Perhaps the “original” form of naturalistic behavioral intervention is incidental teaching (IT), which was developed in the 1960s to increase the language skills of underprivileged children in preschool settings (Hart and Risley 1968) . According to Hart and Risley (1982), IT includes first arranging the child’s natural environment to include toys and activities of interest that are visible to the child but not accessible without assistance or support. The teacher’s behavior in IT follows the following steps: (1) wait for the child to initiate (likely to obtain help, permission, attention, and/or approval), (2) respond to the child’s initiation immediately, (3) verify if unsure of what the child wants after initiation, (4) request the child to elaborate the initiation, (5) prompt and (if necessary) model the elaborated response if the child does not comply, and (6) verbally confirm the child’s request while providing access to the preferred toy or activity contingent on the child’s elaborated response. Hart and Risley reported impressive improvement in many aspects of the speech of the preschool children in their studies (Hart and Risley 1968, 1974; Risley and Hart 1968). Further, they found the children showed generalization of these language improvements that was superior to the generalization of language skills taught during highly structured training sessions (Hart and Risley 1980).

In the 1980s, researchers began in earnest to investigate the utility of IT for the treatment of communication deficits in children with ASD (Green 2001). Early research documented the successful use of IT to teach acquisition and generalization of receptive object labels (McGee et al. 1983), reading (McGee et al. 1986), and sign language (Schepis et al. 1982). In addition, comparison studies examining the difference between teaching expressive use of prepositions and labels to children with ASD using IT procedures versus using traditional structured teaching procedures revealed greater generalization and more spontaneous use of the prepositions and labels taught with IT procedures (McGee et al. 1985; Miranda-Linne and Melin 1992). Despite these promising findings, many children with ASD have difficulty spontaneously initiating interactions with others, even to request preferred items (Mirenda and Iacono 1988). IT learning rates may be slower for these children because there may be limited opportunities for learning if the teacher is required to wait for the child’s initiation. (Rogers-Warren and Warren 1980) .

In order to address this limitation, researchers have begun to adapt the original IT paradigm to better address the specific needs of children with ASD (Haring et al. 1987; Mirenda and Iacono 1988). For example, the modified incidental teaching sessions (MITS) procedure, developed by Charlop-Christy and Carpenter (2000), addresses decelerated child learning as a result of a lack of spontaneous child initiations by allowing mand-model adult initiations (e.g., adult encourages the child to request a preferred object by asking “What do you want?” and then models the appropriate complex response if needed). Charlop-Christy and Carpenter (2000) also modified traditional IT by utilizing parents as implementers of the MITS in the home environment in order to further facilitate generalization. In a comparison study of traditional IT, MITS, and DTT, these authors used a multiple baseline design across subjects in combination with an alternating treatment design to determine the effect of MITS training on three children with ASD. They found that all children showed better acquisition and generalization of the target behaviors learned in the MITS condition as compared to the traditional IT and DTT conditions (Charlop-Christy and Carpenter 2000) .

Originally developed as a specified combination of research-supported interventions to teach language to children with language delay, milieu teaching (MT) was quickly accessed as an intervention option to address communication deficits in children with ASD (Mancil 2009). MT comprises four well-established teaching techniques: (1) modeling and then correcting the child’s behavior if necessary, (2) the mand-model technique described previously, (3) time delay, as the adult will wait for a predetermined amount of time after presenting the stimulus for the child to respond, and (4) all components of IT (Kaiser et al. 1991; Mancil 2009). Kaiser et al. (1992) reported MT to be effective for teaching language skills ranging from single words to complex phrases to children and adolescents with language delay, intellectual disability, or ASD. In addition, the acquired language skills generalized across various settings and trainers in several of the studies (Kaiser et al. 1992).

However, as with other naturalistic behavioral approaches, traditional MT has been modified over time to better accommodate the specific needs of children with ASD. Enhanced milieu teaching (EMT) consists of all of the components of traditional MT with the addition of increased language expansions, turn taking, and following the child’s lead. Hancock and Kaiser (2002) implemented a multiple baseline across subjects design to examine the effects of EMT on four children with ASD. All the children showed increases in their language targets that were maintained at a 6-month follow-up. The children also increased the complexity and diversity of their language. In addition, three out of four of the children generalized their acquired language skills to their home environment. In order to address the needs of children with ASD who have not developed vocal language, prelinguistic milieu teaching (PMT) utilizes the techniques of MT to specifically teach valuable preverbal communication skills such as gestures, vocalizations, and coordinated eye gaze (Fey et al. 2006; Yoder and Stone 2006). In addition, Franco (2008) used PMT to successfully teach school-age children with ASD to increase the frequency and clarity of their communication skills in a home setting.

