Rudel (1976) were used in tor, the speeded naming factor, the phono- data analysis. logical awareness factor, and the vocabulary factor. Data analyses Analysis of variance was employed to com- Age-standardized z scores were created for pare groups of individuals classified according each endophenotype by adjusting the individ- to each of the three clinical dimensions on the

Table 1. Correlation Matrix for Factor Scores of Endophenotypes

Oral motor Speeded Phonological Phonological skills naming awareness memory Vocabulary

Oral motor skills 1.00 .116 −.077 −.097 −.142 Speeded naming 1.000 −.114 −.188∗ −.163∗ Phonological awareness 1.000 .489∗∗ .353∗∗ Phonological memory 1.000 .440∗∗ Vocabulary 1.000

∗p < .05; ∗∗p < .001 118 TOPICS IN LANGUAGE DISORDERS/APRIL–JUNE 2011 factor scores. These classifications included SD = 16.47) than the recovered group (mean severity (moderate SSD, mild-moderate SSD, = 101.26, SD = 14.99), the mild group (mean mild SSD, recovered SSD, and no SSD) de- = 105.18, SD = 14.16), and the typically scribed earlier, comorbidity of RD (SSD alone developing group (mean = 106.52, SD = without RD, SSD + LI without RD, SSD + RD, 15.65). No group differences were observed and SSD + LI + RD), and clinical subtype (no on PIQ. SSD, SSD only, SSD + LI, and CAS). The 168 Comparisons of severity groups on the fac- children with SSD were grouped differently in tor scores revealed that the groups differed each of the three clinical dimensions. When on the phonological memory factor, with the examining the comorbidity of SSD with RD, moderate SSD presenting with lower scores the SSD alone group (n = 74) was split into on phonological memory, F(4, 209) = 12.05, SSD alone without RD (n = 67) and SSD with p < .0001, than the recovered SSD and mild RD (n = 7). The SSD + LI group (n = 94) was SSD groups, and lower scores on vocabulary, split into the SSD + LI without RD (n = 54) F(4, 212) = 4.55, p < .0001, than the recov- and the SSD + LI + RD (n = 40). When group- ered SSD group (Table 3). The no SSD group ing the participants by clinical subtypes, chil- differed from the recovered SSD group, the dren with a diagnosis of CAS (n = 41) formed mild-moderate SSD group, and the moderate a subgroup. All children with CAS except one SSD group on phonological memory, and from presented with comorbid LI. Thus, in the clin- the moderate group on vocabulary. Phono- ical subtype analyses, 73 children remained logical awareness showed a tentative differ- in the SSD alone group and 53 children re- ence among severity groups, F(4, 172) = 2.10, mained in the SSD + LI group. Pairwise dif- p = .083, though none of the pairwise com- ferences between groupings were examined parisons was statistically significant at α = .05. using Tukey’s test with α = .05.

