999905

research-article2021
AUT0010.1177/1362361321999905AutismLaister et al.

Original Article

Autism

Social-communicative gestures at 1­–13

© The Author(s) 2021
Article reuse guidelines:
baseline predict verbal and nonverbal sagepub.com/journals-permissions
DOI: 10.1177/1362361321999905
https://doi.org/10.1177/1362361321999905

gains for children with autism receiving journals.sagepub.com/home/aut

the Early Start Denver Model

Dominik Laister1,2 , Magdalena Stammler1, Giacomo Vivanti3,4

and Daniel Holzinger1,2,5

Abstract
In children with autism spectrum disorder, atypical gesture use is a core deficit with consequences for social learning, social
interaction, and language development. Little is known about the relevance of early gesture use in predicting developmental
outcomes of children receiving early interventions targeting social-communicative behaviors such as the Early Start Denver
Model. We found that the parent-rated “Gestural Approach Behavior” subscale of the Pervasive Developmental Disorder
Behavior Inventory was predictively associated with developmental changes after 1 year of intervention as assessed by the
Mullen Scales of Early Learning. This subscale was as strong a predictor as the Mullen nonverbal development quotient
before intervention. Our findings suggest that children who use more gestures for social communication might be better
equipped to respond to the learning opportunities offered by the Early Start Denver Model.

Lay abstract
Although there is growing evidence of the effectiveness and importance of certain early intervention programs for
children with autism spectrum disorders, little is known about predictive information before intervention to search
for the most accurate therapeutic approach for the individual child and his family. In children with autism spectrum
disorder, atypical gesture use is one core deficit with consequences for the development of social interaction and
language, but there is little knowledge about the relevance of early gesture use in predicting developmental outcomes
of children receiving early interventions targeting social-communicative behaviors such as the Early Start Denver Model.
In this study, we found that the parent-rated “Gestural Approach Behavior” subscale of the Pervasive Developmental
Disorder Behavior Inventory was predictively associated with clinically assessed developmental changes after 1 year of
intervention. This subscale was as strong a predictor as nonverbal development before intervention. Our findings suggest
that children who use more gestures in daily life might be better equipped to respond to learning opportunities offered
by early interventions targeting social communication strategies such as the Early Start Denver Model. Furthermore, we
conclude that the parent-rated questionnaire might be a valuable and economic set of questions with high relevance for
clinical assessments.

Keywords
autism spectrum disorder, early childhood, Early Start Denver Model, gestures, predictors, social communication

1
The vast heterogeneity among children with autism spec- Hospital of St. John of God, Austria
2
trum disorder (ASD) poses a challenge to the evaluation of Johannes Kepler University Linz, Austria
3
Drexel University, USA
treatment effectiveness and prediction of clinical outcome 4
La Trobe University, Australia
for the individual learner (Kim et al., 2016; Loth et al., 5
University of Graz, Austria
2017). The variability documented across child character-
Corresponding author:
istics, family factors (Barbaresi et al., 2006; Waterhouse, Dominik Laister, Institute of Neurology of Senses and Language,
2013), and response to different interventions (Howlin Hospital of St. John of God, Seilerstaette 2, 4021 Linz, Austria.
et al., 2011; Hudry et al., 2018; Sherer & Schreibman, Email: [email protected]
2 Autism 00(0)

2005; Vivanti et al., 2018) points to the importance of substantial variability in intervention response (e.g.
identifying factors associated with variability in interven- Contaldo et al., 2020; Sinai-Gavrilov et al., 2020; Vivanti
tion outcomes to match the child’s unique profile to the et al., 2013, 2019). Knowledge about factors associated
treatment approach with the greatest likely benefit. This with such variability is limited, although chronological age
study focuses on the identification of baseline characteris- at intake has been shown to be a putative predictor of
tics associated with response to the low-intensity provision ESDM outcomes (Rogers et al., 2012; Vivanti et al., 2019;
of the Early Start Denver Model (ESDM; Rogers & see also Smith et al., 2015 for similar findings in response
Dawson, 2010). to other behavioral interventions). Furthermore, research
on ESDM and other interventions based on the NDBI
framework has documented associations between inter-
ESDM
vention gains and milder autism symptoms, milder social
The ESDM is a manualized, comprehensive, early inter- deficits, better language, and higher cognitive functioning
vention program for children with ASD aged 1–4 years (e.g. Paynter et al., 2018; Sandbank et al., 2020). Greater
(Rogers & Dawson, 2010). Teaching strategies are based intervention intensity has been pointed out as another
on a Naturalistic Developmental Behavioral Intervention influencing factor by several researchers (Landa, 2018). In
(NDBI) approach (Schreibman et al., 2015), which includes addition, preliminary evidence suggests larger gains in
the implementation of behavioral and developmental prin- children with more advanced skills in imitation, object
ciples during naturalistic routines and play activities play, and joint attention (Vivanti et al., 2013). Although
(Vivanti & Stahmer, 2020). Objectives are individualized several factors associated with outcomes in ESDM pro-
and child progress is monitored through rigorous data col- grams have been identified, the current knowledge about
lection procedures. Teaching episodes are built on the predictors is insufficient. This study addressed this gap in
child’s interests and embedded in naturalistic and intrinsi- the literature by investigating the role of social-communi-
cally rewarding joint activities. Theoretical foundations of cative gestures at pre-intervention, as measured by paren-
the ESDM include a focus on the role of active experiential tal report, in gains for children receiving the ESDM.
learning, early interaction, and social motivation for learn-
ing and development (Dawson et al., 2004; Rogers &
Pennington, 1991). Within this context, skill acquisition
Social-communicative gestures
and management of unwanted behaviors are promoted Social-communicative gestures as an early expression of
using operant conditioning principles and behavioral strate- social communication are of particular interest in the context
gies derived from Pivotal Response Training (PRT; of atypical language development. Lobban-Shymko et al.
Ingersoll & Schreibman, 2006; Koegel et al., 1987). Early (2017), building on research on the role of social interaction
social learning processes are specifically targeted by in language development (Bruner, 1981; Tomasello &
ESDM, based on research highlighting the importance of Schröder, 2009), argued for a careful distinction between
early social orienting and early participation in dyadic social-communicative gestures (such as pointing to show to
exchanges in the development of children with ASD initiate joint attention and engagement or to share interests)
(Dawson et al., 2004, 2010; Fein et al., 2013; Rogers & and functional gestures (used to regulate the environment
Pennington, 1991). According to this notion, as infants with and to have personal needs met). In infants with ASD, acqui-
ASD are less responsive to social stimuli and initiate social sition of these gestures does not appear to follow the pattern
interactions less frequently, social learning experiences of typically developing children (Paparella et al., 2011).
during critical periods of brain plasticity are limited, which Whereas in children with ASD social-communicative ges-
affects the development of neural specialization and behav- tures are more delayed and less often used than functional
ioral expertise in the social communication domain. The gestures, it is the other way round in typically developing
ESDM is designed to mitigate or prevent this cascade effect children (Maljaars et al., 2011; Watson et al., 2013). The pre-
by addressing early emerging symptoms of autism during verbal use of gestures has been shown to be intricately linked
critical periods of brain plasticity in early childhood to language development in typical development (Hall et al.,
(Rogers & Dawson, 2010; Vivanti & Rogers, 2014). 2013; Murillo & Belinchón, 2012). In children with ASD,
use of gestures has been found to predict language and,
beyond that, ASD symptom severity (Chawarska et al., 2009;
Variability and predictors of Ellis Weismer & Kover, 2015; Lobban-Shymko et al., 2017;
outcomes in ESDM Yoder et al., 2015). Social-communicative behaviors are
The effectiveness of ESDM has been documented across intertwined with a child’s ability to relate to others. In accord-
various delivery formats and levels of intensity (Dawson ance with social motivation models of ASD (Chevallier
et al., 2010; Holzinger et al., 2019; Rogers, Estes, Lord, et al., 2012), limited social use of gestures is a symptom of an
et al., 2019; Rogers, Estes, Vismara, et al., 2019; Vivanti, underlying early onset social attention deficit, which leads to
Paynter, et al., 2014; Xu et al., 2018). However, there is a deprivation of social learning experiences that are required
Laister et al. 3

