The Effects of Prelinguistic Milieu Teaching Implemented in Classrooms for Preschoolers With or at Risk for Autism Spectrum Disorder

Abstract

Much research exists supporting the efficacy of naturalistic behavioral interventions on increasing social communication skills for children with autism spectrum disorder (ASD); however, these evidence-based interventions are not consistently utilized in preschool classrooms. Prelinguistic Milieu Teaching was used to teach early intentional communication (i.e., purposeful and coordinated use of vocalizations, gestures, and eye contact) to three preschool students with or at risk for ASD. The present study extends prior research demonstrating the effects of PMT in increasing intentional communication through implementation in a preschool special education classroom, measurement of collateral gains related to PMT targets, and measurement of maintenance and generalization of gains. Results indicate students increased their rates of intentional communication upon introduction of PMT. These gains maintained over time for two students. Present study results have implications for future research and practice regarding the efficacy and feasibility of implementing PMT in preschool classrooms.

Keywords

autism spectrum disorder intentional communication prelinguistic milieu teaching

Children with autism spectrum disorder (ASD) exhibit deficits in several behaviors that comprise prelinguistic intentional communication (IC), including nonword vocalizations (e.g., Plumb & Wetherby, 2013), communicative gestures (e.g., Shumway & Wetherby, 2009), and eye gaze (e.g., Wetherby et al., 2007). Prelinguistic communicative behaviors are early developing skills that are impacted in ASD. Research exists supporting the relation of prelinguistic IC (i.e., the purposeful and coordinated use of vocalizations, gestures, and eye contact to communicate) to several developmental outcomes, including spoken language (e.g., Toth et al., 2006) and play (e.g., Yoder & Stone, 2006).

Early Social Communication Interventions for Children With ASD

Limitations of “highly structured teaching approaches” Schreibman et al. (2015, p. 2414) for young children with ASD and the movement toward providing instruction in natural settings (e.g., home, school) have prompted a shift toward more naturalistic and developmental interventions that incorporate behavioral learning principles (Schreibman et al., 2015; Sowden et al., 2011). Prelinguistic Milieu Teaching (PMT; Yoder & Warren, 1998) is a manualized intervention that combines behavioral principles and naturalistic strategies to target early developing behaviors (e.g., eye gaze, gestures, vocalizations) used to request and share attention. Initial investigations suggested PMT led to improvements in IC for the purpose of requesting in children with developmental disabilities (Warren et al., 1993; Yoder et al., 1994). A randomized controlled trial resulted in effects on IC that were greater for children assigned to PMT as compared with children assigned to Responsive Small Group, in which a trainer played alongside participants and responded to any communicative acts without using demands or prompts to elicit child responses. These findings were moderated by high maternal responsivity (Yoder & Warren, 1998) leading to the addition of a responsivity education component (RE) in subsequent investigations (Fey et al., 2006, 2013; Yoder & Stone, 2006; Yoder & Warren, 2002). Evidence suggests PMT’s impact spans beyond targeted prelinguistic social communication outcomes, as improvements following PMT have been observed for spoken language (e.g., Fey et al., 2013; Yoder & Stone, 2006) and variables related to play (McDuffie et al., 2012).

Limited support exists for the efficacy of PMT implemented in school settings. McCathren (2000) investigated a teacher’s ability to implement PMT and effects on IC for a 3-year-old student with a developmental disability. Results suggested the teacher-implemented PMT with fidelity, with associated child gains in IC. Further investigation of PMT in the classroom is warranted. Despite legal mandates, research suggests few teachers use evidence-based instruction with students with ASD (Morrier et al., 2011) and core deficits are infrequently targets of instruction (Wong & Kasari, 2012). To better promote the use of research-supported interventions in schools, investigations should be conducted in settings where the intervention will be used (Kasari & Smith, 2013). This study examined the following: (1) Does PMT implemented in a classroom increase the use of IC in preschoolers demonstrating ASD symptomatology? (2) Do increases in targeted behaviors generalize to other people and activities and maintain over time? (3) Do spoken language and object interest change following intervention? and (4) Do changes in collateral behaviors generalize to other people and activities and maintain over time?

Method

Participants

Three preschool-aged students and their teachers were recruited from two classrooms in the southeastern United States. Inclusion criteria for students included (a) child age between 3 and 6 years; (b) eligibility for special education based on state requirements (i.e., child is identified as a child with a disability, negatively impacting educational performance); (c) fewer than 10 spontaneous communicative words produced during preintervention assessments; (d) Individualized Education Program (IEP) objectives related to functional communication; and (e) no known hearing or uncorrected vision impairment. In addition, participants were identified as demonstrating characteristics of ASD as reported by the child’s teacher on the Gilliam Autism Rating Scale, Third Edition (GARS-3; Gilliam, 2013). The Childhood Autism Rating Scale, Second Edition (CARS-2-ST [Standard edition]; Schopler et al., 2010) was also administered by the first author. CARS-2-ST scores in the mild-moderate or severe range of symptoms of ASD indicated confirmation of GARS-3 scores. To be included, teachers were required to have a student enrolled in the study and consent to participate. Pseudonyms are used to identify all participants.