Comorbid conditions and associations RESULTS with endophenotypes Severity groupings and associations Table 4 lists the comorbid conditions re- with endophenotypes ported by the parents for participants in each As shown in Table 2, the five severity group. LI (n = 100) was the most frequently groups (moderate SSD, mild-moderate SSD, reported comorbid condition, followed by RD mild SSD, recovered SSD, and no SSD) dif- (n = 53). Forty-two children (38 in the moder- fered significantly in gender, χ 2 = 12.3, ate group, 2 in the mild-moderate group, and p = .002, and age, F = 21.14, p < .0001. The 2 in the recovered group) were reported to SSD groups comprised more males than fe- have been diagnosed with CAS. The moder- males, with the moderate group 72% male, the ate group had the highest rates of comorbid mild-moderate group 50% male, and the mild disorders. group 65% male; whereas the typical group Group comparisons according to comorbid was 48% male. The recovered SSD group was condition revealed differences in phonologi- older (mean age in years = 6.31, SD = 0.97) cal awareness, F = 15.19, p < .0001; phono- than the other SSD groups (mild group mean logical memory, F = 17.53, p < .0001; and vo- = 5.35, SD = 1.31; mild-moderate mean = cabulary, F = 13.89, p < 0.0001 (Table 5). For 5.13, SD = 1.31 years; and moderate mean = phonological awareness, the children with all 4.86, SD = 1.36). The groups also differed on three comorbid conditions had significantly the composite language scores of the TOLD- lower scores than the children with SSD only, P:3 or the Clinical Evaluation of Language Fun- and the children with only SSD + LI or SSD damentals 3rd Edition (CELF-3; Semel, Wiig, + RI. In addition, children with SSD + LI & Secord, 1995), with the moderate SSD had significantly lower scores on these mea- group performing more poorly (mean = 86.98 sures than children with SSD + RD. For Subtyping Using Endophenotypes 119 value .0001 .0001 .0001 .0001 .0001 < < < < < p F 2 performance IQ intelligence quotient; = Percent Consonant Correct–Revised; PIQ = 45 54 23 16 75 237 No SSD Recovered SSD Mild SSD Mild–Moderate SSD Moderate SSD Total Eta 5.54 (1.76) 6.31 (0.97) 5.35 (1.31) 5.13 (1.61) 4.86 (1.36) 5.52 (1.52) 0.15 8.85 80.17 (20.60) 59.25 (26.74)94.53 (5.63) 48.4397.08 (25.01) (4.24) 94.37 (3.27) 97.40 (1.81) 26.87 (20.78) 88.76 (4.69) 91.96 (2.51) 10.61 (12.34) 83.08 (6.23) 45.80 (34.19) 0.64 89.34 (3.32) 89.28 71.76 (8.34) 75.95 (8.43) 85.81 (11.55) 0.73 89.31 133.41 (10.60) 0.75 147.72 107.22 (14.86) 101.26 (14.99) 105.18 (14.16) 100.33 (15.64) 86.98 (16.47) 98.42 (17.39) 0.22 10.62 efg abcdfghij Goldman-Fristoe Test of Articulation (Goldman & Fristoe, 1986); PCC-R = Comparison of SSD Severity Groups on Demographics and Speech, Language, and PIQ Measures dgi bcdefgij ) bcdefgij speech sound disorder. SD GFTA ( Score percentile = PCC-R PIQ standard score 104.25Language (15.34) standard 102.63 (14.40) 100.60 (17.19) 105.25 (15.73) 98.23 (12.46) 102.40 (14.47) 0.03 1.00 0.4076 n Male/femaleAge mean years 22/23 38/16 15/8 8/8 54/21 146/91 15.19 0.0096 PCC GFTA Mild SSD differs from mild–moderate SSD. No SSD differs from mild SSD. No SSD differs from moderate SSD. Recovered SSD differs from moderate SSD. No SSD differs from mild–moderate SSD. Recovered SSD differs from mild SSD. No SSD differs from recovered SSD. Recovered SSD differs from mild–moderate SSD. Mild SSD differs from moderate SSD. Mild–moderate SSD differs from moderate SSD. SSD Table 2. Note. a b c d e f g h i j 120 TOPICS IN LANGUAGE DISORDERS/APRIL–JUNE 2011

Table 3. Comparison of SSD Severity Groups’ Performance on Endophenotype Factor Scores Mean (SD)

Mild- No Recovered Mild moderate Moderate SSD SSD SSD SSD SSD Eta2 Fp

Oral motor n = 38 n = 47 n = 17 n = 12 n = 53 0.03 1.17 .325 skills N = 167 −1.29 (9.9) −0.99 (6.8) 0.60 (4.9) −2.62 (8.16) 1.34 (7.5) Speeded n = 35 n = 45 n = 17 n = 11 n = 50 0.00 0.12 .975 naming N = 158 −2.31 (31.7) 1.36 (27.3) 1.04 (36.8) −2.90 (31.7) 1.93 (38.9) Phonological n = 42 n = 49 n = 16 n = 14 n = 52 0.05 2.10 .084 awareness N = 173 0.35 (1.7) 0.35 (1.8) 0.02 (1.6) −0.57 (2.7) −0.58 (2.2) Phonological n = 44 n = 54 n = 23 n = 16 n = 73 0.19 12.05 <.001 memoryacdgi N = 210 3.00 (4.3) 0.16 (5.0) 0.58 (3.4) −1.87 (5.0) −2.65 (4.0) Vocabularydg n = 45 n = 54 n = 23 n = 16 n = 75 0.08 4.55 .002 N = 213 1.40 (5.3) 0.49 (5.1) 0.85 (4.7) −0.99 (5.8) −2.37 (5.7)

Note. Cells contain number of observations and mean (standard deviation) for each factor score. SSD = speech sound disorders. aNo SSD differs from recovered SSD. bNo SSD differs from mild SSD. cNo SSD differ from mild-moderate SSD. dNo SSD differ from moderate SSD. eRecovered SSD differ from mild SSD. fRecovered SSD differ from mild–moderate SSD. gRecovered SSD differ from moderate SSD. hMild SSD differ from mild–moderate SSD. iMild SSD differ from moderate SSD. jMild-moderate SSD differ from moderate SSD.

