Skip to main content
NCBI home page
Journal of Speech, Language, and Hearing Research : JSLHR logoLink to Journal of Speech, Language, and Hearing Research : JSLHR
. 2022 Mar 1;65(4):1465–1477. doi: 10.1044/2021_JSLHR-21-00458

Caregiver Language Input Supports Sentence Diversity in Young Children With Autism Spectrum Disorder

Elysha Clark-Whitney a, Claire Brito Klein a, Pamela A Hadley b, Catherine Lord c, So Hyun Kim a,✉
PMCID: PMC9499362  PMID: 35230878

Abstract

Purpose:

Sentence diversity is a measure of early language development that has yet to be applied to individuals with autism spectrum disorder (ASD). The primary aim of this study was to identify whether children with ASD show change in sentence diversity over 6 months of treatment with Naturalistic Developmental Behavioral Intervention (NDBI). The secondary aim was to examine possible predictors of changes in children's sentence diversity, including caregiver use of NDBI strategies, naturally occurring instances of caregiver Toy Talk, and child characteristics.

Method:

Fifty children with ASD (ages 2–4 years) and their caregivers, who were receiving NDBI, engaged in two 10-min video-recorded play interactions, 6 months apart. Child speech was transcribed and coded for sentence diversity. Caregiver input was transcribed and coded for naturally occurring Toy Talk. Zero-inflated negative binomial mixed models were used to explore predictors of change in child sentence diversity.

Results:

Children's sentence diversity improved over time. Changes in caregiver NDBI strategy use and caregiver baseline Toy Talk were significant predictors of changes in sentence diversity, as were baseline age, nonverbal ratio IQ, and child sex. Additionally, a significant interaction of caregiver baseline Toy Talk and change in caregiver NDBI strategies emerged; the effect of caregiver baseline Toy Talk on children's sentence diversity change was stronger when NDBI strategy use improved.

Conclusions:

Sentence diversity is a developmentally sensitive measure of language development in ASD. NDBI strategies that facilitate reciprocal social communication, combined with input composed of declarative sentences with noun or third-person pronoun subjects, may provide optimal support for children's sentence development.

The toddler and preschool years are typically a period of rapid growth in language abilities. The first instances of child-like sentences generally develop between 24 and 26 months of age (Klee & Gavin, 2010; Lee, 1974) with the emergence of the tense and agreement morphemes needed for forming adultlike sentences appearing shortly thereafter (Hadley et al., 2014). In children with autism spectrum disorder (ASD), this trajectory of language development is often delayed or impaired. For instance, the sentence structures used by young children with ASD are less varied and complex as compared to those of children developing typically and children with other developmental delays, even when the children perform equally on measures of amount of language, utterance length, and diversity of vocabulary (Eigsti et al., 2007; Rappin & Allen, 1988; Scarborough et al., 1991). These grammatical deficits have broader implications for language development in ASD, given that understanding and using a range of grammatical structures supports subsequent vocabulary development (Naigles, 1990). Despite the importance of early syntax, however, research focusing on early language development in ASD has typically utilized general indices such as vocabulary size or mean length of utterance (MLU) without investigating core elements of sentence structure specifically.

The Sentence-Focused Framework for Understanding Language Development in ASD

In this study, we employed the sentence-focused framework (see the works of Hadley, 2014, 2020, for an overview) to better understand the transition from words to diverse sentences in children with ASD. The sentence-focused framework identifies four developmental steps in early language development: words, verbs, child-like sentences, and adult sentences. According to this framework, children with a strong underlying representation of sentence structure can produce simple sentences with more diverse subjects and verbs (Hadley, 2014). Hadley (2020) further identified intransitive verbs as playing a key role in the emergence of third-person subject diversity (e.g., it, he, dog, tower). This, in turn, facilitates the emergence of tense and agreement marking (e.g., The baby drinks milk. vs. The baby drank milk; The dog is sleeping. vs. The bears are sleeping.). Thus, by emphasizing the importance of diverse verbs and subjects in sentences, the sentence-focused framework links vocabulary diversity and grammatical development for early language learners.

In turn, Hadley et al. (2018) demonstrated that measures of sentence diversity can be used to characterize variability in early sentence development. Sentence diversity is operationalized as the number of unique combinations of sentence subjects and main verbs in child sentences. Sentence diversity is characterized as a measure of “structurally specific” lexical diversity because it reflects the number of different words a child uses in a specific syntactic structure. Hadley et al. (2018) recommended the assessment of sentence diversity for children transitioning from word combinations to sentences and showed that sentences generally become more diverse as utterances get longer and more complex. However, these associations among utterance length, sentence diversity, and grammatical complexity may not hold for children with language disorders (e.g., Hadley, 2014, 2020; McKenna & Hadley, 2014).

Investigating sentence diversity is therefore an important first step in understanding individual trajectories of grammatical development in ASD, particularly given that verb diversity was recently identified as a better predictor of long-term outcomes for individuals with ASD than noun diversity or MLU (LeGrand et al., 2021). However, this investigation focused on noun and verb diversity in any syntactic context, not specifically in the subject or main verb positions of child sentences. Thus, researchers have yet to apply measures of sentence diversity to analyzing the language of young children with ASD. Additionally, while child-level factors including nonverbal IQ (Ellis Weismer & Kover, 2015; Ellis Weismer et al., 2010; Luyster et al., 2008; Thurm et al., 2007) and ASD symptom severity (Charman et al., 2005; Ellis Weismer & Kover, 2015; Luyster et al., 2008; Paul et al., 2008) have consistently been shown to be key predictors of overall language outcomes in ASD, research to date has not investigated their role in variability in the development of sentence diversity in ASD.

The Role of Caregivers in Supporting Children's Sentence Diversity

In addition to child-level factors, caregiver input is known to play a key role in supporting the grammatical development of young children with ASD (Crandall et al., 2019; Fusaroli et al., 2019). For example, when caregivers expose their children to a richer variety of verbs, children's own verb vocabulary is increased (Crandall et al., 2019). Similarly, caregiver MLU has been shown to predict children's later MLU as well as the variety of words children produce, even when controlling for children's baseline expressive language abilities (Fusaroli et al., 2019). These results demonstrate that the frequency, diversity, and complexity of caregivers' language input may be a powerful predictor of language development for children with ASD (Crandall et al., 2019; Fusaroli et al., 2019).

Less attention has focused on sentence diversity in caregivers' language input; however, there is emerging evidence indicating that caregiver sentence diversity promotes the transition from words to diverse sentences for typically developing toddlers. Rispoli et al. (2018) demonstrated that sentence diversity in naturally occurring caregiver input to 21-month-old toddlers with mean length utterance of less than 1.25 was a significant predictor of the toddlers' sentence diversity 9 months later. They also noted that subject diversity is more limited than verb diversity in caregiver input. However, Hadley and colleagues have also demonstrated that caregiver subject diversity can be increased with brief adult education and coaching by introducing caregivers to Toy Talk strategies (Hadley, Rispoli, Holt, Papastratakos, et al., 2017; Hadley & Walsh, 2014) built on the foundation of responsive interaction strategies (Roberts & Kaiser, 2011; Weitzman et al., 2014). Toy Talk consists of two simple strategies. The first strategy, “talk about the toys,” helps caregivers comment on the toys their children are playing with by describing the toy's properties (e.g., The baby is hungry), actions (e.g., The tower fell down), locations (e.g., The ball is under the table), and so forth. The second strategy, “give the object its name,” helps caregivers use noun subjects instead of pronoun subjects (e.g., baby, tower, ball NOT she, it). Together, the two strategies increase the frequency and diversity of nouns in the subject position of caregiver sentences, thus facilitating children's exposure to greater sentence diversity. Moreover, Hadley and colleagues (Hadley, Rispoli, & Holt, 2017; Hadley, Rispoli, Holt, Papastratakos, et al., 2017) demonstrated that caregivers' noun subject diversity following the brief, caregiver-mediated Toy Talk intervention was a significant predictor of children's growth in sentence diversity as well as tense and agreement morphemes from 21 to 30 months for toddlers who were slow to talk but otherwise developing typically. Toy Talk intervention has yet to be used with caregivers of children with autism; however, the potential for translation is promising given that caregiver-mediated intervention is a recommended core component of early intervention for ASD (Maglione et al., 2012; National Autism Center, 2015).