Pivotal response training (PRT) , sometimes referred to in the literature as pivotal response treatment, is a child-directed therapy well supported as an established evidence-based practice intervention for children with ASD (e.g., Humphries 2003; National Autism Center 2009; National Research Council 2001; National Professional Development Center on Autism Spectrum Disorders 2011). PRT is an expansion of the natural language paradigm (NLP; Gillett and LeBlanc 2007; Laski et al. 1988), a language intervention originally developed as a combination of traditional operant and naturalistic procedures for children with ASD. NLP is presented in a naturalistic teaching paradigm where child choice, task variation, flexible prompting, and reinforcement of child attempts are valued. Comparisons of the NLP and more traditional structured teaching revealed the NLP to provide improved generalization and spontaneity of child language gains (Koegel et al. 1987).

PRT targets the “pivotal” behaviors of motivation and responsivity to multiple environmental cues (Koegel et al. 1989). Pivotal behaviors are so-named because improvements in these behaviors are likely to affect change in collateral behaviors, which may improve overall child response to treatment and minimize the treatment time required to learn new skills. Motivation is a pivotal behavior for children with ASD because increasing child motivation during intervention will likely lead to increased language, play, and social gains (e.g., increasing a child’s motivation to respond to social approaches can reduce avoidance of social learning opportunities). It is often challenging to access high levels of child motivation during treatment. PRT seeks to address limited child motivation levels directly (Koegel et al. 1999; Schreibman 1988). For example, PRT targets child motivation by consistently following the child’s lead including when to play, what to play, and how to play. All intervention is provided in the context of a child’s individual interests and preferences. In addition, difficult tasks are interspersed with easier tasks and children are reinforced for good attempts at the target behavior to help create behavioral momentum.

Children with ASD also exhibit difficulties with “stimulus overselectivity” (Lovaas et al. 1979, 1971; Schreibman 1988), an attentional deficit wherein the child does not respond to simultaneous stimulus input. Stimulus overselectivity may limit a child’s ability to generalize previously learned behavior out of the intervention context (Schreibman 1997). Fortunately, research supports that many children with ASD who display overselectivity can in fact learn to respond to compound stimuli if taught a series of conditional discriminations (i.e., discrimination tasks requiring response to simultaneous multiple cues; Koegel and Schreibman 1977; Schreibman et al. 1982, 1977). PRT addresses stimulus overselectivity directly as another pivotal behavior by teaching children to respond to multiple cues simultaneously, thus removing a barrier to generalization.

Subsequent research has identified self-initiation and self-management as pivotal behaviors in addition to motivation and responsivity to multiple cues. Self-initiation is considered a pivotal behavior as it allows a child with ASD to spontaneously approach others to manipulate her environment for a variety of reasons (e.g., to obtain items or engage in social interaction), providing the child with a wealth of learning opportunities. Education in self-management allows an individual with ASD to monitor her own behavior and provide appropriate consequences in order to manage her own behaviors. Individuals with ASD have effectively utilized self-management to increase desirable target behaviors (e.g., language, play, and social skills) and decrease undesirable target behaviors (e.g., stereotypic behavior; Cowan and Allen 2007; Matson et al. 1996).

PRT is typically implemented during play but is a highly flexible intervention that can be used throughout the child’s day wherever there is a learning opportunity. As noted, the key components of PRT focus on increasing the pivotal behaviors of motivation and responsivity to multiple cues. These components include how to deliver the opportunity for a target behavior by: (1) providing the child clear instructions or questions that are relevant to the current task or activity, (2) interspersing easy tasks with difficult tasks, (3) allowing child choice of activity or objects used during sessions, (4) taking turns with the child, and (5) presenting learning opportunities that require the child to respond to simultaneous multiple stimuli (i.e., conditional discriminations). PRT also addresses how to respond to the target behavior once it occurs by: (6) utilizing direct or natural reinforcement, (7) ensuring reinforcement is contingent, and (8) reinforcing appropriate attempts to respond (Koegel et al. 1989). Multiple single-subject and long-term outcome studies have confirmed that PRT facilitates the functional use of language for many children with ASD, including those with little to no functional speech (Humphries 2003; Koegel et al. 1987, 1999a, b; Laski et al. 1988). Moreover, when compared with more structured behavioral techniques, the specific components of PRT facilitate relatively greater increases in verbalizations and spontaneous language use (Delprato 2001; Koegel et al. 1988; Koegel and Williams 1980; Williams et al. 1981).

