Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2023 Sep 24;59(1):413–425. doi: 10.1111/1460-6984.12956

Prelinguistic communication complexity predicts expressive language in initial minimally verbal autistic children

Min Liu 1,2, Nancy C Brady 3, Olivia Boorom 3, Kandace Fleming 4, Jiaojiao Yue 1, Qiaoyun Liu 1,✉
PMCID: PMC13504684  PMID: 37743638

Abstract

Background

Prelinguistic communication complexity refers to the use of different communication forms such as eye gaze, gestures and vocalisations and the degree to which these forms are coordinated and how directed to a communication partner. To date, little is known about the relationship between prelinguistic communication complexity and expressive language in minimally verbal autistic children.

Aims

To test the hypothesis that prelinguistic communication complexity predicts expressive language 12 months later in autistic children and explore whether there are any differences in specific prelinguistic intentional communicative behaviours that are related to later expressive language levels.

Methods & Procedures

This longitudinal study examined 37 minimally verbal autistic children (29–71 months old). The Communication Complexity Scale (CCS) was used to measure participants’ prelinguistic communication behaviours, which were extracted from a semi‐structured play interaction at Time 1. The Chinese Communicative Development Inventory (CCDI) was used to examine participants’ expressive language at Time 1 and Time 2 (12 months later). According to Time 2 vocabulary size, participants were divided into two groups: Low CCDI, between 0 and 62 words, and High CCDI, more than 100 words. Linear regression was used to examine the relationship between early prelinguistic communication complexity and later expressive language. Binary logistic regression was used to determine which of the early communication behaviours were uniquely significantly related to later expressive language levels.

Outcomes & Results

There was a significant positive relationship between prelinguistic communication complexity and expressive language 12 months later, even after controlling for age and concurrent language. Findings revealed a group difference in the frequency of gesture and vocalisation combinations between the Low and High CCDI groups at Time 1. Gesture‐vocalisation combinations also predicted better expressive language levels at Time 2.

Conclusions & Implications

Our findings suggest that it may be beneficial to incorporate different complex communication behaviours into prelinguistic intervention targets for minimally verbal autistic children. The CCS hierarchies can be used as a reference for the intervention goals of minimally verbal autistic children. These findings highlight the importance of targeting gesture and vocalisation combinations when autistic children transition from single prelinguistic communication behaviours to multimodal behaviours.

What this paper adds

What is already known on this subject

  • Children use eye gaze, gestures and vocalisations to communicate with others before they learn spoken language. There is strong evidence suggest that the frequency of prelinguistic communication predicts later linguistic achievements in autistic children. However, less is known about whether prelinguistic communication complexity also predicts later language and which specific behaviours are most predictive of language outcomes.

What this study adds

  • Minimally verbal autistic children who exhibit more complex prelinguistic communication behaviours have better expressive language 12 months later. Gestures combined with vocalisations predict better expressive language in minimally verbal autistic children.

What are the clinical implications of this work?

  • When identifying intervention targets for minimally verbal autistic children, the clinicians may reference the prelinguistic communication behaviours from the CCS. The gesture and vocalisation combinations are the key behaviours when targets transit from single form to two‐form behaviours.

Keywords: autistic children, expressive language, minimally verbal, multimodal behaviour, prelinguistic communication complexity

INTRODUCTION

Early expressive language is a predictor of long‐term social‐communicative, cognitive and adaptive outcomes for autistic children (Arnett et al., 2020; LeGrand et al., 2021). Delays in language acquisition are common in autistic children and approximately 65% of autistic children are minimally verbal before 5 years old (Maltman et al., 2021). Targeting prelinguistic communication skills during early intervention is one of the ways to improve expressive language for autistic children (Mohamadi et al., 2022). Identifying specific prelinguistic communication skills that predict expressive language leads to tailoring intervention targets to improve autistic children's long‐term outcomes.

Prelinguistic communication behaviours refer to the use eye gaze, gestures and vocalisations to communicate with others. Children's prelinguistic communication behaviours often lead to adult responses that may further enhance language development (Albert et al., 2023; van der Klis et al., 2023). Contingent adult responses such as verbal labels facilitate the development of expressive language by helping children map words to referents (Olson & Masur, 2015; Tamis‐LeMonda et al., 2014).

A number of studies have explored the relationship between early prelinguistic communication behaviours and expressive language in neurotypical children and autistic children. For typically developing (TD) children, studies suggest that frequency or number of gestures (e.g., pointing, showing) are positively associated with expressive language (Choi et al., 2021; Manwaring et al., 2019; Stewart et al., 2021), and early frequency or quantity of vocalisations (e.g., babbling) predict later expressive language (Donnellan et al., 2020; Werwach et al., 2021). These relationships were also found in autistic children. For example, there is evidence that the number of gestures predict later expressive language (Dimitrova et al., 2020; Ramos‐Cabo et al., 2022), proportion of communicative vocalisations and consonant inventory are associated with later expressive language (Gerhold et al., 2020; McDaniel et al., 2020; Yoder et al., 2015), and there was a trend for a positive association between the frequency of communicative eye gaze and later expressive language (Hahn et al., 2019) in autistic children. However, this body of research has primarily focused on the number of individual early prelinguistic communication skills and their relationship to later expressive language. By contrast, less attention has been paid to the role of prelinguistic communication complexity and its relationship to expressive language development.

Prelinguistic communication complexity refers to the degree that different communication forms such as eye gaze, gestures and vocalisations are coordinated or combined with each other and how these forms are directed to another person (Brady et al., 2018; Salley et al., 2020). Prelinguistic communication includes single and multimodal behaviours. Single communicative behaviours refer to use of gestures, eye gaze or vocalisation alone to communicate and multimodal communicative behaviours refer to combine two or more forms (i.e., gesture + vocalisation, gesture + gaze, gaze + vocalisation and gesture +gaze + vocalisation) (Murillo et al., 2021). Only a few studies have explored the relationship between communication complexity and concurrent expressive language (Brady et al., 2012; Salley et al., 2020). We do not yet know whether prelinguistic communication complexity predicts later expressive language development specifically in autistic children. Thus, this study focuses on the relationship between prelinguistic communication complexity and later expressive language in autistic children.