for the acquisition of verbal and nonverbal skills. As children Participants

with ASD now tend to be diagnosed at a younger age, when
they are preverbal, there is a need for nonverbal early mark- Fifty-six children with ASD (51 male and 5 female) aged
ers of ASD that can be reliably observed and used to guide 29–60 months (mean (M) = 41.96 and standard deviation
intervention decisions. (SD) = 7.53) and their parents enrolled in this study.
Participants were recruited from a community-based early
intervention program provided by our institute. They had
Rationale and goals for the current been referred to the institute for diagnostic assessment by
study pediatricians, pre-school teachers, therapists, or parents
between March 2014 and June 2018. Participants met
The ESDM has a focus on early interaction and social com-
Diagnostic and Statistical Manual of Mental Disorders
munication, and parent involvement is a key part of the
(4th ed.; DSM-IV) criteria for ASD according to diagnos-
approach. Consequently, it is plausible that early social-
tic assessments based on a multidisciplinary evaluation
communicative gestures as observed by parents at intake
process that included use of the Autism Diagnostic
might be related to outcomes of children receiving the
Observation Schedule–Second Edition (ADOS-2; Lord
ESDM. In addition, as the ESDM focuses on following the
et al., 2000), general developmental assessment using the
child’s lead to build learning opportunities on the child’s
Mullen Scales of Early Learning (MSEL; Mullen, 1995),
spontaneous initiative, it is plausible that children who use
and parental interviews using a questionnaire battery
gestures more frequently to approach others offer more
including the PDDBI (Cohen & Sudhalter, 2005). The
leads that can be followed, resulting in more learning
diagnostic assessments were administered and scored by
opportunities. Against this background, this study exam-
experienced clinical psychologists, clinical linguists, or
ined a community sample of 56 children aged 29–60 months
neuropediatricians trained in the use of these tools. The
who received the ESDM using in a low-intensity format for
ADOS-2 was administered by two clinicians who cross-
12 months to test whether parent-reported baseline social-
checked their scores. Fidelity of the ADOS-2 scoring is
communicative gestures predict verbal and nonverbal gains
constantly monitored through team trainings. The ESDM
during the treatment period.
intervention was offered to all families who met the fol-
Based on the theory underlying ESDM, we expected
lowing criteria: (1) ASD diagnosed by the institute’s diag-
better baseline social-communicative gestures to predict
nostic team, (2) diagnosis before the age of 4.0 years, (3)
greater improvements in both language and nonverbal
nonverbal developmental age above 12 months, (4) par-
development after intervention. Since parents are usually
ents requested early intervention, and (5) sufficient lan-
the main interaction partners of their children and have
guage skills (German or English) of at least one parent.
rich opportunities to observe their children in a variety of
However, in contrast to other studies (Solomon et al.,
everyday situations, we expected them to be reliable
2014), multilingualism was not an exclusion criterion. In
observers of their children’s early social-communicative
fact, multilingual families with at least one parent using a
development.
language other than German in the family represented the
majority (78.6%). Notably, the multilingual families used
Methods 18 different languages, which indicated challenges associ-
Our institute provides diagnostic and intervention services ated both with the diagnostic process and with interven-
for children and adolescents with developmental disorders tion. Presumably, this was due to the program being a
with a focus on communication difficulties, developmental community-based, cost-free service in the province of
disorders of scholastic skills, hearing loss, and ASD for the Upper Austria, where around 24% of all pre-schoolers
state of Upper Austria (1.5 million inhabitants). This study have a family language other than German (Amt der OÖ
is part of a more comprehensive research project on the Landesregierung, 2017). The Austrian system of regular
efficacy of a low-intensity implementation of the ESDM. well-baby checkups for every child allows monitoring of
Ethical approval was granted by the Ethics Committee of children throughout their first years of life, and pediatri-
the Hospital of St. John God in Linz on 23 April 2014. cians are provided with the opportunity to refer high-risk
Signed informed consent was sought from the participants children to our center, which represents the main clinical
(parents) who were informed their data may be used in a center for ASD in Upper Austria. This system of health-
publication. The data set generated during this study is not care gives all families—regardless of their socio-eco-
publicly available due to the sensitive nature of some of nomic, cultural, or educational background—equal access
the data, but will be made available from the correspond- to diagnostic assessments and, should ASD be diagnosed,
ing author upon reasonable request. the option of joining the ESDM early intervention pro-
gram. Typically, uptake of referral by pediatricians is high
among migrant families, which might be another explana-
Community involvement tion for our mainly multilingual sample. Furthermore,
There is no community involved in this study. some European studies have reported an increased
4 Autism 00(0)

Table 1.  Participant characteristics at baseline (before treatment).

N = 56

  M (SD) Range
Chronological age in months (M, SD) 41.96 (7.53) 29–60
Gender (male %) 91.1% (n = 51)  
Family
  Number of siblings 1.18 (1.22) 0–5
  Multilingualism (%) 78.6%  
  Mother’s level of education (% high school diploma or higher) 46.4%  
MSEL (AE; M)
  MSEL total (months) 21.47 (8.48) 9.50–48.75
  Receptive language (months) 17.63 (10.48) 5–49
  Expressive language (months) 17.64 (10.10) 5–48
  Visual reception (months) 26.19 (9.31) 11–54
  Fine motor (months) 25.58 (9.31) 11–47
PDDBI (T-scores)
  Autism Composite Score 49.04 (8.90) 28–69
  Social Approach Behaviors 47.68 (10.12) 30–72

MSEL: Mullen Scales of Early Learning; AE: age equivalent; PDDBI: Pervasive Developmental Behavior Inventory.

prevalence of ASD in the migrant populations (Bolton children 0–68 months of age. It covers the developmental
et al., 2014; Haglund & Källén, 2010; Lehti et al., 2013). domains of visual reception, fine motor skills, receptive
Sample characteristics at baseline are presented in language, and expressive language. The MSEL manual
Table 1. With an average chronological age of 41.96 months, reports good internal, test–retest, and inter-rater reliabili-
participants demonstrated a severe delay in expressive lan- ties, as well as good convergent validity with the Bayley
guage (age equivalent (AE) of 17.64 months and range = Scales of Infant Development (BSID; Albers & Grieve,
5–48 months), and a slightly less pronounced delay in non- 2007) with a correlation of 0.70 between MSEL-Compos-
verbal development (AE of 26.19 and range = 11–54). All ite and BSID. Since we encountered floor effects when
language parameters were assessed in the child’s primary calculating T-scores for our sample, we used developmen-
language, using translation support if necessary. Social- tal quotients (DQs: AE/chronological age × 100) for each
communicative abilities as rated by parents using the subscale, and a DQ total mean score was calculated in
PDDBI Social Approach Behavior subscale indicated an accordance with other authors (Bishop et al., 2011; Rogers
average level of functioning compared with the normative et al., 2012). As a measure of nonverbal DQ, a composite
sample of children with ASD and an average PDDBI score from the age-corrected MSEL subdomains fine
Autism Composite Score (see Table 1). motor skills and visual recognition was used. A verbal DQ
At baseline, contact with childcare institutions was was derived from the Expressive and Receptive Language
highly variable, with 40 children (71.4%) attending pre- scales.
school for a mean of 21.0 h per week (range = 8–35). At
the end of the intervention, 49 children (87.5%) were
attending pre-school for almost the same time per week (M
Measures: predictors of developmental change
= 21.5 h and range = 8–40). Thirteen children received Gestural approach behaviors. The Gestural Approach
additional therapy (e.g. occupational therapy) with an Behaviors subscale of the Social Approach Behavior
average of 26.2 h per year (range = 3–50), and eleven chil- dimension (SOCAPP) of the PDDBI (Cohen & Sudhalter,
dren received additional general early intervention ser- 2005) was considered to contain the most relevant infor-
vices with a mean of 48.3 h per year (range = 2–85). The mation on social-communicative gestures and, thus, to be
average weekly ESDM intervention time over the period potentially predictive of child outcomes. The PDDBI
of 12 months was M = 4.54 h (SD = 0.93). assesses adaptive and maladaptive behaviors in 10 key
dimensions for children with autism and provides an algo-
rithm to generate an autism total score. It is available in
Outcome measures
two forms: for rating by teachers or parents. This study
MSEL.  Child cognitive abilities were assessed at the insti- used the extended parental questionnaire in German. The
tute’s diagnostic center using the MSEL (Mullen, 1995), PDDBI was designed to monitor treatment responsiveness
which is a standardized developmental assessment for and provides age-referenced, normed dimensional scores
Laister et al. 5