The Mullen Scales of Early Learning (MSEL; Mullen, 1995), an assessment measuring cognitive and motor development in children birth to 68 months was administered prior to probes (see Table 1). A 10-min classroom play observation was also conducted prior to the probe condition. The play sample was videotaped and scored to provide descriptions of the child’s play and other behaviors during unstructured play activities in the classroom (e.g., toys used, social interactions, unprompted spoken language). Information collected via the play sample was used to characterize participants, confirm inclusion criteria, and plan intervention goals.

Table 1.

Student Participant Descriptions.

Child	Age	Gender	Race/ethnicity	Eligibility	Mullen (AE in months)	CARS-2 total	GARS-3 Autism Index
Felicia	4;5	Female	White	SDD, SLI	VR = 10 months; FM = 14 months; RL = 8 months; EL = 6 months	47.5	4
Michael	4;3	Male	African American	SDD, SLI	VR = 26 months; FM = 18 months; RL = 6 months; EL = 13 months	37	6
Justin	4;1	Male	White	Autism, SLI	VR = 29 months; FM = 24 months; RL = 34 months; EL = 6 months	43.5	4

Note. AE = age equivalent; CARS-2 = Childhood Autism Rating Scale (2nd ed.); GARS-3 = Gilliam Autism Rating Scale (3rd ed.); SDD = significant developmental delay; SLI = speech–language impairment; VR = visual reception; FM = fine motor; RL = receptive language; EL = expressive language.

Child participants

Child demographic information is presented in Table 1. Justin attended a private, inclusive preschool program that served typically developing children as well as children with developmental disorders. Justin was eligible for special education in the categories of autism and speech–language impairment (SLI). State eligibility criteria for autism requires that, at minimum, documentation shows that the child is experiencing developmental delays, difficulties in social interaction and participation, and deficits in the use of verbal language for social communication. Between the ages of 2 and 3 years, Justin began receiving speech, occupational therapy, and applied behavior analysis services, all of which were continued over the course of the present study. According to parent and teacher report, he did not use spoken language; however, he was learning to communicate using an iPad. During the observation prior to data collection, Justin walked around the classroom and gathered toys. He did not initiate communication toward his teachers or peers nor did he engage in functional play.

Felicia and Michael were students in a public, self-contained special education preschool classroom. Both students were eligible for special education under the categories of significant developmental delay and SLI. Felicia’s parents reported that she communicated using spoken language during early development but experienced a regression around 18 months of age and did not use any spoken language or gestures to communicate. She had received early intervention services, speech therapy, and occupational therapy between the ages of 17 and 36 months but she was no longer receiving services outside of the classroom. She engaged in relatively high rates of eye contact with the teacher and teaching assistants and occasionally pointed at items in the classroom during an initial observation; however, the function of her communication was typically unclear. Felicia engaged in high rates of vocal stereotypy and exhibited little interest in toys. Michael received early intervention services, speech services, and occupational therapy services between the ages of 2 and 4 years. He continued to receive speech and occupational therapy services for the duration of the study. Although parents reported he had a vocabulary of 15 spoken words, he used fewer than 10 spontaneous, functional words across preprobe assessments. Michael infrequently labeled items using spoken language in the classroom; he did not integrate words with other communicative behaviors, nor did he request or use other forms of functional communication.

Teacher participants

Ms. Brown taught in the public self-contained preschool special education classroom attended by Michael and Felicia. She held a master’s degree in special education and had 3 years of preschool special education teaching experience. Mr. Jones was a high school graduate with just under 1 year of experience working as a paraprofessional in the private, inclusive preschool classroom that Justin attended. Mr. Jones was selected for participation in the study because Justin spent more one-on-one time with him than with the teacher. Ms. Brown and Mr. Jones both reported experience teaching children with ASD and communication delays as well as the use of instructional strategies consistent with PMT.

Setting and Materials

Sessions took place in a 1:1 format in an area of the child’s classroom separate from other students. Information obtained through teacher report and initial observations informed selection of preferred toys used during probe, maintenance, and PMT sessions. The primary investigator provided two sets of toys that were rotated across sessions to decrease the likelihood of satiation and increase consistency across different participant’s sessions. Each set included toys that might encourage child commenting (e.g., toys that light up), at least one highly preferred toy for each participant, and pairs of similar toys used to imitate the child’s play (see Table 2). A variety of classroom toys were used during generalization sessions to provide a representative sample of student–teacher interactions. Additional assessments took place in a small room in the child’s school using novel toys unavailable in the child’s classroom or during intervention sessions.

Table 2.

PMT Session Routines and Activities.