In addition to language acquisition, PRT has been shown to be effective for targeting play and social skills in children with ASD (Pierce and Schreibman 1997a, b; Stahmer 1999). Stahmer (1995) taught seven children with ASD to engage in symbolic play behaviors using PRT. She found that after receiving PRT targeting symbolic play, children with ASD engaged in symbolic play at levels comparable to typically developing children of the same language ability. In addition, the children increased their interactions with adults during play. The children also generalized these skills to new toys, adults, and settings and maintained skills over time. Thorp et al. (1995) targeted sociodramatic play using PRT with three children with ASD. Sociodramatic play consists of the interaction of a group of children cooperating to elaborate together around a central theme, and the study found that PRT was effective in increasing this form of play. Importantly, in addition to their improved play skills, the children showed collateral improvements in language and social skills.

Reciprocal imitation training (RIT; Ingersoll and Gergans 2007) draws from naturalistic behavioral approaches such as IT, MT, and PRT to specifically teach imitation skills to children with ASD using contingent imitation (imitation of the child’s movements and vocalizations), linguistic mapping (narrating the child’s play), and imitation training (modeling actions, prompting, and reinforcement). Targeting imitation skills during an early intervention program may help children with ASD learn a wide range of other functional behaviors (Ingersoll 2010b; Schreibman 2005). RIT has been used to successfully teach joint attention (Meindl and Cannella-Malone 2011), play (Stahmer et al. 2003), object imitation (Ingersoll and Schreibman 2006), and descriptive gestures to children with ASD (Ingersoll et al. 2007). In fact, in the RIT studies where the authors targeted object imitation and descriptive gestures, the treatment gains generalized to new therapists, settings, and materials, and maintained at a 1-month follow-up. In addition, although only imitation skills were targeted, Ingersoll and Schreibman (2006) found collateral effects of RIT on language, pretend play, and joint attention skills of the five children with ASD in their study. In a randomized controlled trial, Ingersoll (2010a) found that RIT was effective at increasing object and gesture imitation skills in a group of children with ASD in comparison to a control group, replicating the previous single-subject research findings.

In a very promising direction of application, naturalistic behavioral approaches have been effectively utilized in classroom settings. Lerman et al. (2004) found increases in student communication behaviors after providing teachers with a week-long workshop, approximately 18 h of training, where IT was one component (teachers were also taught other behavioral techniques like preference assessment and direct teaching). Teachers were provided lectures, handouts, role-playing with feedback, and in vivo practice with feedback. All intervention skills learned generalized to other students in the classroom. There is some evidence to show teacher acquisition of IT skills can be accomplished with less training time. Haring et al. (1987) taught teachers some IT procedures with only self-instruction materials including daily preplanned activities. In their study, teacher self-monitoring was enough to increase maintenance and generalization of IT skills in the classroom. Ryan et al. (2008) found that a brief 30-min group IT training that included didactic lecture, modeling, and role-playing with feedback was similarly effective as multiple individualized training sessions for training teachers to use IT to increase student initiations in the classroom. In addition, the teachers rated IT favorably and reported it useful (Ryan et al. 2008).

IT is the primary intervention used at the Walden Early Childhood Program, an inclusion program for children with ASD (McGee et al. 1999). The Walden Early Childhood Program consists of multiple classrooms: toddler, early preschool, preschool, and pre-kindergarten serving typically developing and children with ASD up to 4 years of age. The Walden model has been replicated at multiple sites (McGee et al. 1999). Teachers in all classrooms are trained to utilize IT during all activities throughout the day and the classroom environment is conducive to implementation of IT (e.g., preferred items are observable but inaccessible to the children and gates divide areas of the classroom to facilitate child requests to move from area to area). Children with ASD who have attended the Walden Early Childhood Program have shown good language and social skill outcomes (McGee et al. 1999).

In addition to teachers, paraprofessionals (often described as classroom aides) can effectively use naturalistic behavioral approaches to improve child behaviors in the classroom. Robinson (2011) taught paraprofessionals to utilize PRT using a 45-min modeling session and subsequent video-based feedback sessions. During the video-based feedback sessions, a PRT trainer and paraprofessional watched videotapes of the paraprofessionals utilizing PRT together while the PRT trainer provided feedback. The paraprofessionals greatly improved their PRT implementation and the author also found increased social-communication behaviors of the children with ASD in the target classroom. In addition, Robinson (2011) found good maintenance and generalization of the PRT implementation skills of the paraprofessionals. When surveyed about the PRT training they received, the paraprofessionals reported that they were completely satisfied with the training protocol and felt better able to help their students with ASD after training.