Furthermore, the goals and methods of Toy Talk overlap substantially with those of caregiver-mediated Naturalistic Developmental Behavioral Intervention (NDBI), a group of caregiver-mediated interventions for autism that apply behavioral principles within naturalistic and developmentally appropriate interactions and routines (see the work of Schreibman et al., 2015, for a review). Child language is one of the core targets of NDBI (Schreibman et al., 2015), with caregivers trained to model and reinforce communication across a range of naturalistic activities with their child. Specifically, NDBI emphasizes training in caregiver responsivity, including following the child's lead in play, modeling language that relates to the child's interests, reinforcing and expanding the child's social communication (SC) attempts, and structuring interactions using turn-taking to provide balanced opportunities for child play and SC as well as caregiver modeling of play (Schreibman et al., 2015; Vivanti & Zhong, 2020). These strategies foster a rich caregiver–child interaction context into which Toy Talk could be readily embedded to specifically target sentence diversity. Indeed, child SC skills such as joint attention and imitation have been shown to moderate language learning and response to language intervention in ASD (Bono et al., 2004; Ellis Weismer & Kover, 2015; Luyster et al., 2008; Paul et al., 2013). Thus, while brief instruction in responsive interaction and Toy Talk was sufficient to support grammatical development in typically developing children (Hadley, Rispoli, Holt, Papastratakos, et al., 2017), core SC deficits in ASD suggest that layering Toy Talk on top of more extensive instruction in NDBI strategies that target SC may be important in order to maximize the benefits of Toy Talk in ASD. To date, however, there has not been any empirical evidence demonstrating how NDBI strategies can optimize the benefits of caregiver language input to support growth in sentence diversity for children with ASD.

Current Study

Based on caregiver–child interaction samples from 50 families of young children between 2 and 4 years of age with ASD who were receiving manualized NDBI programs across various research studies, the current study aimed to examine (1) whether young children with ASD receiving various early interventions show growth in sentence diversity; (2) whether caregiver language input, specifically caregivers' use of naturally occurring Toy Talk sentences, at the beginning of the intervention is associated with the development of sentence diversity in young children with ASD above and beyond the effect of caregiver NDBI strategy use; (3) how caregiver implementation of NDBI strategies moderates the relation between caregiver language input and the development of sentence diversity in young children with ASD; and (4) how children's baseline demographic and clinical features (e.g., age, sex, nonverbal IQ, autism symptom severity) predict changes in sentence diversity over the course of intervention. We hypothesized that children with ASD receiving early intervention would show improvements in sentence diversity over time, but with large individual variability in the magnitude of change. We expected that children of caregivers who naturally used Toy Talk sentences more frequently at intervention baseline would show greater improvements in their language over the course of treatment, and this association would be moderated by improvements in caregiver NDBI strategy use. Specifically, we predicted that the impact of caregiver Toy Talk sentences on children's sentence diversity would be stronger when caregivers' use of NDBI strategies improved over the course of treatment. Given past findings, we expected children with higher nonverbal IQ and/or lower SC difficulties to show larger improvements (Ellis Weismer & Kover, 2015; Ellis Weismer et al., 2010; Luyster et al., 2008; Thurm et al., 2007). In light of mixed findings regarding the effects of other demographic features on language outcomes, we did not have specific hypotheses about the effects of child age and sex on sentence diversity.

Method

Participants

Participants in the current study were drawn from a sample of young children with ASD and their caregivers who were receiving treatment as part of larger studies of various NDBI that were conducted at University of Michigan Autism and Communication Disorders Center (Dawson et al., 2010; Kasari et al., 2015; Wetherby et al., 2014). The intervention did not involve specific training on Toy Talk strategies; rather, caregivers' naturally occurring use of Toy Talk sentences was measured. At several timepoints over the course of the study, caregivers and children were asked to engage in play-based interactions for 10 min using standardized toy sets (see Procedure section). These interactions were video-recorded, and caregiver and child language were transcribed from the recording (see Measures section). Dyads were included in the current study if (a) a pair of transcripts from approximately 6 months apart (M = 5.88 months) was available for the dyad, (b) the child was between 2 and 4 years of age inclusive at the date of the first transcript, and (c) the child was not yet using phrases with more than three words and was not combining phrases at the date of the first transcript, as determined using the Overall Level of Language item from the standardized diagnostic assessment, the Autism Diagnostic Observation Schedule (ADOS-2; Lord, Luyster, et al., 2012; Lord, Rutter, et al., 2012; Mazurek et al., 2019)—specifically, children who received the ADOS-2 Toddler Module (n = 16), Module 1 (n = 29), or scored 1–3 for Item A1 (overall level of language) on the ADOS-2 Module 2 (n = 5) were included in the current study. These specific language criteria captured children whose language levels ranged from no words to emerging two- to three-word phrases. These age and language criteria were selected to capture the developmental period during which children are exponentially increasing their use of unique subject–verb (SV) combinations. While typically developing children usually produce diverse combinations by 32 months (Hadley et al., 2018), children with ASD show large variability in the onset and development of word combinations and sentences, with many not producing any language until at least age 3 years (Kim et al., 2014; Pickles et al., 2014). Thus, a wider age range was chosen to capture variability in language development in ASD. This resulted in a sample of 50 children aged between 25 and 55 months at the date of the first transcript (41 male, nine female, M = 35.32 months, SD = 7.73) with clinical best estimate diagnoses of ASD and their caregivers (46 female, four male; see Table 1).

Table 1.

Baseline demographic and clinical features.

Child (n = 50) M (SD) or n (%) Range Caregiver (n) = 50 M (SD) or n (%) Range
Age (months) 35.32 (7.73) 25–55 Age (years; n = 32) a 36.22 (5.87) 25–49
Sex 41 male, 9 female Sex 4 male, 46 female
Race (n = 48) a Race (n = 34) a
 White 35 (72.9%)  White 26 (76.5%)
 Black 6 (12.5%)  Black 6 (17.6%)
 Asian 1 (2.1%)  Asian 1 (2.9%)
 Other 6 (12.5%)  Other 1 (2.9%)
ADOS-2 CSS Education (n = 40) a
 CSS SC 7.00 (2.06) 2–10  BA/BS or above 26 (65%)
 CSS RRB 7.58 (1.97) 1–10  Below BA/BS 14 (35%)
ADOS-2 expressive language Income (n = 33) a
 No spontaneous language 2 (4%)  $35,000 or less 8 (24.2%)
 At least 1 word or approximation 13 (26%)  $36,000–$80,000 10 (30.3%)
 At least 5 words or approximation 11 (22%)  Above $81,000 15 (45.5%)
 Occasional phrases 14 (28%) Relationship to child
 Two or more word phrases 10 (20%)  Mother 45 (90%)
MSEL  Father 4 (8%)
 NV ratio IQ 61.74 (24.21) 23–120  Grandmother 1 (2%)
 V ratio IQ 72.67 (17.18) 37–108
Open in a new tab
a

Race information was unavailable for two children and 16 caregivers; age was unavailable for 18 caregivers; education information was unavailable for 10 caregivers; income information was unavailable for 17 caregivers. ADOS-2 CSS SC = Autism Diagnostic Observation Schedule social communication calibrated severity score; ADOS-2 CSS RRB = Autism Diagnostic Observation Schedule restricted/repetitive behaviors calibrated severity score; MSEL = Mullen Scales of Early Learning; NV = nonverbal; V = verbal; BA/BS = bachelor's degree or equivalent.