Table 4 Percentage of Individuals by SSD Severity Group who Reported a Comorbid Disorder

Severity group Comorbid Controls Recovered Mild SSD Mild–moderate Moderate SSD disorder count (%) SSD count (%) count (%) SSD count (%) count (%)

LI 6 (6.0) 21 (21.0) 9 (9.0) 9 (9.0) 55 (55.0) n = 100 RD 6 (11.3) 16 (30.2) 5 (9.4) 6 (11.3) 20 (37.7) n = 53 CAS 0 (0.0) 2 (4.8) 0 (0.0) 2 (4.8) 38 (90.5) n = 42

Note. CAS = childhood apraxia of speech; LI = language impairment; RD = reading disorder; SSD = speech sound disorder. Subtyping Using Endophenotypes 121

Table 5. Comparison of Groups Based on Comorbid Conditions of LI and RD and Performance on Endophenotype Factor Score Mean (SD)

Comorbid conditions

SSD alone SSD + LI SSD + No RD No RD SSD + RD LI + RD N = 67 N = 54 N = 7 N = 40 Fp

Oral motor skills n = 53 n = 33 n = 6 n = 37 0.04 .988 −.11 (5.7) −.14 (8.8) 0.68 (3.5) 0.26 (7.5) Speeded naming n = 53 n = 31 n = 5 n = 34 0.48 .695 −0.22 (31.1) −2.61 (31.6) 0.28 (20.9) 6.91 (40.8) Phonological awarenessab n = 54 n = 37 n = 6 n = 34 15.19 <.0001 0.66 (1.4) −0.20 (1.6) 1.88 (3.3) −1.77 (2.2) Phonological memoryac n = 67 n = 52 n = 7 n = 40 17.53 <.0001 1.23 (4.0) −2.49 (3.9) 1.56 (2.8) −4.13 (4.7) Vocabularyd n = 67 n = 54 n = 7 n = 40 13.89 <.0001 2.14 (4.9) −2.90 (5.0) −1.58 (4.3) −3.08 (5.1)

Note. LI = language impairment; RD = reading disorder; SSD = speech sound disorder. aSSD + LI + RD differs from SSD alone, SSD + LI, and SSD +RD. bSSD + LI differs from SSD +RD. cSSD + LI differs from SSD alone. dSSD + LI + RD differs from SSD alone. Cells contain number of observations and mean (standard deviation) for the factor score. In Table 5, we distinguish subtypes of SSD alone and SSD + LI as to whether or not they have comorbid RD. phonological memory, children with all three had significantly higher scores than the SSD + disorders again had significantly lower scores LI or CAS subtypes, the SSD only subtype had than children with SSD alone or SSD + RD and significantly higher scores than the SSD + LI SSD + LI; and, in this case, children with SSD and CAS subtypes, and the SSD + LI subtype + LI had significantly lower scores than chil- had significantly higher scores than CAS sub- dren with SSD alone. Finally, for vocabulary, type (p < .05). Thus, all four clinical subtypes children with all three disorders scored signif- were distinguishable from each other using icantly lower than children with SSD only. the phonological memory factor. Finally, for vocabulary, children with no SSD had signifi- Clinical subtypes and associations cantly higher scores than the SSD + LI or CAS with endophenotypes subtypes, and the SSD subtypes had signifi- + Comparison of the clinical subtypes on fac- cantly higher scores than the SSD LI or CAS < tors scores revealed differences in phonologi- subtypes (p .05). cal awareness, F = 11.084, p < .0001; phono- logical memory, F = 35.443, p < .0001; and DISCUSSION vocabulary, F = 18.344, p < .0001 (Table 6). For phonological awareness, the CAS subtype The goal of the present study was to de- had significantly lower scores than both chil- termine whether endophenotypes of oral mo- dren with no SSD and the SSD only subtype, tor skills, phonological awareness, phonolog- and the SSD + LI subtype had lower scores ical memory, speeded naming, and vocab- than the SSD alone subtype (p < .05). For ulary distinguish children grouped accord- phonological memory, children with no SSD ing to common clinical classifications or 122 TOPICS IN LANGUAGE DISORDERS/APRIL–JUNE 2011

Table 6. Comparison of Clinical Subtypes of SSD and Their Performance on Endophenotype Factor Score Mean (SD)

Clinical subtypes

Endophenotype No SSD SSD alone SSD + LI CAS measures N = 69 N = 73 N = 53 N = 42 Fp

Oral motor skills n = 58 n = 58 n = 42 n = 29 0.864 .461 N = 187 −1.05 (8.6) 0.23 (5.2) −1.12 (7.36) 1.26 (9.3) Speeded naming n = 52 n = 57 n = 39 n = 27 1.015 .387 N = 175 −2.57 (27.9) −0.23 (30.5) −2.66 (32.2) 9.64 (41.4) Phonological awarenessab n = 61 n = 59 n = 45 n = 27 11.34 <.0001 N = 192 0.25 (2.0) 0.80 (1.7) −0.58 (2.0) −1.56 (−.02) Phonological memorycd n = 68 n = 73 n = 52 n = 41 35.44 <.0001 N = 234 2.93 (4.2) 1.29 (3.9) −2.18 (4.7) −4.44 (3.23) Vocabularyc n = 69 n = 73 n = 53 n = 42 18.34 <.0001 N = 237 1.46 (4.7) 1.83 (5.0) −2.05 (5.2) −4.11 (4.5)

