A Systematic Review of Play-Integrated AAC Interventions for Minimally Verbal and Nonspeaking Children

Abstract

This systematic review explored past experimental research on augmentative and alternative communication (AAC) interventions for minimally and nonspeaking autistic and intellectually/developmentally disabled children, with a focus on interventions which integrated play. Databases were searched between 2021 and 2025; included studies (n = 14) implemented AAC interventions for children ages 3–8 years old with a primary outcome of social communication. Studies were coded for AAC outcomes and play variables. Results indicate a temporal shift in intervention frameworks and level of play integration. Limitations include an expanded and broad definition of what constituted play-integrated interventions and restrictions on descriptions of the participants included. Implications for practice and future research are discussed. The primary meta-analysis was preregistered on PROSPERO (CRD42021292027).

Complex communication needs (CCN) is a term used to describe individuals with significant language needs who do not use spoken language as their primary communication mode and who may benefit from nonverbal means of communication, such as augmentative and alternative communication (AAC). This can include individuals who are nonverbal and do not demonstrate consistent verbal expressive words or minimally verbal and can use some spoken words but still demonstrate word production levels below the 10th percentile relative to age (Koegel et al., 2020). CCN can be associated with a wide range of conditions, including intellectual and developmental disabilities (IDDs) and autism spectrum disorder (ASD). According to the American Association on Intellectual and Developmental Disabilities, an IID is a condition characterized by significant limitations in both intellectual functioning and adaptive behavior and encompasses labels such as intellectual disability, ASD, and multiple disabilities (Schalock et al., 2021). ASD is a neurodevelopmental disorder categorized by impaired social skills and delays in the development of social communication skills and restrictive and repetitive behavior (American Psychiatric Association, 2013). Current prevalence of children in the United States diagnosed with a developmental disability is 8.56%, according to the Department of Health and Human Services (Zablotsky et al., 2023). Recent studies estimate that 30% of children with ASD and significant communication delays are designated as minimally verbal or nonspeaking (MV/NS; Lord et al., 2004; Norrelgen et al., 2014; Wodka et al., 2013).

Children with IDD and ASD may experience altered or delayed development of language and social communication (American Psychiatric Association, 2013). Likewise, autistic children may experience delays in the development of play skills (Chen et al., 2019; Shepherd et al., 1994). As play is a key way in which children connect with others, including their peers, children with IDD may face isolation, which can further the developmental gap between them and their peers (Wolfberg et al., 2012). For these reasons, many interventions for young children with disabilities target social communication skills and play skills through the use of play-based interventions (Vivanti & Zhong, 2020).

AAC for Children with ASD and/or IDDs

AAC is a system of communication that provides individuals with an alternative means of communication by replacing or supplementing speech that is absent or unintelligible (American Speech-Language-Hearing Association [ASHA], 2025; Steinbrenner et al., 2020). AAC includes unaided and aided types and encompasses a broad range of modalities from gestures to using an electronic device. Unaided communication modes do not require the speaker to use additional materials and/or devices, such as gestures or sign language. Aided AAC utilizes additional materials and/or electronic equipment and can range from low-tech systems (e.g., exchange-based picture systems and communication books) to high-tech systems (e.g., speech-generating devices [SDGs] and software applications on a phone/tablet; ASHA, 2025; Johnston & Cosbey, 2012; Mirenda, 2019).

Multiple prior meta-analyses and a review of reviews, or mega-review, evaluated the implementation of AAC interventions and several relevant communication outcomes for MV/NS individuals, including those with ASD and/or IDD (Crowe et al., 2022; Ganz et al., 2023a, 2023b). Of note, the primary meta-analysis via which the included articles were gleaned primarily included participants with ASD, with or without IDD (Reichle et al., under review). One of the prior meta-analyses found overall positive effects across communication modes and functions for participants with ASD/IDD (Ganz et al., 2023b); however, there were no clear moderators related to age, diagnoses (ASD, IDD, or both), or possible prerequisite skills (Ganz et al., 2023b); communication mode prior to intervention was found to be a moderator for one effect size metric, with the highest effects for low-tech AAC, but this was not a moderator for the other metric (log response ratio). Another found overall positive effects for individuals with ASD/IDD when considering effects for AAC-related outcomes only, again, with no clear moderators related to type of communication mode or communicative functions (e.g., joint attention, behavior regulation, and social interaction), even when controlling for participant characteristics (Ganz et al., 2024). An additional meta-analysis with ASD/IDD participants (Ganz et al., 2023a) reported that, for AAC-related and other communication outcomes, there was no statistically significant difference between implementation of behavioral or contrived versus naturalistic strategies, except when not controlling for participant and instructional variables for one of the metrics (Tau(AB)), which found higher effects for behavioral AAC interventions for participants with ASD. Conversely, when speech-generating AAC interventions for the broader population of MV/NS children were reviewed, almost 20 studies were found to have successfully implemented AAC interventions within naturalistic settings (Gevarted & Zamora, 2018).

Included articles across AAC reviews and primary studies provided sparse information regarding race, ethnicity, and primary language (Crowe et al., 2022; Ganz et al., 2023b); thus, it is unclear to what populations these results are generalizable. That said, the participants were skewed toward young children, with a median age of 5. Given that most of the participants of AAC research with this population were under age 10 (Ganz et al., 2023b), it appears that AAC interventions are effective in improving social communication outcomes in many young children with ASD and/or IDD, although results are highly variable across included studies within several meta-analyses and potential moderator variables (Ganz et al., 2023a, 2023b, 2024). A more fine-grained review of behavioral versus naturalistic strategies, including play-based activities, with younger children is warranted.

Models of Intervention

Autistic children often receive various interventions designed to improve developmental outcomes, including social communication (Fuller & Kaiser, 2019). Traditional approaches have used behavior analytic techniques derived from the work of Ivar Lovaas (1987). These traditional behaviorist approaches often use highly didactic and contrived methods of teaching, such as discrete trial training (DTT) (Smith, 2001). In contrast, naturalistic intervention models use preexisting routines and child-directed activities (e.g., play) to provide teaching (Dubin & Lieberman-Betz, 2020). Naturalistic developmental behavioral interventions (NDBIs) are a hybrid approach of these frameworks. NDBIs integrate behavior analytic principles into the intervention by introducing a measure of contrivance to child-directed activities or routines (Schriebman et al., 2015).