Procedure

Caregivers were instructed to interact with their child as they normally would for 10 min while playing with standard sets of age-appropriate toys. The toys were standardized across participants and timepoints and included toys to accommodate varying play levels, ranging from cars and construction toys to toy sets that could be used for pretend play (e.g., picnic food set). Providing dyads with a range of toys across different play levels was intended to create opportunities for the child to engage in activities involving the various types of toys and for the dyads to comment on those toys during the interactions; however, toys were not specifically selected to directly elicit the production of diverse SV combinations. The free-play nature of the interaction allowed for the observation of children's spontaneous language skills, as recommended for capturing young children's optimal language use (Hadley et al., 2018). Interactions were video-recorded. Twelve sessions from 10 dyads were recorded at home, whereas all other sessions were recorded in the clinic. Results were consistent when the dyads whose observations were conducted at home were excluded from all analyses (data available upon request). Within the 6 months prior to the first play session (M = 1.81 months), children also completed diagnostic and cognitive assessments.

Trained and reliable coders transcribed caregiver and child speech from interaction videos using Systematic Analysis of Language Transcripts (SALT) software, 2012 Research Version (Miller & Iglesias, 2012; see Measures section). Transcripts were coded for child SV combination use and caregiver use of Toy Talk sentences during the transcription process. Coders then reviewed the transcription a second time to check for accuracy (e.g., word choice, spelling) and review sentence diversity and Toy Talk codes. The two coders initially completed the self-paced transcription trainings available on the SALT website. They then independently transcribed and coded five videos and then reviewed the transcripts together to establish consensus. Coders established reliability with each other for child SV combination and caregiver Toy Talk coding before proceeding with independent coding; coders were considered reliable after achieving three consecutive transcripts with scores no more than two apart (or no more than 80% apart for codes with a total incidence of less than 10 per transcript) on at least 80% of sentence diversity and Toy Talk codes per transcript. Reliability was monitored throughout the transcription and coding process via regular consensus coding sessions, and interrater reliability was calculated using a randomly selected sample of 20 videos. Average intraclass correlation coefficients for sentence diversity and Toy Talk codes as well as SALT-calculated variables used in analysis ranged from 0.92 to 0.99. Coders were masked to video timepoint.

Measures

SALT Coding of Language Features

Child and caregiver speech from the recorded interactions was transcribed using SALT software (Miller & Iglesias, 2012). SALT transcription conventions were used for utterance segmentation and transcript formatting (e.g., marking of bound morphemes, overlapping speech). Given that our interest was in children's functional word usage, rather than correct pronunciation, root identification was used to mark word approximations (e.g., disor as an approximation of dinosaur; mo as an approximation of more) as the full word, when approximations had at least one clear syllable of the full word. Codes were also used to indicate word approximations and nonword vocalizations (i.e., communicative sounds that were neither approximations nor full words). Noncommunicative speech, including clearly undirected vocalizations, singing, and echolalia, was transcribed using comment notation and was excluded from all analyses that focused only on spontaneous, functional language use. SALT automated analyses were used to calculate child MLU in morphemes as well as number of different words used during the 10-min baseline language sample.

Child sentence diversity. During transcription, coders tallied the unique SV combinations children produced, based on conventions developed by Hadley et al. (2014) and Hadley, Rispoli, Holt, Papastratakos, et al. (2017). Codes were summed for each subject type (i.e., first, second, and third person) as well as the total number of unique (different) SV combinations across all subject types. Given that the number of SV combinations that children produced for each individual subject type was quite low (the maximum number of unique SVs produced in each subject type ranged from one to 11, compared to the total number of unique SV, which ranged from zero to 34), only the total number of unique SVs was used in analyses.

Caregiver Toy Talk. While transcribing, coders noted caregivers' naturally occurring use of Toy Talk sentences according to the definitions from Hadley, Rispoli, Holt, Papastratakos, et al. (2017). A Toy Talk sentence was defined as an active declarative sentence in which the caregiver used a noun or third-person pronoun subject and described the object's state, property, action, location, or possession. All instances of caregiver Toy Talk were noted (i.e., when caregivers used the same Toy Talk sentence multiple times, each instance was counted in the total).

Child Autism Symptoms

Autism symptoms were assessed using the ADOS-2 (Lord, Luyster, et al., 2012). The ADOS-2 is a standardized, semistructured assessment in which the assessor interacts with the child using specific presses in order to observe the child's SC, interaction, and play. ADOS-2 calibrated severity scores (CSS) for SC were used to estimate children's level of SC impairment (Esler et al., 2009; Gotham et al., 2009; Hus et al., 2014). Higher CSS values reflect greater severity of autism symptoms.

Child Cognitive Ability

Children's cognitive abilities were measured using the Mullen Scales of Early Learning (Mullen, 1995). Nonverbal ratio IQ (NVIQ) was calculated from scores on the Visual Reception and Fine Motor subscales, and verbal ratio IQ (VIQ) was calculated from scores on the Receptive Language and Expressive Language subscales.

Caregiver NDBI Strategy Use

Caregiver's implementation of NDBI strategies was measured using the Measure of NDBI Strategy Implementation–Caregiver Change (MONSI-CC; Vibert et al., 2020). Trained and reliable coders scored caregiver NDBI strategy use based on the same 10-min caregiver–child interaction recording that was used for language transcription. The current sample was drawn from the initial validation sample for the MONSI-CC (see the work of Vibert et al., 2020, for psychometric data on the reliability and validity of the instrument). Coders were masked to timepoint and other subject information. MONSI-CC total scores were utilized for the current study, with higher scores indicating more proficient use of NDBI strategies.

Analytic Approach

Given that many children (70%) were not yet using SVs at T1, zero-inflated negative binomial mixed models (“glmmTMB” package for R Studio; Brooks et al., 2017) were used. The outcome variable was the number of unique SVs children used across both Time 1 and Time 2. To explore Aim 1, timepoint and caregiver NDBI strategy use were included as predictors, as well as demographic and clinical features as covariates (Model 1). Baseline language level was not included as a covariate as the outcome variable encompasses the children's SVs at both T1 and T2. Type of NDBI, caregiver education (bachelor's degree or higher vs. lower than bachelor's degree), and child race (White vs. non-White) were initially included in Model 1 as controls; however, as they were not significant, they were removed for parsimony. To explore Aim 2, caregiver baseline Toy Talk was added into the above model (Model 2), and to explore Aim 3, the interaction of caregiver baseline Toy Talk and caregiver NDBI strategy use change was then also added into the model (Model 3). Baseline Toy Talk was used because Toy Talk was not specifically targeted in the current interventions and therefore was not expected to change over time, whereas NDBI strategies were the main intervention target. Indeed, caregivers' use of Toy Talk did not change significantly between Time 1 and Time 2 (t = 0.57, p = .57). The Akaike information criterion (AIC) for each model was compared to identify whether adding caregiver baseline Toy Talk and the interaction term improved the fit of the model above and beyond the initial model predicting child SV use based on caregiver NDBI strategy use change and demographic and clinical features, using the AICcmodavg package for R Studio (Mazerolle, 2020).