Although multiple studies have explored the relationship between prelinguistic behaviours and later language, they have tended not to distinguish different types of prelinguistic behaviours (McDaniel et al., 2019; Shumway & Wetherby, 2009). For example, one study integrated all of the different multimodal communication behaviours together for analysis (Shumway & Wetherby, 2009). A recent study systematically examined whether single and multimodal communicative behaviours distinguish autistic children with different concurrent expressive language abilities (Murillo et al., 2021). But we know little about what specific single and multimodal prelinguistic communication behaviours may be associated with higher levels of later expressive language skill in autistic children. In the present study, we further examine whether there were any differences in specific types of prelinguistic communication behaviours between children who developed small versus large vocabulary sizes 12 months later.

The relationship between prelinguistic communication complexity and concurrent language

Prelinguistic communication complexity provides information on the developmental changes in prelinguistic communication behaviours from simple to complex. It is important to investigate the relationship between complexity and expressive language, since it can provide the evidence to support different targeted prelinguistic intervention goals. Previous research has explored the relationship between prelinguistic communication complexity and concurrent expressive language. For children with intellectual disabilities, there is a moderate relationship between communication complexity and expressive language from the Mullen Scales of Early Learning (Brady et al., 2012). Additionally, TD infants’ communication complexity also demonstrates a positive relationship with their expressive and receptive vocabulary, as measured by the MacArthur‐Bates Communicative Development Inventory (MCDI) (Salley et al., 2020). Based on the relationship between concurrent language and communication complexity, we hypothesised that autistic children who have higher prelinguistic communication complexity will have better expressive vocabulary 12 months later.

One tool developed to measure communication complexity is the Communication Complexity Scale (CCS) (Brady et al., 2018). The CCS takes into account the complexity of prelinguistic communication. Brady and colleagues developed the CCS to systematically measure the continuum of expressive prelinguistic communication skills in individuals with minimal verbal skills. Previous studies assessed more than 200 children with intellectual disabilities including autistic children between 10 months and 16 years of age (Brady et al., 2018, 2012). They confirmed that the CCS had high test‐retest and interrater reliability (0.83–0.99) and moderate concurrent validity (0.40–0.44).

The CCS is a 12‐point scale which can measure increasingly complex behaviours reflecting emerging levels of communication achievement and it has been used in various populations. For example, the CCS was used to measure the social communication levels in infants with elevated likelihood of autism (Hahn et al., 2017). The CCS also captured differences in TD infants’ prelinguistic complexity from 6 to 12 months, whereby older infants’ prelinguistic communication behaviours were more complex (Salley et al., 2020). Further, the CCS was sensitive to subtle changes in prelinguistic communication complexity in autistic children receiving communication interventions (Brady et al., 2020). Children who experienced an intervention focused on social communication significantly increased scores on the CCS over a 6‐month period. These studies show that measuring complexity could add valuable information about children's prelinguistic communication that may be difficult to capture using other assessment tools.

The relationship between single and multimodal prelinguistic communication and concurrent language

There is currently not enough evidence to support the relationship between specific single and multimodal prelinguistic communication behaviours and expressive language. Murillo and colleagues looked at differences in specific types of single and multimodal communication behaviours of two group autistic children whose expressive language at different levels (Murillo et al., 2021). They divided 11 autistic children aged between 28 to 79 months into low‐vocabulary (12 or less words) and high‐vocabulary (100 words or more) groups based on expressive vocabulary which was measured by the MCDI.

They compared whether there were differences among three single forms of communication behaviours (i.e., gesture, eye gaze and vocalisation) and four multimodal behaviours (i.e., gesture + gaze, gesture + vocalisation, gaze + vocalisation and gaze + gesture + vocalisation) between the two groups of autistic children. For the single form behaviours, autistic children with different language levels showed different patterns. Autistic children in the low MCDI group used more gestures than those in the high MCDI group. However, the pattern for vocalisations was reversed, children in the high MCDI group produced more vocalisations than children in the low MCDI group. There was no difference in the use of eye gaze between two groups. For the multimodal behaviours, there was no significant difference in the frequency of behaviours associated with eye gaze between the two groups of autistic children. This may be due to their low use of eye gaze as a communicative resource in both groups. Additionally, although the two groups of autistic children differed in gestural and vocal single behaviours, they combined the two behaviours similar, resulting in equally gesture and vocalisation combinations in both groups.

However, the sample size in this study was small (n = 11), making it difficult to generalise these conclusions. Additional research is needed to replicate these findings measuring both single and multimodal behaviours and extend this work by examining how these different types of prelinguistic communication predict later language outcomes in autistic children.

Current study

This study focused on the relationship between prelinguistic communication skills and later expressive language in minimally verbal autistic children (i.e., less than 20 functional spoken words). The first aim was to test the hypothesis that higher prelinguistic communication complexity predicts better expressive language 12 months later. The second aim was to explore whether there were any differences in specific types of prelinguistic communication behaviours between children who developed small versus large vocabulary sizes 12 months later.

METHODS

Participants

Thirty‐seven minimally verbal autistic children aged 29–71 months—expressive vocabulary <20 on the Chinese Communicative Development Inventory—Mandarin Version (CCDI)—participated. Twenty‐nine were boys and 8 were girls (see Table 1). Families were contacted through facilities serving autistic children as part of a larger study on prelinguistic complexity assessment (Liu et al., 2023). There were 67 autistic children in the previous study. At the follow‐up, 26 participants had left the institution where they were recruited and could not be contacted. Parents of four of the participants were contacted but did not provide data on their children's expressive language. Families were contacted if their children met the inclusion criteria: (a) diagnosed with autism spectrum disorder by expert clinicians based on the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition; (b) chronological age between 2;0 and 5;11 years; (c) expressed less than 20 functional spoken words at the time of study enrollment confirmed by parent report; (d) demonstrated no severe sensory and motor impairments; and (e) monolingual and native language was Mandarin Chinese. Parents volunteered to participated in the study and signed informed consent. The study was approved by East China Normal University Ethics Committee (certificate number HR 590–2020). We used the Childhood Autism Rating Scale (CARS)—Chinese Version to confirm autism diagnoses. The reliability and validity of the CARS in Chinese context are adequate (Lu et al., 2004). Participants with a CARS score ≥30 was included in the study (Lu et al., 2004).

TABLE 1.

Description of participants.