Use of Play-Based Communication Interventions

In the examination of young children’s communicative behavior, play and social interactions that occur during play warrant additional consideration. In typically developing children, play skills, like social communication skills, tend to develop along a typical trajectory following a predictable timeline (Casby, 2003). From the perspective of social development, children typically advance from solitary play to parallel play, associative play, and cooperative play, reflecting increasingly advanced social skills (Parten, 1932). Cognitively, play is reflective of cognitive development as children gradually advance from engaging in simple body play to pretend play, which is reflective of symbolic thinking, and games with rules (Piaget, 1952; Vygotsky, 2016). Play is the primary medium through which young children explore their world, foster relationships, and learn (Barnett, 1990; Yogman et al., 2018). Moreover, play is a crucial way in which children acquire and practice language skills (Moreno, 2016; Weisberg et al., 2013). This is, in part, because play provides a natural social context wherein children can communicate with their peers or with adults. Play provides a context wherein children can negotiate with each other, describe objects and actions, and build on each other's communication (Levy, 1984).

Naturalistic interventions, such as NDBIs, represent a paradigm of intervention where the interventions are implemented in preexisting routines, such as play; are child directed; and integrate the child's interests and preferences (Dufek et al., 2024; Schreibman et al., 2015). Past reviews on play-based interventions for social communication for children with ASD have indicated that naturalistic interventions and NDBIs can be effective for teaching these skills (Deniz et al., 2024; O’Keeffe & McNally, 2023). These interventions include enhanced milieu teaching (EMT; Kaiser & Hester, 1994); joint attention, symbolic play, engagement, and regulation intervention (JASPER; Kasari et al., 2006); prelinguistic milieu teaching (Yoder & Warren, 1998); and the Early Start Denver Model (ESDM) (Rogers & Dawson, 2010). EMT (Kaiser & Hester, 1994) and JASPER (Kasari et al., 2006) are two child-directed naturalistic interventions that use play to teach target developmental skills: communication is emphasized in EMT, while joint attention and social play are emphasized in JASPER. Naturalistic interventions also appear effective in teaching communication skills to learners with CCN, including IDD and ASD (Gevarter & Zamora, 2018). Additionally, many naturalistic interventions and NDBI explicitly teach play skills or include play-related outcomes as dependent measures (Schreibman et al., 2015; Vivanti & Zhong, 2020). These outcomes may include toy play, symbolic play, measures of the cognitive or social level of play, and related skills. For example, JASPER explicitly teaches joint attention and symbolic play, ESDM teaches sustained engagement and social communication with play (Jhou & Chu, 2022; Rogers et al., 2017), and reciprocal imitation training (RIT) (Ingersoll & Schreibman, 2006) uses imitation to teach functional play skills and expand play actions (Ingersoll, 2010; Meter & Sarah, 2016).

State of the Body of Literature on What AAC Interventions Have Used Play

As social communication is integral to engaging in social forms of play, it stands to reason that integrating play into AAC interventions for children who are MV/NS would enhance communication skills and provide opportunities for social interaction. Play can be a useful context for teaching children to use AAC and can promote further peer interactions (Gevarter et al., 2020, 2022; Kent-Walsh et al., 2015; Olive et al., 2007; Sun et al., 2022). While the use of the play context in social communication interventions for young children with ASD and IDD has been well explored (Deniz et al., 2024; Gibson et al., 2021; O’Keefe & McNally, 2023), the applicability of the play context for AAC interventions is largely underexplored. Perchance the general dearth of literature on the applicability of AAC in the play context is due to the nature of many of past and current AAC interventions and, ergo, the extant research on AAC. The use of highly structured and discrete teaching episodes in the form of DTT, repetitive implementation of teaching episodes with consistent reinforcement of target skills, is a widely adopted practice used in many interventions, including communication interventions, for individuals with ASD and IDD (Dagmawi et al., 2023; Frank-Crawford et al., 2024; Smith, 2001). DTT procedures in communication studies often have trainers provide a preferred item, such as a toy, following a correct response (Smith, 2001). This reflects a broader trend of interventions operating from a traditionally behaviorist framework. Traditional behavioral interventions are those interventions that follow the paradigm set forth by Lovaas (1987) wherein the intervention is delivered in a controlled setting following the three-part contingency of behavior.

Given that play is theorized to have a critical role in the development of social communication skills, and vice versa (see Piaget, 1952; Vygotsky, 2016), the use of play to develop children's social communication skills should be emphasized. Play provides an opportunity for children to be exposed to rich and complex language and creates a context for natural language exchanges between children and their play partners (Levy, 1984). Several play-based and naturalistic intervention models have been used to teach social communication skills to children with IDD.

Reviews have examined play-based interventions for supporting social communication in children with ASD, ages 2 to 13 years, across general (Gibson et al., 2021), school-based (O’Keefe & McNally, 2023), and parent-mediated contexts (Deniz et al., 2024). Definitions of play-based interventions varied, with some including behavioral methods (Gibson et al., 2021) (such as interventions that used DTT or positive reinforcement techniques) and others excluding them (Deniz et al., 2024). Two reviews (Deniz et al., 2024; O’Keefe & McNally, 2023) noted a predominance of naturalistic approaches (see Schreibman et al., 2015). Inclusion criteria ranged from any intervention involving play (Gibson et al., 2021) to those where play was the central component (O’Keefe & McNally, 2023). All reviews included studies judged by reviewers to be play-based, regardless of whether the source articles provided explicit labeling of the intervention as play-based. All reviews assessed social communication outcomes, with two also reporting on child autonomy, highlighting the use of guided approaches (interventions where the intervention agent set limits or where interventions were both child-led and interventionist-led) (Gibson et al., 2021; O’Keefe & McNally, 2023). Findings on effectiveness were mixed, though Deniz et al. (2024) reported a moderate, statistically significant improvement in social communication such as social responsiveness, joint attention, and child interactions with others.