Table 2.  Gestural approach behaviors, PDDBI parent rating form (Cohen & Sudhalter, 2005, p. 12).

Gestural approach behaviors (score these items whether or not the child has language)
117. Shows recognition of pictures of familiar people, animals, etc. by spontaneously pointing at or 0 1 2 3 ?
touching the pictures when he/she sees them
118. Spontaneously nods yes and/or shakes head no correctly when asked questions 0 1 2 3 ?
119. Moves arm(s)/hand(s) in beckoning motion to signal others to come to him or her 0 1 2 3 ?
120. Gestures and looks back and forth to others in order to share the experience with them (e.g. 0 1 2 3 ?
sees airplane and points at it while looking back to others and then back to the airplane)

for Approach/Withdrawal Problems, and Receptive/ correlates with the Childhood Autism Rating Scale (CARS;
Expressive Social Communication Abilities. Schopler et al., 1980; r = −0.60), the Vineland Adaptive
The Gestural Approach Behaviors subscale of the Behavior Scales (Sparrow et al., 1984; Communication: r
PDDBI measures the following gestural abilities used = 0.52; Socialization: r = 0.58) and the Griffiths scales
interactively by the child: (1) pointing to show (“Shows (Griffiths, 1984; General intelligence quotient (IQ) r =
recognition of pictures of familiar people, animals, etc. by 0.63). In summary, the SOCAPP Gestural Approach
spontaneously pointing at or touching the pictures when he Behaviors subscale refers specifically to the social use of
or she sees them”), in contrast to (2) pointing to initiate gestures, whereas the SOCAPP composite score can be
joint attention (“Gestures and looks back and forth to oth- interpreted as strengths/weaknesses in the use of a large
ers in order to share the experience with them (e.g. sees variety of social communication behaviors as compared to
airplane and points at it while looking back to others and an age-referenced norm sample of children with ASD.
then back to the airplane)”), (3) facial expressions (head
nodding/shaking) of agreement/rejection to yes/no ques-
tions, and (4) manual gestures to ask others to approach
Measures: controlling predictor variables
(see all items in Table 2). These skills are relevant to the To control for potentially confounding variables, we
theoretical framework of ESDM, and have been previ- included child’s age, multilingualism of family, mother’s
ously indicated as plausible predictors of early interven- educational level, and pre-school time during ESDM inter-
tion (Vivanti et al., 2017; e.g. Vivanti, Prior, et al., 2014). vention in our regression analysis. To control for IQ and
Each of the four behaviors is rated on a 4-point Likert-type language at baseline, we included MSEL nonverbal and
scale (0 = does not show behavior, 1 = rarely shows verbal DQs.
behavior, 2 = sometimes/partly shows behavior, and 3 =
usually/typically shows behavior).
Intervention
Social approach behaviors. The Social Approach Dimen- Implementation of the ESDM.  The intervention team con-
sion (SOCAPP) of the PDDBI (Cohen & Sudhalter, 2005) sisted of 12 persons from various disciplines (speech-lan-
assesses a variety of social communication skills that are guage therapy, occupational therapy, clinical psychology,
challenging for children with ASD and thus considered to and special education) who had been trained and certified
be highly discriminative between ASD and other develop- in the ESDM model between 2013 and 2018. Training
mental disorders. The scale contains the following nine consisted of introductory and intensive courses provided
subdomains: visual social approach behaviors (e.g. “Pays by certified ESDM trainers, followed by individual video-
attention to other’s face when given instructions or when based supervision according to the ESDM certification
asked questions”), positive affect behaviors (e.g. “Smiles process. The process until certification as ESDM therapist
when praised”), responsiveness to social inhibition cues lasted between 7 and 13 months (Rogers & Dawson, 2010).
(e.g. “Completely stops inappropriate behavior for at least In 2016, a member of the intervention team went through
a day when warned or punished”), social play behaviors the certification process of becoming an ESDM trainer,
(e.g. “Selects his or her own toys to play with and allows which allowed her to instruct other members of the team.
other(s) to play along”), imaginative play behaviors (e.g. The ESDM trainer also became supervisor of quality of
“Shows more complex imaginative toy play (e.g. feeds implementation: the intervention procedure for all children
doll, makes Superman™ toy fly)”), empathy behaviors enrolled in ESDM intervention was routinely supervised
(e.g. “Tries to physically or verbally comfort others when and reviewed according to the specific fidelity procedures
they are sad”), social interaction behaviors (e.g. “Seeks included in the ESDM manual (Rogers & Dawson, 2010).
affection (e.g. hugs, kisses) from caregivers or siblings”),
social imitative behaviors (e.g. “Can imitate tongue click- Intervention procedure. Participants meeting the criteria
ing”), and gestural approach behaviors (Table 2). SOCAPP described above were referred to the Autism Centre, where
has an excellent reliability score (α = 0.91) and highly the intervention team carried out assessments using the
6 Autism 00(0)

Table 3.  Participant functioning at baseline (T1) and at the end of intervention after 1 year (T2).

N = 56 T1 T2 MΔ (SD) t-test; p-value (Bonferroni–

Holm corrected)
M (SD) M (SD)
MSEL
  Receptive language (AE) 17.56 (10.54) 27.98 (15.52) 10.42 (8.87) t(49) = −8.31; p < 0.001**
  Expressive language (AE) 17.64 (10.10) 25.85 (14.28) 8.21 (7.91) t(52) = −7.55; p < 0.001**
  Visual reception (AE) 26.08 (9.54) 37.18 (14.77) 11.10 (8.79) t(50) = −9.01; p < 0.001**
  Fine motor (AE) 25.33 (7.94) 34.42 (11.29) 9.10 (6.79) t(51) = −9.67; p < 0.001**
  MSEL total (AE) 21.42 (8.64) 31.34 (13.12) 9.92 (6.86) t(46) = −9.91; p < 0.001**
  MSEL nonverbal (DQ) 62.75 (17.66) 66.71 (23.56) 3.96 (14.04) t(50) = −2.01; p = 0.050*
  MSEL verbal (DQ) 42.18 (22.10) 49.75 (26.82) 7.56 (14.20) t(49) = −3.77; p < 0.001**

MSEL: Mullen Scales of Early Learning; AE: age equivalent; DQ: developmental quotient.
*p < 0.05; **p < 0.01.