Low group (N = 27) High group (N = 10)
Variable Mean SD Mean SD t df p
CA 47.89 9.60 43.10 7.28 1.428 35 0.162
CARS 36.74 3.10 35.20 2.74 1.383 35 0.175
Gender Female Male Female Male
7 20 1 9
Open in a new tab

Abbreviations: CA, chronological age in months; CARS, Childhood Autism Rating Scale.

Procedures and materials

Participants and their parents visited the clinic two times 12 months apart. Between visits, subjects received general early intervention, which included fine and gross motor, cognition, receptive and expressive language or social communication. Intervention sessions lasted around 1 to 2 h per day. During the first visit (Time 1), children interacted in a videorecorded semi‐structured play interaction with researchers. During both the first visit and second visit (Time 2), parents were asked to complete the CCDI and report the words that their children can say. The CCDI is a useful and convenient tool to gather information about expressive vocabulary in young children between 8 and 30 months old or children with developmental delay at a specific moment (Tardif et al., 2008). The CCDI had good psychometric properties and the expressive vocabulary subtest showed high test‐retest reliability (0.93–0.98) and moderate concurrent validity (0.62–0.69) (Liang et al., 2001).

During the semi‐structured interaction, we provided opportunities to initiate communication. These activities were modelled after the activities described by Brady and colleagues (2018, 2012) but translated to be culturally and linguistically appropriate for children in China (Liu et al., 2023). The researcher was trained by the developer to administer the scripts. We conducted the play interaction using a series of standard toys in the assessment room of children's clinics. The researchers first engaged the child in a warmup activity to ensure that they were comfortable with researchers and the play procedures. Next, we administered 10 interactive scripts and the interaction took place at a table with the child sitting next to the researcher. Five interactive scripts (i.e., sealed bubbles, wind‐up toys, spinning tops, battery operated duck toys and food items) were designed to evoke behavioural regulation communication, which refers to communication used to regulate another's behaviour, such as obtaining an object, getting them to act or preventing someone from doing an activity. The other five interactive scripts (i.e., spider in block container, musical instruments, wiggly in marker box, books with blank or altered pages and switch activated toys) were used to motivate joint attention communication, which refers to directing the attention of others to an object or event by commenting or sharing attention (Bruner, 1981). Behavioural regulation scripts mainly evoke requests by creating obstacles. For example, during the wind‐up toy activity the researcher gives the participant a broken wind‐up toy to set up an opportunity for the participant to request help. Joint attention scripts mainly use novel stimuli to encourage the participant to share interests and experiences with the researcher. For example, when the participant is drawing, the researcher slips an unusual toy into the marker box when the participant is not looking, to see if the participant may show the toy to the researcher after seeing it. The entire semi‐structured play interaction session was video recorded for further coding.

During the first visit, the CCDI was used to confirm whether participants were minimally verbal. Parents were asked to select which words their child uses by completing the CCDI: Words and Gestures vocabulary list which includes 411 words. The initial expressive vocabulary size of the 37 participants ranged from 0 to 12 words according to this measure. After 1 year, some participants had a substantial increase in vocabulary, so parents completed the CCDI: Words and Sentences vocabulary list which includes the same words as Words and Gestures plus more, totalling 799 words. The CCDI raw scores at Time 2 ranged from 0 to 569 words.

Coding

All participants’ communicative behaviours from the semi‐structured play interaction were coded using ELAN software (Lausberg & Sloetjes, 2009). We derived the prelinguistic communicative behaviour complexity score (CCS scores) and the frequency of intentional communication behaviours from these codes.

CCS scores

The prelinguistic communicative behaviour complexity score was coded using the CCS (Brady et al., 2018). The CCS is a 12‐point scale which reflects a developmental continuum of prelinguistic early verbal communication behaviours. Scores of 1–5 indicate pre‐intentional communicative behaviours, such as a visual or body orientation toward a person or object. Scores 6–10 reflect intentional pre‐symbolic communication, for example, when an individual uses gestures and eye gaze to initiate communication behaviour. Scores 11–12 are used for intentional early symbolic communication, that is, when an individual uses words to communicate. Higher scores demonstrate more complex communication behaviours in terms of the numbers of behaviours. For example, a score of 6 indicates triadic orientation between object and person, whereas a score of 9 indicates a triadic eye gaze with a gesture or vocalisation. The complete scale is shown in Table 2 (originally published in Brady et al., 2018). Coders identified the highest scoring communicative behaviour for each activity from the CCS scripted interaction. Each participant received 10 item‐level scores and then we averaged the three highest scores for each participant to derive an optimal complexity score (see Brady et al., 2018 for further explanation). For example, if a participant's three highest scores were as follows: 8 for sealed bubbles, 7 for wind‐up toys and 7 for switch activated toys, his/her complexity score would be 7.33.

TABLE 2.

CCS scores.

Number Definition Communication level
0 No response
1 Alerting—a change in behaviour or stops doing a behaviour Pre‐intentional
2 Single orientation only—on an object, event or person; can be communicated through vision, body orientation or other means
3 Single orientation only + 1 other potentially communicative behaviour (PCB)
4 Single orientation only + more than 1 PCB
5 Dual orientation—shift in focus between a person and an object, between a person and an event using vision, body orientation, etc. (without PCB)
6 Triadic orientation (e.g., eye gaze or touch from object to person and back) Intentional nonsymbolic
7 Dual orientation + 1 PCB (e.g., dual focus + gesture)
8 Dual orientation + 2 or more PCB (e.g., dual focus + gesture + vocalisation, switch closure)
9 Triadic orientation + 1 PCB (e.g., triadic + vocalisation)
10 Triadic orientation plus more than 1 PCB (e.g., triadic plus vocalisation and differential switch closure)
11 One‐word verbalisation, sign or AAC symbol selection Intentional symbolic
12 Multiword verbalisation, sign or AAC symbol selection
Open in a new tab

Abbreviations: AAC, augmentative and alternative communication; CCS, Communication Complexity Scale; PCB, potentially communicative behaviour.