To further explore any interaction effects (Aim 3), caregivers were classified into worsening, stable, and improving NDBI strategy change groups, based on whether MONSI-CC scores changed by more than half a standard deviation from Time 1 to Time 2. The slope of the relation between caregiver Toy Talk and child SV use as predicted by the overall mixed model was then calculated separately for each NDBI strategy change group (“NDBI Worsening,” “Stable,” and “Improving”) using the “emmeans” R Studio package (Lenth et al., 2018). The model compares the slope of linear trends by a predictor, such as a group, using estimated marginal means from the overall mixed model. We then examined whether the slope for each of the three groups was significantly different from 0 based on the confidence intervals of the slopes (UCLA Statistical Consulting Group, 2021). To explore Aim 4, the effect of T1 child demographic and clinical features (sex, baseline age, baseline ADOS CSS SC, baseline NVIQ) on child SV use was examined in the model with the best fit (i.e., lowest AIC value).

We also plotted the individual trajectories of children's SV use between T1 and T2 to further examine variability within the children. Subsequently, among the 35 children who used no SVs at T1, we identified those who maintained no use of SVs at T2 (“SV Stable,” n = 22) and those who showed improvements in SV (“SV Improving,” n = 13). Post hoc t tests were conducted between these two groups to determine whether children who initiated use of unique SV combinations and those who did not differed significantly at baseline on any child clinical features or on caregiver use of Toy Talk sentences.

Results

Changes Over Time and the Effect of NDBI Strategy Use Change (Aim 1)

Model 1 revealed that children as a whole showed improvement in sentence diversity over the course of intervention, with higher average unique SV use at Time 2 (estimated marginal mean = 3.72, SD = 5.78) as compared to Time 1 (estimated marginal mean = 1.18, SD = 2.61). Furthermore, out of 35 children who were not yet using SVs at Time 1, 13 of them (26%) initiated use of SVs between the timepoints. Time was a significant predictor of child-unique SVs (z = 2.58, p < .01). Caregiver change in NDBI strategy use was also a significant predictor of child-unique SVs (z = 2.76, p < .01). The AIC for Model 1 was 340.62 (see Table 2).

Table 2.

Zero-inflated negative binomial mixed models predicting child sentence diversity.

Model 1 source Estimate Standard error z value p value AIC
(Intercept) −9.98 2.65 −3.76 < .001***
Caregiver NDBI strategy change 0.05 0.02 2.76 < .010**
Timepoint 1.09 0.42 2.58 < .006**
Baseline child age 0.05 0.03 1.70 .089
Baseline child NVIQ 0.07 0.02 3.99 < .001***
Baseline child CSS SC 0.02 0.12 0.20 .843
Child sex 0.75 0.68 1.10 .270
Model 1 overall 340.62
Model 2 source
Estimate
Standard error
z value
p value
AIC
(Intercept) −11.81 2.74 −4.31 < .001***
Baseline caregiver Toy Talk 0.07 0.03 2.64 .008**
Caregiver NDBI strategy change 0.04 0.02 1.98 .048*
Timepoint 1.19 0.39 3.09 .002**
Baseline child age 0.06 0.03 2.10 .036*
Baseline child NVIQ 0.07 0.02 3.73 < .001***
Baseline child CSS SC 0.17 0.13 1.29 .197
Child sex 1.07 0.67 1.62 .106
Model 2 overall 333.67
Model 3 source
Estimate
Standard error
z value
p value
AIC
(Intercept) −12.74 2.60 −4.89 < .001***
Baseline caregiver Toy Talk 0.06 0.02 2.57 .010*
Caregiver NDBI strategy change −0.02 0.02 −0.90 .367
Timepoint 1.45 0.37 3.91 < .001***
Baseline child age 0.06 0.03 2.38 .018*
Baseline child NVIQ 0.07 0.02 4.58 < .001***
Baseline child CSS SC 0.20 0.12 1.61 .107
Child sex 1.55 0.64 2.40 .016*
Baseline Caregiver Toy Talk × NDBI Strategy Change interaction 0.01 0.01 2.52 .012*
Model 3 overall 331.73
Open in a new tab

Note. Bold italicized predictors were added into each model. AIC = Akaike information criterion; NDBI = Naturalistic Developmental Behavioral Intervention; NVIQ = nonverbal ratio IQ; CSS SC = calibrated severity scores for social communication.

*

p < .05.

**

p < .01.

***

p < .001.

Effect of Caregiver Toy Talk (Aim 2)

Adding caregiver baseline Toy Talk into the model (Model 2) revealed that baseline caregiver Toy Talk significantly predicted children's improvement in unique SVs over time (z = 2.64, p < .05). The AIC for Model 2 was 333.67, indicating a better fit than Model 1.

Moderating Role of NDBI Strategy Use on the Relation Between Caregiver Toy Talk and Child Sentence Diversity (Aim 3)

The interaction of baseline caregiver Toy Talk and changes in caregiver NDBI strategy use was added in Model 3, and the interaction term was significant (z = 2.52, p < .05). The main effect of caregiver Toy Talk also remained significant in this model. The AIC of this model was 331.73, which was lower than that of Model 1 or 2. Model 3 was therefore identified as the model with best fit, carrying 72% of the cumulative model weight.

Baseline caregiver Toy Talk and child SVs were then plotted separately by the three caregiver NDBI strategy change groups (“NDBI Worsening,” “NDBI Stable,” and “NDBI Improving” groups) based on the overall model using estimated marginal means (see Figure 1). This post hoc analysis demonstrated that the impact of caregiver Toy Talk on child sentence diversity was stronger when caregivers' use of NDBI strategies improved over the course of treatment. Specifically, the slope for the “NDBI Improving” group was significantly different from 0 (β = 0.56, 95% CI [0.23, 0.90]), whereas the slopes for the “Stable” (β = 0.15, 95% CI [−0.36, 0.65]) and “Worsening” (β = −0.01, 95% CI [−0.45, 0.43]) groups were not significantly different from 0. This suggests that the association between caregiver baseline Toy Talk and child unique SVs is stronger for dyads in which caregivers' use of NDBI strategies improved over time, as opposed to those who worsened or stayed stable in their use of NDBI strategies over the 6 months of intervention.

Figure 1.

Figure 1.

Open in a new tab

Interaction between changes in caregiver NDBI strategy use and baseline Toy Talk in predicting estimated marginal means of child total subject–verb (SV) combinations. MONSI–CC = Measure of NDBI Strategy Implementation–Caregiver Change; NDBI = Naturalistic Developmental Behavioral Intervention.

Effects of Demographic and Clinical Features (Aim 4)

Given that Model 3 was identified as the model with best fit, child demographic and clinical features were further explored using this model. Child sex (z = 2.40, p < .05; m female change = 3.00; m male change = 2.44), baseline child age (z = 2.38, p < .05; with older children showing larger improvements), and baseline child NVIQ (z = 4.58, p < .001; with children with higher IQs showing larger improvements) were significant predictors, whereas baseline ADOS-2 SC CSS was not a significant predictor (z = 1.61, p = .11).