ESDM curriculum checklist on referral and then every longitudinal design was implemented between the two
3 months. Learning objectives were subsequently defined groups.
according to the curriculum checklist based on the obser- All data analyses were conducted by means of the SPSS
vation of the child’s abilities and including parent informa- 26 statistical software. In the first step, paired sample tests
tion. During assessment, two or more interventionists were were carried out to determine the differences in develop-
present, while intervention was delivered using a one-to- mental status pre- and post-treatment for the whole group.
one format. Two or three different interventionists pro- In the second step, t-tests for independent samples were
vided three weekly ESDM intervention sessions (90 min performed to check for differences between low and high
each), usually one at the child’s home, one at pre-school (if responders in terms of language gains. To control for Type
the child was in childcare and if the pre-school manage- 1 error within multiple comparisons, Bonferroni–Holm
ment approved), and one at the therapy center. Child pro- correction was used to adjust significance levels. Bivariate
gress was monitored through rigorous data collection, as correlations between predictors and outcome variables
required by the ESDM. were examined using Pearson’s correlation to obtain addi-
tional information for the regression models. Independent
Involvement of parents. Prior to intervention, 90% of the predictor variables were checked for multicollinearity, with
parents took part in a parent-training workshop comprising excellent variance inflation factor and tolerance values for
six 2-h meetings per week. During the 1:1 intervention all integrated variables. Separate hierarchical multiple
period, parents had to take part in at least one of the weekly regressions were used to compare the amount of variance
appointments at which interventionists offered parent explained by controlling and specific predictor variables.
coaching. The coaching procedure was based on the
ESDM parent-coaching approach (Rogers & Dawson,
2012), although fidelity data were not taken for parent Results
implementation of the ESDM strategies at home. Parents
could also use these coaching sessions to seek help in
Developmental gains
working on specific issues with their child with autism, for As shown in Table 3, a series of paired sample t-tests indi-
instance, on problematic behaviors or self-care issues. cates significant gains for all MSEL domains over the
12-month intervention period. Participants gained on aver-
age 9.92 months (SD = 6.86) in the MSEL total score
Design and statistical analysis (t(46) = −9.91; p < 0.001), with considerable individual
This study used a within-subject longitudinal (pre- and variability (range = 0.86–31.50 months). t-tests for the
post-) design to measure response to ESDM treatment over age-corrected MSEL nonverbal (t(50) = −2.01; p = 0.050)
a 12-month intervention period. Standardized develop- and the MSEL verbal (t(49) = −3.77; p < 0.001) DQ dif-
mental child information and theory-driven child-related ferences showed significant changes between T1 and T2.
predictors were measured by standardized instruments at In additional analyses, we found that parent-reported
T1. Developmental assessments were repeated at T2 (post- autism symptom severity (PDDBI Autism Composite
intervention) with an average of 12.86 months (range = Score) was significantly lower at T2 (T-score reduction by
10–20 and SD = 1.97) intervention period. A median split 7.21 points (SD = 8.58, p < 0.001)), and parents reported
of verbal AE gains was used to derive two groups: high a significant gain in social approach behaviors between T1
responders and low responders, and a between-subject and T2 (t(55) = −4.82; p < 0.001).
Laister et al. 7

Table 4.  Intercorrelations of all predictor variables.

Predictors 1 2 3 4 5 6 7 8
1. Age 1 −0.17 −0.10 0.13 −0.23 0.00 0.29* 0.27*
2. Multilingualism −0.17 1 0.09 −0.13 0.13 0.29* −0.02 0.09
3. Mother’s level of education −0.10 0.09 1 0.22 0.08 0.05 0.03 0.01
4. Pre-school time (weekly) 0.13 −0.13 0.22 1 0.02 −0.05 0.09 0.00
5. MSEL nonverbal (DQ) −0.23 0.13 0.08 0.02 1 0.65** 0.45** 0.58**
6. MSEL verbal (DQ) 0.00 0.29* 0.05 −0.05 0.65** 1 0.46** 0.69**
7. PDDBI Social Approach Behaviors (T) 0.29* −0.02 0.03 0.09 0.45** 0.46** 1 0.70**
8. PDDBI Gestural Approach Behaviors (RS) 0.27* 0.09 0.01 0.00 0.58** 0.49** 0.70** 1

MSEL: Mullen Scales of Early Learning; DQ: developmental quotient; PDDBI: Pervasive Developmental Behavior Inventory; T: T-score; RS: raw score.
*p < 0.05; **p < 0.01.

Table 5.  Correlations between putative predictors and MSEL verbal/nonverbal age equivalent gains.

Predictors MSEL nonverbal Δ MSEL verbal Δ MSEL total Δ

Age −0.05 −0.12 −0.06
Multilingualism −0.07 −0.14 −0.12
Mother’s level of education 0.01 0.14 0.10
Pre-school time/weekly 0.08 0.08 0.09
MSEL nonverbal (DQ) 0.36* 0.55** 0.50**
MSEL verbal (DQ) 0.37* 0.42** 0.48**
PDDBI Social Approach Behaviors (T) 0.38* 0.32* 0.39*
PDDBI Gestural Approach Behavior (RS) 0.43** 0.42** 0.48**

MSEL: Mullen Scales of Early Learning; DQ: developmental quotient; PDDBI: Pervasive Developmental Behavior Inventory; T: T-score; RS: raw score.
*p < 0.05; **p < 0.01.

Outcome predictors First, a hierarchical multiple regression analysis (see all

regression models in Table 6) was conducted to determine
Intercorrelations among predictor variables are summarized
whether social approach or gestural approach behaviors
in Table 4. The MSEL DQs used in the analysis were highly
improved the prediction of MSEL nonverbal or verbal AE
correlated with each other (r = 0.65; p < 0.01), as were the
gains when added to a set of controlling variables (step 1:
PDDBI subdomains (r = 0.70; p < 0.01). MSEL DQs also
age, mother’s level of education, multilingualism, time at
showed high correlations with PDDBI parent-reported ges-
pre-school during intervention) in step 2. Within this set of
tural approach behaviors (MSEL nonverbal DQ: r = 0.58;
p < 0.01 and verbal DQ: r = 0.69; p < 0.01). controlling variables, no significant predictors were deter-
Correlations between all controlling and putative predic- mined (variable set: R2= 0.07/0.04 for verbal and nonver-
tors with MSEL AE gains are summarized in Table 5. bal AE gains). In line with our hypotheses, adding social
Moderate to high correlations with MSEL AE changes were approach behaviors led to significant variance explanation
found between clinical dimensions (nonverbal and verbal for verbal and nonverbal AE gains (21%). Interestingly,
baseline DQs: r = 0.36–0.55) and parent-assessed social- gestural approach behavior added as much as 21% and
communicative behaviors (social approach and gestural 19% of variance explanation of verbal and nonverbal AE
approach behaviors: r = 0.32–0.48). Gestural approach gains, respectively.
behaviors were found to correlate significantly more As expected, nonverbal and verbal DQs at T1 were sig-
strongly with MSEL developmental AE gains than the much nificant predictors of verbal AE gains with significant
more comprehensive SOCAPP. Notably, correlations regression models, explaining a total variance of 29%–
between the social-communicative use of gestures (gestural 36% when added to the first-step controlling set.
approach behaviors) and MSEL AE gains were very similar Surprisingly, no significant regression model was found
in strength as MSEL DQs at baseline. No significant corre- for nonverbal AE gain when nonverbal or verbal DQs at
lations with MSEL AE gains were found for age, multilin- T1 were entered as predictors.
gualism, mother’s level of education or pre-school time. In a three-step analysis, adding baseline MSEL verbal
Since only a small number of children participated in other and nonverbal DQs to the set of control variables led to
therapeutic or non-specialized intervention services, these increased explanation of variance of MSEL verbal but not
variables were not integrated into the model. of nonverbal AE gains. We subsequently controlled
8 Autism 00(0)

Table 6.  Hierarchical multiple regression predicting verbal/nonverbal age equivalent gains.