Note. CAS = childhood apraxia of speech; LI = language impairment; SSD = speech sound disorders. aCAS differs from no SSD and SSD alone. bSSD + LI differs from SSD alone. cCAS and SSD + LI differs from no SSD and SSD alone. dCAS differs from SSD + LI. Cells contain number of observations and mean (standard deviation) for the factor score. Table 6 represents clinical subtypes that we have previously reported. In this table we distinguish children with CAS from children with SSD alone or SSD + LI. All children with CAS except for one, have comorbid LI. dimensions (severity, comorbid conditions, RD and clinical diagnostic categories of SSD clinical subtype) and thus provide informa- alone, SSD + LI, and CAS. All three clini- tion on underlying basis of these clinical dis- cal classifications were associated with differ- tinctions. Three endophenotypes, phonolog- ences in phonological memory. The results ical memory, phonological awareness, and indicate that difficulty in holding onto speech vocabulary, were associated with severity of sounds in memory long enough to form robust SSD, the presence of the comorbid conditions phonological representations is an important of LI and RD, and our previously reported clin- aspect of SSD in young children (Baddeley, ical subtypes of SSD alone, SSD + LI, and CAS. 1998; Preston & Edwards, 2010). Kenney, Findings supported a Multiple Deficit Model Barac-Cikoja, Finnegan, Jeffries, and Ludlow proposed by Pennington (2006) and others (2006) suggest that this memory deficit may (McGrath et al., 2007, in press; Pennington & persist into adulthood for individuals with his- Bishop, 2009; Peterson, Pennington, Shriberg, tories of SSD. Deficits in phonological short- & Boada, 2009). term memory also have been linked to RD and These findings provide insights into the LI (Kamhi & Catts, 1986; Kamhi, Catts, Mauer, core deficits associated with SSD and comor- Apel, & Gentry, 1988). bid conditions of LI and RD, as well as sup- A second endophenotype, phonological port for current clinical diagnostic categories. awareness, has been proposed as a skill that The endophenotypes of phonological mem- underlies SSD, LI, and RD (Peterson, Pen- ory, phonological awareness, and vocabulary nington, Shriberg, & Boada, 2009; McGrath best distinguished subtypes based on sever- et al., in press). Phonological awareness dis- ity of SSD, comorbid conditions of LI and tinguished groups based on both the presence Subtyping Using Endophenotypes 123 of comorbid LI or RD and our previous clinical disorders (SSD + LI + RD) had poorer vocabu- classifications of children into subgroups with lary skills than the other three groups. Vocab- SSD only, SSD + LI, and CAS. Children with ulary may be related to phonological memory. SSD + LI and those with CAS had lower scores The learning of new words requires that the on our measure of phonological awareness child hold the phonological forms for the new than children with SSD alone or no SSD. This word in while a phonologi- is contrary to the findings of Bird, Bishop, and cal template for the word is created and stored Freeman (1995) who reported that children (Baddeley, Gathercole, & Papagno, 1998). with expressive phonology disorders scored Poor phonological memory and vocabulary poorly on phonological awareness whether may result in poor language abilities that im- or not they had additional LI. One possible pact reading and spelling skills. explanation for this discrepancy is the age The two other endophenotypes, oral motor of the cohort that we examined. Phonolog- skills and speeded naming, were not related ical awareness is a skill that emerges from to severity of SSD, patterns of comorbidity, or 3 to 4 years of age with some skills such as clinical subtypes. Speech sound disorder of phoneme segmentation not established until unknown origin has historically been viewed 6 to 7 years of age (McLeod, 2009). Specific as a problem in the motor execution of ar- phonological awareness tasks might be diffi- ticulatory movements (functional articulation cult for all children with SSD in the age range disorders). General deficits in motor perfor- of our study (4–7 years) but may be more dis- mance may be an indicator of an underlying criminating at older ages when most typically neurodevelopmental immaturity rather than developing children have mastered phonolog- a true motor disorder. This neurodevelop- ical awareness. Comparisons of children ex- mental immaturity may explain the high rates hibiting different patterns of comorbidity re- of comorbidity of SSD, LI, RD, with motor vealed that children with co-occurring LI or incoordination (Bishop, 2002; Hill, 2001, RD had poorer phonological awareness skills Visser, 2003). More recently poor oral motor than those with SSD alone. Interestingly, in skills have been considered characteristic this study, children with