Individual trajectories from T1 to T2 using estimated marginal means from the overall model (see Figure 2) revealed large variability across individual children in their use of unique SV combinations over 6 months of intervention, with 54% of children showing no change or a decrease (three children decreased from one to zero combinations and one child decreased from four combinations to one) in unique SV combination use, whereas 46% of children showed improvements in unique SV combinations ranging from 1 to 21 (M = 5.83 change, SD = 5.02).

Figure 2.

Figure 2.

Open in a new tab

Number of unique subject–verb (SV) combinations over the course of the 6-month intervention. Each line represents change in the total number of SV combinations for one child. For children who showed no change over the course of a 6-month intervention, the light blue dot (post) represents their SV combination use at both timepoints. Stable and improving groups included only the subset of children (n = 35) who produced zero SV combinations at the first transcript.

Independent-samples t tests comparing “SV Stable (n = 22; those who had 0 SV combination at both T1 and T2)” and “SV Improving (n = 13; those who had 0 SV combination at T1 but who show improvements in the use of SVs at T2)” groups revealed that children in the “SV Improving” group had higher baseline NVIQ (“SV Stable”: M = 62.52, SD = 14.34; “SV Improving”: M = 76.71, SD = 16.21; p < .001) and lower baseline ADOS-2 CSS SA (“SV Stable”: M = 7.77, SD = 1.80; “SV Improving”: M = 6.00, SD = 1.55; p < .001), and baseline caregiver Toy Talk was higher for the “SV Improving” group (“SV Stable”: M = 5.59, SD = 4.95; “SV Improving”: M = 12.31, SD = 10.67, p < .01). While children across these two groups were not using SV combinations at baseline, children in the “SV Improving” group showed superior language skills more generally in the baseline language sample, including significantly higher MLU-morphemes (“SV Stable”: M = .94, SD = .49; “SV Improving”: M = 1.29, SD = .49, p < .05) and number of different words (“SV Stable”: M = 8.45, SD = 10.78; “SV Improving”: M = 21.92, SD = 16.91, p < .01).

Discussion

The current study investigated the development of sentence diversity in young children with ASD receiving NDBIs. The results indicate that sentence diversity is a useful measure of language development in ASD, given that it was sensitive to change over a relatively short 6-month period of intervention. In concert with recent findings that early verb diversity is a stronger predictor of adult language outcomes in ASD compared to noun diversity and MLU (LeGrand et al., 2021), the current study underscores the importance of measuring sentence diversity as a complement to other language measures, in order to capture a more comprehensive picture of early language development in individuals with ASD. While for the purposes of the current study, language was transcribed and coded using SALT, Hadley et al. (2018) note that recording all SV combinations produced by a child during a language sample can be done easily in real time for children who are producing sentences at low rates, such as those in the current study. Sentence diversity thus represents a practical means of measuring early sentence development, which places relatively low burden on the clinician and can be calculated quickly to facilitate ongoing decision-making (Hadley et al., 2018).

Caregivers' baseline use of Toy Talk was a predictor of overall improvements in sentence diversity above and beyond the contribution of caregivers' NDBI strategy use and also distinguished those children who began using SV combinations between Time 1 and Time 2 from those who were still not using SV combinations after 6 months. It is important to note that caregivers were not trained to use Toy Talk as part of the current intervention. As such, caregivers' use of Toy Talk sentences did not change over the course of intervention, and thus naturally occurring baseline Toy Talk was conceptualized as indicative of ongoing caregiver input throughout the intervention period. Nevertheless, the importance of caregivers' use of Toy Talk sentences in supporting children's sentence development suggests that caregiver-mediated language intervention, such as Toy Talk, is a viable means to support children with ASD primarily using single words and short phrases to make a successful transition to using sentences. Importantly, such intervention has the potential to support more robust change in children's language as compared to the effects of caregivers' naturally occurring Toy Talk in the current study. This has important implications for our understanding of how best to support language acquisition in early intervention for young children with ASD. Currently, many evidence-based NDBIs recommend a “one-word up” strategy for language input from caregivers (i.e., caregivers are instructed to expand language by adding one word to the children's utterances; Schreibman et al., 2015). However, a recent study found that when children demonstrate early language delays, caregivers simply increasing their use of object labels in an attempt to provide additional support is not effective in facilitating language development (Roemer et al., 2021). Thus, our findings, combined with the results from past studies, suggest that it is not just the length of caregivers' utterances that is important for language acquisition, but the provision of lexically rich input through simple sentences with diverse subjects that promotes children's ability to combine words into diverse sentences (Hadley, 2020). Therefore, leveraging Toy Talk to model lexically rich sentences with diverse subjects may be especially effective in improving sentence development in children with ASD.

Caregivers' improvement in use of NDBI interactive strategies significantly moderated the effect of Toy Talk on child sentence diversity in the current study. In fact, the best-fitting model included the interaction of caregiver Toy Talk and NDBI strategy use, indicating that the combination of NDBI strategy improvement and caregiver Toy Talk has added value in explaining variability in children's sentence diversity beyond the main effect of each. Improvements in caregivers' use of NDBI strategies such as following the child's lead in play, structuring turn-taking, and scaffolding the child's play and communication likely supported child–caregiver engagement and facilitated children's development of SC behaviors such as joint attention, imitation, and reciprocal interaction over the intervention period. Given that previous research has shown that SC deficits interfere with language learning in ASD (Ellis Weismer & Kover, 2015; Luyster et al., 2008; Paul et al., 2013), the improvements in SC facilitated by improvements in caregivers' NDBI strategy use likely played an important role in children's ability to attend to and make use of caregivers' Toy Talk sentences. Indeed, in the current study, baseline SC symptom severity was an important predictor of onset of SV combination use. As such, the current findings may be explained by a cascading effect in which NDBI strategy implementation can support reciprocal caregiver–child engagement and child SC skills, which in turn can enhance children's attention to the linguistic input from their caregiver, maximizing the contribution of caregivers' ongoing use of Toy Talk sentences throughout the period of NDBI. Taken together, the current findings therefore suggest that caregiver-mediated NDBI could provide a strong foundation for more targeted language intervention using Toy Talk for children with ASD and delays in early grammatical development. Future research that integrates caregiver training in NDBI strategies and Toy Talk into a comprehensive language intervention program would be beneficial in evaluating the efficacy of this combined approach.