In past years, some AAC intervention research shifted toward what may be considered play-based. This may be in line with the increased recognition and use of NDBIs in early intervention. In past years, play-based NDBIs have been adapted to include AAC use (Bourque & Goldstein, 2020; Olive et al., 2007; Schreibman et al., 2015; Wright et al., 2013). While the use of the play-based and naturalistic interventions is not novel for teaching social communication skills, the adaptation of these interventions for AAC is new. Milieu teaching methods have been combined with modified versions of exchange-based aided AAC (Gevarter et al., 2020). Further, the natural language paradigm has been successfully used to teach SGD use, and an intervention package consisting of EMT combined with the JASPER intervention has been used to teach sign language to autistic children (Wright et al., 2013). Moreover, naturalistic interventions and NDBIs have shown promise in teaching communication skills to children with IDD, including co-occurring ASD (Akemoglu et al., 2020; Kaiser & Roberts, 2013). This review concerns AAC interventions that incorporated play, games, or toys in the delivery of the intervention. Similar to previous reviews, this review also focused on interventions the authors considered to be play-based or were known to incorporate play such as EMT or JASPER.

Purpose and Research Questions

While social communication is well-defined, play, on the other hand, is difficult to define. There is no agreed-upon definition of play (Ebrele, 2014; Smith & Volstedt, 1985). To our knowledge, no review to date has examined the aspects of play that have been integrated into AAC interventions or how they are implemented in the play context. The primary research question is as follows: when examining children with ASD and/or IDD, how has play been used or integrated in AAC interventions, using single-case experimental designs, in studies that examine social communication skills as the primary outcome? This analysis seeks to address the following subquestions:

1. Regarding the characteristics of the literature base for play-integrated interventions to teach social communication skills for MV/NS children using AAC:

   1. What single-case experimental study designs have been employed (e.g., multiple baseline and multiple probe)?
   2. What manualized interventions or intervention procedures have been used, and were play and communication partners mentioned?
   3. What target outcomes (e.g., requests, responses, initiations, tacts, communicative functions, and play skills) have been evaluated in the intervention studies?
   4. What modes of AAC have been taught within play-integrated studies?
2. What elements or variables of play have been integrated into the interventions?

   1. Who were the participants’ play and communication partners, or were the participants engaged in solitary play?
   2. Was a preference assessment employed?
   3. Were play materials or toys selected from materials that were already present, or were they specifically provided for the intervention?
   4. What types of play materials or toys were used?
   5. Did play take place during the delivery of the interventions (i.e., instruction) or only after a child's communicative attempt (i.e., reinforcement)?
   6. What proportion of studies involved implementation of directed play (i.e., interventionist directed) versus emergent play (i.e., child directed)?
   7. What frameworks (i.e., naturalistic vs. behavioral/structured) have these interventions used?

Methodology

The current systematic review is a secondary analysis derived from a much broader primary meta-analysis conducted by Reichle et al. (under review). The primary meta-analysis was preregistered on PROSPERO (CRD42021292027). The primary meta-analysis examined treatment intensity parameters and overall communication outcomes across a wide range of AAC interventions for individuals with ASD and/or IDD. Its focus was not on play, nor on how AAC interventions incorporated play elements. That project included all eligible AAC intervention studies that met What Works Clearinghouse (WWC) design standards, regardless of whether the interventions involved play.

In contrast, the current review represents a distinct and narrower investigation, examining only those studies from the primary meta-analysis that explicitly incorporated play elements into AAC interventions. Furthermore, while the primary meta-analysis used quantitative synthesis methods to evaluate communication outcomes and moderators, the current review is descriptive in nature: it analyzes how play was integrated, what play variables were included, and whether play-based outcomes were measured. These variables were not coded in the primary meta-analysis and required the development of an additional coding framework specific to play. Because questions about play in AAC interventions were outside the scope and analytic aims of the primary project and because the play-related variables required distinct coding procedures, inclusion criteria, and research questions, it was methodologically necessary to present these findings separately. Doing so allows for a more targeted synthesis of the role of play in AAC interventions—an area largely underexplored in the broader AAC literature—and avoids conflating the aims and outcome metrics of the two projects. Given the need to describe processes of both the primary meta-analysis and the secondary analysis, more details can be found regarding the primary analysis search terms, selection criteria, and interrater reliability (IRR) on the Open Science Framework (https://osf.io/r4hgj/).

Procedures: Summary of Primary Meta-Analysis

Primary Meta-Analysis Inclusion Criteria

Articles were included if they (a) reported outcomes of an intervention that included aided and/or unaided AAC, (b) included at least one participant with an IDD and/or ASD, (c) reported results of an independent variable that involved AAC, (d) utilized a single-case experimental design, and (e) was published in English. Inclusion was not limited by year of publication or report status (i.e., gray literature was included).

Primary Meta-Analysis Information Sources and Search Process

For the primary meta-analysis, the literature search was conducted by an experienced research librarian, who also served as Head of the Center for Systematic Reviews and Research Syntheses at Texas A&M University Libraries and who had two decades of experience conducting systematic reviews and meta-analyses in the field of education. To identify relevant studies, a search of the literature was conducted using the following interfaces and databases: EBSCO (ERIC, APA PsycINFO, Academic Search Ultimate, CINAHL); Ovid (Medline); Web of Knowledge (Conference Proceedings Citation Index—Social Science & Humanities and Sciences); and ProQuest (PQDissertations). The searches were conducted in 2021, 2022, 2023, and 2025, along with a forward- and backward-citation search in 2023. Limiters for dates were used, from 1980 to the present. Limiters did not restrict results to peer-reviewed documents, and gray literature was included. Search terms involved the concepts of CCN, autism, and interventions for these learners.

Primary Meta-Analysis Selection Process

Title/Abstract and Full-Text Review Processes

Using the aforementioned inclusion criteria, two special education doctoral students and a professor of special education conducted title/abstract screening, followed by full-text review of remaining articles. Screening for both phases was conducted using Covidence (Veritas Health Innovation, 2021), a web-based collaboration software platform, with the blind mode enabled.