MSEL verbal gain MSEL nonverbal gain

  B SE B β B SE B β
Controlling set (single step) R2 = 0.07; p = 0.58 R2 = 0.04; p = 0.80
 Age −0.20 0.16 −0.20 −0.13 0.14 −0.15
  Mother’s level of education 0.32 0.93 0.06 −0.49 0.818 −0.10
 Multilingualism −2.69 2.68 −0.16 −1.69 2.47 −0.11
  Time at pre-school during intervention 0.08 0.15 0.09 0.08 0.13 0.10
Two-step analyses
  Controlling set + MSEL nonverbal DQ R2 = 0.36; p = 0.005* R2 = 0.19; p = 0.13
0.24 0.06 0.57** 0.16 0.06 0.40*
  Controlling set + MSEL verbal DQ R2 = 0.29; p = 0.02* R2 = 0.19; p = 0.15
0.17 0.05 0.50** 0.14 0.05 0.42*
  Controlling set + social approach domain (RS) R2 = 0.21; p = 0.11 R2 = 0.21; p = 0.10
0.16 0.06 0.40* 0.15 0.05 0.42*
  Controlling set + gestural approach behavior (RS) R2 = 0.28; p = 0.03* R2 = 0.23; p = 0.07
4.54 1.37 0.48** 3.78 1.24 0.45**
Three-step analyses
 Controlling set + MSEL nonverbal DQ + MSEL R2 = 0.41; p = 0.004** R2 = 0.22; p = 0.15
verbal DQ 0.15 0.07 0.37* 0.09 0.08 0.23
0.11 0.06 0.31*** 0.08 0.07 0.26
 Controlling set + MSEL nonverbal DQ + gestural R2 = 0.38; p = 0.008** R2 = 0.25; p = 0.08
approach behavior (RS) 0.19 0.08 0.46* 0.08 0.08 0.19
1.62 1.73 0.18 2.70 1.64 0.32***

SE: standard error; MSEL: Mullen Scales of Early Learning; DQ: developmental quotient; RS: raw score.
*p < 0.05; **p < 0.01; ***p = ~0.10.

whether the highly predictive gestural approach behavior differences between the two groups in terms of develop-
scale adds further variance explanation when introduced mental profile (see Table 7).
after the set of control variables (age, mother’s level of Interestingly, no significant group differences were
education, multilingualism, and time at pre-school) and found for the parent-reported Social Approach Behavior
the nonverbal DQ at T1. As expected, gestural approach domain (t(48) = −1.10; p = 1.00), but in line with its high
increased the prediction of verbal AE gains only minimally predictive value for outcome as reported above, we con-
(by 2%) after the highly predictive nonverbal DQ at T1 firmed significant group differences between low and high
had been entered. However, gestural approach added responders for the Gestural Approach Behavior subscale
another 6% of variance explanation of nonverbal AE gain (t(48) = −2.85; p = 0.045*) at T1. Severity of Autism
and was determined as the major predictor in this model Symptoms (PDDBI Composite Score) at baseline did not
that explained slightly more variance than the three-step differ significantly between the two subsamples. Further
model with verbal DQ at T1 entered as third predictor. single-step regression analyses to predict language AE
gains within both subsamples showed that social-commu-
nicative gestures (R2 = 0.23) and SOCAPP (R2 = 0.17)
Differences between low and high responders
were significant predictors exclusively in the group of low
in terms of language outcomes responders, which was characterized by substantial delays
Finally, we sought to find baseline differences between in language (MSEL verbal AE around 14 months) and non-
children with low and high AE language gains in response verbal cognition (MSEL nonverbal AE around 21 months).
to intervention. To this end, the intervention group was
divided by median split of MSEL verbal developmental
Discussion
age gains (in months; median = 7.69; range = −0.67 to
28.50) into low and high responders. There were no differ- This study examined the theory-driven hypothesis that early
ences between the two groups concerning age, multilin- social-communicative gestures predict verbal and nonver-
gualism in the family, mother’s level of education, or bal AE gains in children with ASD receiving community-
pre-school time during intervention. Because of high cor- based ESDM. As ESDM focuses on building learning
relations among all MSEL subscales and between MSEL opportunities on the child’s social initiations, the ability to
language and PDDBI subscales, we expected large use social-communicative gestures was hypothesized to
Laister et al. 9

Table 7.  Baseline characteristics of low versus high responders (median split of MSEL verbal AE gains).

Low responders (N = 25) High responders (N = 25) t-test; p-value (Bonferroni–

Holm corrected)
  M (SD) M (SD)
Chronological age (M, SD) 41.56 (6.99) 41.72 (8.22) 1.00
Pre-school time/weekly 21.25 (8.26) 23.04 (8.95) 1.00
MSEL total (AE) 16.98 (5.53) 25.83 (8.81) 0.002**
MSEL receptive language (AE) 13.76 (8.81) 21.36 (10.92) 0.06
MSEL expressive language (AE) 13.84 (8.78) 30.92 (10.23) 0.06
MSEL visual reception (AE) 20.92 (5.55) 30.88 (10.64) 0.002**
MSEL fine motor (AE) 21.75 (5.84) 28.44 (8.68) 0.025**
MSEL nonverbal (DQ) 53.24 (14.44) 71.32 (16.79) <0.002**
MSEL verbal (DQ) 33.55 (18.94) 50.82 (21.98) <0.036*
PDDBI Autism Composite Score (T) 50.44 (9.17) 47.96 (9.04) 1.00
PDDBI Social Approach Behaviors (T) 46.32 (10.75) 49.44 (9.20) 1.00
PDDBI Gestural Approach Behavior (RS) 0.70 (0.75) 1.32 (0.79) 0.045*

MSEL: Mullen Scales of Early Learning; AE: age equivalent; DQ: developmental quotient; PDDBI: Pervasive Developmental Behavior Inventory; T:
T-score; RS: raw score.
*p < 0.05; **p < 0.01.

facilitate verbal and nonverbal learning. To our knowledge, variance was smaller than for the verbal gain prediction
this is the first study to analyze the use of gestures in every- model. Gestural approach behavior was the predictor con-
day life—measured by the PDDBI Gestural Approach sub- tributing most to explaining the variance (23%), closely fol-
scale—in terms of its predictive value for child-developmental lowed by social approach behavior (21%) and verbal and
outcome following early intervention. As detailed below, nonverbal baseline levels (both 19%) after accounting for
the prediction that social-communicative gestures at base- the controlling variables. As hypothesized, the use of social-
line contributed to AE gains for children receiving the communicative gestures—and, to a smaller degree, social-
ESDM appeared to be supported by our data. communicative skills in general (SOCAPP)—predicted
verbal and nonverbal AE gains after ESDM intervention.
Interestingly, after accounting for baseline nonverbal skills,
Factors associated with verbal gains
age, family variables, and time at pre-school, Gestural
Children demonstrated significant AE gains in verbal and Approach Behavior explained an additional significant pro-
nonverbal development after about 12 months of low-inten- portion of variance in nonverbal development.
sity ESDM intervention. Child nonverbal and verbal charac-
teristics assessed at baseline (after controlling for child age,
Role of social-communicative use of gestures
family variables, and time at pre-school) contributed to a
significant proportion of verbal gains during the interven- Our finding is in line with the literature supporting the role
tion period. Nonverbal developmental level was the single of social-communicative gestures for language develop-
highest predictive variable (cf. Paynter et al., 2018), fol- ment in typically developing children (Hall et al., 2013). In
lowed by verbal skills. Social-communicative gestures at addition, our sample was characterized by profound lan-
baseline as reported by parents using the PDDBI Gestural guage and severe cognitive delays, with a majority of chil-
Approach Behaviors subscale were found to significantly dren from non-German speaking families. Therefore, early
predict verbal development to the same extent as baseline social-communicative use of gestures might be a factor pre-
language skills. The predictive value of the comprehensive dicting gains in children receiving ESDM with low lan-
PDDBI Social Approach Behaviors scale at baseline was guage levels at baseline. Concerning the quality of parent
lower, and the model (including the control set) did not reports, there are indications that their assessments of social-
reach significance. Similarly, social-communicative ges- communicative gestures using the PDDBI subscale of
tures at baseline were significantly lower for low responders Gestural Approach Behaviors are valid. Our pre-interven-
compared to high responders (based on a median split tion analyses demonstrate strong correlations with nonver-
according to language AE gains), whereas social approach bal DQ (r = 0.58) and—as expected—even stronger
behaviors were not discriminative. correlations with verbal DQ (r = 0.69) as assessed by stand-
ardized instruments (MSEL). These results suggest that
individual differences in nonverbal social development can
Factors associated with nonverbal gains be reliably assessed by parents observing their children’s
Nonverbal AE gains during the intervention period were use of gestures. Parental rating of social use of gestures was
smaller but still significant. The total amount of explained the single strongest predictor of nonverbal development
10 Autism 00(0)

