out of the mouths of babes - vocal production in infant siblings of children with asd

Out of the mouths of babes: vocal production ininfant siblings of children with ASD

Rhea Paul,1 Yael Fuerst,2 Gordon Ramsay,1,3 Kasia Chawarska,1 and Ami Klin1

1Yale Child Study Center, USA; 2Southern Connecticut State University, USA; 3Haskins Laboratories, USA

Background: Younger siblings of children with autism spectrum disorders (ASD) are at higher risk foracquiring these disorders than the general population. Language development is usually delayed inchildren with ASD. The present study examines the development of pre-speech vocal behavior in infantsat risk for ASD due to the presence of an older sibling with the disorder. Methods: Infants at high risk(HR) for ASD and those at low risk, without a diagnosed sibling (LR), were seen at 6, 9, and 12 months aspart of a larger prospective study of risk for ASD in infant siblings. Standard clinical assessments wereadministered, and vocalization samples were collected during play with mother and a standard set oftoys. Infant vocal behavior was recorded and analyzed for consonant inventory, presence of canonicalsyllables, and of non-speech vocalizations, in a cross-sectional design. Children were seen again at 24months for provisional diagnosis. Results: Differences were seen between risk groups for certain vocalbehaviors. Differences in vocal production in the first year of life were associated with outcomes in termsof autistic symptomotology in the second year. Conclusions: Early vocal behavior is a sensitive indi-cator of heightened risk for autistic symptoms in infants with a family history of ASD.Keywords: Autism, speech, vocalization, infant siblings.

Autism spectrum disorders (ASD) are developmentaldisabilities marked by severe deficits in reciprocalsocial interaction and communication, as well asrepetitive and restricted patterns of interests andbehavior (APA, 1994; Volkmar, Lord, Bailey, Schultz,& Klin, 2004). Children who qualify for diagnoses ofASD in the second year of life consistently demon-strate significant delay in the development ofexpressive language (Chawarska, Klin, Paul, &Volkmar, 2007; Chawarska & Volkmar, 2005; Paul,Chawarska, Cicchetti, & Volkmar, 2008; Wetherby,Watt, Morgan, & Shumway, 2007). Other studies ofearly vocal behavior in infant siblings of childrenwith ASD (Landa & Garrett-Mayer, 2006; Landa,Holman, & Garrett-Mayer, 2007; Iverson & Wozniak,2007; Zwaigenbaum et al., 2005) have reported bothdelays on standard tests of expressive language, andless mature syllable production (Dawson, Osterling,Meltzoff, & Kuhl, 2000). In addition, several studieshave shown that preschoolers with ASD producehigher levels of non-speech-like, atypical vocaliza-tions than peers with typical development (Olleret al., 2010; Schoen, Paul, & Chawarska, 2010;Sheinkopf, Mundy, Oller, & Steffens, 2000; Wether-by et al., 2007, 2004).

Risk for ASD is known to be elevated in bothparents and siblings of affected children (Bailey etal., 1995; Dawson et al., 2007; Folstein & Rutter,1977, 1988; Landa et al., 1992; Piven, Palmer,Jacobi, Childress & Arndt, 1997; Santangelo, Fol-stein, & Tager-Flusberg, 1999). In addition, LeCouteur et al. (1996) found that, in twin pairs dis-

cordant for autism, concordance for languageimpairment and a broader autism phenotype (BAP)was much greater in monozygotic pairs thandizygotic pairs, consistent with a strong geneticcomponent. Although definitions are still evolving,Micali, Chakrabarti, and Fombonne (2004) present adefinition of BAP based on epidemiological find-ings, which includes social and communicationimpairments, as well as repetitive and stereotypedbehaviors. Social abnormalities have been found toconsist of impaired friendships, impaired social play,odd behavior and impaired conversation. Develop-mental language abnormalities are seen particularlyin pragmatic language areas (Landa et al., 1992).Obsessional and repetitive behaviors can includecircumscribed interests, rigidity, obsessions/compulsions and repetitive interests and activities.Micali et al. (2004) define BAP as a constellation ofthese abnormalities at a subthreshold level ofseverity that would qualify an individual for neitherASD nor another diagnosis, such as developmentaldelay or anxiety disorder.

Recent research on ASD has included ongoingmonitoring of infants with an older sibling with ASD.This research focus stems from the literaturereviewed above, which suggests that genetic factorsplay a significant role in the incidence of ASD (Baileyet al., 1995; Folstein & Rutter, 1988; Le Couteuret al., 1996; Rutter, 2005). In addition, awareness ofrisk for disorders associated with BAP, even in theabsence of the full-blown syndrome (Szatmari et al.,2000), contributes to this trend. Studies document-ing infant siblings’ development are aimed at earlieridentification of these potential problems, and earlierConflict of interest statement: No conflicts declared.

Journal of Child Psychology and Psychiatry 52:5 (2011), pp 588–598 doi:10.1111/j.1469-7610.2010.02332.x

� 2010 The Authors. Journal of Child Psychology and Psychiatry � 2010 Association for Child and Adolescent Mental Health.Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA 02148, USA

provision of treatment in order to maximize outcomepotential.

In the present report, we describe findings on earlyvocal behaviors in a group of infant siblings with thisrisk, and compare these behaviors to those of age-mates from low-risk families. In doing so, we wouldlike to argue that the direct and detailed analysis ofvocal behaviors in infants at risk for ASD will providericher information than scores on standard tests,such as Mullen (1995), or than retrospective parentreport of early vocal behavior. First, these analysescan tell us not only that high-risk infants are delayedrelative to peers, but also in precisely what aspects ofpre-speech behavior, whether it be the acquisition ofspecific consonants, syllable shapes, or prosodiccontours. This level of detail may be illuminating inbetter understanding how the path to spoken lan-guage is affected in high-risk infants. Second,although parents are sensitive to certain milestonesof speech development (first canonical syllables[Oller, Eilers, Neal, & Cobo-Lewis, 1998], firstwords), they are not generally attuned to the specificconsonants their children are using or the relativedistribution of consonant and syllable types in infantproduction. Finally, retrospective parent reports,which have been used in other studies of earlyspeech development (e.g., Gernsbacher, Sauer,Geye, Schweigert, & Goldsmith, 2008) are subject toerrors of recall and bias. By directly examining thematurity of prelinguistic vocalizations, we test thehypothesis that infant siblings of children with ASDwill show precursors to delays in expressivelanguage that are consistent with the development ofASD or BAP at higher rates than will low-risk infants,and we examine the relation of these delays in thefirst year of life to the presence of autistic symp-tomatology in the second year.

Vocal production in typical development

Oller, Eilers, Neal, and Schwartz (1999), Stark(1980), Stark, Bernstein, and Demorest (1993), andLocke (1993) have described the course of vocaldevelopment in the first year of life. Major milestonesinclude:

• The appearance of first consonant-like produc-tions in vocal play at 4–6 mo.