Participant	Routine/activity
Felicia	Pin toy: put pin toy on student’s hands and arms; removed pin toy and paused to evoke communication. Trampoline: said “123 jump!” and held student’s hands as she jumped on a trampoline; paused and let go to evoke communication. Snack: provided preferred snacks in a closed, transparent container upon communication. Playdoh: took turns playing with playdoh; blocked access to evoke communication. Bouncing: held student’s hands as she bounced on a ball; paused to evoke communication. Brush: lightly brushed student’s hair or pressed a brush onto her hands and arms; paused to evoke communication Spinning lights: blocked access to switch to evoke communication.
Michael	Pretend food: took turns feeding and pretending to eat food; blocked access to food to evoke communication. Number puzzles: took turns removing and inserting puzzle pieces; blocked access to pieces or puzzle to evoke communication. Bouncy balls: took turns pushing a button to make balls in a toy bounce; blocked access to button to evoke communication. Pin toy: placed foam numbers on a pin toy and paused for child to communicate prior to providing access.
Justin	Books: read story and paused to evoke communication before continuing. iPad: took turns playing various games. Shape sorter: took turns placing shapes into the sorter. Number and letter puzzles: took turns removing and inserting puzzle pieces; blocked access to pieces or puzzle to evoke communication. Spinning lights: blocked access to switch to evoke communication. Stuffed animals: chased student around classroom with stuffed animals; paused and withheld stuffed animals to evoke communication. Dump truck: took turns putting items into a dump truck and dumping them out; blocked access to truck and/or crashed into truck to evoke communication.

Response Definitions and Direct Observation Recording Procedures

Target behaviors included unprompted IC and two-component pre-IC acts. Each occurrence of the target behavior was coded during 1-min intervals (minute-by-minute event recording) from video recorded sessions using INTERACT software (Mangold, 2015). IC was defined as any communicative act containing a gesture or vocalization combined with coordinated attention to an object and person or use of spoken words used for the purposes of requesting or initiating joint attention (IJA). Coordinated attention occurred when the child directed their attention to the adult and to an object/event of interest, such as through a gaze shift. IC would be coded if a child pointed to a toy high on a shelf, looked at the adult, and vocalized “ah ah!” IC would not be coded if the child looked at the toy and vocalized “ah ah,” but did not look to the adult.

Because increases in developmental variables for children with significant delays may not occur immediately upon introduction of the independent variable, impacting inferences of a functional relation (Lieberman et al., 2010), a variable sensitive to incremental changes in participants’ communicative behaviors was created. Two-component pre-IC acts (i.e., combined use of two communicative behaviors that did not clearly demonstrate coordinated attention to an object and person) were also coded and incorporated into intervention decisions by aggregating with the number of IC acts to create a single variable sensitive to change. A pre-IC act would be coded if the child pointed to a picture in a book and vocalized without showing attention to the adult through eye contact or a gaze shift. A pre-IC act would not be coded if the child touched the picture without vocalizing or showing attention to the adult. An aggregate of the rate of unprompted IC acts and two-component pre-IC acts, referred to as child communicative acts, was the primary variable used to make decisions regarding introduction of PMT to participants and mastery criteria.

All IC acts, components of IC acts, and collateral behaviors (i.e., spoken language [spontaneous, intelligible spoken words not immediately preceded by an adult model] and object interest [number of different toys the child played with using non-imitative, differentiated actions]) and the time at which they occurred were coded during probe, intervention, generalization, and maintenance conditions. During probe, generalization, and maintenance conditions, only unprompted responses were recorded. During intervention sessions, an average of one teaching episode occurred per minute. As such, PMT targets were recorded as prompted or unprompted. A prompted response included any IC act or component performed by the child within 3 s of any adult prompt (e.g., verbal, model, gesture, physical). An unprompted response included any IC act or component that was not preceded within 3 s by an adult prompt. The same procedures used during probe, generalization, and maintenance sessions were used to code object interest during PMT sessions.

Experimental Design and Procedures

Four conditions (probe, intervention, generalization, and maintenance) were implemented within the context of a multiple probe across participants design (Ledford & Gast, 2018). Repeated measurement in the probe condition and staggered intervention introduction across participants provided intersubject replication of effect, allowing for detection of threats to internal validity. The primary investigator conducted all pre- and postintervention assessments and maintenance sessions. Probe sessions and PMT instructional sessions were conducted by the primary investigator twice per week. Two special education master’s students with experience teaching communication to young children with disabilities were trained in PMT by the primary investigator to serve as secondary clinicians. Each secondary clinician conducted PMT sessions once per week to promote across-person generalization. Across conditions, a research assistant video recorded all sessions for data collection purposes. This study was approved by the university’s institutional review board.