after ESDM intervention and a strong predictor of verbal Conclusion and future directions
AE gains, explaining about the same variance of verbal AE
gains as verbal skills at baseline assessed by a standardized Social communication deficits are a core feature of ASD.
instrument. Notably, implementation of the very short sec- Atypical social-communicative use of gestures is one of
tion of the PDDBI questionnaire (four questions) in the fam- the earlier expressions of these deficits, which can be
ily context resulted to be highly feasible in a sample with a simply and reliably assessed by parents before their
high proportion of family languages other than German. child starts to use expressive language. Previous find-
ings have suggested that baseline imitation, joint atten-
tion, and functional play—three early emerging preverbal
Implications abilities linked to language development—might be
Our findings point to several clinical implications. First, as associated with change for children receiving the ESDM.
use of gestural communication was associated with change This study contributes further support to the notion that
in children receiving ESDM, it is critical to include these preverbal social-communicative skills might play a role
early emerging behaviors as intervention targets. Second, in response to the ESDM by suggesting that children
due to our specific sample characteristics (and the signifi- who use social-communicative gestures more often
cant predictive results for the group of low responders), might benefit more from receiving the ESDM. Future
our finding that social-communicative gestures are a research should substantiate this finding by testing
robust predictor of response to ESDM intervention may whether gains in the social-communicative use of ges-
apply particularly to children with highly restricted lan- tures mediate treatment-related changes in children
guage skills and may therefore be relevant to intervention receiving the ESDM using a randomized controlled
planning for young children and those with severe lan- design. Based on this preliminary body of evidence, it
guage deficits. As age at ASD diagnosis decreases, paren- appears that children who engage more frequently in
tal observations of early social-communicative use of preverbal social-communicative behaviors, such as imi-
gestures as rated with this key set of questions from the tation, gaze following, functional play, and, as shown in
PDDBI subscale should be considered as possibly relevant this study, social-communicative gestures, might be
to decision-making in the care process. Furthermore, in good candidates for receiving the ESDM (and most
predicting ESDM outcomes, social-communicative ges- likely, other interventions based on similar principles).
tures were as relevant as language skills assessed by a For children with more limited skills in these dimen-
standardized instrument. Gesture use should be considered sions, it is imperative that these skills are targeted early
as an alternative or additional factor when lack of child and intensively as they might play a pivotal role in
cooperation or foreign family languages makes language responding to the learning opportunities provided by the
assessment difficult. The higher predictive value of social ESDM. In addition, it is critical that future research
gesture assessment compared to a comprehensive ques- investigates whether children who do not respond opti-
tionnaire on a variety of social communication behaviors mally to one intervention (e.g. ESDM) might be more
(SOCAPP) further supports its clinical usability. responsive to a different intervention, using comparative
designs that examine moderators and mediators of
response to different treatments.
Limitations
The findings of this study must be viewed considering sev- Acknowledgements
eral limitations. First, in the absence of a control group, it The authors thank all participating families for their patience and
was not possible to isolate outcomes after ESDM interven- great effort in supporting their children, and the intervention and
tion from outcomes that would have occurred naturally in diagnostic team of the Institute of Neurology of Senses of the
Hospital of St. John of God, who assisted with recruitment and
the course of development. In addition, since—as per the
data collection. The authors also thank Christoph Weber, who
clinic’s standard procedures—children are seen by the same
provided valuable input on statistical procedures, and William
clinicians over time, they were not blinded for the assess- Barbaresi and Johannes Fellinger, who contributed valuable
ments. Furthermore, our finding of a higher predictive value thoughts on formulating the argumentation line.
of baseline social gestures compared to a composite with
other social communication skills (SOCAPP) might be lim- Funding
ited to children with profound language delay. Finally, since
The author(s) received no financial support for the research,
the whole sample received low-intensity ESDM, we do not
authorship, and/or publication of this article.
know whether the predictive validity of the use of early
social-communicative gestures would still hold for a more
ORCID iDs
intense mode of intervention delivery, particularly as teach-
ing the use of gestures to express basic communicative Dominik Laister https://orcid.org/0000-0002-3868-8344
functions is included in the ESDM strategies. Daniel Holzinger https://orcid.org/0000-0001-6634-6117
Laister et al. 11