• The predictable expansion of the consonant rep-ertoire; Shriberg (1993) classified the consonants

of English into three groups, based on their typicalorder of acquisition in the first year (see Table 1).

• The emergence of canonical babble at 6–10 mo.;i.e., the use of consonant-vowel syllables withspeech-like voice quality (Oller et al., 1999).

• The production of closed syllables (conso-nant-vowel-consonant) by the end of the first yearof life (Levitt & Utman, 1992).

Studies of typically developing children (e.g., Oller,Wieman, Doyle, & Ross, 1976; Schwartz & Leonard,1982; Stoel-Gammon & Cooper, 1984; Vihman,Ferguson, & Elbert, 1986) have shown that canoni-cal babble in the second half of the first year of life isgenerally continuous with the development of firstsounds and words in speech, and babbling behaviorat the end of the first year predicts later speech andlanguage production in children with both typicaldevelopment (Jensen, Boggild-Andersen, Schmidt,Ankerhus, & Hansen, 1988; Leonard, Schwartz,Morris, & Chapman, 1981; Menyuk, Liebergott, &Schultz, 1986; Stoel-Gammon, 1998; Storkel &Morrisette, 2002; Vihman & Greenlee, 1987) andthose with delayed language (Paul & Jennings, 1992;Roberts, Rescorla, Giroux, & Stevens, 1998; Thal,Oroz, & McCaw, 1995). Thus investigating thedevelopment of pre-speech vocalization in the latterhalf of the first year may provide some insight intothe transition to language in this population.

In the present report, we examine the forms of bothspeech-like and non-speech vocalizations in samplesof infants at high and low risk for ASD, seen at 6, 9,and 12 months of age, using a cross-sectionaldesign. In this way we aim to investigate the vocaltools children at risk for ASD are bringing to the taskof learning to communicate. The following hypothe-ses will be tested:

1. HR infants will produce fewer consonants andcanonical syllable shapes, and more non-speechvocal behavior than LR peers.

2. Less frequent and mature pre-speech vocalbehavior in HR infants during the first year of lifewill be associated with greater likelihood of theappearance of autistic symptomotology in thesecond year.

MethodParticipants

Infants were seen as part of a longitudinal study of thedevelopment of children at high and low risk for ASD.The study was approved by the Institutional ReviewBoard of the Yale School of Medicine for human subjectresearch, and all families signed informed consentsprior to participation. Participants were assigned to theHigh Risk (HR) group if a full sibling had received adiagnosis of ASD by an experienced professional ormultidisciplinary team. Many of the proband siblings’families had participated in research on toddlers withASD at our center, had been diagnosed by our team and

Table 1 Shriberg’s (1993) classification of developmentalorder and approximate age of emergence* of consonantacquisition in English

Early 8: /m/ /b/ /j/ /n/ /w/ /d/ /p/ /h/Middle 8: /t/ /g/ /k/ /g/ /f/ /v/ /tS/ /dZ/Late 8: /S/ /h/ /s/ /z/ /ð/ /æ/ /Z/ /r/

*Based on Stoel-Gammon (1998); Locke (1989, 1993).

Vocalizations in infant siblings 589

� 2010 The Authors. Journal of Child Psychology and Psychiatry � 2010 Association for Child and Adolescent Mental Health.

were engaged in ongoing follow-up. Other proband oldersiblings’ families were referred through ongoing rela-tionships with local pediatric practices, local autismadvocacy organizations, and word of mouth from fami-lies already participating in our HR infant follow-up.Diagnoses of older siblings were evaluated by review ofclinical reports of previous evaluations and confirmedby means of the Autism Diagnostic Interview-Revised(ADI-R; Lord, Rutter, & Le Couteur, 1994) conductedwith parents by trained clinicians. Infants wereassigned to the Low Risk (LR) group if no sibling in thefamily had received a diagnosis of autism or any otherdevelopmental disorder. Families for this group wererecruited from local pediatric practices, advertisementsin parenting media, and word of mouth.

Exclusionary criteria for both HR and LR groups in-cluded less than 32 weeks’ gestation, known sensorydeficits such as significant vision or hearing loss,known genetic syndrome, or known neurological orsignificant chronic medical disorder.

For the longitudinal study from which the presentstudy draws data, families are invited for multiple visitsduring the infants’ first year and for follow-up at 24 and36 mo. During these visits, a variety of experimentalprocedures on visual and auditory processing wereadministered. From 6 months onward, visits alsoincluded clinical characterization measures.

Participants in the present report include those whoparticipated in vocalization sample collection at the6- (HR: n = 28; LR: n = 20), 9- (HR: n = 37; LR: n = 29),and/or 12- (HR: n = 38; LR: n = 31) month visit. Inaddition, a subgroup of these who had contributedvocalization samples have by now reached age 2 andparticipated in a 24-month follow-up (HR: n = 24; LR: n =21). The remaining children had not reached 24 monthsat the time of this writing. It should be noted thatalthough some subjects participated in more than onevisit, not all families attended every visit. The number offamilies who participated in all four visits (n = 14) was toosmall to allow for treating the sample longitudinally;therefore data were analyzed cross-sectionally.

Participants who attended the 24-month visitreceived a consensus provisional diagnosis based onclinical observations and standard assessments con-ducted by experienced clinicians blind to the partici-pants’ risk status. Clinicians examined all informationcollected at the visit and conferred one of the followingconsensus provisional diagnoses: ASD, broad autismphenotype (BAP), non-autistic developmental delay(DD), or no clinical diagnosis (i.e., typical development).

Procedures

Standard measures. Standard characterizationmeasures included the Autism Diagnostic ObservationScale–Toddler Module (ADOS-T; Lord, Rutter, DiLavore,& Risi, 2002; Luyster et al., 2009) and theMullen Scalesof Early Learning (MSEL; Mullen, 1995). It should benoted that at each visit at which the MSEL wasadministered (6, 12, and 24 mo.), there was no signifi-cant difference between the HR and LR groups in termsof the Visual Reception scale. Although the HR group’sscores tended to be lower, as Table 2 shows, the groupswere statistically equivalent in terms of this measure ofnonverbal cognition. T

able

2Mean(ands.d.)MSEL1T-scoresatth

reevisitsforhigh-vs.low-riskgroups

6Month

s12Month

s24Month

s

VR2

*FM

3GM

4RL5

*EL6

VR2

FM

3GM

4*RL5

*EL6

VR2

FM

3GM

4*RL5

EL6

HR

43.5

(8.7)

40.1

(7.9)

44.2

(8.7)

43.6

(9.9)

39.0

(5.3)

53.4

(10.4)

57.0

(10.6)

45.5

(11.3)

41.1

(9.7)

40.8

(10.7)

54.9

(12.6)

52.9

(11.9)

47.3

(4.9)

53.0

(9.8)

58.6

(15.2)

LR

46.4

(8.6)

45.4

(7.2)

44.8

(7.2)

46.5

(8.2)

42.2

(5.8)

58.1

(9.4)

56.3

(10.7)

45.0

(14.0)

46.1

(7.2)

47.4

(12.0)

59.7

(11.4)

54.9

(6.9)

49.5

(7.5)

61.6

(10.2)

54.7

(13.0)

HR

=HighRiskgroup;LR

=Low

Riskgroup.