General procedures

Probe, intervention, and maintenance sessions were 20 min each and conducted with the primary investigator or secondary clinician. Generalization sessions were 10 min each and conducted with the teacher. Data were collected concurrently during the probe condition for all participants, during which time generalization was measured at least once per participant. PMT intervention was introduced for the first participant once data were stable in the probe condition. Probe data for Participants 2 and 3 were slightly variable, but did not show a trend in the therapeutic direction when the intervention was introduced to Participant 1. Once Participant 1 met predetermined criteria (i.e., visible and stable increase in level or trend of intentional communicative acts), intervention began for the second participant following stable probe responding for at least three consecutive data points. Intervention was introduced to the remaining participant in the same manner. Generalization was measured twice per participant during intervention. Maintenance data were collected 3 weeks after intervention completion.

Probe condition

Probe sessions occurred in the classroom with the primary investigator, with at least one probe taking place with a secondary clinician and a generalization session taking place with the child’s teacher. During probe sessions, the clinician engaged in play with the child and several preferred toys available in the classroom. Communication was not directly prompted or encouraged. The clinician maintained engagement with the child through use of strategies such as imitating the child’s play and responding contingently to the child’s initiations.

Intervention

Each child participated in two intervention sessions per week with the primary investigator, and one session per week with one of the two secondary clinicians. PMT sessions were conducted 1:1 3 days per week for 6 weeks using activities embedded in play. During intervention sessions, milieu teaching strategies were used to teach prelinguistic communicative behaviors. Environmental arrangement was used to create opportunities for communication (e.g., requesting an unreachable toy). Following the child’s lead (e.g., providing natural consequences for the child’s requests) and social routines (e.g., pushing a toy back and forth) were employed to maintain child engagement and provide contexts for communication during highly motivating interactions. Linguistic mapping was used (i.e., putting words to the child’s nonverbal communication; Fey et al., 2017), which may support incidental language acquisition. For example, if a child pointed to a toy car out of reach and looked to the adult, the adult would say “want car,” and retrieve the toy for the child.

The primary investigator and secondary clinicians targeted specific child IC behaviors through a series of teaching episodes within each session. During PMT sessions, the clinician followed the child’s lead to determine interest in an activity or toy. For each material available, a specific routine was created to elicit communication (see Table 2). When children demonstrated interest in a target toy/activity, the clinician initiated the planned social routine, and implemented teaching episodes targeting IC behaviors. Teaching episodes consisted of a hierarchy of prompts, including time delay (i.e., waiting 3 s for a child to perform an IC act, verbally directing the child to perform an IC act, modeling the target act, and physically guiding a child to perform an IC act when appropriate), which were systematically applied to teach IC acts. Initially, prompts were used to teach IC for the purpose of requesting. Upon consistent, independent child requesting, the clinician modeled IJA using gestures the child had been using for the function of requesting. Teaching episodes were implemented approximately once per minute during intervention sessions (Fey et al., 2013). Clinicians moved to a new toy or activity when children expressed interest through picking up or actively engaging with that toy. If children remained unengaged, clinicians directed children’s attention to a target toy/activity. Mastery criterion for successful completion of PMT was two consecutive sessions during which a child independently initiated two IC acts per minute, as children exhibiting this rate of IC are considered ready for interventions targeting linguistic communication. Over the course of the 6-week intervention, at least one generalization session was conducted per month to measure each child’s use of social communication behaviors taught during PMT using only classroom materials with the teacher or paraprofessional. More details regarding PMT procedures can be found in Fey et al. (2017).

Maintenance

Three weeks after PMT completion, Justin and Michael received an additional PMT session with the primary clinician to collect follow-up data. Generalization data were collected for Michael during maintenance using generalization session procedures. The second teacher was unavailable 3 weeks post-PMT, which precluded measurement of generalization during follow-up for the other participant. Maintenance data were not collected for Felicia due to the school year ending.

Interobserver Agreement and Procedural Fidelity

Interobserver agreement (IOA) was calculated for each child and teacher behavior. The primary investigator served as the primary coder and trained two research assistants, one of whom was blind to study hypotheses, to 75% reliability criterion on the coding system, given the complexity of the data collection system (Cooper et al., 2007). The primary and secondary coders independently coded 33% of Justin’s probe sessions, 29% of Michael’s probe sessions, and 25% of Felicia’s probe sessions. IOA for the intervention condition was coded for 22% of Justin’s and Felicia’s sessions, and 25% of Michael’s intervention sessions. In addition, IOA was coded for 25% to 33% of generalization sessions randomly selected for each participant. If IOA fell below 75% for two consecutive sessions, the secondary coder was provided with refresher training on the coding system.