SSD + LI performed of CAS or childhood dysarthria with most more poorly on tests of phonological aware- children with speech delay demonstrating ness than children with SSD + RD. This may normal motor skills (Shriberg, 2010). It is be due to the small number of children in the possible that oral motor skills may not vary SSD + RD group (n = 6),suggestingthatitis systematically with the clinical dimensions unusual to find children with SSD + RD with- examined here, though may be useful in out comorbid LI. In addition, children with characterizing other dimensions of SSD, such SSD + RD and no LI may differ in the underly- as the age of resolution of speech disorder ing etiology of the RD. For example, Rvachew or the type of speech therapy that is most & Grawburg (2006) reported that both speech effective in treating the disorder. perception and vocabulary predict phonolog- Rapid Naming of Colors (also known ical awareness skills. Children with SSD + RD as rapid serial naming) draws on articula- may have deficits in speech perception that tory, vocabulary, and cognitive skills and were not examined in this study. has been related to RD, LI and attention- Another endophenotype that distinguished deficit/hyperactivity disorder (Peterson et al., clinical dimensions of SSD was vocabulary. 2009). In this study, however, speeded nam- As expected, children with moderate SSD ing was not associated with differences in performed more poorly than children with severity, patterns of comorbidity, or clini- mild or mild to moderate SSD on vocabu- cal subtypes. Lahey, Edwards, and Munson lary. Children with CAS and those with SSD (2001) also failed to find a relation of pro- + LI also performed more poorly than chil- cessing speed to the severity of SLI. Similar dren with SSD alone. Children with all three results are reported by Raitano et al. (2004) in 124 TOPICS IN LANGUAGE DISORDERS/APRIL–JUNE 2011 children with SSD and by Young et al. (2002) vary across development. As children in the in adults with histories of SSD. In contrast, no SSD group were typically developing sib- Peterson et al. (2009) found that children lings of the participants with SSD, with some with SSD with poor speeded naming skills sharing of genetic risks for SSD, it may also be were more likely to have RD. Thus, reports useful in future studies to include an unrelated of deficits in speeded naming in children with control group. A further limitation is that the SSD are mixed, potentially because of the mul- study did not include a group of children with tiple motor and cognitive skills involved in LI alone. Such a group might help to disentan- this task. gle the relationships of the endophenotypes In summary, the results suggest that clinical to SSD versus LI. dimensions of SSD are related to a constella- Other limitations of the study are related tion of deficits rather than to singular impair- to measurement issues. Comorbid conditions ments and that comorbidities of SSD are best of LI, RD, and CAS were determined by par- explained in terms of the joint contribution ent report. Future studies may directly test for of multiple endophenotypes to the child’s these conditions to verify the accuracy of the disorder. These findings support the Multi- parent’s report. Some participants had miss- ple Deficit Model proposed by Pennington ing data because they were either too young (2006) that hypothesizes a multifactorial eti- for the test battery or unable to do the test, ology for complex developmental disorders as in the case of the phonological awareness such as SSD. In this model, whether or not measures. an individual presents with a disorder and co- morbid clinical impairments depends on the risk and protective factors shared among dis- CONCLUSIONS orders. Despite these limitations, this first study of Limitations and future directions the association of endophenotypes and sub- Study limitations include participant and types of SSD provides insight as to individ- measurement issues. The participants repre- ual differences in children with SSD, suggests sented an age range (4–7 years old) dur- approaches to refine assessment of these dis- ing which rapid development of speech and orders, and helps to guide the development language takes place. Significant differences of interventions that target core areas of defi- were observed in age between the recovered ciency. Furthermore, common underlying en- SSD group and the other SSD groups. Because dophenotypes for SSD, LI, and RD may have clinical classifications may change with age, common underlying genetic influences. Iden- longitudinal follow-up would be useful in de- tifying these genes and their role in neurolog- termining the extent to which the endophe- ical development will help us better under- notypic differences observed in this study stand the etiological bases of these disorders.