1MullenSca

lesofEarlyLearn

ing(M

ullen,1995)T-scores(m

ean

50;s.d.,

10);

2VisualReception

(nonverb

alproblem

solving)Scale;3FineMotorScale;4GrossMotorScale;5Receptive

LanguageScale;6ExpressiveLanguageScale.

*Signifi

cantdiffere

ncebetw

een

High-andLow-R

iskgro

upsatp<.05.

590 Rhea Paul et al.


Provisional clinical diagnosis. Provisional diagno-ses were conferred at the 24-month visit by consensusjudgment from two highly experienced clinicians. Theaverage ADOS-T total scores for the HR children whoreceived diagnoses of ASD, BAP, or no ASD symptomsappear in Figure 1. Clinical diagnoses were alwaysarrived at by consensus among experienced clinicians.

ASD. Provisional diagnosis of ASD was based on Cha-warska et al.’s (Chawarska, Klin, Paul, Macari, &Volkmar, 2009) model employing data both from theADOS-T and the pattern of MSEL (Mullen, 1995) verbaland nonverbal scores, as well as clinical observationand judgment. HR children receiving this diagnosis hada significant number of points on the ADOS-T (at least5), had a pattern of verbal and nonverbal performanceon the MSEL suggestive of that seen in toddlers withASD (i.e., poorer verbal than nonverbal skills, poorerreceptive than expressive language scores on theMSEL), and qualitatively demonstrated more severe,more atypical, and more consistently aberrant socialbehavior than children in the other diagnostic groups.

BAP. HR children receiving a provisional diagnosis ofBAP also evidenced some elevation on ADOS-T scores(all those diagnosed with BAP had ADOS-T scores of atleast 8). However, the abnormalities in their socialinteractions, the severity of their atypicality and theconsistency of their aberrant behaviors were less pro-nounced than those diagnosed with ASD. Clinicaljudgment based on extensive observation and review ofparental report was used to assess this. Childrenassigned to the BAP category also did not exhibit thecharacteristic pattern of MSEL scores seen in ASD (i.e.,Visual Reception>Express Language>Receptive Lan-guage). Although some children who were diagnosed asBAP had atypical language behaviors, such as echolaliaor scripting, none had language delay alone, in theabsence of other social difficulties, nor did any languagedelays they showed qualify them for a diagnosis ofdelayed language development, according to Connecti-cut Birth-to-Three criteria (i.e., more than one standarddeviation below the mean in both expressive and

receptive language, or more than two s.d. below themean in either one).

No ASD symptoms. Children received no diagnosis (i.e.,typical development) if their social behavior was judgedto be within the typical range for age (ADOS-T scoresranged from 3 to 9), and they showed no significantdelays on the MSEL. Developmental delay (DD) wasconferred when a child scored more than one standarddeviation below average on at least one scale of theMSEL, but showed no significant social deficits onobservation or by parent report. Only one HR child inthis cohort was assigned to this category.

Vocalization sample collection. Samples were col-lected from a timed parent–child interaction. Parentswere asked to play and interact with the child with astandard set of toys for five minutes, with no warm-upperiod. Vocalization samples were recorded using aShure SM58 omni-directional microphone placed on astand on the floor close to the child. A Marantz CDR300compact disk recorder captured the vocalizations ondigital media.

Transcription. Infant vocalizations were separatedinto two categories, speech-like or non-speech. Thespeech-like category included vocalizations character-ized by the production of consonants and/or vowelsthat could be represented by phonetic symbols andcontained speech-like vocal quality such that theyresembled typical babble. The non-speech categoryincluded vocalizations characterized by the productionof non-speech resonance and vocal quality (e.g., squeal,yells, growls) without recognizable consonants.

The first 50 speech-like vocalizations produced weretranscribed using broad phonemic transcription. Forparticipants who did not produce 50 speech-likevocalizations within the time frame, all speech-likevocalizations produced were transcribed. It should benoted that at 6 mo. approximately 95% of participantsin both risk groups produced fewer than 50 speech-likevocalizations; at 9 months this figure was approxi-mately 90%; and at 12 months, approximately 80%.The same five-minute segments transcribed for speech-like vocalizations were also coded for non-speechvocalizations. Vegetative sounds including coughing orburping were not included in analyses of either speech-like or non-speech vocalizations.

Coding. Audiofiles were edited so that speech-like andnon-speech productions were copied onto separate filesand stored. Coders were blind to participants’ age andrisk group assignment.

Frequency of vocalization. Frequency of speech-likevocalizations (up to 50) was tallied for each participant.All non-speech vocalizations occurring within the sametime period as the speech-like sample were tallied. Non-speech vocalizations were separated into utterancesbased on breath groups or a pause of greater than onesecond. Total number of vocal productions for eachparticipant was derived by summing the number ofspeech-like and non-speech vocalizations. The twovocalization types were then coded separately. Average

ADOS-T Total Algorithm Score

0

2

4

6

8

10

12

14

16

18

No Dx ASD BAP

Diagnosis

Mea

n A

DO

S-T

Scor

e

Figure 1 ADOS-T total algorithm scores for three groups of HRinfants at 24 months. *Bars represent one standard deviationaround the group mean



frequencies of each vocalization type for each risk groupat each age appear in Table 3.

Coding of speech-like vocalizations. Any utterancethat could not be confidently transcribed after fourplaybacks was eliminated. The following coding wascompleted for each participant:

Consonant inventory. Inventories of singleton conso-nants were assembled for each participant. Consonantswere then divided into three categories as outlined byShriberg (1993; see Table 1). The total number of con-sonant types and the number of consonant types ineach of the three developmental categories were com-puted for each participant.

Percent canonical syllables. The number of syllablescontaining the basic Consonant–Vowel (CV) structureassociated with true babbling (Oller et al., 1999) wastallied for each participant and divided by the totalnumber of speech-like vocalizations to derive this met-ric.

Non-speech coding. Proportion of non-speech pro-duction. The percentage of non-speech vocalizationswas computed for each participant by counting thenumber of non-speech vocalizations and dividing by thesum of the number of speech-like plus non-speechvocalizations produced. The following categories of non-speech vocalizations were used to identify non-speechproductions (Sheinkopf et al., 2000):

1. Delight: Laughing or giggling.2. Distress: Crying, whining or fussing.3. Atypical: High-pitched squeals, low-pitched growls,

yells, grunts.