IOA for rate of target behaviors was calculated using the exact count-per-interval method, in which the number of agreements (i.e., instances in which both observers recorded the same frequency of behavior per minute interval) were divided by number of agreements plus disagreements and multiplied by 100. During the probe condition, IOA for communicative acts ranged from 75% to 97% with a mean agreement of 84%. IOA was also adequate for collateral gains during the probe condition (spoken language IOA range 90%–100%, M = 97%; object interest IOA range 80%–100%, [M = 94%]). Average agreement for communicative acts during treatment was 79% (70%–95%), 92% (80%–100%) for object interest, and 92% (53%–100%) for spoken language across participants. Potential factors contributing to the lower IOA include coding behaviors such as gaze shifts and gestures in authentic classroom contexts where participants were free to move around and engage with a variety of materials. For all generalization sessions, average agreement ranged from 75% to 100% across teacher variables. Average agreement during generalization was 93% (range = 90%–95%) for communicative acts, 100% for spoken language, and 92% (75%–100%) for object interest.

Procedural fidelity data were collected to demonstrate evidence of adherence and differentiation across conditions (Ledford & Wolery, 2013). Data were collected by the same independent observers for the probe and intervention sessions used for reliability coding, plus additional sessions to ensure fidelity was assessed for at least 20% of sessions conducted by each clinician. Thirty-five percent of all probe sessions were coded for fidelity. In the probe condition, clinician engagement with the child and/or toys was measured using 15-s momentary time sampling due to the continuous nature of the behavior (Bakeman & Quera, 2011; Lane & Ledford, 2014). Event recording was used to measure prompts for communication, linguistic mapping, initiating routines, and environmental arrangement. Correct probe procedures required the clinician to engage with the child, but to refrain from implementing procedures congruent with PMT (i.e., prompting, linguistic mapping, initiating routines, and environmental arrangement). The percentage of intervals with correctly implemented probe procedures ranged from 97% to 100% across all interventionists and participants. In addition, no components of PMT were observed during probe sessions coded for procedural fidelity, indicating evidence of differentiation between conditions.

Procedural fidelity for PMT sessions was collected on accurate implementation of teaching episodes (Fey et al., 2006). Each teaching episode was coded for appropriate use of prompting, linguistic mapping, and providing the natural consequence of the child’s communicative act. Procedural fidelity was calculated for 22% to 28% of PMT sessions for each participant. The percentage of correctly implemented teaching episodes ranged from 85% to 100% (M = 90%) across interventionists. Prompting hierarchies were used appropriately in 99% of teaching episodes. High percentages of correctly implemented linguistic mapping (M = 98%) and provision of natural consequences (M = 100%) were also observed across teaching episodes.

Information on clinician use of various PMT strategies (e.g., initiation of routines, environmental arrangement, following the child’s lead) was collected. Clinicians followed the child’s lead for an average of 97% of each session (range = 92%–100%) and attempted to initiate an average of 18 routines per session (range = 3–31). On average during sessions, clinicians linguistically mapped child communicative acts 11 times (range = 1–19), provided natural consequences 23 times (range = 1–73), and arranged the environment to promote communication 9 times (range = 1–21). As there is no recommended amount of PMT strategies that should be used in each session, information about clinician strategy use served as a standard with which to compare teacher strategy use.

Social Validity

Social validity was measured in three ways. First, children’s IEP objectives were reviewed to see how many were targeted through PMT. Second, one of the teachers completed a survey postintervention assessing her attitudes toward PMT (e.g., effectiveness, intrusiveness in the classroom). Finally, teacher use of PMT strategies was assessed via direct observation across probe and intervention conditions to assess incidental acquisition of strategies, which will inform feasibility of training teachers to implement PMT. Teacher behaviors collected as part of social validity measures were operationally defined based on behaviors compatible (i.e., environmental arrangement, linguistic mapping, following the child’s lead, use of prompts, and initiation of routines) and incompatible (i.e., directing child) with PMT strategies. Data on following the child’s lead were collected using 15-s interval momentary time sampling. All other teacher behaviors were coded using minute-by-minute event recording. Teacher behavior data were collected during generalization sessions across all conditions. Rate of teacher behaviors during probe, intervention, and maintenance conditions was aggregated to compare strategy use across conditions, as the number of generalization sessions varied across children.

Results

Data collected on target child behaviors during baseline, intervention, and maintenance sessions were analyzed using visual analysis. Specifically, the level, trend, and stability of data points for each target behavior were compared within and across conditions. Identification of a functional relation was ascertained by examining stability of baseline and absence of trend in the therapeutic direction within and across participants, with an increase in level of the dependent variable (DV) upon introduction of PMT across all three participants.