Intentional single and multimodal communication behaviour coding

In addition to CCS coding, we described the different type of behaviours used in each intentional communication act. For this coding, each intentional communication act was identified and then the type of act (described later) was determined. Although the CCS differentiates eye gaze into dyadic gaze and triadic gaze, participants in this study showed very few triadic gaze shifts combined with gestures and/or vocalisations. The eye gaze in two‐form and three‐form combinations is not necessarily triadic; it can be dyadic. We calculated the frequency of prelinguistic communicative behaviours from raw total counts of each prelinguistic communicative behaviour. Prelinguistic intentional communication behaviours consist of communication forms such as eye gaze, gestures and vocalisations and their combinations. We considered six types of prelinguistic intentional communication behaviours (see Table 3). Since our semi‐structured play session was designed to assess triadic coordination behaviour, for the purposes of this coding scheme, vocalisations alone were not considered to be single form communication behaviours because prelinguistic vocalisations were often ambiguous or unclear in their intentionality. Therefore, we coded two types of single form communicative behaviours, which include triadic eye gaze (TGz) and gesture (Gs) behaviours. There were three types of two‐form behaviours: gesture and vocalisation combinations (GsVo), gesture and eye gaze combinations (GsGz) and eye gaze and vocalisation combinations (GzVo). Finally, we coded one type of three‐form prelinguistic communicative behaviour, gesture, eye gaze and vocalisation combinations (GsGzVo).

TABLE 3.

Types of intentional communication behaviours.

Type Definition
Single‐form TGz Triadic eye gaze (e.g., eye gaze from object to person and back)
Gs Gesture (e.g., give object to person)
Two‐form GsVo Gesture and vocalisation combination (e.g., give object to person and vocalise within 3 s)
GsGz Gesture and eye gaze combination (e.g., give object to person and eye contact with person within 3 s)
GzVo Eye gaze and vocalisation combination (e.g., eye gaze from object to person and vocalise within 3 s)
Three‐form GsGzVo Gesture, eye gaze and vocalisation combination (e.g., give object to person with eye contact and vocalisation within 3 s)
Open in a new tab

Interobserver reliability

A trained master's student in speech and hearing science served as secondary coder and independently coded a random sample of 20% of all coded sessions for both communication complexity and types of communicative behaviours. The primary coder was blind to which sessions would be coded for reliability. Two‐way random effects single measures intraclass correlation coefficients with absolute agreement was used to index communication complexity agreement between the two observers. Interobserver reliability was 0.92 for prelinguistic communicative behaviour complexity. Weighted Kappa was used to analyse the agreement of the intentional communicative behaviour codes. The Kappa interobserver agreement was 0.92.

Data analysis

The Statistical Package for the Social Sciences (SPSS) version 28 was used for all analyses. The analysis method assumes multivariate normality, and multivariate normality is more likely when all variables are normally distributed (Tabachnick & Fidell, 2014). Variables showing univariate skewness>|1| or kurtosis>|3| were transformed. Transformations followed the principles laid out by Tabachnick and Fidell (2014), and a constant was added to each score to avoid taking the log of zero.

First, we used linear regression to examine the predictive relationship of prelinguistic communication complexity on later expressive vocabulary. We obtained zero‐order correlations among age, concurrent expressive vocabulary (Time 1), CCS scores and 12 months later expressive vocabulary (Time 2). Then we used simple linear regression to examine whether the prelinguistic communication complexity index and Time 1 expressive vocabulary were significantly related to 12 months later expressive vocabulary without controlling for any other factors. We subsequently used hierarchical regression to examine whether the prelinguistic communication complexity index could predict 12 months later expressive vocabulary after controlling for age and concurrent expressive vocabulary (Time 1). Second, we conducted independent t tests to compare the frequency of types of intentional communicative behaviours according to CCDI levels at Time 2 using forward likelihood ratio binary logistic regression to explore if these differences are related to later expressive vocabulary levels. The a priori alpha level established for regression statistical significance was p<0.05, and significance of t tests was corrected using Bonferroni‐Holm (Armstrong, 2014).

RESULTS

Preliminary analyses

The CCS scores demonstrated a normal distribution. CCDI scores were positively skewed in the first and second visit, but were corrected with log10 transformation (see Table 4).

TABLE 4.

The skewness and kurtosis of raw data and transformed data.

Variable N Minimum Maximum Mean (SD) Skewness (SE) Kurtosis (SE)
CCS 37 6.33 9 7.612 (0.651) 0.214 (0.388) −0.229 (0.759)
CCDI Time 1 37 0 12 1.730 (2.864) 2.028 (0.388) 4.179 (0.759)
CCDI Time 2 37 0 569 86.919 (147.674) 2.316 (0.388) 4.837 (0.759)
Transformed CCDI Time 1 37 0 1.11 0.267 (0.360) 0.916 (0.388) −0.614 (0.759)
Transformed CCDI Time 2 37 0 2.76 1.355 (0.786) 0.078 (0.388) −0.860 (0.759)
Open in a new tab

Abbreviations: CCDI, Chinese Communicative Development Inventory; CCS, Communication Complexity Scale.

A visual examination of the distribution of raw CCDI scores revealed two clusters, which we used to divide the participants into two groups: low expressive vocabulary group (between 0 to 62 words) and high expressive vocabulary group (more than 100 words) (see Figure 1). No participants scored between 62 and 100. We used independent samples t tests to test the differences in age between low group (n = 27) and high group (n = 10). The mean and SD of low and high CCDI group were 47.89 (9.60) and 43.10 (7.28) months, respectively. The results showed that there was no significant difference in the age of participants (t (35) = 1.428, p = 0.162, d = 0.529).

FIGURE 1.

FIGURE 1

Open in a new tab

Simple boxplot of Chinese Communicative Development Inventory (CCDI) Time 2 by group.

Prediction of prelinguistic communication complexity on later expressive language

The results of the zero‐order correlations (Pearson's r) showed that CCS scores from Time 1 and transformed CCDI scores at Time 2 were significantly correlated (r = 0.497, n = 37, p<0.01). Transformed CCDI scores at Time 1 and Time 2 were also significantly correlated (r = 0.564, n = 37, p<0.001). CCS scores at Time 1 and transformed CCDI scores at Time 1 were not significantly correlated (r = 0.218, n = 37, p = 0.194). Child age was not significantly correlated with CCS scores and transformed CCDI scores at Time 1 and Time 2 (see Table 5).

TABLE 5.

Pearson's r values among zero‐order predictors of language outcome.