It is important to note that only about half the children in the current study demonstrated improvements in sentence diversity between timepoints and that 22 children did not begin using SV combinations by Time 2. It is possible that, while toddlers who are developing typically have been shown to improve in sentence diversity over a period of 6 months (Hadley, Rispoli, Holt, Papastratakos, et al., 2017), a longer or more intensive intervention might be required to see changes in sentence diversity among a subset of young children with ASD, particularly given that sentence diversity was not a core target in the current intervention. In addition, as the current study did not exclude children whose language was below a certain threshold, it is possible that many of these children were not developmentally ready to initiate SV combination use. In fact, Hadley et al. (2018) recommend assessment of sentence diversity among children with parent-reported vocabulary size of at least 100 words including at least 20 verbs, who are regularly combining words and have an MLU between 1.50 and 3.0. Those children in the current study who did not initiate use of SV combinations between timepoints (“SV Stable” group) had baseline language levels significantly below these recommended indicators of readiness for SV combination use and significantly lower than those who used no SV combinations at baseline but did initiate use between timepoints (“SV Improving” group). Beyond this, given that children with ASD may lag behind their same-age peers in sentence diversity and complexity even when MLU is equivalent, further studies are needed to examine child demographic and clinical features that may be predictors of treatment response for interventions targeting the development of sentence diversity. Consistent with previous studies of language development in ASD (Ellis Weismer & Kover, 2015; Ellis Weismer et al., 2010; Luyster et al., 2008; Thurm et al., 2007), children with higher baseline NVIQ showed greater improvement in sentence diversity in the current study. Thus, cognitive abilities may be an important predictor of treatment response, consistent with findings from other NDBIs (Klinger et al., 2020). In contrast, baseline SC severity was not predictive of progress in sentence diversity within the overall sample, but distinguished those who started to use SV combinations between Time 1 and Time 2 from those who did not. This finding might indicate that for children with limited vocabulary abilities, social skills like joint attention and imitation play a particularly critical role in facilitating the child's ability to engage in and learn from reciprocal interactions (Bono et al., 2004; Ellis Weismer & Kover, 2015; Luyster et al., 2008; Paul et al., 2013). Consistent with typical patterns of language development (Eriksson et al., 2012; Hulle et al., 2004), girls showed greater improvements in sentence diversity over the 6-month period, though this finding should be interpreted with caution given that only 18% of the children were female. Past research has shown inconsistent effects of age on intervention response in ASD (Klinger et al., 2020), whereas in the current study, older children demonstrated more growth in sentence diversity.

Limitations and Future Directions

Our study leveraged data from a subset of children who were receiving treatment from larger NDBI studies; therefore, examining additional potential moderators of intervention response was beyond its scope. Future research should aim to identify indicators that a child is likely to benefit from a targeted grammatical intervention such as Toy Talk, as well as the optimal timing for commencing such intervention, including indicators of treatment readiness in areas such as vocabulary size and MLU, nonverbal skills, and SC skills. These findings would support clinicians in making informed treatment decisions to maximize intervention benefits.

Importantly, the treatment programs received by participants in the current study were not designed to improve caregiver Toy Talk specifically. Therefore, the current study was limited to examining the association between caregivers' naturally occurring use of Toy Talk sentences at baseline and children's sentence diversity outcomes. Future research should also explore how language input changes over the course of caregiver-mediated language intervention, as caregivers specifically learn and implement NDBI strategies as well as Toy Talk. In the current study, all instances of caregivers' naturally occurring Toy Talk (i.e., both unique and repeated) were counted. However, given that training caregivers on use of Toy Talk has been shown to increase the diversity of caregiver input (Hadley, Rispoli, Holt, Papastratakos, et al., 2017; Hadley & Walsh, 2014), future studies should also explore how Toy Talk sentences with unique subjects effect child sentence diversity. Furthermore, while caregivers were instructed to play with their child as they normally would during the video-recorded interaction samples, it is important to note that brief video recordings often capture more lexically rich input compared to that which children receive in everyday contexts (Bergelson et al., 2019). Thus, future studies that train caregivers on Toy Talk might benefit from the collection of natural language samples more frequently and across a variety of contexts to understand how caregivers integrate Toy Talk strategies into their everyday interactions, as well as the specific mechanisms of change in child sentence diversity. More frequent measurement of child SC symptoms over the course of NDBI and Toy Talk intervention would also be key to understanding these mechanisms.

Conclusions

The current study demonstrated that sentence diversity (Hadley et al., 2018) is a sensitive measure of developmental change over relatively brief periods of treatment for young children with ASD who are producing simple sentences. This study also documented that naturally occurring differences in caregivers' use of input sentences with diverse third-person noun subjects and verb combinations (“Toy Talk”) promoted sentence diversity in young children with ASD, above and beyond the level of caregiver NDBI strategy use. Of particular importance was the finding that the impact of Toy Talk sentences was more robust when caregiver use of NDBI strategies improved over the course of intervention. Thus, targeted language interventions are likely to be optimized when used in conjunction with NDBI strategies. This finding emphasizes the value of the prelinguistic developmental skills that children build through NDBIs, such as joint attention, in increasing opportunities for learning and language development. Together, leveraging NDBI intervention, Toy Talk for caregivers, and monitoring of children's sentence diversity may provide a feasible and empirically sound approach to improving language outcomes in children with ASD.

Author Contributions

Elysha Clark-Whitney: Data curation (Equal), Formal analysis (Equal), Writing – original draft (Lead), Writing – review & editing (Equal), Claire Brito Klein: Conceptualization (Supporting), Data curation (Equal), Formal analysis (Equal), Methodology (Supporting), Writing – original draft (Supporting), Writing – review & editing (Equal), Pamela A. Hadley: Conceptualization (Supporting), Supervision (Supporting), Writing – review & editing (Supporting), Catherine Lord: Methodology (Supporting), Project administration (Equal), Supervision (Supporting), Writing – review & editing (Supporting), So Hyun Kim: Conceptualization (Lead), Funding acquisition (Lead), Methodology (Lead), Supervision (Lead), Writing – original draft (Supporting), Writing – review & editing (Supporting).

Acknowledgments

This study was funded by the National Institute of Mental Health (1R01 MH114925), awarded to author So Hyun Kim. The authors would like to thank all of the families who participated in this study for their time. They would also like to thank their data manager Shanping Qui for her efforts coordinating this data set, as well as to the research assistants who coded the interaction videos.

Funding Statement

This study was funded by the National Institute of Mental Health (1R01 MH114925), awarded to author So Hyun Kim.