Primary Meta-Analysis WWC Design Standards Review

The remaining documents were evaluated using the WWC Study Review Guidelines 4.1, version 2 (2021), with two coders reviewing 53% of the experiments in the included articles. Coding was conducted using an online cloud-based spreadsheet document with separate documents for each coder to ensure blind coding. Each study was examined for evidence quality for basic design standards. See WWC for details regarding the quality standards (https://ies.ed.gov/ncee/wwc/handbooks).

Primary Meta-Analysis IRR Training and Process

The same coders participated throughout the title/abstract and full-text screening and WWC quality review phases. Both coders received training in blind-coding procedures from two professors of special education and speech-language pathology, who had experience and expertise in conducting meta-analyses and systematic reviews. The coders reviewed and discussed the criteria, followed by practice coding. Training continued until IRR reached 80% for practice coding of the criteria before progressing to independent coding. During the title/abstract screening phase, two coders independently reviewed 100% of the documents, ensuring a double-coding procedure. During the full-text review, two coders individually reviewed and compared codes for over 30% of the documents. Any coding disagreements were reviewed as a group with one or both co-principal investigators (co-PIs) present to discuss and reach a consensus.

During coding of WWC design quality standards, the first coder completed the online training for single-case design standards (4.1 module version) through the Institute of Education Sciences and developed PowerPoint tutorials to train the second coder, under the guidance and supervision of co-PIs. Part of training included the discussion and coding of one article together as a group (coders and co-PIs) for each design type (i.e., multiple baseline and alternating treatment). This was followed by practice coding, during which approximately four to five studies were coded each week until IRR reached 80% for each group of study design type before progressing to independent coding. All phases (i.e., training, practice coding, and independent coding) were completed for all experiments within a design type before proceeding to the next. For instance, the co-PIs oversaw training for the coders, followed by practice coding and then independent coding for all of the multiple-baseline studies before moving to alternating treatment designs and so on. For independent coding, batches of 8–15 articles were assigned to each rater, with 53% of the experiments overlapping. To address disagreements between coders, co-PIs met with the coders to discuss disagreements and reach consensus. This process was consistent across practice and independent coding phases.

Procedures: Current Secondary Systematic Review

Secondary Systematic Review Selection Criteria

Inclusion criteria. Following the guidelines from the primary meta-analysis (Reichle et al., under review), the inclusion criteria noted above (i.e., included outcomes of aided and/or unaided AAC, included at least one participant with an IDD and/or ASD, reported results of an AAC intervention, involved a single-case experimental design, and was published in English) remained the same in this review, with the additions of reporting play outcomes, AAC usage, and communication outcomes. Articles that did not meet any of these criteria were excluded.

Secondary Systematic Review Data Items

The following information was extracted from each article. We coded the data based on the following variables: study design, outcomes measured (communication), type of intervention implemented, level of child control, naturalistic or behavioral framework, type of AAC and mode, and study quality (meeting WWC). We also coded the following play variables: play partner, preference assessment performed, play materials, whether play took place during instruction or delivery of intervention, whether play took place only after the child's communicative attempt or as reinforcement, and whether play materials were provided by the researcher or from materials already possessed by the participant. The code categories and definitions of each descriptor are available in Table A1 in Appendix A.

Secondary Systematic Review Data Collection Process and IRR

During the article inclusion criteria, two doctoral students trained in coding procedures and definitions and practiced coding until a minimum of 80% agreement was reached. Five articles were randomly selected as practice articles for the inclusion criteria. Once coders had achieved the established criterion of five articles above 80% or more, they began to independently code articles for inclusion.

At the article coding level, the same two coders completed the IRR coding procedures. The final 14 articles that were included in this review following the title/abstract review were randomly selected for coding. The criterion was set at 50% of articles at 80% agreement or above. Once the criterion was reached, coders started to independently code each dependent variable for IRR. IRR was completed for 50% (n = 7) articles. A third coder, an assistant professor, then reviewed and compared the data and calculated the IRR results for each dependent variable. IRR was calculated by calculating the number of agreements per variable/total number of variables dual coded × 100.

Results

Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines (Page et al., 2021), a flowchart in Figure 1 summarizes the study selection process for the treatment intensity meta-analysis and play review. Records were identified through database searches and additional sources (n = 2841). Following the removal of duplicates (n = 1010), titles and abstracts were screened (n = 1831) to further exclude studies that did not meet criteria (n = 1652). After exclusion of studies that did not meet design criteria (n = 40) and studies that did not meet title/abstract criteria (n = 182), 71 articles were selected to assess relevance to treatment intensity parameters in AAC and play-based interventions for school-aged learners with autism and IDDs. After further exclusion of two articles due to the absence of usable data (n = 2), full-text articles were then assessed for eligibility based on predefined inclusion and exclusion criteria (n = 69). These final set of studies meeting inclusion criteria was retained for quantitative synthesis in the meta-analysis and qualitative synthesis in the play review. Of these, 14 studies were selected for this study based on our inclusion/exclusion criteria, excluding the others (n = 55). Details can be found regarding the primary analysis selection process on the Open Science Framework (https://osf.io/r4hgj/).

Figure 1.

PRISMA diagram. Source: Page et al. ( 2021). This work is licensed under CC BY 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/

Secondary Systematic Review IRR Results

At the article selection stage, the IRR results showed that 100% of articles were double-coded and agreement was 99%. During the coding stage, IRR results across variables averaged 97.95%.

Study Characteristics

Participant demographic information can be found in Table 1. The characteristics of each included article pertaining to Question 1 are summarized in Table A1. Research Question 1 evaluated the included studies’ characteristics. Regarding the single-case experimental designs implemented, 50% (n = 7) of the studies used multiple baseline designs (Alzrayer et al., 2021; Dogoe et al., 2010; Iacono et al., 1993; Ingersoll et al., 2007; Lorah et al., 2019), 43% (n = 6) used multiple probe designs (Alzrayer et al., 2020; Choi et al., 2010; Gevarter et al., 2020; Kent-Walsh et al., 2015), and 7% (n = 1) used an alternating treatment design (Couper et al., 2014).

Table 1.

Participant Demographic Information.