Reliability. Two trained raters, blind to subject ageand risk status, independently scored a 10% randomlyselected sample of the vocalizations on all coding cate-gories. Point-to-point reliability exceeded 80% on allcoding categories.

Analytic strategy

MSEL scores were examined at each age-level for whichdata had been collected (6, 12, and 24 mo.), and one-way analysis of variance at each age was carried out to

compare scores on the five MSEL subtests between thetwo risk groups. These data appear in Table 2. Two-wayanalysis of variance, with age (6, 9, 12 mo.) and riskgroup (HR, LR) as independent variables, was used toanalyze data on each of the eight vocalization behaviorsstudied (total # vocalizations, # speech-like vocaliza-tions, # consonants, # early consonants, # middleconsonants, # late consonants, % canonical syllables,% non-speech vocalizations). Main effects for age andrisk group, as well as the interaction of age and riskgroup, were examined. In cases where significantinteraction effects were found, pair-wise comparisonsbetween risk groups at each age-level (6, 9, and 12 mo.)were conducted, and significant pair-wise differences,along with effect sizes (Cohen, 1988), are reported. Datainput to these analyses are presented in Table 3.

Correlations between MSEL scores and vocal pro-ductions were also examined within the HR group at thetwo ages for which MSEL scores were available (6 and12 mo.). Discriminant function analysis was then usedto predict placement in groups with and withoutautistic symptoms at a 24-month follow-up visit for theHR group only (since prediction of outcome for HRchildren was the question of interest), using the cate-gories of vocal production on which risk group differ-ences were found in the first year data.

ResultsTotal frequency of vocalizations

ANOVA indicated that there was a main effect for age(F(2) = 3.8; p = .02), but not for risk, and no signifi-cant age · risk interaction for total number ofvocalizations. These findings suggest that all infantsproduced more vocalizations as they got older, butthere was no significant difference between riskgroups in frequency of vocal production, when bothspeech and non-speech vocalizations were com-bined. Thus, differences between groups cannot beattributed to less frequent vocal behavior on the partof the HR group.

Frequency of speech-like vocalizations

There was a significant main effect for age (F(2) = 8.3;p = .0003) and risk group (F(1) = 7.9; p = .006) but nosignificant interaction for frequency of speech-like

Table 3 Mean (and s.d.) vocal productions in two risk groups

AgeRiskgroup

# Totalvocalizations(speech +

non-speech)# Speech-likevocalizations

Total #consonants

# Earlyconsonants

# Middleconsonants

# Lateconsonants

% Canonicalsyllables

% Non-speechproductions

6 mo. HR 36.5 (27.4) 15.0 (16.0) 2.5 (2.9) 1.5 (1.7) .5 (1.1) .5 (.9) 4.5 (7.0) 60.2 (28.6)LR 39.2 (27.8) 18.3 (15.3) 3.7 (3.9) 2.1 (2.2) 1.1 (1.4) .5 (.9) 8.3 (11.5) 57.0 (25.2

9 mo. HR 25.7 (19.6) 15.1 (16.2) 2.7 (3.0) 1.9 (1.7) .5 (.9) .4 (.8) 8.7 (11.5) 50.6 (31.7)LR 31.1 (18.6) 22.6 (16.4) 5.6 (3.9) 3.2 (2.1) 1.3 (1.4) 1.1 (1.1) 20.5 (17.6) 34.0 (24.2)

12 mo. HR 35.2 (18.1) 12.6 (16.8) 6.26 (3.6) 3.8 (2.2) 1.6 (1.2) .8 (.9) 27.1 (22.9) 36.5 (28.4)LR 39.1 (18.1) 31.9 (15.7) 7.3 (3.7) 4.5 (2.1) 1.8 (1.5) 1.0 (1.1) 28.8 (19.0) 19.3 (16.1)

HR = High Risk group; LR = Low Risk group.Significant difference between groups at p < .05.



vocalization. This main effect suggests that speech-like productions increased with age for all infants,and the HR children produced fewer speech-likevocalization at all ages.

Percent canonical syllables

There were significant main effects for age (F(2) =29.6; p < .0001) and risk group (F(1) = 6.0; p = .001)and a significant age · risk interaction in terms of theproportion of speech-like productions that consistedof canonical syllables. Main effects suggest thatcanonical syllable production increased with ageoverall, and that there is a difference in canonicalsyllable production between risk groups. Exploringthe interaction effect with post-hoc pair-wise com-parisons, using the Bonferroni correction reveals asignificant difference between risk groups at the9-month level only (p = .04; Cohen’s d = .79) with amedium to large effect size.

Consonant inventories

Significantmain effectswere found for both age (F(2) =20.1; p < .0001) and risk group (F(1) = 11.4; p = .0009)in terms of the total number of consonants produced.All infants produced more consonants as they gotolder, and those in the LR group produced moreconsonants at each age than those in the HR group.

Consonant classes. To look in more detail at theproduction of consonants in this sample, we dividedthe consonants produced into the three developmen-tal classes presented by Shriberg (1993; see Table 1).

Results for each of the three classes were similar.In each case, there were significant main effects forage (Early C: F(2) = 24.1; p < .0001; Middle C: F(2) =11.2; p < .0001; Late C: F(2) = 3.7; p = .03) and riskgroup (Early C: F(1) = 8.7; p = .004; Middle C: F(1) =9.0; p = .003; Late C: F(1) = 5.4; p = .02). The inter-action effect for late consonants approached signifi-cance at p < .09. Pair-wise comparisons revealedsignificant differences between groups at 9 months(p < .03; d = .78, large) on this variable.

Non-speech vocalizations

The proportion of non-speech production out of totalvocalizations also showed main effects for age (F(2) =20.1; p < .0001) and risk status (F(1) = 11.0; p = .001)with no significant interaction. Both groupsdecreased the proportion of non-speech productionwith age, but the HR infants tended to produce morenon-speech at each age.

Correlations between MSEL scores and vocalproduction in HR infants

At 6 months there were no significant correlationsbetween any of the MSEL scores and vocal produc-

tion data, suggesting that vocal production issomewhat independent of developmental level at thisage. At 12 months, there were no significant corre-lations between MSEL Gross Motor, Visual Recep-tion, or Receptive Language scores and any of thevocalization variables. There was a significant cor-relation, however, between the number of Earlyconsonants produced and both MSEL Fine Motor (r =.43, p = .007) and Expressive Language (r = .49, p =.002) scores. Early consonants were still the mostfrequent consonant class at 12 months, and thecorrelations between these two MSEL scores and thetotal number of consonants produced at 12 monthswas also marginally significant (r = .38, p = .018; andr = .40, p = .014, respectively). No other correlationsreached significance at p < .01.