1 2 3 4
1. CCS 1
2. Age −0.095ns 1
3. Transformed CCDI Time1 0.218ns 0.189ns 1
4. Transformed CCDI Time2 0.497 * 0.011ns 0.564 ** 1
Open in a new tab

Note: All p values are two tailed.

Abbreviations: CCDI, Chinese Communicative Development Inventory; CCS, Communication Complexity Scale; ns, nonsignificant result.

* p<0.01, ** p<0.001.

We performed linear regression analysis with transformed CCDI scores at Time 2 as the dependent variable and CCS scores at Time 1 as the independent variable. Simple linear regression analyses suggested that CCS scores significantly predicted expressive vocabulary 12 months later (β = 0.497, p<0.01), accounting for 24.7% of the variance in later word use in our sample (see Table 6). Hierarchical regression analyses showed that CCS scores also continued to significantly predict expressive vocabulary 12 months later after controlling for age and concurrent expressive vocabulary (Time 1) (β = 0.386, p<0.01), accounting for 13.9% of the variance in later expressive vocabulary. The CCS scores in combination with concurrent vocabulary and age accounted for 46.7% of the variance in later expressive vocabulary (see Table 7).

TABLE 6.

Simple linear regression model for predictor (CCS) of Time 2 transformed CCDI.

Coefficient R 2 ΔR 2 B SEB β t p
Model 1 0.247 0.247 <0.01
Constant −3.213 1.354 −2.373 <0.05
CCS 0.600 0.177 0.497 3.386 <0.01
Open in a new tab

Abbreviations: CCDI, Chinese Communicative Development Inventory; CCS, Communication Complexity Scale.

TABLE 7.

Hierarchical regression models for predictors (CCS, transformed CCDI Time 1, age) of Time 2 transformed CCDI.

Coefficient R 2 ΔR 2 B SEB β t p
Model 1 0.328 0.328 <0.01
Constant 1.409 0.572 2.465 <0.05
Age −0.008 0.012 −0.099 −0.690 >0.05
Transformed CCDI Time 1 1.272 0.313 0.583 4.069 <0.001
Model 2 0.467 0.139 <0.001
Constant −2.303 1.367 −1.684 >0.05
Age −0.004 0.011 −0.044 −0.339 >0.05
Transformed CCDI Time 1 1.066 0.291 0.488 3.658 <0.001
CCS 0.466 0.159 0.386 2.932 <0.01
Open in a new tab

Abbreviations: CCDI, Chinese Communicative Development Inventory; CCS, Communication Complexity Scale.

Comparison of prelinguistic single and multimodal intentional communicative behaviours frequency between different later language ability groups

We conducted six independent samples t tests to compare the frequency of prelinguistic intentional communicative behaviours according to CCDI level. The Bonferroni–Holm correction was used for correcting multiple tests p values. Descriptive and statistics data are reported in Table 8. The p values for the t tests are ranked from low to high. The results showed that there was a significant group difference in the frequency of GsVo combinations (t (35) = −2.918, p = 0.003, d = 1.080). The p value of GsVo combinations was less than Bonferroni–Holm α which was 0.0083. There was no significant group difference in the frequency of three forms of communicative behaviour—GsGzVo combinations (t (12) = −1.956, p = 0.037, d = 0.865). The p value of GsGzVo combinations was greater than Bonferroni‐Holm α which was 0.01. There were also no group differences in triadic eye gaze, gesture, GsGz combinations, GzVo combinations and total behaviours. From these results, it can be seen that autistic children with high CCDI scores produce more GsVo combinations compared with the low CCDI group.

TABLE 8.

Descriptive and inferential statistics for communication behaviours frequency.

Low group (N = 27) High group (N = 10)
Variable Mean SD Mean SD t df p
GsVo 1.19 1.798 3.30 2.359 −2.918 35 0.003
GsGzVo 0.19 0.622 0.80 0.919 −1.956 12 0.037
Gs 4.22 2.833 3.20 1.989 1.045 35 0.152
TGz 0.33 0.832 0.10 0.316 0.858 35 0.198
GzVo 0.59 1.248 0.30 0.675 0.700 35 0.244
GsGz 0.41 0.797 0.30 0.949 0.346 35 0.366
Total 6.93 2.827 8.00 3.432 −0.969 35 0.170
Open in a new tab

Abbreviations: Gs, gesture; GsGz, gesture and eye gaze combination; GsGzVo, gesture, eye gaze and vocalisation combination; GsVo, gesture and vocalisation combination; TGz, triadic eye gaze.

Then, we used binary logistic regression to determine how the GsVo combinations uniquely predicted later expressive vocabulary group. A Hosmer–Lemeshow test showed that the overall model was a well‐fitted model, χ 2(3) = 5.654, p>0.05, suggesting that frequency of GsVo combinations had a significant effect on the odds of membership in the high CCDI group. The regression coefficient for GsVo combinations was significant, B = 0.486, odds ratio = 1.626, p < 0.05, indicating that for one unit increase in frequency of GsVo combinations, the odds of being in the high CCDI group would increase by 62.6% (see Table 9).

TABLE 9.

Binary logistics regression for communication behaviour influence vocabulary levels.

Variable β SE Wald χ 2 p OR (95% CI)
GsVo 0.486 0.209 5.419 <0.05 1.626 (1.080–2.448)
Open in a new tab

Abbreviations: CI, confidence interval; GsVo, gesture and vocalisation combination; OR, odds ratio.

DISCUSSION

Prelinguistic communication complexity predicts later expressive language

The first aim of the present study was to test the hypothesis that higher prelinguistic communication complexity predicts better expressive language 12 months later. Our findings showed that prelinguistic communicative complexity predicted expressive language 12 months later and it continued to have a positive relationship after controlling for age and concurrent CCDI vocabulary (Time 1). The present work extends previous studies showing that the number of prelinguistic communication behaviours have a positive relationship with expressive language in autistic children (e.g., Kilili‐Lesta et al., 2022), and that the complexity of prelinguistic communicative behaviours is associated with concurrent expressive language (e.g., Brady et al., 2012). It is possible that greater communication complexity earlier in childhood promotes more successful social interactions with caregivers and interaction partners, which in turn results in greater responsiveness and more opportunities for language learning. Future work should explore potential mechanisms driving this relationship.