Author and Year	Number of Participants	Age Range of Participants	Race of Participants	Diagnoses of Participants	Home Language of Participants	Participants’ Previous Experience with AAC
Alzrayer et al., 2020	Three	Five to eight years old	One participant indicated being White, one participant indicated being Black, and one participant did not specify.	One was identified with ASD, and two were identified as having a developmental disability.	Two indicated English as their home language, and one indicated Spanish as their home language.	Not specified
Alzrayer et al., 2021	Three	Four to six years old	Not specified	ASD with inclusion criteria specifying no cooccurring conditions	Not specified	One participant has past experience with sign language.
Choi et al., 2010	Four	Six to nine years old	Two identified as White, and two identified as Black.	Three were identified as having ASD, and one was identified as having a severe developmental disability.	Not specified	Not specified
Couper et al., 2014	Nine	Four to twelve years old	Not specified	All participants were identified as having ASD	Not specified	Not specified
Dogoe et al., 2010	Three	Three to five years old	Not specified	All participants were identified as having ASD, and one participant identified as having a cooccurring intellectual disability.	Not specified	Not specified
Gevarter et al., 2020 1	Five	Three to five years old	Two participants identified as Hispanic, one participant identified as Asian, and two participants were not specified.	All participants identified as having ASD.	English	Not specified
Iacono et al., 1993	Two	Three to four years old	Not specified	Intellectual disability with cooccurring hearing loss	Not specified	One participant had past experience with sign language.
Ingersoll et al., 2007	Five	Two to four years old	Not specified	All participants were described as having ASD.	Not specified	Not specified
Kent-Walsh et al., 2015	Three	Four to six years old	All participants identified as White.	One participant was diagnosed with childhood apraxia, one participant was diagnosed with childhood apraxia with cooccurring developmental delay and sensory processing disorder, and one participant was identified with Down syndrome.	Not specified	Not specified
Law et al., 2018	Two	Two to three years old	Both participants identified as Asian.	Both participants were diagnosed with ASD.	Not specified	Not specified
Lorah et al., 2019	Three	Three to four years old	Not specified	All participants were identified as having ASD.	Not specified	Not specified
Lorah et al., 2019	Five	Three to five years old	Not specified	All participants were diagnosed with ASD.	Not specified	Two participants were identified as using natural gestures, and one participant was identified as using sign language.
Muharib et al., 2019	Three	Six to eight years old	Tw participants identified as Black, and one participant identified as White.	One participant was diagnosed with ASD, and two participants were diagnosed with developmental disability.	Not specified	One participant was identified as having used SGD.
Wright et al., 2013	Four	Twenty three months to 29 months old	All participants identified as White.	All participants were diagnosed with Down syndrome.	English	One participant was identified as having experience with sign language.

Intervention targets were typically focused on improving AAC use for requesting, responses, initiations, labeling, communicative functions, and play skills (100%, n = 14; Couper et al., 2014; Dogoe et al., 2010; Lorah et al., 2019; Muharib et al., 2019). Five studies noted that the play partner was the interventionist (36%, n = 5) while nine studies reported solitary play (64%, n = 9). One study mentioned the parent as the play partner (7%, n = 1), and one study mentioned both parent and interventionist to be play partners (7%, n = 1). One study did not describe the play partner (7%, n = 1). Only one study included data relating to play outcomes (Ingersoll et al., 2007). The included studies involved implementation of the following AAC modes: 71% (n = 10) included high-tech AAC (e.g., SGDs; Alzrayer et al., 2021; Choi et al., 2010; Gevarter et al., 2020; Kent-Walsh et al., 2015; Lorah et al., 2019); 21% (n = 3) of the studies included low-tech AAC (e.g., picture exchange/Picture Exchange Communication System [PECS]; Couper et al., 2014; Dogoe et al., 2010; Gevarter et al., 2020); and 21% (n = 3) included no tech (e.g., gestures and manual sign language; Ingersoll et al., 2007; Law et al., 2018; Wright et al., 2013); percentages add up to more than 100% because some studies used more than one type of AAC mode.

Play Elements and Variables in AAC Interventions

The characteristics of each included article pertaining to Question 2 are summarized in Table 2. Research Question 2 examined how play elements and variables were incorporated into AAC interventions, including directionality, framework, play partners, implementation of a preference assessment, type of play objects used, the timing of play implementation during sessions, and the duration of play during sessions. A notable shift toward child-directed (29%, n = 4; Alzrayer et al., 2021; Gevarter et al., 2020; Law et al., 2018; Wright et al., 2013) and naturalistic (50%, n = 7; Alzrayer et al., 2021; Gevarter et al., 2020; Law et al., 2018; Wright et al., 2013) approaches was observed in more recent studies, particularly those conducted after 2010. A majority of the studies (71%, n = 10) (71%) involved interventionist-led approaches. Some of these studies (14%, n = 2) integrated both child-led and interventionist-led activities in the intervention (Ingersoll et al., 2007; Kent-Walsh et al., 2015), while half of the studies (50%, n = 7) adopted a behavioral framework, and two (14%, n = 2) of these involved both behavioral and naturalistic frameworks (Law et al., 2018; Lorah et al., 2019). Five studies used the naturalistic framework (36%, n = 5). Regarding communicative and play partners, half of the studies (50%, n = 7) involved play with an adult (i.e., interventionist, teacher, and parent), and many of the studies incorporated solitary play as a reinforcement (43%, n = 6), with limited emphasis on cooperative or pretend play. One study did not specify any information on communication and play partners (Alzrayer et al., 2020). Six studies used play as a reinforcement (43%, n = 6), i.e., after the child's communicative attempt.

Whether or not preference assessments were conducted was not stated in the majority of the studies (57%, n = 8). Open-ended questionnaires were used in 29% (n = 4), free-operant assessments in 21% (n = 3), multiple stimuli with replacement in 7% (n = 1), and multiple stimuli without replacement in 7% (n = 1). Toys and materials were frequently used to motivate engagement or elicit specific communication behaviors in many studies. Almost all of the studies included low-tech toys (93%, n = 13), while only one included high-tech toys (Muharib et al., 2019). Most studies (64%, n = 9) did not report whether they used already existing objects as play materials or provided play materials specifically for the study. In 29% of the studies (n = 4), toys were provided by the researchers, and only one (7%, n = 1) study explicitly stated that the toys were already present in the research context (Alzrayer et al., 2020). Regarding timing, play took place during the delivery of the interventions (i.e., instruction) for 57% of the studies (n = 8) and after a child's communicative attempt (i.e., reinforcement) for 43% of studies (n = 6).