Follow-up data at 24 months

Provisonal diagnostic assignments. As of thiswriting, 24 HR and 21 LR participants from whomvocalization data were collected have reached theirsecond birthday and have been seen for follow-upevaluations at 24 months of age. One child of the 21in the LR risk who participated in the 24-month visitwas given a provisional ASD diagnosis; 2 othersreceived provisional diagnoses of other, non-autisticdevelopmental delay (it should be noted that parentsof LR children who chose to participate in this studymay have done so because of underlying concernsabout their child’s development or behavior; 7% ofparents of LR children who enrolled in the study at 6mo. of age indicated that they had some develop-mental concerns about their infants). In the HRgroup, 7 (29%) received a provisional diagnosis ofASD; 6 (25%) others were seen to exhibit somesymptoms without meeting full clinical criteria for aprovisional diagnosis of BAP; 1 (4%) showed non-autistic developmental delay; and 10 (42%) wereconsidered to merit no clinical diagnosis. Discrimi-nant function analyses included only HR children,since this is the group for which we aimed to dis-criminate outcomes on the basis of early vocal data.

It should be noted that diagnoses for high-riskinfants at 24 months must be considered provi-sional. Although several research groups (e.g., Cha-warska et al., 2009; Cox et al., 1999; Lord, 1995;Turner & Stone, 2007) have reported relative stabil-ity of ASD diagnoses in 2-year-olds in the generalpopulation, HR infants show more varying symp-toms (Ozonoff et al., 2009; Rogers, 2009). Thus thefact that nearly a third of HR infants appear to have adiagnosis of ASD at 24 months, and another quarterappear to show BAP, does not mean that thesediagnoses will inevitably remain stable. Moreover,this figure is not unlike findings of other studies ofHR siblings seen for assessment at 24 months. Forexample, both Zwaigenbaum et al. (2005) and Sul-livan et al. (2007) found that 30% of an HR samplemet criteria for ASD at 24 months.



Discriminating outcomes in terms of autistic symp-toms. In order to determine whether any aspects ofprelinguistic vocal development were associated withthe appearance of autistic symptoms at the24-month visit, discriminant function analyses wereperformed. Participants in the HR group who tookpart in the 24-month visit were subdivided into twogroups: those with autistic symptoms (n = 14); i.e.,those to whom clinicians assigned a provisionaldiagnosis of either ASD or BAP; and those whoreceived no diagnosis or a diagnosis of non-autisticdevelopmental delay (n = 11). The seven vocal mea-sures on which risk group differences were found(# speech-like vocalizations, # consonants, # earlyconsonants, # middle consonants, # late conso-nants, % canonical syllables, % non-speech vocal-izations) were entered step-wise into each of thediscriminant function analysis.

Classification based on 6-month vocal data. Discrimi-nant function analysis (SPSS 16.0) revealed thatonly the number of middle consonant types producedby the 17 HR children seen at both 6 months and 24months was significant in classifying participantswith and without autistic symptoms at 24 mo., witha canonical correlation of .47 (p < .04). Seventy-fourpercent of subjects were correctly classified on thebasis of this variable. One-way ANOVA comparingthe 6 mo. MSEL scores of the children who did anddid not show symptoms of ASD at 24 monthsrevealed a significant difference (F(1, 21); p < .05; d =.88) in Expressive Language scores only, with chil-dren who later showed no symptoms scoring higher(Mean: 41.1; s.d., 3.8) than those who developedsymptoms (Mean: 36.4, s.d. 4.5).

Classification based on 9-month vocal data. Enteringthe same vocal measures collected at the 9-monthvisit in a step-wise fashion to predict placement forthe 19 HR children seen at both 9 months and at 24months revealed that only the number of late conso-

nant types produced at nine months was significantin classifying participants (canonical correlation =.53; p < .03), with 77% of subjects were correctlyclassified on the basis of this variable. No MSEL dataare available at this age.

Classification based on 12-month vocal data. Finally,entering the same vocal measures collected for the20 HR children seen at both 12 and 24 months in astep-wise fashion to predict placement in the groupwith or without autistic symptoms at 24 months,discriminant function analysis revealed that onlytotal number of different consonant types producedwas significant in classifying participants (canonicalcorrelation = .43; p < .03), with 65% of subjectscorrectly classified on the basis of this variable. One-way ANOVA comparing the 12 mo. MSEL scores ofthe children who did and did now show symptoms of

ASD at 24 months revealed no significant differenceson any of the MSEL scales.

DiscussionApart from showing that all infants progress with agein all the pre-speech variables studied, these dataalso allow us to address the hypotheses stated at theoutset. First, it does appear that infants at risk forASD show differences in some prelinguistic vocalbehaviors relative to LR peers; specifically, we foundthat:

• on average, HR infants produced significantlyfewer speech-like vocalizations and more non-speech vocalization than LR peers;

• on average, HR children produced significantlyfewer consonant types than LR peers;

• HR infants produced significantly fewer canonicalsyllable shapes than LR peers, particularly at 9months of age.

Second, differences in vocal production in the firstyear of life were associated with outcomes in terms ofautistic symptomatology in the second year for theHR group.

Differences seen in pre-speech vocal production ofHR infants cannot be ascribed to a general decre-ment in developmental level, as their MSEL VisualReception scores do not differ from those of LR peersat any time point measured. Rather, they appear toreflect a specific delay in acquiring age-appropriatesounds and syllable shapes in the first year, and intransitioning from the diverse range of sound pro-duction typical of all young babies to a morerestricted range dominated by speech-like sounds atthe first year’s end. These delays are captured notonly by the detailed analysis of sound productionreported here; they are hinted at by MSEL Expres-sive Language scores, which are significantly lowerthan those of the LR group throughout the first year,although these scores are within the normal range.Moreover, Expressive Language scores on the MSELcorrelate with vocal production of consonants at 12months. This is likely a reflection of the fact thatmost items used to score Expressive Language onthe MSEL at the 12-month level relate to consonantproduction. It is also noteworthy, however, that finemotor skills are also related to consonant productionat this age.

These findings can be interpreted to suggest thatpre-speech sound development is a sensitive indi-cator of the rate and degree to which infants at riskfor ASD are following the developmental path tolanguage acquisition and, by extension, to expres-sive communication. Motor delays cannot be ruledout as a source of the pattern seen here, especially inlight of the correlation between Fine Motor scores onthe MSEL and consonant production at 12 months.Despite this correlation, though, findings from theMSEL suggest that, apart from an apparent mild



delay in Fine Motor performance at 6 months, nosignificant motor delays persist into the second halfof the first year.