REFERENCES

Adams, M., & Gathercole, S. E. (1995). Phonological work- Childhood apraxia of speech (position statement). ing memory and speech production in preschool chil- Rockville, MD: Author. dren. Journal of Speech and Hearing Research, 38, Aram, D. M., & Hall, N. E. (1989). Longitudinal follow-up 403–414. of children with preschool communication disorders: American Speech-Language-Hearing Association (2004). Treatment implications. School Psychology Review, Preferred practice patterns for the profession of 18, 487–501. speech-language pathology [preferred practice pat- Baddeley, A. (1998). Recent developments in working terns]. Retrieved April 11, 2011, from http://www. memory. Current Opinion in Neurobiology, 8(2), asha.org/policy. 234–238. American Speech-Language-Hearing Association. (2007). Baddeley, A., Gathercole, S., & Papagno, C. (1998). The Subtyping Using Endophenotypes 125

phonological loop as a language learning device. Psy- Fisher, S. E., & DeFries, J. C. (2002). Developmental chological Review, 105, 158–173. : Genetic dissection of a complex cognitive Bernthal, J. E., Bankson, N. W., & Flipsen, P. (2009). Ar- trait. Nature Review Neuroscience, 3, 767–780. ticulation and phonological disorders: Speech sound Fletcher, J. (1977). Fletcher Time-by-Count Test of Di- disorders in children, sixth edition. New York: Pear- adochokinetic Syllable Rate. Tigard, OR: C.C. Publi- son. cations. Bird, J., Bishop, D. V. M., & Freeman, N. H. (1995). Flipsen, P., Hammer, J. B., & Yost, K. M. (2005). Phonological awareness and development in Measuring severity of involvement in speech delay: children with expressive phonological impairments. Segmental and whole-word measures. American Journal of Speech and Hearing Research, 31, 446– Journal of Speech-Language Pathology, 14, 462. 298–312. Bishop, D. V. (2002). Motor immaturity and spe- Gardner, M. F. (1990). Expressive One Word Picture Vo- cific speech and language impairment: Evidence cabulary Test-Revised. Novato, CA: Academic Ther- for a common genetic basis. American Journal of apy Publications. Medical Genetics, 114, 56–63. Gathercole, S. E., & Baddeley, A. D. (1990). Phonological Bishop, D. V., & Adams, C. A. (1990). A prospective memory deficits in language disordered children–Is study of the relationship between specific language there a causal connection Journal of Memory and impairment, phonological disorders and reading retar- Language, 29(3), 336–360. dation. Journal of Child Psychology and Psychiatry, Goldman, R., & Fristoe, M. (1986). The Goldman-Fristoe 31, 1027–1050. Test of Articulation. Circle Pines, MN: American Guid- Bishop, D. V., & Edmundson, A. (1987). Specific language ance Services. impairment as a maturational lag: evidence from lon- Gottesman, I. I., & Gould, T. D. (2003). The endophe- gitudinal data on language and motor development. notype concept in psychiatry: Etymology and strate- Developmental Medicine and Child Neurology, 29, gic intentions. American Journal of Psychiatry, 160, 442–459. 636–645. Bishop, D. V., & Snowling, M. J. (2004). Developmental Hill, E. L. (2001). Non-specific nature of specific language dyslexia and specific language impairment: Same or impairment: a review of the literature with regard to different? Psychological Bulletin, 130, 858–886. concomitant motor impairments. International Jour- Campbell, T. F., Dollaghan, C. A., Rockette, H. E., nal of Language and Communication Disorders, 36, Paradise, J. L., Feldman, H. M., Shriberg, L. D., et al. 149–171. (2003). Risk factors for speech delay of unknown ori- Inoue, K., & Lupski, J. R. (2003). Genetics and genomics gin in 3-year-old children. Child Development, 74, of behavioral and psychiatric disorders. Current Opin- 346–357. ion in Genetics Development, 13, 303–309. Castellanos, F. X., & Tannock, R. (2002). Neuroscience Kamhi, A. G., & Catts, H. W. (1986). Toward an under- of attention-deficit/hyperactivity disorder: the search standing of developmental language and reading dis- for endophenotypes. Nature Review Neuroscience, orders. Journal of Speech and Hearing Disorders, 51, 3, 617–628. 337–347. Catts, H. (1986). Speech production/phonological Kamhi, A. G., Catts, H. W., Mauer, D., Apel, K., deficits in reading disordered children. Journal of & Gentry, B. F. (1988). Phonological and spatial Learning Disabilities, 19, 504–508. processing abilities in language and reading impaired Colledge, E., Bishop, D. V., Dale, P. S., Happe, F. G., children. Journal of Speech and Hearing Disorders, Koeppen-Schomerus, G., Price, T. S., et al. (2002). 53(3), 316–327. The structure of language abilities at 4 years: A twin Kenney, M. K., Barac-Cikoja, D., Finnegan, K., Jeffries, study. Developmental Psychology, 38, 749–757. N., & Ludlow, C. L. (2006). Speech perception and Denckla, M. B., & Rudel, R. G. (1976). Naming of object- short-term memory deficits in persistent developmen- drawings by dyslexic and other learning disabled chil- tal speech disorder. Brain and Language, 96, 178– dren. Brain and Language, 3, 1–15. 190. Dodd, B. (Ed.) (2005). Differential diagnosis and treat- King, R. R., Jones, C., & Laskey, E. (1982). In retrospect: ment of children with speech disorder. (2nd ed.). A fifteen-year follow-up report of speech-language dis- Philadelphia: Whurr. ordered children. Language Speech and Hearing Ser- Dunn, L. M., & Dunn, L. M. (1997). Peabody Picture vices in Schools, 13, 24–32. Vocabulary Test–3rd ed. Circle Pines, MN: American Kovas,Y.,Hayiou-Thomas,M.E.,Oliver,B.,Dale,P.S., Guidance Services, Inc. Bishop, D. V., & Plomin, R. (2005). Genetic influences Elbro, C., Borstrom, I., & Peterson, D. K. (1998). Pre- in different aspects of language development: The eti- dicting dyslexia from kindergarten: The importance ology of language skills in 4.5-year-old twins. Child of distinctness of phonological representations of Development, 76, 632–651. lexical items. Reading Research Quarterly, 33(1), Lahey, M., Edwards, J., & Munson, B. (2001). Is process- 36–60. ing speed related to severity of language impairment? 126 TOPICS IN LANGUAGE DISORDERS/APRIL–JUNE 2011