Discussion

Findings were limited to implementation of AAC interventions with a play component across the 14 single-case studies included in this review, almost all of which selected a multiple-baseline or multiple-probe design. They all evaluated impacts on AAC use for young children with ASD and/or IDD, mostly those with ASD with or without intellectual disabilities, with primarily high-tech AAC modes implemented and a few with low-tech and unaided AAC. Most evaluated and taught requesting (communication) skills, but only one study explicitly targeted play-based outcomes as a dependent variable (Ingersoll et al., 2007). This anomaly underscores a critical gap in the literature, as most interventions prioritized communication outcomes (e.g., requests and AAC use) rather than assessing play behaviors or the role of play in achieving these outcomes. The limited emphasis on play-based outcomes may reflect a historical bias in intervention research toward measurable, task-specific behaviors rather than broader developmental gains.

A temporal trend was observed, with most of the child-directed and naturalistic AAC interventions emerging after 2010. Only one child-directed study was conducted before this period (Ingersoll et al., 2007), reflecting a historical reliance on structured, adult-led practices. This shift aligns with broader trends in early intervention and special education, emphasizing the developmental and relational benefits of child-centered approaches. Older studies were predominantly implemented in contrived contexts and rooted in behavioral methodologies, such as DTT, which often relegated play to secondary roles or used it as a means to achieve specific communication goals. Structured, adult-directed formats limit opportunities for child-led exploration or spontaneous interaction. This contrasts with more recent interventions adopting NDBIs, which emphasize child agency and the social nature of play.

Half of the interventions involved adult communication partners, and almost half included in this review utilized solitary-play contexts, with limited exploration of cooperative, parallel, or pretend play. This trend is particularly striking given that cooperative and pretend plays are strongly associated with social communication development. The reliance on solitary play may stem from behavioral intervention frameworks, which prioritize individual responses and structured learning over dynamic, peer-mediated interactions (Frank-Crawford et al., 2024). Play types were often restricted to functional or manipulative play, with limited representation of imaginative, symbolic, or cooperative play. This narrow focus may limit the generalizability of findings, as different types of play provide unique opportunities for social communication development.

The researchers reported how they selected reinforcers, such as via a preference assessment, for less than half of the studies, including preferred play items, and most did not report whether the material selected were or were not already present in the participants’ natural settings; use of objects familiar to the children may have resulted in faster acquisition or longer-term maintenance of learned skills. In some studies, objects were preselected by the researchers or practitioners. In others, children chose their materials, that is, a more naturalistic approach. The influence of choice on communication outcomes was rarely discussed, representing another gap in the literature. Further, all but one of the studies (Muharib et al., 2019) involved low-tech toys. Interestingly, the only study to incorporate high-tech toys used a naturalistic framework. The timing and role of play within AAC interventions varied significantly across studies, with some interventions incorporating play before or after communication attempts, while others integrated play within tasks. For instance, while child-directed play was emphasized in naturalistic approaches, structured studies often used play as a reward rather than as an active teaching context. Of note, all studies that used a naturalistic framework or a hybrid framework (Alzrayer et al., 2021; Gevarter et al., 2020¹; Ingersoll et al., 2007; Kent-Walsh et al., 2015; Law et al., 2018; Lorah et al., 2019; Wright et al., 2013) used play during instruction. In contrast, only one of the studies that used a strictly behavioral framework (Iacono et al., 1993) implemented play as part of the instruction process, while the rest used play as a reward.

Limitations

The present study has several limitations. Regarding the evidence base, it must be acknowledged that this review is a secondary analysis of a larger meta-analysis that analyzed the effects of various factors related to treatment intensity in AAC interventions. This review focuses on AAC interventions that involved elements of play. The included evidence base is limited in that it includes a limited number of studies that used play-inclusive frameworks. Moreover, due to this small number of studies, we used a broad definition of what constitutes play-based interventions in the present study, to include those that use play as a reward, solitary play, or toys in the delivery of the intervention. Furthermore, we did not code for the duration of the play activity, and our coding for the point at which play was introduced was evaluated only if play was used as an instruction, reward, or both. Additionally, there are limitations of the review process. First, this analysis included only articles that met WWC standards, with or without reservations, which further limited the articles included. Given the exploratory nature of this analysis, we believe that excluding studies that did not meet these standards limited the findings. Second, this review explored these interventions only for a limited population, including individuals with ASD and IDD, who are MV/NS. Additionally, similar to past reviews on the use of play-based social communication interventions (see Deniz et al., 2024; Gibson et al., 2021), the majority of the participants in the included studies had a primary diagnosis of ASD, while fewer studies focused on children with IDD (Alzrayer et al., 2020; Choi et al., 2010; Iacono et al., 1993; Kent-Walsh et al., 2015; Muharib et al., 2019; Wright et al., 2013). While members of this population are often recipients of AAC interventions, they are not the only recipients of AAC interventions. Third, no risk-of-bias assessment was conducted, although systematic reviews infrequently include them and there is no agreed-upon measure of bias for single-case designs because those that exist for group design are unlikely to be appropriate for single-case designs (Ganz et al., 2023b; Shadish et al., 2015). That said, publication bias was addressed by the inclusion of gray literature in this review. Fourth, we did not extract information from the articles on sensory processing difficulties or other interventions received by included participants, which may have had helpful information as it may contribute toward the selection of AAC mode or play materials used in the interventions.

Research Implications

To our knowledge, this is the first review of the use of play-based or play-integrated interventions to teach the acquisition of communication skills, primarily AAC use outcomes, using AAC systems for children with ASD and/or IDD. Previous research has explored interventions for the acquisition of social communication skills for children with ASD broadly (Gibson et al., 2021), when delivered by parents (Deniz et al., 2024), and in the school setting (O’Keefe & McNally, 2023).