Intentional Communication

During the probe condition, Justin’s rate of communicative acts per minute ranged from 0.05 to 0.07 (M = 0.06). Upon introduction of PMT, Justin exhibited an immediate albeit small increase in rate of communicative acts (M = 1.15, range = 0.10–2.50). His engagement in communicative acts was highly variable across sessions. Justin’s rate of engagement in communicative acts decreased slightly at follow-up; however, it remained higher than rates observed during probe. Justin’s rate of communicative acts during the generalization probe was similar to that observed during probe sessions (0.10 per minute). His rate of engagement in communicative acts increased during PMT generalization sessions (M = 1.9, range = 1.30–2.50). Michael’s rate of communicative acts during the probe condition was variable yet below mastery criteria across all sessions. He engaged in an average rate of 0.74 communicative acts per minute (range = 0.59–0.90). Michael’s rate of communicative acts sharply increased immediately upon introduction of PMT and remained high across the entire condition (M = 1.94, range = 1.15–2.80). Michael met mastery criteria by the 12th session. At follow-up, Michael continued to demonstrate a mastery level rate of communicative acts. Generalization across individuals as represented by Michael’s rate of communication during sessions with his teacher was not observed. Felicia exhibited an average rate of 0.23 communicative acts per minute (range = 0.10–0.76) during the probe condition. There was a slight decreasing trend in communicative acts initially across probe sessions, which stabilized during the five sessions conducted prior to PMT implementation. Upon introduction of PMT, Felicia’s rate of communicative acts per minute immediately increased and continued to slowly increase across the majority of sessions (M = 1.20, range = 0.42–2.06). Felicia’s improvement did not generalize across individuals (see Figure 1).

Figure 1.

Participant rate of independent communicative acts across conditions.

Collateral Gains

Improvements in communication and play variables not directly targeted were assessed via repeated measurement across conditions (see Figure 2). The average rate of spoken language during the probe condition ranged from 0.00 (Justin and Felicia) to 0.75 (Michael). Michael exhibited an immediate yet small increase in spoken language upon PMT introduction, averaging 1.49 spoken words per minute during the PMT condition and continued to engage in high rates of spoken language at follow-up. Justin engaged in some instances of spoken language during two PMT sessions; otherwise both his and Felicia’s rate of spoken language remained at probe condition levels throughout the study. The number of objects with which students engaged was measured during probe and PMT sessions. Average object interest ranged from approximately 4 (Michael and Felicia) to 6 (Justin) during the probe condition. All participants exhibited levels of object interest similar to the probe sessions during PMT implementation.

Figure 2.

Participant collateral gains across conditions.

Social Validity

All students had IEP goals consistent with PMT targets. Several of the student’s objectives were addressed during PMT sessions (e.g., request highly preferred items using verbalizations and/or gestures; communicate for the purpose of commenting). In addition to reviewing children’s IEP goals, Michael and Felicia’s teacher filled out a social validity survey following the intervention. Ms. Brown reported that PMT sessions did not interfere with typical classroom activities, she wanted to use the strategies again, and other teachers would benefit from using PMT. Ms. Brown reported she believed PMT improved Michael and Felicia’s social communicative behaviors, but that Felicia appeared to benefit more from PMT than Michael.

Teacher behaviors were measured during generalization sessions and aggregated across conditions to provide information about teacher behaviors consistent and inconsistent with PMT. Ms. Brown and Mr. Jones followed the students’ leads for the majority of each session (average = 89.4% and 81.67%, respectively); however, they also engaged in high levels of directing students’ behavior (average = 13.3 and 24.7 times per session, respectively). Both teachers prompted communication several times per session (average = 38.7 and 12.67 times, respectively) and attempted to engage students in routines (average = 4.7 and 1.7 times per session, respectively). Teachers engaged in minimal environmental arrangement (average = 2.9 and 1.7 times per session, respectively), linguistic mapping (average = 1.6 and .3 times per session, respectively), and provision of natural consequences (average = 3.1 and 1.7 times per session, respectively).

Discussion

The current study examined the efficacy and feasibility of using PMT with preschoolers with or at risk for ASD in their classrooms. All three participants increased their engagement in communicative acts during PMT. Michael’s rate of IC immediately increased upon PMT introduction and he reached mastery criteria within 12 sessions. Justin and Felicia’s improvements were more gradual. Both exhibited small increases in their communicative acts when they began PMT; however, they required several sessions before exhibiting relatively consistent higher rates of IC. A secondary goal of the current study was to examine collateral behaviors that may have improved with PMT implementation. Aside from Michael, who began the study with some spoken language, students did not exhibit increases in spoken language over the course of the study. It is possible that a certain level of prelinguistic communication must be achieved prior to observing collateral improvements in linguistic communication, as children in previous research demonstrating such gains typically exhibited more postintervention IC than exhibited by Justin and Felicia (e.g., Whalen et al., 2006). All three children exhibited varying levels of object interest throughout probe and PMT sessions, suggesting PMT had no impact on the number of objects with which the children played. Importantly, sessions with higher object interest were frequently those in which it was difficult to engage the student with a given object for enough time to create routines and subsequent opportunities to communicate. In contrast, low levels of object interest within a given session often reflected sustained interest in few objects, enabling more opportunities to communicate and therefore higher rates of communication. Perhaps more informative indicators of overall gains in play would include sustained engagement with objects and variety of actions used.