References Fein, D., Barton, M., Eigsti, I. M., Kelley, E., Naigles, L., Schultz,
R. T., Stevens, M., Helt, M., Orinstein, A., Rosenthal, M.,
Albers, C. A., & Grieve, A. J. (2007). Test review: Bayley,
Troyb, E., & Tyson, K. (2013). Optimal outcome in individ-
N. (2006). Bayley scales of infant and toddler develop-
uals with a history of autism. Journal of Child Psychology
ment–third edition. San Antonio, TX: Harcourt assess-
and Psychiatry, and Allied Disciplines, 54(2), 195–205.
ment. Journal of Psychoeducational Assessment, 25(2),
https://doi.org/10.1111/jcpp.12037
180–190.
Griffiths, R. (1984). The abilities of young children: A compre-
Amt der OÖ Landesregierung. (2017). Kindertagesheimstatistik
hensive system of mental measurement for the first eight
2018/19: Bericht [Statistical report for day-care centers for
years of life. Association for Research in Infant and Child
children in Upper Austria]. https://www.land-oberoesterre-
Development.
ich.gv.at/Mediendateien/Formulare/DokumenteAbt_Stat/
Haglund, N. G., & Källén, K. B. (2010). Risk factors for autism and
Bericht_Kindertagesheimstatsitik_2016_17_Internet.pdf
Asperger syndrome: Perinatal factors and migration. Autism:
(accessed 27 February 2021).
The International Journal of Research and Practice, 15(2),
Barbaresi, W. J., Katusic, S. K., & Voigt, R. G. (2006). Autism:
163–183. https://doi.org/10.1177/1362361309353614
A review of the state of the science for pediatric primary
Hall, S., Rumney, L., Holler, J., & Kidd, E. (2013). Associations
health care clinicians. Archives of Pediatrics & Adolescent
among play, gesture and early spoken language acqui-
Medicine, 160(11), 1167–1175.
sition. First Language, 33(3), 294–312. https://doi.
Bishop, S. L., Guthrie, W., Coffing, M., & Lord, C. (2011).
org/10.1177/0142723713487618
Convergent validity of the Mullen Scales of Early Learning
Holzinger, D., Laister, D., Vivanti, G., Barbaresi, W. J., &
and the differential ability scales in children with autism
Fellinger, J. (2019). Feasibility and outcomes of the Early
spectrum disorders. American Journal on Intellectual and
Start Denver Model implemented with low intensity in a
Developmental Disabilities, 116(5), 331–343. https://doi.
community setting in Austria. Journal of Developmental
org/10.1352/1944-7558-116.5.331
and Behavioral Pediatrics, 40(5), 354–363. https://doi.
Bolton, S., McDonald, D., Curtis, E., Kelly, S., & Gallagher, L.
(2014). Autism in a recently arrived immigrant population. org/10.1097/DBP.0000000000000675
European Journal of Pediatrics, 173(3), 337–343. https:// Howlin, P., Charman, T., & Ghaziuddin, M. (2011): The SAGE
doi.org/10.1007/s00431-013-2149-6 handbook of developmental disorders. SAGE Publications
Bruner, J. S. (1981). The social context of language acquisition. Ltd.
Language & Communication, 1(2–3), 155–178. https://doi. Hudry, K., McConachie, H., Le Couteur, A., Howlin, P.,
org/10.1016/0271-5309(81)90010-0 Barrett, B., Slonims, V., & PACT Consortium. (2018).
Chawarska, K., Klin, A., Paul, R., Macari, S., & Volkmar, F. Predictors of reliable symptom change: Secondary anal-
(2009). A prospective study of toddlers with ASD: Short- ysis of the Preschool Autism Communication Trial.
term diagnostic and cognitive outcomes. Journal of Child Autism & Developmental Language Impairments, 3 (28),
Psychology and Psychiatry, 50(10), 1235–1245. https://doi. 2396941518764760.
org/10.1111/j.1469-7610.2009.02101.x Ingersoll, B., & Schreibman, L. (2006). Teaching reciprocal imi-
Chevallier, C., Kohls, G., Troiani, V., Brodkin, E. S., & Schultz, tation skills to young children with autism using a naturalis-
R. T. (2012). The social motivation theory of autism. tic behavioral approach: Effects on language, pretend play,
Trends in Cognitive Sciences, 16(4), 231–239. https://doi. and joint attention. Journal of Autism and Developmental
org/10.1016/j.tics.2012.02.007 Disorders, 36(4), 487–505. https://doi.org/10.1007/s10803-
Cohen, I. L., & Sudhalter, V. (2005). PDD Behavior Inventory 006-0089-y
(PDDBI). Psychological Assessment Resources. Kim, S. H., Macari, S., Koller, J., & Chawarska, K. (2016).
Contaldo, A., Colombi, C., Pierotti, C., Masoni, P., & Muratori, Examining the phenotypic heterogeneity of early autism spec-
F. (2020). Outcomes and moderators of Early Start Denver trum disorder: Subtypes and short-term outcomes. Journal
Model intervention in young children with autism spectrum of Child Psychology and Psychiatry, and Allied Disciplines,
disorder delivered in a mixed individual and group setting. 57(1), 93–102. https://doi.org/10.1111/jcpp.12448
Autism: The International Journal of Research and Practice, Koegel, R. L., O’Dell, M., & Koegel, L. K. (1987). A natural
24(3), 718–729. https://doi.org/10.1177/1362361319888344 language teaching paradigm for nonverbal autistic chil-
Dawson, G., Rogers, S., Munson, J., Smith, M., Winter, dren. Journal of Autism and Developmental Disorders, 17,
J., Greenson, J., Donaldson, A., & Varley, J. (2010). 187–199.
Randomized, controlled trial of an intervention for toddlers Landa, R. J. (2018). Efficacy of early interventions for infants
with autism: The Early Start Denver Model. Pediatrics, and young children with, and at risk for, autism spectrum
125(1), e17–e23. https://doi.org/10.1542/peds.2009-0958 disorders. International Review of Psychiatry, 30(1), 25–39.
Dawson, G., Toth, K., Abbott, R., Osterling, J., Munson, J., Estes, https://doi.org/10.1080/09540261.2018.1432574
A., & Liaw, J. (2004). Early social attention impairments in Lehti, V., Hinkka-Yli-Salomäki, S., Cheslack-Postava, K.,
autism: Social orienting, joint attention, and attention to dis- Gissler, M., Brown, A. S., & Sourander, A. (2013). The
tress. Developmental Psychology, 40(2), 271–283. https:// risk of childhood autism among second-generation migrants
doi.org/10.1037/0012-1649.40.2.271 in Finland: A case–control study. BMC Pediatrics, 13(1),
Ellis Weismer, S., & Kover, S. T. (2015). Preschool language Article 171. https://doi.org/10.1186/1471-2431-13-171
variation, growth, and predictors in children on the autism Lobban-Shymko, J., Im-Bolter, N., & Freeman, N. (2017). Early
spectrum. Journal of Child Psychology and Psychiatry, social communicative skills as predictors of symptom sever-
56(12), 1327–1337. https://doi.org/10.1111/jcpp.12406 ity in autism spectrum disorder. Autism & Developmental
12 Autism 00(0)