Journal of Speech, Language and Hearing Research, Newcomer, P., & Hammill, D. (1997). Test of 44, 1354–1361. Language Development–Primary Third Edition Lewis, B. A., Freebairn, L. A., Hansen, A. J., Iyengar, S. (TOLD: P-3). Austin, TX: PRO-ED Inc. K., & Taylor, H. G. (2004). School-age follow-up of Panagos, J. (1974). Persistence of the open syllable re- children with childhood apraxia of speech. Language, interpreted as a symptom of a . Jour- Speech and Hearing Services in the Schools, 35, 122– nal of Speech and Hearing Disorders, 9, 23–31. 140. Pennington, B. F. (1997). Using genetics to dissect cog- Lewis, B. A., Freebairn, L. A., Hansen, A. J., Stein, C. M., nition. American Journal of Human Genetics, 60, Shriberg, L. D., Iyengar, S. K., et al. (2006). Dimen- 13–16. sions of early speech sound disorders: A factor an- Pennington, B. F. (2006). From single to multiple deficit alytic study. Journal of Communication Disorders, models of developmental disorders. Cognition, 101, 39,139–157. 385–413. Lewis, B. A., Freebairn, L. A., Hansen, A. J., Taylor, H. Pennington, B., & Bishop, D. V. M. (2009). Relations G., Iyengar, S., & Shriberg, L. D. (2004). Family pedi- among speech, language, and reading disorders. An- grees of children with suspected childhood apraxia nual Review of Psychology, 60, 283–306. of speech. Journal of Communication Disorders, Peterson, R. L., Pennington, B. F., Shriberg, L. D., & Boada, 37(2), 157–175. R. (2009). What influences literacy outcome in chil- Lewis, B. A., Freebairn, L. A., & Taylor, H. G. (2000). Aca- dren with speech sound disorder? Journal of Speech, demic outcomes in children with histories of speech Language, and Hearing Research, 52, 1175–1188. sound disorders. Journal of Communication Disor- Preston, J., & Edwards, M. L. (2010). Phonological aware- ders, 33, 11–30. ness and types of sound errors in preschoolers with Lewis, B. A., O’Donnell, B., Freebairn, L. A., & Taylor, H. speech sound disorders. Journal of Speech, Lan- G. (2002). Spoken language and written expression– guage, and Hearing Research, 53, 44–60. Interplays of delays. American Journal of Speech Lan- Raitano, N. A., Pennington, B. F., Tunick, R. A., Boada, guage Pathology, 7, 66–73. R, & Shriberg, L. D. (2004). Pre-literacy skills of sub- McGrath, L. M., Pennington, B. F., Shanahan, M. A., groups of children with speech sound disorders. J. Santerre-Lemmon, L. E., Barnard, H. D., Willcutt, E. Child Psychology and. Psychiatry, 45, 821–835. G., DeFries, J. C., & Olson, R. K. (2010, December Robbins, J., & Klee, T. (1987). Clinical assessment of 3). A multiple deficit model of and oropharyngeal motor development in young children. attention-deficit/hyperactivity disorder: Searching for Journal of Speech and Hearing Disorders, 52, 271– shared cognitive deficits. Journal of Child Psychology 277. and Psychiatry. (Epub ahead of print). Rvachew, S., & Grawberg, M. (2006). Correlates of McGrath, L. M., Pennington, B. F., Willcutt, E. phonological awareness in preschoolers with speech G., Boada, R., Shriberg, L. D., & Smith, S. D. sound disorders. Journal of Speech Language and (2007). Gene x environment interactions in speech Hearing Research, 49, 74–87. sound disorder predict language and preliteracy Semel, E., Wiig, E. H., & Secord, W. A. (1995). outcomes. Developmental Psychopathology, 19(4), Clinical Evaluation of Language Fundamentals–3rd 1047–1072. Edition (CELF-III). San Antonio, TX: The Psychologi- McLeod, S. (2009). Speech sound acquisition. In cal Corp. Bernthal, J. E., Bankson, N.W., & Flipsen, P. (Eds.). Ar- Shriberg, L. D. (1993). Four new speech and prosody- ticulation and phonological disorders: Speech sound voice measures for genetic research and other studies disorders in children (6th ed.). New York: Allyn & Ba- in developmental phonological disorders. Journal of con. Speech and Hearing Research, 36, 105–140. Metsala, J. L. (1999). Young children’s phonological Shriberg, L. D. (1994). Five subtypes of developmental awareness and nonword repetition as a function of vo- phonological disorders. Clinics in Communication cabulary development. Journal of Educational Psy- Disorders, 4, 38–53. chology, 91, 3–19. Shriberg, L. D. (2010). Childhood speech sound disor- Montgomery, J. W., Magimairaj, B. M., & Finney, M. C. ders: From postbehaviorism to the postgenomic era. (2010). Working memory and specific language im- In Paul, R, & Flipsen, P. (Eds.), Speech sound disor- pairment: An update on the relation and perspectives ders in children. San Diego: Plural Publishing. on assessment and treatment. American Journal of Shriberg, L. D., & Austin, D. (1998). The speech-language Speech-Language Pathology, 19, 78–94. connection (pp. 73–117). Baltimore, MD: Paul H Nathan, L., Stackhouse, J., Goulandris, N., & Snowling, Brookes. M. J. (2004). The development of early literacy Shriberg, L. D., Austin, D., Lewis, B. A., McSweeny, J. L., skills among children with speech difficulties: A & Wilson, D. L. (1997). The percentage of consonants test of the “critical age hypothesis.” Journal of correct (PCC) metric: Extensions and reliability data. Speech, Language, and Hearing Research, 47, Journal of Speech Language and Hearing Research, 377–391. 40, 708–722. Subtyping Using Endophenotypes 127