Clinical Implications

Results provide preliminary support for the use of PMT to teach prelinguistic communicative behaviors to preschoolers with symptoms of ASD in the classroom. Benefits to implementation in the classroom included the availability of a wide array of toys with which the students were familiar and may have had preexisting routines or preferred ways of playing, thus creating opportunities for requesting. Similarly, the importance of spending time creating routines at the onset of PMT and conducting frequent preference assessments throughout PMT implementation to ensure student motivation and the opportunity to engage in IC cannot be overemphasized. Teachers spend a significant amount of time with the students and likely have more insight into the students’ preferences than a clinician who may only see them a few times per week. As such, students may respond more quickly to PMT implemented by teachers, who can apply PMT strategies within routines that are already being used in the classroom. These potential benefits of classroom-based, teacher-implemented PMT paired with preliminary evidence of at least some degree of teacher engagement in strategies congruent with PMT provide support for training teachers in PMT.

Possible drawbacks to implementation in a classroom include the introduction of variables that may have been implicated in the participants’ variability in responding, including distractions and interruptions to the typical schedule. Although it is not possible to analyze statistically within the current study design, variable responding may also be attributable to preintervention characteristics (e.g., rate of IJA), which has been the case in previous research conducted in other settings (e.g., Yoder & Stone, 2006). Information about preintervention characteristics, including the student’s tolerance of change and distractibility, may be helpful in adapting the intervention to individual students to increase likelihood of responding (e.g., conducting shorter, more frequent sessions; conducting sessions during times when the student is less likely to be distracted).

Information about challenging behaviors exhibited by students may be helpful for intervention planning. It may be useful to conduct a functional behavior assessment prior to PMT implementation to inform additional strategies that may be necessary during intervention sessions (e.g., use of extinction; reducing time delay before prompting to ensure use of IC is easier than engagement in challenging behavior). A previous study examining the effectiveness of milieu therapy combined with functional communication training conducted functional analyses as part of preintervention assessment (Mancil et al., 2009). Mancil and colleagues found positive intervention effects for concurrent acquisition of communication skills and reduction of challenging behaviors.

Limitations and Directions for Future Research

Although positive effects were observed for all participants, changes in the dependent variable were delayed for two participants, so the possibility that improvements were due to factors outside of the intervention cannot be excluded. However, delayed responding has been reported in prior PMT research (Yoder et al., 1994) and may relate to the need to measure other behaviors required to teach IC within the PMT framework (e.g., play skills and routines; individual components of IC). In addition, maintenance sessions were limited to two participants, and procedural fidelity data were not collected during maintenance; interpretation of these data should be made with caution. Another limitation of the present study involves the individuals implementing the intervention. Implementation by trained graduate students rather than natural implementers (i.e., teachers) limits the generality of findings. Future studies where teachers are trained to implement PMT with their students are necessary to better understand the effects of classroom-implemented PMT on the IC of preschoolers with ASD. Finally, the current study cannot be classified as providing strong evidence of PMT effects according to the What Works Clearinghouse (WWC) evidence standards because the probe condition for one of the participants only contained three data points. Despite limitations, the present study provides preliminary evidence for the efficacy and feasibility of implementing PMT in a preschool classroom. These findings can be used to inform future research on PMT implemented in the school environment, including training teachers to implement PMT and individualizing intervention components based on preintervention child characteristics.

Footnotes

Authors’ Note

A.H.D. is now with Nemours/Alfred I. DuPont Hospital for Children, Division of Behavioral Health, Department of Pedia-trics, Delaware; A.Z. is now with Cherokee Health Systems, Tennessee.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported through funding from the Dissertation Grant Award Program of the Society for the Study of School Psychology.

References

Bakeman

Quera

(2011). Sequential analysis and observational methods for the behavioral sciences. Cambridge University Press.

Cooper

J. O.

Heron

T. E.

Heward

W. L.

(2007). Applied behavior analysis. Pearson Prentice Hall.

Fey

M. E.

Warren

S. F.

Bredin-Oja

S. L.

Yoder

P. J.

(2017). Responsivity education/prelinguistic milieu teaching. In McCauley

R. J.

Few

M. E.

(Eds.), Treatment of language disorders in children (2nd ed.) (pp. 57–84). Brookes Publishing.

Fey

M. E.

Warren

S. F.

Fairchild

Sokol

Yoder

P. J.

(2006). Early effects of responsivity education/prelinguistic milieu teaching for children with developmental delays and their parents. Journal of Speech, Language, and Hearing Research, 49, 526–547.

Fey

M. E.

Yoder

P. J.

Warren

S. F.

Bredin-Oja

S. L.

(2013). Is more better? Milieu communication teaching in toddlers with intellectual disabilities. Journal of Speech, Language, and Hearing Research, 56, 679–693.

Gilliam

(2013). Gilliam Autism Rating Scale 3rd edition manual. Pearson Assessments.

Kasari

Smith

(2013). Interventions in schools for children with autism spectrum disorder: Methods and recommendations. Autism, 17, 254–267.