This review represented a departure from previous reviews on play-based social communication interventions for autistic children regarding inclusion criteria. Whereas past reviews excluded studies that used behavioral interventions (Deniz et al., 2024) or required more rigid and theory based definitions of play as part of inclusion criteria (Gibson et al., 2021), this study used a broader definition of play, including studies that explicitly mentioned play in the provision of the intervention, studies with manualized interventions known to be play-based (EMT, JASPER, etc.), and relevant to said expanded definition, studies that incorporated the use of toys, because the focus of this project was on AAC, most of which did not explicitly target play skills. However, it must be noted that definitions of play vary based on discipline; most definitions of play include the following elements: a sense of nonliterality (pretense), intrinsic motivation, engaged in freely chosen activities, flexible nature, and a manifestation of joy (Larsen, 2015; Smith & Vollstedt, 1985). Given that some studies incorporated pretend play and some used play as a reinforcer, we believe that all included studies incorporated some if not all of these elements. Possibly due to these expanded criteria, this review included many studies that used behavioral frameworks (Alzrayer et al., 2020; Choi et al., 2010, Couper et al., 2014; Dogoe et al., 2010; Iacono et al., 1993; Lorah et al., 2019; Muharib et al., 2019). Recent reviews on play-based social communication interventions that included AAC for children with ASD/IDD have not reported the types of framework used in the included studies. Furthermore, this review examined the levels of social play that were incorporated into the interventions following Parten's (1932) theory, one of the most referenced theoretical frameworks for play development. The majority of interventions utilized solitary play as reinforcement. This may be relevant as autistic children often display delays in progressing through the social dimensions of play.

Similar to past reviews (Deniz et al., 2024; Gibson et al., 2021; O’Keefe & McNally, 2023), this review did include studies that utilized NDBIs (e.g., Alzrayer et al., 2021; Gevarter et al., 2020; Wright et al., 2013). By nature of their developmental framework, NDBIs may be well suited to the acquisition of social communication skills, including AAC use. These common elements of NDBIs include natural teaching episodes, a level of child control, and often the use of play (Schreibman et al., 2015), and many NDBIs are play-based or explicitly teach play skills (Vivanti & Zhong, 2020). Notably, the studies that included NDBIs in this review were more recent (Alzrayer et al., 2020; Gevarter et al., 2020) than many of the behaviorally based studies. Similar to how past reviews coded for the level of child control (Gibson et al., 2021), the current review examined child control in the interventions, finding that the studies that incorporated a naturalistic framework often included a child-directed or mixed directed approach as opposed to a strictly interventionist-directed approach (e.g., Law et al., 2018; Wright et al., 2013).

Implications for Practitioners

The findings from this review hold several implications for practitioners. Several studies made use of preference assessments in selecting the target toys or other materials (Alzrayer et al., 2020, 2021; Choi et al., 2010; Couper et al., 2014; Gevarter et al., 2020; Muharib et al., 2019). This indicates that play and play materials are motivating for the child and can be used to increase interest in the interventions. Given the delays and difficulties in learning communication skills, including AAC, in MV/NS children with ASD and IDD, initial instruction in AAC should certainly involve evaluation and selection of materials that are favored by the learner. Likewise, several studies follow naturalistic frameworks that use child choice and child control (Alzrayer et al., 2021; Gevarter et al., 2020; Ingersoll et al., 2007; Kent-Walsh et al., 2015; Law et al., 2018; Wright et al., 2013). This may indicate that by using play as a medium for the delivery of the AAC intervention, interventionists, whether they be researchers, parents, or teachers, can utilize child choice and initiation and grant children control in the intervention, thereby increasing child autonomy (Schuck et al., 2022). Given the need for communication across almost all settings and contexts and challenges in generalization of skills within new settings or contexts, communication partners, and materials, it is particularly important to target AAC instruction across a range of natural activities, especially in play, which is a crucial developmental context for young learners.

Future Research

While this review, to our knowledge, is the first to explore the integration of play within AAC interventions and associated communication outcomes, limitations regarding article inclusion and coding make future work critically important. Future single-case studies involving AAC may benefit from a stronger emphasis on play. Moreover, future reviews should include group design studies alongside single-case studies. In the current review, only one article included play-based outcomes as a dependent variable (Ingersoll et al., 2007). Finally, future work may focus on other populations or broader populations by including individuals who display more vocal communication, occasional vocal communication, or individuals with speech and language impairments associated with other conditions.

Conclusion

In summary, while AAC interventions for children with ASD and IDD who are MV/NS have begun to integrate play elements, the literature remains sparse regarding explicit play-based outcomes. This review underscores the need for more comprehensive research exploring the interplay between AAC and play, particularly in fostering social communication and engagement. While most AAC interventions historically utilized structured, adult-directed methodologies, the review highlighted a gradual shift toward play-based and naturalistic approaches. These newer interventions align with the principles of NDBIs, emphasizing child engagement, social interaction, and play as a medium for learning. Future reviews should explore the incorporation of varying elements of NDBI frameworks.

Appendix

Appendix A

Table A1.

Study Design Variables.