Does this detailed analysis of vocal production ininfants at risk for ASD contribute any informationbeyond what is measured by the more general MSELlanguage scales? Potential answers to this questionmay be seen in the changing pattern of relationshipbetween the vocal performance of the HR and LRgroups throughout the first year of life, as well as thefindings of the discriminant function analyses. Thatis, detailed description of vocal production revealsthat the areas of difference consistently reflect slowerprogress on the part of the HR infants in moving fromone phase of pre-speech development to the next. By9 months – when LR children are developing a rangeof tools necessary to begin the transition from pre-speech to first speech production – HR childrenproduce, on average, a lower rate of production ofcanonical syllables, the kind that will eventually beused as first words, and a smaller repertoire of con-sonants, particularly the Late group, those justemerging at this developmental level. By 12 months,the participants with HR, as a group, are not signif-icantly different from LR peers in production of theconsonant and syllable types their LR peers learnedsome months earlier (although they continue totrend toward less frequent production), but theycontinue, as they have throughout the first year, toproduce a higher proportion of other kinds of vocal-izations that do not contribute to their transition tofirst words. As we have argued previously (Schoen,Paul, Chawarska, Klin, & Volkmar, 2008), LRinfants, tend, toward the end of the first year of life,to match their productions more and more closely toambient speech models. HR infants, on the otherhand, show less winnowing of the other kinds ofsounds characteristic of the first year of life. Thusthe pattern of pre-speech development seen in theHR group suggests both slower transition from onephase of pre-speech development to the next, and areduced tendency to hone sound productionincreasingly closely to models produced by others inthe environment as the child nears the first birthday.This more detailed examination suggests that lowMSEL Expressive Language scores have differentsources at different ages: slow acquisition of pre-speech behaviors during the first year, and a slowertransition away from non-speech vocal productionnear the first birthday. Moreover, although MSELExpressive Language scores at 6 mo. appear to besomewhat sensitive to differences between childrenwho will and will not go on to show ASD symptoms at24 mo., the same is not true at 12 mo., even thoughMSEL Expressive Language is generally correlatedwith consonant production at this age. It wouldappear that although the MSEL Expressive Lan-guage score captures some of the decreased pre-speech behavior present at 6 mo. in HR infants whogo on to develop symptoms, it is less sensitive to the

transitional behaviors at 12 months that are asso-ciated with later symptom appearance.

It is important to recall that the high prevalence ofautistic symptoms in the HR does not necessarilymean that all symptomatic children will remain onthe autism spectrum. Thus HR/LR group differencesreported here may be driven by the children whohave significant vulnerabilities for ASD. The discri-minant function analyses are telling in this regard.At each age-level, the metric most strongly associ-ated with placement in the outcome group with orwithout autistic symptoms is the production of newlyemerging speech behaviors. That is, at 6 months it isthe frequency of middle consonant production, theconsonants just emerging at this age, that predictsgroup placement; at 9 months, it is the frequency oflate consonants, again those at the leading edge ofpre-speech development. At 12 months, even thoughgroup differences in consonant production are nolonger significant, it is the total number of conso-nants, which can be interpreted to reflect theamount of word-like babble, that is associated withoutcome.

What might this pattern suggest about spokenlanguage development in children at risk for ASD? Itwould seem, first, to indicate that it is not only thedevelopment of communicative intention and lan-guage function that is vulnerable in the HR popula-tion; the acquisition of basic vocal ‘tools’ for thedevelopment of speech is also affected, even beforethey are recruited into the service of linguistic com-munication. Prior to the onset of language, babblingis used in typical development both as a means of‘practicing’ the motor patterns associated withspeech production and as an early form of interac-tion, through back-and-forth babbled ‘conversa-tions’ with family members (Locke, 1989). Perhapsvocal development is vulnerable in HR infantsbecause they are less apt to ‘practice’ sound-makingin order to match it more closely to the sounds theyhear others in their environment producing, as wellas because they are less captivated by the turn-taking exchanges in which babbled productionsallow them to participate. Thus, the differencesobserved here may be a result of vulnerabilities forsocial disability, and not their source. It is not thecase, however, that HR infants are simply ‘quieter’than LR peers, since the total number of vocaliza-tions produced never differed between the two riskgroups at any age. Differences appear not in soundproduction, but in the kinds of sounds produced,with LR children and HR children who do not developASD symptoms producing more vocal behavior at the‘cutting edge’ of development at each phase.

Limitations

Although delayed language development is a corefeature of ASD, it is not specific to ASD, and thelack of a non-autistic developmentally delayed



contrast group is a limitation of the present study. Itcould be the case that the differences in preverbalbehavior in the HR infants are associated with amore general or more language-specific delay thanwith ASD. Although our discriminant functionanalysis argues for a connection between preverbalvocal behavior and autistic symptoms rather thangeneral delay, and ANOVA results revealed therewere no differences in either nonverbal or receptivelanguage skills at any age between HR children whodid and did not develop ASD symptoms by 24months, nonetheless, a direct contrast of preverbalvocal production in children with non-autisticdevelopmental delays in the first year of life andthose at risk for ASD is clearly warranted. In addi-tion, a true longitudinal design would provide amore direct answer to the questions raised in thisstudy, which was limited to a cross-sectionaldesign.

Clinical implications

Although the limitations of the present study preventdrawing a definitive clinical connection betweenearly speech behavior and symptoms of ASD in HRinfants, the findings do suggest that children show-ing delays in speech acquisition, whether measuredby direct observations of consonant inventories andcanonical syllables, or more indirect measures suchas the MSEL Expressive Language scale, warrantextra vigilance during the second half of their firstyear. This vigilance may include more frequentmonitoring of a range of developmental milestones.In addition, parent counseling that guides families to

maximize the infant’s experience in listening to andmaking speech sounds, such as that proposed byKuhl and Meltzoff (1996), may be considered alongwith interventions that encourage the developmentof joint attention and communicative intent. Thefinding that pre-speech vocal development is a sen-sitive indicator of heightened risk for the appearanceof ASD symptoms in HR infants provides an additionto the range of tools emerging from the infant siblingliterature in ASD for the important task of identifyingchildren at the earliest possible point in develop-ment. Early identification, combined with advice tofamilies that helps intensify and focus their naturalinteractions with these infants on activities aimed atoptimizing early interpersonal experiences, mayallow us alter their developmental trajectory for thebetter.

AcknowledgementsPreparation of this paper was supported by NIMH Aut-ism Center of Excellence grant P50 MH81756; ResearchGrant P01-03008 funded by the National Institute ofMental Health (NIMH); MidCareer Development AwardK24 HD045576 funded by NIDCD to Rhea Paul; as wellas by the Simons Foundation. We wish to thank thefamilies who participated in this research.

Correspondence toRhea Paul, Communication Disorders Section, YaleChild Study Center, 40 Temple St. # 6B, New Haven, CT06510, USA; Tel: 203 785 3388; Fax: 203 785 3705;Email: [email protected]

Key points

• Infants with older siblings diagnosed with ASD show increased risk for the syndrome.• Pre-speech vocalizations in the first year of life are related to subsequent language development.• Vocal productions of infants at risk for ASD were studied and found to show differences from those of low-risk

peers at particular points in the first year of life.• These differences were related to the appearance of autistic symptoms in the second year.• These findings may contribute to early identification and intervention efforts in ASD.• Findings contribute to understanding the trajectory of language growth in infants at risk for ASD.