Lane

J. D.

Ledford

J. R.

(2014). Using interval-based systems to measure behavior in early childhood special education and early intervention. Topics in Early Childhood Special Education, 34, 83–93.

Ledford

J. R.

Gast

D. L.

(2018). Single case research methodology: Applications in Special Education and Behavioral Sciences (3rd ed.). Routledge.

10.

Ledford

J. R.

Wolery

(2013). Procedural fidelity: An analysis of measurement and reporting practices. Journal of Early Intervention, 35, 173–193.

11.

Lieberman

R. G.

Yoder

P. J.

Reichow

Wolery

(2010). Visual analysis of multiple baseline across participants graphs when change is delayed. School Psychology Quarterly, 25, 28–44.

12.

Mancil

G. R.

Conroy

M. A.

Haydon

T. F.

(2009). Effects of a modified milieu therapy intervention on the social communicative behaviors of young children with autism spectrum disorders. Journal of Autism and Developmental Disorders, 39, 149–163.

13.

Mangold. (2015). INTERACT (Version 14) [Computer software]. Mangold International GmbH. www.mangold-international.com

14.

McCathren

R. B.

(2000). Teacher-implemented prelinguistic communication intervention. Focus on Autism and Other Developmental Disabilities, 15, 21–29.

15.

McDuffie

A. S.

Lieberman

R. G.

Yoder

P. J.

(2012). Object interest in autism spectrum disorder: A treatment comparison. Autism, 16, 398–405.

16.

Morrier

M. J.

Hess

K. L.

Heflin

L. J.

(2011). Teacher training for implementation of teaching strategies for students with Autism Spectrum Disorders. Teacher Education and Special Education, 34, 119–132.

17.

Mullen

(1995). Mullen scales of early learning. American Guidance Service.

18.

Plumb

A. M.

Wetherby

A. M.

(2013). Vocalization development in toddlers with autism spectrum disorder. Journal of Speech, Language, and Hearing Research, 56, 721–734.

19.

Schopler

Bourgondien

Love

Wellman

(2010). Childhood Autism Rating Scale–Second edition (CARS-2). Pearson.

20.

Schreibman

Dawson

Stahmer

A. C.

Landa

Rogers

S. J.

McGee

G. G.

Kasari

Ingersoll

Kaiser

A. P.

Bruinsma

McNerney

Wetherby

Halladay

(2015). Naturalistic developmental behavioral interventions: Empirically validated treatments for Autism Spectrum Disorder. Journal of Autism and Developmental Disorders, 45, 2411–2428.

21.

Shumway

Wetherby

A. M.

(2009). Communicative acts of children with autism spectrum disorders in the second year of life. Journal of Speech, Language, and Hearing Research, 52, 1139–1156.

22.

Sowden

Perkins

Clegg

(2011). Context and communication strategies in naturalistic behavioural intervention: A framework for understanding how practitioners facilitate communication in children with ASD. Child Language Teaching and Therapy, 27, 21–38.

23.

Toth

Munson

Meltzoff

A. N.

Dawson

(2006). Early predictors of communication development in young children with autism spectrum disorder: Joint attention, imitation, and toy play. Journal of Autism and Developmental Disorders, 36, 993–1005.

24.

Warren

S. F.

Yoder

P. J.

Gazdag

G. E.

Kim

Jones

(1993). Facilitating prelinguistic communication skills in young children with developmental delay. Journal of Speech and Hearing Research, 36, 83–83.

25.

Wetherby

A. M.

Watt

Morgan

Shumway

(2007). Social communication profiles of children with autism spectrum disorders late in the second year of life. Journal of Autism and Developmental Disorders, 37, 960–975.

26.

Whalen

Schreibman

Ingersoll

(2006). The collateral effects of joint attention training on social initiations, positive affect, imitation, and spontaneous speech for young children with autism. Journal of Autism and Developmental Disorders, 36, 655–664.

27.

Wong

Kasari

(2012). Play and joint attention of children with autism in the preschool special education classroom. Journal of Autism and Developmental Disorders, 42, 2152–2161.

28.

Yoder

P. J.

Stone

W. L.

(2006). Randomized comparison of two communication interventions for preschoolers with autism spectrum disorders. Journal of Consulting and Clinical Psychology, 74, 426–435.

29.

Yoder

P. J.

Warren

S. F.

(1998). Maternal responsivity predicts the prelinguistic communication intervention that facilitates generalized intentional communication. Journal of Speech Language and Hearing Research, 41, 1207–1219.

30.

Yoder

P. J.

Warren

S. F.

(2002). Effects of prelinguistic milieu teaching and parent, speech, responsivity education on dyads involving children with intellectual disabilities. Journal of Language, and Hearing Research, 45, 1158–1174.

31.

Yoder

P. J.

Warren

S. F.

Kim

Gazdag

G. E.

(1994). Facilitating prelinguistic communication skills in young children with developmental delay II: Systematic replication and extension. Journal of Speech and Hearing Research, 37, 841–851.