References

  1. Bergelson, E. , Amatuni, A. , Dailey, S. , Koorathota, S. , & Tor, S. (2019). Day by day, hour by hour: Naturalistic language input to infants. Developmental Science, 22, Article No. e12715. https://doi.org/10.1111/desc.12715 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bono, M. A. , Daley, T. , & Sigman, M. (2004). Relations among joint attention, amount of intervention and language gain in autism. Journal of Autism and Developmental Disorders, 34(5), 495–505. https://doi.org/10.1007/s10803-004-2545-x [DOI] [PubMed] [Google Scholar]
  3. Brooks, M. E. , Kristensen, K. , van Benthem, K. J. , Magnusson, A. , Berg, C. W. , Nielsen, A. , Skaug, H. J. , & Mächler, M. Bolker, B. M. (2017). GlmmTMB balances speed and flexibility among packages for zero-inflated generalized linear mixed modeling. The R Journal, 9(2), 378–400. https://doi.org/10.32614/RJ-2017-066 [Google Scholar]
  4. Charman, T. , Taylor, E. , Drew, A. , Cockerill, H. , Brown, J.-A. , & Baird, G. (2005). Outcome at 7 years of children diagnosed with autism at age 2: Predictive validity of assessments conducted at 2 and 3 years of age and pattern of symptom change over time. The Journal of Child Psychology and Psychiatry, 46(5), 500–513. https://doi.org/10.1111/j.1469-7610.2004.00377.x [DOI] [PubMed] [Google Scholar]
  5. Crandall, M. C. , McDaniel, J. , Watson, L. R. , & Yoder, P. J. (2019). The relation between early parent verb input and later expressive verb vocabulary in children with Autism Spectrum Disorder. Journal of Speech, Language, and Hearing Research, 62(6), 1787–1797. https://doi.org/10.1044/2019_JSLHR-L-18-0081 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Dawson, G. , Rogers, S. , Munson, J. , Smith, M. , Winter, J. , Greenson, J. , Donaldson, A. , & Varley, J. (2010). Randomized, controlled trial of an intervention for toddlers with autism: The Early Start Denver Model. Pediatrics, 125(1), e17–e23. https://doi.org/10.1542/peds.2009-0958 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Eigsti, I.-M. , Bennetto, L. , & Dadlani, M. B. (2007). Beyond pragmatics: Morphosyntactic development in autism. Journal of Autism and Developmental Disorders, 37(6), 1007–1023. https://doi.org/10.1007/s10803-006-0239-2 [DOI] [PubMed] [Google Scholar]
  8. Ellis Weismer, S. , & Kover, S. T. (2015). Preschool language variation, growth, and predictors in children on the autism spectrum. The Journal of Child Psychology and Psychiatry, 56(12), 1327–1337. https://doi.org/10.1111/jcpp.12406 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Ellis Weismer, S. , Lord, C. , & Esler, A. (2010). Early language patterns of toddlers on the autism spectrum compared to toddlers with developmental delay. Journal of Autism and Developmental Disorders, 40, 1259–1273. https://doi.org/10.1007/s10803-010-0983-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Eriksson, M. , Marschik, P. B. , Tulviste, T. , Almgren, M. , Pérez Pereira, M. , Wehberg, S. , Marjanovič-Umek, L. , Gayraud, F. , Kovacevic, M. , & Gallego, C. (2012). Differences between girls and boys in emerging language skills: Evidence from 10 language communities. British Journal of Developmental Psychology, 30(2), 326–343. https://doi.org/10.1111/j.2044-835X.2011.02042.x [DOI] [PubMed] [Google Scholar]
  11. Esler, A. N. , Hus Bal, V. , Guthrie, W. , Wetherby, A. , Ellis Weismer, S. , & Lord, C. (2009). The Autism Diagnostic Observation Schedule, toddler module: Standardized severity scores. Journal of Autism and Developmental Disorders, 45(9), 2704–2720. https://doi.org/10.1007/s10803-015-2432-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fusaroli, R. , Weed, E. , Fein, D. , & Naigles, L. (2019). Hearing me hearing you: Reciprocal effects between child and parent language in autism and typical development. Cognition, 183, 1–18. https://doi.org/10.1016/j.cognition.2018.10.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Gotham, K. , Pickles, A. , & Lord, C. (2009). Standardizing ADOS scores for a measure of severity in autism spectrum disorders. Journal of Autism and Developmental Disorders, 39(5), 693–705. https://doi.org/10.1007/s10803-008-0674-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hadley, P. A. (2014). Approaching early grammatical intervention from a sentence-focused framework. Language, Speech, and Hearing Services in Schools, 45(2), 110–116. https://doi.org/10.1044/2014_LSHSS-14-0017 [DOI] [PubMed] [Google Scholar]
  15. Hadley, P. A. (2020). Exploring sentence diversity at the boundary of typical and impaired language abilities. Journal of Speech, Language, and Hearing Research, 63(10), 3236–3251. https://doi.org/10.1044/2020_JSLHR-20-00031 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hadley, P. A. , McKenna, M. M. , & Rispoli, M. (2018). Sentence diversity in early language development: Recommendations for target selection and progress monitoring. American Journal of Speech-Language Pathology, 27(2), 553–565. https://doi.org/10.1044/2017_AJSLP-17-0098 [DOI] [PubMed] [Google Scholar]
  17. Hadley, P. A. , Rispoli, M. , & Holt, J. K. (2017). Input subject diversity accelerates the growth of tense and agreement: Indirect benefits from a parent-implemented intervention. Journal of Speech, Language, and Hearing Research, 60(9), 2619–2635. https://doi.org/10.1044/2017_JSLHR-L-17-0008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hadley, P. A. , Rispoli, M. , Holt, J. , Fitzgerald, C. , & Bahnsen, A. (2014). Growth of finiteness in the third year of life: Replication and predictive validity. Journal of Speech, Language, and Hearing Research, 57, 887–900. https://doi.org/10.1044/2013_JSLHR-L-13-0008 [DOI] [PubMed] [Google Scholar]
  19. Hadley, P. A. , Rispoli, M. , Holt, J. K. , Papastratakos, T. , Hsu, N. , Kubalanza, M. , & McKenna, M. M. (2017). Input subject diversity enhances early grammatical growth: Evidence from a parent-implemented intervention. Language Learning and Development, 13(1), 54–79. https://doi.org/10.1080/15475441.2016.1193020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hadley, P. A. , & Walsh, K. M. (2014). Toy Talk: Simple strategies to create richer grammatical input. Language, Speech, and Hearing Services in Schools, 45(3), 159–172. https://doi.org/10.1044/2014_LSHSS-13-0055 [DOI] [PubMed] [Google Scholar]
  21. Hulle, C. A. V. , Goldsmith, H. H. , & Lemery, K. S. (2004). Genetic, environmental, and gender effects on individual differences in toddler expressive language. Journal of Speech, Language, and Hearing Research, 47(4), 904–912. https://doi.org/10.1044/1092-4388(2004/067) [DOI] [PubMed] [Google Scholar]
  22. Hus, V. , Gotham, K. , & Lord, C. (2014). Standardizing ADOS domain scores: Separating severity of social affect and restricted and repetitive behaviors. Journal of Autism and Developmental Disorders, 44(10), 2400–2412. https://doi.org/10.1007/s10803-012-1719-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kasari, C. , Gulsrud, A. , Paparella, T. , Hellemann, G. , & Berry, K. (2015). Randomized comparative efficacy study of parent-mediated interventions for toddlers with autism. Journal of Consulting and Clinical Psychology, 83(3), 554–563. https://doi.org/10.1037/a0039080 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kim, S. H. , Paul, R. , Tager-Flusberg, H. , & Lord, C. (2014). Language and communication in autism. In Volkmar F. R., Paul R., Rogers S. J., & Pelphrey K. A. (Eds.), Handbook of autism and pervasive developmental disorders (4th ed., pp. 230–262). Wiley. [Google Scholar]
  25. Klee, T. , & Gavin, W. (2010). LARSP reference data for 2- and 3-year-old children. University of Canterbury Research Repository. [Google Scholar]