Language Impairments. Advance online publication. https:// Rogers, S. J., Estes, A., Vismara, L., Munson, J., Zierhut, C.,
doi.org/10.1177/2396941517743418 Greenson, J., Dawson, G., Rocha, M., Sugar, C., Senturk,
Lord, C., Risi, S., Lambrecht, L., Cook, J. E. H., Leventhal, B. D., Whelan, F., & Talbott, M. (2019). Enhancing low-inten-
L., DiLavore, P. C., Pickles, A., & Rutter, M. (2000). The sity coaching in parent implemented Early Start Denver
autism diagnostic observation schedule-generic: A stand- Model intervention for early autism: A randomized compar-
ard measure of social and communication deficits associ- ison treatment trial. Journal of Autism and Developmental
ated with the spectrum of autism. Journal of Autism and Disorders, 49(2), 632–646. https://doi.org/10.1007/s10803-
Developmental Disorders, 30(3), 205–223. https://doi. 018-3740-5
org/10.1023/A:1005592401947 Rogers, S. J., & Pennington, B. F. (1991). A theoretical approach
Loth, E., Charman, T., Mason, L., Tillmann, J., Jones, E. J. to the deficits in infantile autism. Development and
H., Wooldridge, C., Ahmad, J., Auyeung, B., Brogna, Psychopathology, 3(2), 137–162. https://doi.org/10.1017/
C., Ambrosino, S., Banaschewski, T., Baron-Cohen, S., S0954579400005204
Baumeister, S., Beckmann, C., Brammer, M., Brandeis, Sandbank, M., Bottema-Beutel, K., Crowley, S., Cassidy, M.,
D., Bölte, S., Bourgeron, T., Bours, C., . . . Buitelaar, J. Feldman, J. I., Canihuante, M., & Woynaroski, T. (2020).
K. (2017). The EU-AIMS Longitudinal European Autism Intervention effects on language in children with autism: A
Project (LEAP): Design and methodologies to identify and project AIM meta-analysis. Journal of Speech, Language,
validate stratification biomarkers for autism spectrum dis- and Hearing Research, 63(5), 1537–1560. https://doi.
orders. Molecular Autism, 8, 24. https://doi.org/10.1186/ org/10.1044/2020_JSLHR-19-00167
s13229-017-0146-8 Schopler, E., Reichler, R. J., DeVellis, R. F., & Daly, K. (1980).
Maljaars, J., Noens, I., Jansen, R., Scholte, E., & van Berckelaer- Toward objective classification of childhood autism:
Onnes, I. (2011). Intentional communication in nonverbal Childhood Autism Rating Scale (CARS). Journal of Autism
and verbal low-functioning children with autism. Journal and Developmental Disorders, 10(1), 91–103. https://doi.
of Communication Disorders, 44(6), 601–614. https://doi. org/10.1007/BF02408436
org/10.1016/j.jcomdis.2011.07.004 Schreibman, L., Dawson, G., Stahmer, A. C., Landa, R., Rogers,
Mullen, E. M. (1995). Mullen Scales of Early Learning. American S. J., McGee, G. G., & Halladay, A. (2015). Naturalistic
Guidance Service. developmental behavioral interventions: Empirically vali-
Murillo, E., & Belinchón, M. (2012). Gestural-vocal coordina- dated treatments for autism spectrum disorder. Journal of
tion: Longitudinal changes and predictive value on early Autism and Developmental Disorders, 45(8), 2411–2428.
lexical development. Gesture, 12(1), 16–39. https://doi. Sherer, M. R., & Schreibman, L. (2005). Individual behavio-
org/10.1075/gest.12.1.02mur ral profiles and predictors of treatment effectiveness for
Paparella, T., Goods, K. S., Freeman, S., & Kasari, C. (2011). The children with autism. Journal of Consulting and Clinical
emergence of nonverbal joint attention and requesting skills Psychology, 73(3), 525–538.
in young children with autism. Journal of Communication Sinai-Gavrilov, Y., Gev, T., Mor-Snir, I., Vivanti, G., &
Disorders, 44(6), 569–583. https://doi.org/10.1016/j.jcom- Golan, O. (2020). Integrating the Early Start Denver
dis.2011.08.002 Model into Israeli community autism spectrum disor-
Paynter, J., Trembath, D., & Lane, A. (2018). Differential out- der preschools: Effectiveness and treatment response
come subgroups in children with autism spectrum disor- predictors. Autism: The International Journal of
der attending early intervention. Journal of Intellectual Research and Practice, 24(8), 2081–2093. https://doi.
Disability Research: JIDR, 62(7), 650–659. https://doi. org/10.1177/1362361320934221
org/10.1111/jir.12504 Smith, T., Klorman, R., & Mruzek, D. W. (2015). Predicting out-
Rogers, S. J., & Dawson, G. (2010). Early start Denver model for come of community-based early intensive behavioral inter-
young children with autism: Promoting language, learning, vention for children with autism. Journal of Abnormal Child
and engagement. Guilford Press. Psychology, 43(7), 1271–1282. https://doi.org/10.1007/
Rogers, S. J., & Dawson, G. (2012). Early start for your child s10802-015-0002-2
with autism: Using everyday activities to help kids connect, Solomon, R., van Egeren, L. A., Mahoney, G., Quon Huber, M. S.,
communicate, and learn. Guilford Publications. & Zimmerman, P. (2014). Play Project Home Consultation
Rogers, S. J., Estes, A., Lord, C., Munson, J., Rocha, M., Winter, J., intervention program for young children with autism spec-
Greenson, J., Colombi, C., Dawson, G., Vismara, L. A., Sugar, trum disorders: A randomized controlled trial. Journal of
C. A., Hellemann, G., Whelan, F., & Talbott, M. (2019). A mul- Developmental and Behavioral Pediatrics, 35(8), 475–485.
tisite randomized controlled two-phase trial of the Early Start https://doi.org/10.1097/DBP.0000000000000096
Denver Model compared to treatment as usual. Journal of the Sparrow, S. S., Balla, D. A., Cicchetti, D. V., & Harrison, P. L.
American Academy of Child and Adolescent Psychiatry, 58(9), (1984). Vineland adaptive behavior scales. Circle Pines,
853–865. https://doi.org/10.1016/j.jaac.2019.01.004 MN: American Guidance Service.
Rogers, S. J., Estes, A., Lord, C., Vismara, L., Winter, J., Tomasello, M., & Schröder, J. (2009). Die Ursprünge der men-
Fitzpatrick, A., Guo, M., & Dawson, G. (2012). Effects schlichen Kommunikation [The origins of human’s commu-
of a brief Early Start Denver Model (ESDM)-based nication]. Suhrkamp.
parent intervention on toddlers at risk for autism spec- Vivanti, G., Dissanayake, C., Duncan, E., Feary, J., Capes, K.,
trum disorders: A randomized controlled trial. Journal Upson, S., Bent, C. A., Rogers, S. J., & Hudry, K. (2019).
of the American Academy of Child and Adolescent Outcomes of children receiving Group-Early Start Denver
Psychiatry, 51(10), 1052–1065. https://doi.org/10.1016/j. Model in an inclusive versus autism-specific setting: A
jaac.2012.08.003 pilot randomized controlled trial. Autism: The International
Laister et al. 13

Journal of Research and Practice, 23(5), 1165–1175. Vivanti, G., Kasari, C., Green, J., Mandell, D., Maye, M., &
https://doi.org/10.1177/1362361318801341 Hudry, K. (2018). Implementing and evaluating early inter-
Vivanti, G., Dissanayake, C., Zierhut, C., & Rogers, S. J. vention for children with autism: Where are the gaps and
(2013). Brief report: Predictors of outcomes in the Early what should we do? Autism Research, 11(1), 16–23.
Start Denver Model delivered in a group setting. Journal Vivanti, G., & Stahmer, A. C. (2020). Can the Early Start Denver
of Autism and Developmental Disorders, 43(7), 1717–1724. Model be considered ABA practice? Behavior Analysis
https://doi.org/10.1007/s10803-012-1705-7 in Practice. Advance online publication. https://doi.
Vivanti, G., Feary, J., Duncan, E., Zierhut, C., Dawson, G., & Rogers, org/10.1007/s40617-020-00474-3
S. J. (2017). G-ESDM treatment strategies. In G. Vivanti, E. Waterhouse, L. H. (2013). Rethinking autism: Variation and
Duncan, G. Dawson, & S. J. Rogers (Eds.), Implementing the complexity (1st ed.). Academic Press.
group-based Early Start Denver Model for preschoolers with Watson, L. R., Crais, E. R., Baranek, G. T., Dykstra, J. R., &
autism (Vol. 23, pp. 71–86). Springer International Publishing. Wilson, K. P. (2013). Communicative gesture use in infants
https://doi.org/10.1007/978-3-319-49691-7_6 with and without autism: A Retrospective Home Video
Vivanti, G., Paynter, J., Duncan, E., Fothergill, H., Dissanayake, Study. American Journal of Speech-Language Pathology,
C., & Rogers, S. J. (2014). Effectiveness and feasibility 22(1), 25–39. https://doi.org/10.1044/1058-0360(2012/11-
of the Early Start Denver Model implemented in a group- 0145)
based community childcare setting. Journal of Autism and Xu, Y., Yang, J., Yao, J., Chen, J., Zhuang, X., Wang, W.,
Developmental Disorders, 44(12), 3140–3153. https://doi. Zhang, X., & Lee, G. T. (2018). A Pilot study of a
org/10.1007/s10803-014-2168-9 culturally adapted early intervention for young chil-
Vivanti, G., Prior, M., Williams, K., & Dissanayake, C. (2014). dren with autism spectrum disorders in China. Journal
Predictors of outcomes in autism early intervention: Why of Early Intervention, 40(1), 52–68. https://doi.
don’t we know more? Frontiers in Pediatrics, 2, Article 58. org/10.1177/1053815117748408
https://doi.org/10.3389/fped.2014.00058 Yoder, P., Watson, L. R., & Lambert, W. (2015). Value-added
Vivanti, G., & Rogers, S. J. (2014). Autism and the mirror neuron predictors of expressive and receptive language growth in
system: Insights from learning and teaching. Philosophical initially nonverbal preschoolers with autism spectrum dis-
Transactions of the Royal Society of London. Series B, orders. Journal of Autism and Developmental Disorders,
Biological Sciences, 369(1644), 20130184. https://doi. 45(5), 1254–1270. https://doi.org/10.1007/s10803-014-
org/10.1098/rstb.2013.0184 2286-4