ReferencesAmerican Psychiatric Association & American Psychiatric

Association. Task Force on DSM-IV. (1994). Diagnostic andstatistical manual of mental disorders: DSM-IV (4th edn).Washington, DC: American Psychiatric Association.

Bailey, A., Le Couteur, A., Gottesman, I., Bolton, P., Simonoff,E., Yuzda, E., & Rutter, M. (1995). Autism as a stronglygenetic disorder: Evidence from a British twin study.Psychological Medicine, 25, 63–77.

Chawarska, K., Klin, A., Paul, R., & Volkmar, F. (2007). Autismspectrum disorder in the second year: Stability and changein syndrome expression. Journal of Child Psychology andPsychiatry, 48, 128–138.

Chawarska, K., Klin, A., Paul, R., Macari, S., & Volkmar, F.(2009). A prospective study of toddlers with ASD: Short-term

diagnostic and cognitive outcomes. Journal of Child Psychol-ogy and Psychiatry, 50, 1235–1245.

Chawarska, K., & Volkmar, F. (2005). Autism in infancy andearly childhood. In F. Volkmar, R. Paul, A. Klin, & D.J.Cohen (Eds.), Handbook of autism and pervasive develop-mental disorders (3rd edn, vol. 1, pp. 223–246). New York:Wiley.

Cohen, J. (1988). Statistical power analysis for the behavioralsciences (2nd edn). Hillsdale, NJ: Lawrence Erlbaum Asso-ciates.

Cox, A., Klein, K., Charman, T., Baird, G., Baron-Cohen, S.,Swettenham, J., et al. (1999). Autism spectrum disorders at20 and 42 months of age: Stability of clinical and ADI-RDiagnosis. Journal of Child Psychology and Psychiatry, 40,719–732.



Dawson, G., Estes, A., Munson, J., Schellenberg, G., Bernier,R., & Abbott, R. (2007). Quantitative assessment of autismsymptom-related traits in probands and parents: BroaderPhenotype Autism Symptom Scale. Journal of Autism andDevelopmental Disorders, 37, 523–536.

Dawson, G., Osterling, J., Meltzoff, A., & Kuhl, P. (2000). Casestudy of the development of an infant with autism from birthto two years of age. Journal of Applied DevelopmentalPsychology, 21, 299–313.

Folstein, S., & Rutter, M. (1977). Infantile autism: A geneticstudy of 21 twin pairs. Journal of Child Psychology andPsychiatry, 18, 297–321.

Folstein, S.E., & Rutter, M.L. (1988). Autism: Familial aggre-gation and genetic implications. Journal of Autism andDevelopmental Disorders, 18, 3–30.

Gernsbacher, M.A., Sauer, E.A., Geye, H.M., Schweigert, E.K.,& Goldsmith, H.H. (2008). Infant and toddler oral- andmanual-motor skills predict later speech fluency in autism.Journal of Child Psychology and Psychiatry, 49, 43–50.

Iverson, J.M., & Wozniak, R.H. (2007). Vatiation in vocal-motordevelopment in infant siblings of children with autism.Journal of Autism and Developmental Disorders, 37, 158–170.

Jensen, T., Boggild-Andersen, B., Schmidt, J., Ankerhus, J., &Hansen, E. (1988). Prenatal risk factors and first yearvocalizations: Influence on preschool language and motorperformance. Developmental Medicine and Child Neurology,30, 153–161.

Kuhl, P., & Meltzoff, A. (1996). Infant vocalizations in responseto speech: Vocal imitation and developmental change. Jour-nal of the Acoustical Society of America, 100, 2425–2438.

Landa, R., & Garrett-Mayer, E. (2006). Development in infantswith autism spectrum disorders: A prospective study. Jour-nal of Child Psychology and Psychiatry, 47, 629–638.

Landa, R., Piven, J., Wzorek, M., Gayle, J., Chase, G., &Folstein, S. (1992). Social language use in parents of autisticindividuals. Psychological Medicine, 22, 245–254.

Landa, R.J., Holman, K.C., & Garrett-Mayer, E. (2007). Socialand communication development in toddlers with early andlater diagnosis of autism spectrum disorders. Archives ofGeneral Psychiatry, 64, 853–864.

Le Couteur, A., Bailey, A., Goode, S., Pickles, A., Robertson, S.,Gottesman, I., et al. (1996). A broader phenotype of autism:The clinical spectrum in twins. Journal of Child Psychologyand Psychiatry, 37, 785–801.

Leonard, L., Schwartz, R., Morris, B., & Chapman, K. (1981).Factors influencing early lexical acquisition: Lexical orien-tation and phonological composition. Child Development, 52,882–887.

Levitt, J., & Utman, G. (1992). From babbling towards thesound systems of English and French: A longitudinal two-case study. Journal of Child Language, 19, 19–49.

Locke, J. (1993). The child’s path to spoken language. Cam-bridge, MA: Harvard University Press.

Locke, J.L. (1989). Babbling and early speech: Continuity andindividual differences. First Language, 9, 191–205.

Lord, C. (1995). Follow-up of two-year-olds referred for possi-ble autism. Journal of child Psychology and Psychiatry, 36,1365–1382.

Lord, C., Rutter, M., DiLavore, P.C., & Risi, S. (2002). AutismDiagnostic Observation Schedule (ADOS). Los Angeles, CA:Western Psychological Services.

Lord, C., Rutter, M., & Le Couteur, A. (1994). AutismDiagnostic Interview–Revised: A revised version of a diag-nostic interview for caregivers of individuals with possiblepervasive developmental disorders. Journal of Autism andDevelopmental Disorders, 24, 659–685.

Luyster, R., Guthrie, W., Gortham, K., Risi, S., DiLavore, P., &Lord, C. (2009). The Autism Diagnostic Observation Sche-dule–Toddler module: Preliminary findings using a modifiedversion of the ADOS. Paper presented at the IMFAR.

Menyuk, P., Liebergott, J., & Schultz, M. (1986). Predictingphonological development. In B. Lindstrom, & R. Zetter-strom (Eds.), Precursors of early speech (pp. 70–93). NewYork: Stockton Press.

Micali, N., Chakrabarti, S., & Fombonne, E. (2004). The BroadAutism Phenotype: Findings from an Epidemiological Sur-vey. Autism, 8, 21.

Mullen, E. (1995). Mullen Scales of Early Learning. CirclePines, MN: American Guidance Serivce, Inc.

Oller, D.K., et al. (2010). Proceedings of the National Academyof Sciences, USA. doi:10.1073/pnas.1003882107 (2010).Retrieved July 29, 2010.