Shriberg, L. D., & Kwiatkowski, J. (1982). Phonolog- Stein, C. M., Schick, C. M., Taylor, J. H., Shriberg, ical disorders III: A procedure for assessing sever- G. H., Millard, L. D., Kundtz-Kluge, C., et al. ity of involvement. Journal of Speech and Hearing (2004). Pleiotropic effects of a chromosome 3 lo- Disorders, 47, 256–270. cus on speech-sound disorder and reading. American Shriberg, L. D., & Kwiatkowski, J. (1988). A follow-up Journal of Human Genetics, 74, 283–297. study of children with phonologic disorders of un- Torgesen, J. K., & Wagner, R. K. (1992). Language abili- known origin. Journal of Speech and Hearing Disor- ties, reading acquisition, and developmental dyslexia: ders, 53, 144–155. Limitations and alternative views. Journal of Learn- Shriberg, L. D., & Kwiatkowski, J. (1994). Developmental ing Disabilities, 25, 577–581. phonological disorders I: A clinical profile. Journal of Visser, J. (2003). Developmental coordination disorder: Speech and Hearing Research, 37, 1100–1126. a review of research on subtypes and comorbidities. Shriberg, L. D., Kwiatkowski, J., Best, S., Hengst, J., Human Movement Science, 22(4–5),479–493. & Terselic-Weber, B. (1986). Characteristics of chil- Wagner, R., Torgesen, J., & Rashotte, C. A. (1999). Com- dren with phonologic disorders of unknown ori- prehensive test of phonological processing. San Anto- gin. Journal of Speech and Hearing Disorders, 51, nio, TX: Pro-Ed. 140–161. Wechsler, D. (1989). Wechsler Preschool and Primary Shriberg, L. D., Lewis, B. A., Tomblin, J. B., McSweeny, J. Scale of Intelligence, Revised. San Antonio, TX: The L., Karlsson, H. B., & Scheer, A. R. (2005).Toward di- Psychological Corporation. agnostic and phenotype markers for genetically trans- Wechsler, D. (1991). Wechsler Intelligence Scale for Chil- mitted speech delay. Journal of Speech, Language, dren, Third Edition. The Psychological Corporation, and Hearing Research, 48, 834–852. San Antonio, TX: The Psychological Corporation. Shriberg, L. D., Tomblin, J. B., & McSweeny, J. L. (1999). Windsor, J., & Hwang, M. (1999). Testing the generalized Prevalence of speech delay in 6-year-old children and slowing hypothesis in specific language impairment. comorbidity with language impairment. Journal of Journal of Speech, Language, and Hearing Research, Speech Language and Hearing Research, 42, 1461– 42, 1205–1218. 1481. Wise, J. C., Sevcik, R. A., Morris, R. D., Lovett, M. W., Sices, L., Taylor, H. G., Freebairn, L., Hansen, A., & & Wolf, M. (2007). The relationship among recep- Lewis, B. A. (2007). The relationship between speech tive and expressive vocabulary, listening comprehen- sound disorders and early literacy skills in preschool- sion, pre-reading skills, word identification skills, and aged children: Impact of comorbid language impair- reading comprehension by children with reading dis- ment. Journal of Developmental and Behavioral Pe- abilities. Journal of Speech, Language, and Hearing diatrics, 28(6), 438–447. Research, 50, 1093–1109. Stackhouse, J., & Wells, B. (1997). Children’s speech and Young, A. R., Beitchman, A. R., Johnson, J. H., Atkinson, literacy difficulties. London: Whurr. C., Escobar, L., Douglas, M., et al. (2002). Young adult Stein, C. M., Lu, Q., Elston, R. C., Freebairn, L. A., Hansen, academic outcomes in a longitudinal sample of early A. J., Shriberg, L. D., et al. (2011). Heritability estima- identified language impaired and control children. tion for traits with developmental trajectories. Behav- Journal of Child Psychology and Psychiatry,43, 635– ior Genetics, 41(2).184–191. 645.