Author and Year	Study Design	Outcome(s) Measured (AAC Use, Communicative Functions, and Play)	Intervention	Level of Child Control	Framework	Type of AAC and Mode	Study Quality (Meeting WWC Standards)
Alzrayer et al., 2020	MPD	SGD and PECS AAC use, multiple symbol requesting, vocal requests	PECS and SGD with modified PECS protocol	Interventionist directed	Behavioral	High tech (SGD)	Meets without reservations
Alzrayer et al., 2021	MBD	AAC use, vocalizations, augmented requests	Natural Language Paradigm + time delay	Child directed	Naturalistic	High-tech (SGD via iPad)	Meets with reservations
Choi et al., 2010	MPD	AAC use, rejecting, re-requesting	Mand training using DTT with progressive time delay and gestural prompting	Interventionist directed	Behavioral	High tech, low tech (SGD, PE)	Meets with reservations
Couper et al., 2014	ATD	AAC use, requesting	DTT with physical and verbal prompting	Interventionist directed	Behavioral	High tech, low tech (PECS, SGD)	Meets without reservations
Dogoe et al., 2010	MBD	AAC use	PECS phases I–III	Interventionist directed	Behavioral	Low tech (PECS)	Meets with reservations
Gevarter et al., 2020 1	MPD	AAC use, number of vocabulary words	Combination intervention package using milieu methods and PECS protocol	Child directed	Naturalistic	High tech (SGD)	Meets without reservations
Iacono et al., 1990	MBD	Unimodal and multimodal AAC use	Behavioral intervention using stimuli, prompts, and models	Interventionist directed	Behavioral	High tech, no tech (SGD, manual sign language)	Meets with reservations
Ingersoll et al., 2007	MBD	AAC use, imitation, play skills	Reciprocal imitation training	Both	Naturalistic	No tech (gestures)	Meets without reservations
Kent-Walsh et al., 2015	MPD	AAC use, answering yes/no	Aided modeling—using concentrated modeling during interactive play context	Both	Naturalistic	High tech (SGD)	Meets with reservations
Law et al., 2018	MBD	AAC use, verbal utterances	Behavior modeling training for parent-mediated NDBI	Child directed	Behavioral and naturalistic	No tech (gesture)	Meets without reservations
Lorah et al., 2019	MBD	AAC use, requesting	Within-stimulus discrimination training with prompt fading	Interventionist directed	Behavioral and naturalistic	High tech (SGD via iPad)	Meets with reservations
Lorah et al., 2019	MBD	AAC use, labeling	DTT with time delay and physical prompting to teach sentence frame discrimination	Interventionist directed	Behavioral	High tech (SGD)	Meets with reservations
Muharib et al., 2019	MPD	AAC use, requesting	Backward training with LTM prompting, time delay, and differential reinforcement	Interventionist directed	Behavioral	High tech (SGD via iPad)	Meets with reservations
Wright et al., 2013	MPD	AAC use	EMT + JASPER to teach sign language	Child directed	Naturalistic	No tech (manual sign language)	Meets with reservations

Note. ADT = alternating treatment design; MBD = multiple baseline design; MPD = multiple probe design. AAC = augmentative and alternative communication; PECS = Picture Exchange Communication System; SGD = speech-generating device.

Table A2.

Play Elements and Variables.

Author and Year	Play Partner	Preference Assessment Performed	Play Materials	Play Took Place During Instruction or Delivery of Intervention	Play Only Took Place After Child's Communicative Attempt or as Reinforcement	Play Materials Provided or Already Present
Alzrayer et al., 2020	Not specified	Free operant	Low tech (e.g., toys and books)	No	Yes	Materials already present
Alzrayer et al., 2021	Teacher	Free operant and open-ended questionnaire	Low tech (e.g., balls, books, blocks, and trains)	Yes	No	Provided
Choi et al., 2010	Solitary	Open-ended questionnaire	Low tech (e.g., marbles, bubbles, and VELCROⓇ darts)	No	Yes	Not stated
Couper et al., 2014	Solitary	Multiple stimuli with replacement	Low tech (e.g., bubbles, trains, and jigsaw puzzles)	No	Yes	Not stated
Dogoe et al., 2010	Solitary	Not stated	Low tech (e.g., fire trucks and trains)	No	Yes	Not stated
Gevarter et al., 2020 1	Interventionist	Free operant and open-ended questionnaire	Low tech (e.g., farms, race car mats, and ocean mats)	Yes	No	Not stated
Iacono et al., 1990	Interventionist	Not stated	Low tech (e.g., toy stoves with cookware and doctor sets)	Yes	No	Provided
Ingersoll et al., 2007	Interventionist	Not stated	Low tech (e.g., dolls)	Yes	No	Not stated
Kent-Walsh et al., 2015	Interventionist	Not stated	Low tech (e.g., Mickey's clubhouse, monorails, and race cars)	Yes	No	Not stated
Law et al., 2018	Parent	Not stated	Low tech (e.g., bubbles, building blocks, and toy cars)	Yes	No	Provided
Lorah et al., 2019	Solitary	Not stated	Low tech (e.g., puzzle)	Yes	No	Not stated
Lorah et al., 2019	Solitary	Not stated	Low tech (e.g., books, balls, and crayons)	No	Yes	Provided
Muharib et al., 2019	Solitary	Multiple stimuli without replacement and open-ended questionnaire	High tech (e.g., iPad video games)	No	Yes	Not stated
Wright et al., 2013	Interventionist	Not stated	Low tech (e.g., balls, cars with ramps, and baby dolls)	Yes	No	Not stated

Note. Low-tech toys refer to toys that are operated without a battery or electricity. High-tech toys refer to toys that are operated with a battery or electricity.

Appendix B

Table B1.

IRR Percentages.

	Directionality	Framework (Behavioral vs. Naturalistic)	Play as DV	Play Materials Technology (This was Two Columns Collapsed into 1)	Play Took Place During Instruction or Delivery of Intervention	Play Only Took Place after Child's Communication Attempt or as Reinforcement	Play Materials Provided vs if Participants Already had Them (Two Columns Collapsed into 1)
Original labels on IOA coding document	1. Who is directing the intervention?	1. Is the intervention behavioral or ABA based?	1. Did this intervention collect data on play skills? Dependent variable	Low tech/high tech	Is the child playing during the instruction portion or… (last column)	Is the child playing during the instruction portion or… (last column)	1. Were play materials chosen from materials the participant already had?
		2. Is the intervention naturalistic/NDBI?					2. Were play materials supplied for this intervention?
Choi et al., 2010	100%	100%	100%	100%	100%	100%	100%
Couper et al., 2014	100%	0%	100%	100%	100%	100%	100%
Gevarter et al., 2020	100%	100%	100%	100%	100%	100%	100%
Iacono et al., 1990	100%	100%	100%	100%	100%	100%	100%
Kent-Walsh et al., 2015	100%	100%	100%	100%	100%	100%	100%
Law et al., 2018	100%	100%	100%	100%	100%	100%	100%
Wright et al., 2013	100%	100%	100%	100%	100%	100%	100%
Totals	100%	85.70%	100%	100%	100%	100%	100%

Note. IOA = interobserver agreement; ABA = applied behavior analysis; NDBI = naturalistic developmental behavioral interventions.