This finding also provides evidence for existing prelinguistic intervention practices for autistic children. Prelinguistic milieu teaching (PMT) is a widely used approach that focuses on prelinguistic communication behaviours (Peters‐Scheffer et al., 2016; Yoder & Warren, 2002), and it has been used in minimally verbal autistic children (Dubin et al., 2020; Franco et al., 2013). PMT intervention targets mainly have two levels: the first level is single‐form behaviours (i.e., eye gaze, gestures and vocalisations), and the second level is multimodal behaviours (i.e., two‐form combinations and three‐form combinations). The positive relationship between prelinguistic communication complexity and expressive language provides support for treatment approaches such as PMT, which explicitly focusses on goal hierarchies related to communication complexity.

Additionally, these findings suggest that understanding of the development of prelinguistic communication behaviours, such as that provided by the CCS, might be helpful in terms of targets for intervention. For example, targeting more complex communication may be a good proximal target for some children with less complex prelinguistic communication.

Gesture and vocalisation combinations uniquely predict expressive language

The second aim of the study was to explore whether there were any differences in specific types of communicative behaviours across low and high vocabulary achievers. We first compared different types of communicative behaviours between two groups. In order to better assess the differences in specific communicative behaviours of autistic children with different expressive language ability levels, we first analysed the total prelinguistic communicative behaviour frequency between two groups. Consistent with previous work (Murillo et al., 2021), the two groups of autistic children with different CCDI (Time 2) levels produced similar total prelinguistic communicative behaviours in our study.

We are not surprised that autistic children demonstrated low frequencies of single and multimodal eye gaze communicative behaviours because previous studies have also described this outcome (Hahn et al., 2019; Murillo et al., 2021). We also found that single gestures were the most frequent of all communication behaviours in the low CCDI group, and the two groups used similar amounts of single gestures. However, the most frequent behaviour in the high CCDI group was GsVo combinations which was significantly different compared to the low CCDI group. This result is different from the study of Murillo and colleagues. They found no difference in gesture and vocalisation combinations between the high and low expressive language groups of autistic children. There are two possible reasons: First, there were only 11 autistic children in the Murillo's study and 37 autistic children in this study. Second, the criteria for grouping were different. The Murillo's study was based on concurrent expressive language, while the present study was based on expressive language 12 months later.

The difference between two groups of GsVo combinations suggests that this multimodal behaviour may be particularly important for predicting later language. Further analysis showed that gesture and vocalisation combinations are uniquely predictive of expressive language 12 months later in our sample. To some extent, this result supports Fröhlich and colleagues’ view that human language, especially gesture and vocalisation combinations, has multimodal origins (Fröhlich et al., 2019).

Our findings may have clinical implications, providing further direction on how to better implement and target early intervention strategies. Clinicians could consider which multimodal behaviours are most critical to target when focussing on helping children transition from single communicative behaviours to multimodal behaviours. PMT does not make specific suggestions regarding which multimodal behaviours are most beneficial in terms of downstream development. Our findings suggest that the GsVo combinations can be the first target that clinicians can focus on among multimodal behaviours, especially for autistic children who use gestures alone to communicate, as these combinations were most predictive of later language skills. One potential mechanism driving this relationship is parental responsiveness: autistic children who produce fewer gesture and vocalisation combinations may elicit less responsivity and language input from their caregivers, which then affects language development. Caregiver contingent responses support the infant's language development by providing additional input and social exchanges (Fagan & Doveikis, 2019; Lopez et al., 2020). This input from caregivers may help drive further development in the children (Lam‐Cassettari et al., 2021). Studies have also showed that when combined with gestures, children produced more advanced vocalisations (e.g., canonical syllables) with syllable durations and fundamental frequencies more similar to those of mature speech (Murillo & Capilla, 2016). In sum, gesture and vocalisation combinations may be an important intervention target for minimally verbal autistic children.

LIMITATIONS AND FUTURE DIRECTIONS

There are several limitations in this study, which must be taken into account. First, given the variability of the heterogeneity of autism, caution should be taken in generalising the results to apply to autistic children who differ substantially from participating autistic children, who were minimally verbal, Chinese Mandarin‐speaking, monolingual. Due to the small size of the sample in this study which also did not take into account family socioeconomic status, future research should replicate these findings with larger sample size and consider the impact of family socioeconomic status. Second, although we found that prelinguistic communication complexity was predictive of later expressive language, this prediction was correlational in nature. We did not control the possible impact of early intervention. The language growth from the ongoing intervention could affect the relationship between prelinguistic communication complexity and expressive vocabulary. In the future, experimental intervention studies can be carried out to further confirm the causal relationship between improving communication complexity and language outcomes. Further, although we found that two‐form and three‐form multimodal communication behaviours were related to language outcomes, we do not know whether there were some factors, such as mental age, caregiver response and speech abilities that may have a moderating effect. This is also an important research direction in the future.

CONCLUSION

In summary, for a sample of minimally verbal autistic children, the complexity of prelinguistic communicative behaviour was positively associated with expressive language development 12 months later. The production of gesture and vocalisation combinations also predicted later expressive vocabulary. Future research examining interventions for prelinguistic communication complexity may find that these combinations are appropriate proximal targets for promoting expressive language in autistic children who are minimally verbal.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

ACKNOWLEDGEMENTS

We are extremely appreciative of the children and their caregivers for their participation, as well as the education and service institutions for their support. This work was funded by the National Key Research and Developmental Program of China (grant number 2022YFC2705201), the Shanghai Pujiang Program (grant number 2019PJC033), the National Language Commission Research Planning Program of China (YB145‐21) and China Scholarship Council Funding.

DATA AVAILABILITY STATEMENT

Data available on request from the first author.