  26. Klinger, L. G. , Cook, M. L. , & Dudley, K. M. (2020). Predictors and moderators of treatment efficacy in children and adolescents with autism spectrum disorder. Journal of Clinical Child & Adolescent Psychology, 50(4), 517–524. https://doi.org/10.1080/15374416.2020.1833735 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lee, L. (1974). Developmental sentence analysis: A grammatical assessment procedure for speech and language clinicians. Northwestern University Press. [Google Scholar]
  28. LeGrand, K. L. , Weil, L. W. , Lord, C. , & Luyster, R. J. (2021). Identifying childhood expressive language features that best predict adult language and communication outcome in individuals with autism spectrum disorder. Journal of Speech, Language, and Hearing Research, 64(6), 1977–1991. https://doi.org/10.1044/2021_JSLHR-20-00544 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lenth, R. V. , Buerkner, P. , Herve, M. , Love, J. , Riebel, H. , & Singmann, H. (2018). emmeans: Estimated marginal means, aka least-squares means. R package version 1.6.1. https://CRAN.R-project.org/package=emmeans
  30. Lord, C. , Luyster, R. J. , Gotham, K. , & Guthrie, W. (2012). Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) Manual (Part II): Toddler Module. Western Psychological Services. [Google Scholar]
  31. Lord, C. , Rutter, M. , DiLavore, P. C. , Risi, S. , Gotham, K. , & Bishop, S. L. (2012). Autism Diagnostic Observation Schedule: ADOS-2. Western Psychological Services. [Google Scholar]
  32. Luyster, R. J. , Kadlec, M. B. , Carter, A. , & Tager-Flusberg, H. (2008). Language assessment and development in toddlers with autism spectrum disorders. Journal of Autism and Developmental Disorders, 38, 1426–1438. https://doi.org/10.1007/s10803-007-0510-1 [DOI] [PubMed] [Google Scholar]
  33. Maglione, M. A. , Gans, D. , Das, L. , Timbie, J. , & Kasari, C. (2012). Nonmedical interventions for children with ASD: Recommended guidelines and further research needs. Pediatrics, 130(Suppl. 2), S169−S178. https://doi.org/10.1542/peds.2012-0900O [DOI] [PubMed] [Google Scholar]
  34. Mazerolle, M. J. (2020). AICcmodavg: Model selection and multimodel inference based on (Q)AIC(c). R package version 2.3–1. https://cran.r-project.org/package=AICcmodavg
  35. Mazurek, M. O. , Baker-Ericzén, M. , & Kanne, S. M. (2019). Brief report: Calculation and convergent and divergent validity of a new ADOS-2 expressive language score. American Journal on Intellectual and Developmental Disabilities, 124(5), 438–449. https://doi.org/10.1352/1944-7558-124.5.438 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. McKenna, M. M. , & Hadley, P. A. (2014). Assessing sentence diversity in toddlers at-risk for language disorders. SIG 1 Perspectives on Language Learning and Education, 21(4), 159–172. https://doi.org/10.1044/lle21.4.159 [Google Scholar]
  37. Miller, J. , & Iglesias, A. (2012). Systematic Analysis of Language Transcripts (SALT), Research Version 2012 [Computer Software] . SALT Software, LLC. [Google Scholar]
  38. Mullen, E. M. (1995). Mullen Scales of Early Learning. AGS. [Google Scholar]
  39. Naigles, L. (1990). Children use syntax to learn verb meanings. Journal of Child Language, 17(2), 357–374. https://doi.org/10.1017/S0305000900013817 [DOI] [PubMed] [Google Scholar]
  40. National Autism Center. (2015). Findings and conclusions: National standards project, phase 2.
  41. Paul, R. , Campbell, D. , Gilbert, K. , & Tsiouri, I. (2013). Comparing spoken language treatments for minimally verbal preschoolers with autism spectrum disorders. Journal of Autism and Developmental Disorders, 43, 418–431. https://doi.org/10.1007/s10803-012-1583-z [DOI] [PubMed] [Google Scholar]
  42. Paul, R. , Chawarska, K. , Ciccheti, D. , & Volkmar, F. (2008). Language outcomes of toddlers with autism spectrum disorders: A two year follow-up. Autism Research, 1(2), 97–107. https://doi.org/10.1002/aur.12 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Pickles, A. , Anderson, D. K. , & Lord, C. (2014). Heterogeneity and plasticity in the development of language: A 17-year follow-up of children referred early for possible autism. The Journal of Child Psychology and Psychiatry, 55(12), 1354–1362. https://doi.org/10.1111/jcpp.12269 [DOI] [PubMed] [Google Scholar]
  44. Rappin, I. , & Allen, D. A. (1988). Syndromes in developmental dysphasia and adult aphasia. Research Publications–Association for Research in Nervous and Mental Disease, 66, 57–75. [PubMed] [Google Scholar]
  45. Rispoli, M. , Hadley, P. , & Simmons, H. (2018). Simple sentences aren't all the same: Variation in input and acquisition. In Proceedings of the 42nd annual Boston University Conference on Language Development (pp. 673−686). Cascadilla Press. [Google Scholar]
  46. Roberts, M. Y. , & Kaiser, A. P. (2011). The effectiveness of parent-implemented language interventions: A metaanalysis. American Journal of Speech-Language Pathology, 20(3), 180–199. https://doi.org/10.1044/1058-0360(2011/10-0055) [DOI] [PubMed] [Google Scholar]
  47. Roemer, E. J. , Kushner, E. H. , & Iverson, J. M. (2021). Joint engagement, parent labels, and language development: Examining everyday interactions in infant siblings of children with autism. Journal of Autism and Developmental Disorders, 1–20. https://doi.org/10.1007/s10803-021-05099-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Scarborough, H. S. , Rescorla, L. , Tager-Flusberg, H. , Fowler, A. E. , & Sudhalter, V. (1991). The relation of utterance length to grammatical complexity in normal and language-disordered groups. Applied Psycholinguistics,, 12, 23–46. https://doi.org/10.1017/S014271640000936X [Google Scholar]
  49. Schreibman, L. , Dawson, G. , Stahmer, A. C. , Landa, R. , Rogers, S. J. , McGee, G. G. , Kasari, C. , Ingersoll, B. , Kaiser, A. P. , Bruinsma, Y. , McNerney, E. , Wetherby, A. , & Halladay, A. (2015). Naturalistic Developmental Behavioral Interventions: Empirically validated treatments for autism spectrum disorder. Journal of Autism and Developmental Disorders, 45(8), 2411–2428. https://doi.org/10.1007/s10803-015-2407-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Thurm, A. , Lord, C. , Lee, L.-C. , & Newschaffer, C. (2007). Predictors of language acquisition in preschool children with autism spectrum disorders. Journal of Autism and Developmental Disorders, 37, 1721–1734. https://doi.org/10.1007/s10803-006-0300-1 [DOI] [PubMed] [Google Scholar]
  51. UCLA Statistical Consulting Group. (2021). Decomposing, probing, and plotting interactions in R. https://stats.idre.ucla.edu/r/seminars/interactions-r/#s3e
  52. Vibert, B. A. , Dufek, S. , Klein, C. B. , Choi, Y. B. , Winter, J. , Lord, C. , & Kim, S. H. (2020). Quantifying caregiver change across early autism interventions using the Measure of NDBI Strategy Implementation: Caregiver Change (MONSI-CC). Journal of Autism and Developmental Disorders, 50(4), 1364–1379. https://doi.org/10.1007/s10803-019-04342- 0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Vivanti, G. , & Zhong, H. N. (2020). Naturalistic developmental behavioural interventions for children with autism. In Vivanti G., Bottema-Beutel K., & Turner-Brown L. (Eds.), Clinical guide to early interventions for children with autism: Best practices in child and adolescent behavioral health care (pp. 93–130). Springer. https://doi.org/10.1007/978-3-030-41160-2_6 [Google Scholar]
  54. Weitzman, E. , Girolametto, L. & Drake, L. (2014). Hanen programs for parents: Parent-implemented early language intervention. In McCauley R., Fey M., Gillam R. (Eds.), Treatment of language disorders in children (2nd ed., pp. 27–56). Brookes.
  55. Wetherby, A. M. , Guthrie, W. , Woods, J. , Schatschneider, C. , Holland, R. D. , Morgan, L. , & Lord, C. (2014). Parent-implemented social intervention for toddlers with autism: An RCT. Pediatrics, 134(6), 1084–1093. https://doi.org/10.1542/peds.2014-0757 [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Journal of Speech, Language, and Hearing Research : JSLHR are provided here courtesy of American Speech-Language-Hearing Association

RESOURCES