Oller, D.K., Eilers, R., Neal, A.R., & Cobo-Lewis, A.B (1998).Late onset canonical babbling: A possible early marker forabnormal development. American Journal on Mental Retar-dation, 103, 249–263.

Oller, D.K., Eilers, R.E., Neal, A.R., & Schwartz, H.K. (1999).Precursors to speech in infancy: The prediction of speechand language disorders. Journal of Communication Disor-ders, 32, 223–245.

Oller, D., Wieman, L., Doyle, W., & Ross, C. (1976). Infantbabbling and speech. Journal of Child Language, 3, 1–11.

Ozonoff, S., Losif, A.-M., Baguio, F., Cook, I.C., Hill, M.M.,Hutman, T., et al. (2009). A prospective study of theemergence of early behavioral signs of autism. Journal ofthe American Academy of Child and Adolescent Psychiatry,49, 256–266.

Paul, R., & Jennings, P. (1992). Phonological behavior innormal and late talking toddlers. Journal of Speech andHearing Research, 35, 99–107.

Paul, R., Chawarska, K., Cicchetti, D., & Volkmar, F. (2008).Language outcomes in toddlers with ASD: A 2 year follow-up. Autism Research, 1, 97–107.

Piven, J., Palmer, P., Jacobi, D., Childress, D., & Arndt, S.(1997). Broader autism phenotype: Evidence from a familyhistory study of multiple-incidence autism families. Ameri-can Journal of Psychiatry, 154, 185–190.

Roberts, J., Rescorla, L., Giroux, J., & Stevens, L. (1998).Phonological skills of children with specific expressivelanguage impairment (SLI-E): Outcome at age 3. Journal ofSpeech, Language, and Hearing Research, 41, 374–384.

Rogers, S.J. (2009). What are infant siblings teaching us aboutautism in infancy? Autism Research, 2, 125–137.

Rutter, M. (2005). Genetic influences and Autism. In R.P.F.Volkmar, A. Klin, & D. Cohen (Eds.), Handbook of autismand pervasive developmental disorders (vol. 1, pp. 425–452).New York: Wiley & Sons.

Santangelo, S.L., Folstein, S.E., & Tager-Flusberg, H. (1999).Autism: A genetic perspective. In H. Tager-Flusberg (Ed.),Neurodevelopmental disorders (pp. 431–447). Cambridge,MA: The MIT Press.

Schoen, E., Paul, R., & Chawarska, K. (2010). Vocal develop-ment in toddlers with ASD. In R. Paul, & P. Flipsen (Eds.),Child speech sound disorders: In honor of Lawrence D.Shriberg. San Diego, CA: Plural.

Schoen, E., Paul, R., Chawarska, K., Klin, A., & Volkmar, F.(Nov., 2008). Early vocal behavior and expressive languagein toddlers with autism spectrum disorder. Paper presentedat the National Convention of the American Speech, Lan-guage and Hearing Association.

Schwartz, R., & Leonard, L. (1982). Do children pick andchoose? An examination of phonological selection andavoidance in early lexical acquisition. Journal of ChildLanguage, 9, 319–336.

Sheinkopf, J., Mundy, P., Oller, D.K., & Steffens, M. (2000).Vocal atypicalities of preverbal autistic children. Journal ofAutism and Developmental Disorders, 30, 345–354.

Shriberg, L. (1993). Four new speech and prosody-voicemeasures for genetics research and other studies in devel-opmental phonological disorder. Journal of Speech andHearing Disorders, 36, 105–140.



Stark, R. (1980). Stages of speech development in the first yearof life. In G. Yeni-Komshian, J. Kavanagh, & C. Ferguson(Eds.), Child phonology (vol. 1, pp. 73–92). New York:Academic Press.

Stark, R.E., Bernstein, L.E., & Demorest, M.E. (1993). Vocalcommunication in the first 18 months of life. Journal ofSpeech and Hearing Research, 36, 548–558.

Stoel-Gammon, C. (1998). Sounds and words in early languageacquisition: The relationship between lexical and phonolog-ical development. In R. Paul (Ed.), Exploring the speech-language connection (pp. 25–52). Baltimore, MD: Paul H.Brookes.

Stoel-Gammon, C., & Cooper, J. (1984). Patterns of earlylexical and phonological development. Journal of ChildLanguage, 11, 247–271.

Storkel, H.L., & Morrisette, M.L. (2002). The lexicon andphonology: Interactions in language acquisition. Language,Speech, and Hearing Services in Schools, 33, 24–37.

Sullivan, M., Finelli, J., Marvin, A., Garrett-Mayer, E., Bau-man, M., & Landa, R. (2007). Response to joint attention intoddlers at risk for autism spectrum disorder: A prospectivestudy. Journal of Autism and Developmental Disorders, 37,37–48.

Szatmari, P., MacLean, J.E., Jones, M.B., Bryson, S.E.,Zwaigenbaum, L., Bartolucci, G., et al. (2000). The familialaggregation of the lesser variant in biological and nonbio-logical relatives of PDD probands: A family history study.Journal of Child Psychology and Psychiatry, 41, 579–586.

Thal, D., Oroz, M., & McCaw, V. (1995). Phonological andlexical development in normal and late talking toddlers.Applied Psycholinguistics, 16, 407–424.

Turner, L.M., & Stone, W.L. (2007). Variability in outcome forchildren with an ASD diagnosis at age 2. Journal of ChildPsychology and Psychiatry, 48, 793–802.

Vihman, M., Ferguson, C., & Elbert, M. (1986). Phonologicaldevelopment from babbling to speech: Common tendenciesand individual differences. Applied Psycholinguistics, 7, 3–40.

Vihman, M., & Greenlee, M. (1987). Individual differences inphonological development: Ages one and three years. Jour-nal of Speech and Hearing Research, 30, 503–521.

Volkmar, F.R., Lord, C., Bailey, A., Schultz, R.T., & Klin, A.(2004). Autism and pervasive developmental disorders.Journal of Child Psychology and Psychiatry, 45, 135–170.

Wetherby, A., Watt, N., Morgan, L., & Shumway, S. (2007).Social communication profiles of children with autismspectrum disorders late in the second year of life. Journalof Autism and Developmental Disorders, 37, 960–975.

Wetherby, A., Woods, J., Allen, L., Cleary, J., Dickinson, H., &Lord, C. (2004). Early indicators of autism spectrum disor-ders in the second year of life. Journal of Autism andDevelopmental Disorders, 34, 473–493.

Zwaigenbaum, L., Bryson, S., Lord, C., Rogers, S., Carter, A.,Carver, L., et al. (2009). Clinical assessment and manage-ment of toddlers with suspected autism spectrum disorder:Insights from studies of high-risk infants. Pediatrics, 123,1383–1391.

Accepted for publication: 9 August 2010Published online: 11 October 2010



out of the mouths of babes - vocal production in infant siblings of children with asd

Documents