REFERENCES

  1. Albert, R.R. , Ernst, M. & Vallotton, C.D. (2023) Infant vocalizations elicit simplified speech in childcare. Infancy, 28(2), 322–338. 10.1111/infa.12520 [DOI] [PubMed] [Google Scholar]
  2. Armstrong, R.A. (2014) When to use the Bonferroni correction. Ophthalmic and Physiological Optics, 34(5), 502–508. 10.1111/opo.12131 [DOI] [PubMed] [Google Scholar]
  3. Arnett, A.B. , Beighley, J.S. , Kurtz‐Nelson, E.C. , Hoekzema, K. , Wang, T. , Bernier, R.A. & Eichler, E.E. (2020) Developmental predictors of cognitive and adaptive outcomes in genetic subtypes of autism spectrum disorder. Autism Research, 13(10), 1659–1669. 10.1002/aur.2385 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brady, N.C. , Fleming, K. , Romine, R.S. , Holbrook, A. , Muller, K. & Kasari, C. (2018) Concurrent validity and reliability for the communication complexity scale. American Journal of Speech‐Language Pathology, 27(1), 237–246. 10.1044/2017_AJSLP-17-0106 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Brady, N.C. , Fleming, K. , Thiemann‐Bourque, K. , Olswang, L. , Dowden, P. , Saunders, M.D. & Marquis, J. (2012) Development of the communication complexity scale. American Journal of Speech‐Language Pathology, 21(1), 16–28. 10.1044/1058-0360(2011/10-0099) [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Brady, N.C. , Romine, R.E.S. , Holbrook, A. , Fleming, K.K. & Kasari, C. (2020) Measuring change in the communication skills of children with autism spectrum disorder using the communication complexity scale. American Journal on Intellectual and Developmental Disability, 125(6), 481–492. 10.1352/1944-7558-125.6.481 [DOI] [PubMed] [Google Scholar]
  7. Bruner, J. (1981) The social context of language acquisition. Language & Communication, 1(2‐3), 155–178. 10.1016/0271-5309(81)90010-0 [DOI] [Google Scholar]
  8. Choi, B. , Wei, R. & Rowe, M.L. (2021) Show, give, and point gestures across infancy differentially predict language development. Developmental Psychology, 57(6), 851–862. 10.1037/dev0001195 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dimitrova, N. , Mohr, C. , Özçalışkan, Ş. & Adamson, L.B. (2020) Early lateralization of gestures in autism: right‐handed points predict expressive language. Journal of Autism and Developmental Disorders, 50(4), 1147–1158. 10.1007/s10803-019-04347-9 [DOI] [PubMed] [Google Scholar]
  10. Donnellan, E. , Bannard, C. , Mcgillion, M.L. , Slocombe, K.E. & Matthews, D.E. (2020) Infants’ intentionally communicative vocalisations elicit responses from caregivers and are the best predictors of the transition to language: a longitudinal investigation of infants’ vocalizations, gestures, and word production. Developmental Science, 23(1), e12843. 10.1111/desc.12843 [DOI] [PubMed] [Google Scholar]
  11. Dubin, A.H. , Lieberman‐Betz, R.G. , Ayres, K.M. & Zawoyski, A. (2020) The effects of prelinguistic milieu teaching implemented in classrooms for preschoolers with or at risk for autism spectrum disorder. Focus on Autism and Other Developmental Disabilities, 35(2), 79–89. 10.1177/1088357619888917 [DOI] [Google Scholar]
  12. Fagan, M.K. & Doveikis, K.N. (2019) What mothers do after infants vocalize: implications for vocal development or word learning? Journal of Speech Language and Hearing Research, 62(8), 2680–2690. 10.1044/2019_JSLHR-S-18-0136 [DOI] [PubMed] [Google Scholar]
  13. Franco, J.H. , Davis, B.L. & Davis, J.L. (2013) Increasing social interaction using prelinguistic milieu teaching with nonverbal school‐age children with autism. American Journal of Speech‐Language Pathology, 22(3), 489–502. 10.1044/1058-0360(2012/10-0103) [DOI] [PubMed] [Google Scholar]
  14. Fröhlich, M. , Sievers, C. , Townsend, S.W. , Gruber, T. & van Schaik, C.P. (2019) Multimodal communication and language origins: integrating gestures and vocalizations. Biological Reviews, 94(5), 1809–1829. 10.1111/brv.12535 [DOI] [PubMed] [Google Scholar]
  15. Gerhold, K. , Torrington Eaton, C. , Newman, R.S. & Bernstein Ratner, N. (2020) Early phonological predictors of toddler language outcomes. Folia Phoniatrica et Logopaedica, 72(6), 442–453. 10.1159/000503230 [DOI] [PubMed] [Google Scholar]
  16. Hahn, L.J. , Brady, N.C. , McCary, L. , Rague, L. & Roberts, J.E. (2017) Early social communication in infants with fragile X syndrome and infant siblings of children with autism spectrum disorder. Research in Developmental Disabilities, 71, 169–180. 10.1016/j.ridd.2017.10.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hahn, L.J. , Brady, N.C. & Versaci, T. (2019) Communicative use of triadic eye gaze in children with Down syndrome, autism spectrum disorder, and other intellectual and developmental disabilities. American Journal of Speech‐Language Pathology, 28(4), 1509–1522. 10.1044/2019_AJSLP-18-0155 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kilili‐Lesta, M. , Voniati, L. & Giannakou, K. (2022) Early gesture as a predictor of later language outcome for young children with autism spectrum disorder (ASD): a systematic literature review. Current Developmental Disorders Reports, 9, 110–126. 10.1007/s40474-022-00250-8 [DOI] [Google Scholar]
  19. Lam‐Cassettari, C. , Peter, V. & Antoniou, M. (2021) Babies detect when the timing is right: evidence from event‐related potentials to a contingent mother‐infant conversation. Developmental Cognitive Neuroscience, 48, 100923. 10.1016/j.dcn.2021.100923 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Lausberg, H. & Sloetjes, H. (2009) Coding gestural behavior with the NEUROGES‐ELAN system. Behavior Research Methods, 41(3), 841–849. 10.3758/BRM.41.3.841 [DOI] [PubMed] [Google Scholar]
  21. LeGrand, K.J. , 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. 10.1044/2021_JSLHR-20-00544 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Liang, W. , Hao, B. , Wang, S. , Yang, Y. , Zhang, Z. , Zuo, Q. , Tardif, T. & Fletcher, P. (2001) The study of Chinese language and comunicative development‐word part. Chinese Journal of Child Health Care, 9(5), 295–297. [Google Scholar]
  23. Liu, M. , Brady, N.C. , Zhao, H. & Liu, Q. (2023) A Chinese version of the communication complexity scale (CCS): psychometric evaluation in children with minimal verbal skills. Disability and Rehabilitation, 1–8. 10.1080/09638288.2023.2226405 [DOI] [PubMed] [Google Scholar]
  24. Lopez, L.D. , Walle, E.A. , Pretzer, G.M. & Warlaumont, A.S. (2020) Adult responses to infant prelinguistic vocalizations are associated with infant vocabulary: a home observation study. Plos One, 15, e0242232. 10.1371/journal.pone.0242232 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lu, J. , Yang, Z. , Shu, M. & Su, L. (2004) Reliability, validity andlysis of the childhood autism rating scale. China Journal of Modern Medicine, 14(13), 119–123. [Google Scholar]
  26. Maltman, N. , DaWalt, L.S. , Hong, J. & Mailick, M. (2021) Brief report: socioeconomic factors associated with minimally verbal status in individuals with ASD. Journal of Autism and Developmental Disorders, 51(6), 2139–2145. 10.1007/s10803-020-04646-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Manwaring, S.S. , Swineford, L. , Mead, D.L. , Yeh, C.‐C. , Zhang, Y. & Thurm, A. (2019) The gesture – language association over time in toddlers with and without language delays. Autism & Developmental Language Impairments, 4, 1–15. 10.1177/2396941519845545 33912683 [DOI] [Google Scholar]
  28. McDaniel, J. , Woynaroski, T. , Keceli‐Kaysili, B. , Watson, L.R. & Yoder, P. (2019) Vocal communication with canonical syllables predicts later expressive language skills in preschool‐aged children with autism spectrum disorder. Journal of Speech Language and Hearing Research, 62(10), 3826–3833. 10.1044/2019_jslhr-l-19-0162 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. McDaniel, J. , Yoder, P. , Estes, A. & Rogers, S.J. (2020) Predicting expressive language from early vocalizations in young children with autism spectrum disorder: which vocal measure is best? Journal of Speech, Language, and Hearing Research, 63(5), 1509–1520. 10.1044/2020_JSLHR-19-00281 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Murillo, E. , Camacho, L. & Montero, I. (2021) Multimodal communication in children with autism spectrum disorder and different linguistic development. Journal of Autism and Developmental Disorders, 51(5), 1528–1539. 10.1007/s10803-020-04637-7 [DOI] [PubMed] [Google Scholar]
  31. Murillo, E. & Capilla, A. (2016) Properties of vocalization‐and gesture‐combinations in the transition to first words. Journal of Child Language, 43(4), 890–913. 10.1017/S0305000915000343 [DOI] [PubMed] [Google Scholar]
  32. Olson, J. & Masur, E.F. (2015) Mothers’ labeling responses to infants’ gestures predict vocabulary outcomes. Journal of Child Language, 42(6), 1289–1311. 10.1017/S0305000914000828 [DOI] [PubMed] [Google Scholar]
  33. Peters‐Scheffer, N.C. , Huskens, B. , Didden, R. & van der Meer, L. (2016) Prelinguistic milieu teaching. Early intervention for young children with autism spectrum disorder. Switzerland: Springer, pp. 151–175. [Google Scholar]
  34. Ramos‐Cabo, S. , Acha, J. , Vulchanov, V. & Vulchanova, M. (2022) You may point, but do not touch: impact of gesture‐types and cognition on language in typical and atypical development. International Journal of Language & Communication Disorders, 57, 324–339. 10.1111/1460-6984.12697 [DOI] [PubMed] [Google Scholar]
  35. Salley, B. , Brady, N.C. , Hoffman, L. & Fleming, K. (2020) Preverbal communication complexity in infants. Infancy, 25(1), 4–21. 10.1111/infa.12318 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Mohamadi, R. , Sangani, T.M. , Ansari, N.N. & Soleymani, Z. (2022) A preliminary study of telepractice prelinguistic milieu teaching for children with autism spectrum disorders. Journal of Iranian Medical Council, 5(3), 471–477. 10.18502/jimc.v5i3.10943 [DOI] [Google Scholar]
  37. Shumway, S. & Wetherby, A.M. (2009) Communicative acts of children with autism spectrum disorders in the second year of life. Journal of Speech, Language, and Hearing Research, 52(5), 1139–1156. 10.1044/1092-4388(2009/07-0280) [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Stewart, J.R. , Vigil, D.C. & Carlson, R. (2021) Frequency of gesture use and language in typically developing prelinguistic children. Infant Behavior and Development, 62, 101527. 10.1016/j.infbeh.2021.101527 [DOI] [PubMed] [Google Scholar]
  39. Tabachnick, B.G. & Fidell, L.S. (2014) Using multivariate statistics. Boston, MA: Pearson. [Google Scholar]
  40. Tamis‐LeMonda, C.S. , Kuchirko, Y. & Song, L.L. (2014) Why is infant language learning facilitated by parental responsiveness? Current Directions in Psychological Science, 23(2), 121–126. 10.1177/0963721414522813 [DOI] [Google Scholar]
  41. Tardif, T. , Fletcher, P. , Zhang, Z.X. & Liang, W.L. (2008) The Chinese communicative development inventories. Beijing: Peking University Medical Press. [Google Scholar]
  42. van der Klis, A. , Adriaans, F. & Kager, R. (2023) Infants’ behaviours elicit different verbal, nonverbal, and multimodal responses from caregivers during early play. Infant Behavior and Development, 71, 101828. 10.1016/j.infbeh.2023.101828 [DOI] [PubMed] [Google Scholar]
  43. Werwach, A. , Mürbe, D. , Schaadt, G. & Männel, C. (2021) Infants’ vocalizations at 6 months predict their productive vocabulary at one year. Infant Behavior and Development, 64, 101588. 10.1016/j.infbeh.2021.101588 [DOI] [PubMed] [Google Scholar]
  44. Yoder, P. , Watson, L.R. & Lambert, W. (2015) Value‐added predictors of expressive and receptive language growth in initially nonverbal preschoolers with autism spectrum disorders. Journal of Autism and Developmental Disorders, 45(5), 1254–1270. 10.1007/s10803-014-2286-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Yoder, P.J. & Warren, S.F. (2002) Effects of prelinguistic milieu teaching and parent responsivity education on dyads involving children with intellectual disabilities. Journal of Speech, Language, and Hearing Research, 45(6), 1158–1174. 10.1044/1092-4388(2002/094) [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Data available on request from the first author.

Articles from International Journal of Language & Communication Disorders are provided here courtesy of Wiley

RESOURCES