Social validity (i.e., social importance and acceptability; Foster and Mash 1999) is an important consideration for interventions intended for classroom use, as social validity can greatly impact effective dissemination (Stahmer 2007). There is some evidence that naturalistic behavioral approaches such as PRT may be more difficult for teachers to implement than traditional structured teaching techniques (Zandi et al. 2011). However, Lerman et al. (2004) found that teachers in their study showed better acquisition of IT techniques than of direct teaching techniques during their teacher training sessions. Despite mixed research findings regarding the ease of teacher adoption of naturalistic behavioral approaches, successful dissemination of these approaches is essential given their potential for improved child maintenance and generalization of acquired skills. Thus, in order to ease the transition of PRT from the laboratory to the classroom, an adaption of PRT called classroom pivotal response teaching (CPRT; Stahmer et al. 2011) has been developed by intervention researchers in a collaborative relationship with classroom teachers. CPRT is currently under careful study to determine if the adjustments made will simultaneously uphold the integrity and positive student outcomes of the intervention while allowing PRT to be more easily translated into the classroom setting (Stahmer et al. 2011).

Educating parents of children with ASD to implement naturalistic behavioral interventions can be highly beneficial, as parent-implemented naturalistic behavioral approaches may improve child outcome. There are several specific benefits of parent education. First, children spend the majority of their time with their parents, thus providing a wealth of valuable intervention opportunities. Second, training parents saves families and the service system time and costly resources because parents can provide treatment themselves instead of utilizing outside resources. Third, such training can reduce parent frustration levels by allowing parents to better manage their children’s challenging behaviors. Fourth, parent training can improve child generalization and maintenance of skills learned in intervention (Kaiser 1995; Koegel et al. 1996; National Research Council 2001; Symon 2005).

There is a great deal of evidence in the literature to support the effectiveness of a parent education component for the naturalistic behavioral approaches mentioned thus far. Parents have effectively utilized MIT (Charlop-Christy and Carpenter 2000; Kaiser et al. 1995) and EMT (Kaiser et al. 2000) to increase communication by their children with ASD. Parents have also learned to implement PRT to improve communication (Coolican et al. 2010; Minjarez et al. 2011), as well as adaptive functioning (Baker-Ericzén et al. 2007), and joint attention skills (Rocha et al. 2007). Ingersoll and Gergans (2007) successfully trained three parents to use RIT to teach spontaneous object and gesture imitation to their children with ASD.

In addition to well-documented parent and child gains during treatment, most of the above parent implementation studies measured generalization and maintenance of parent and child skills after treatment concluded. Overall, the majority of the parents and children in these studies were able to generalize skills learned from the clinic to the home setting and often these skills were maintained over time for both the parents and the children. Recent research also suggests that parent education in naturalistic strategies can be conducted in less time (Coolican et al. 2010) and more easily (Minjarez et al. 2011) than originally thought. For example, parent education programs typically consist of parent education implemented over many individual sessions. Minjarez et al. (2011) successfully utilized a group training format over 10 weeks to instruct 17 parents to implement PRT to improve the communication skills of their children with ASD. The parents in their study learned to implement PRT with high levels of fidelity and their children showed associated gains in functional language.

Naturalistic behavioral approaches are particularly well suited to parent education programs as they are implemented in a variety of natural settings and circumstances and at different times of day. Thus, it is not surprising there is some evidence that parents may favor such an approach over more traditional structured teaching approaches. For example, Schreibman et al. (1991) found that parents implementing PRT exhibited more positive affect (i.e., parent enthusiasm, interest, and happiness) than parents implementing traditional DTT with their children with ASD when rated by naive observers during a parent education program. The authors propose the parent affect difference may be the result of the more “natural” interaction style of a naturalistic behavioral approach such as PRT over the more contrived style of a traditional structured approach such as DTT. They also suggest the improved parent affect may be due to increased child motivation during intervention sessions that is directly targeted with PRT (Schreibman et al. 1991).

In addition to parents as interventionists, educating peers (peer tutors) to implement naturalistic behavioral approaches for individuals with ASD (peer learners) can be valuable for several reasons. Peers are often readily available in the individual’s natural environment and the availability of peers as interventionists provides more opportunities for learning. Educating peers to be interventionists is relatively easy and educating peers may foster better inclusion, as individuals with ASD will have more experience interacting successfully with their peers. In addition, typical peers may provide better models for age-appropriate language, play, and social skills than other interventionists, and utilizing peers themselves as interventionists may improve generalization and maintenance of language, play, and social skills to community settings such as school where peers will be present (Chan et al. 2009; Pierce and Schreibman 1997a, b).

Typically, these approaches consist of the peer trainer modeling and role-playing the intervention techniques for the peer tutor, providing feedback to the peer tutor, and providing the peer tutor with a visual checklist composed of words or pictures to ensure proper implementation. Then the trainer fades herself from the intervention sessions slowly when the peer tutor is comfortable with the intervention and is reliably implementing the intervention with acceptable fidelity levels. In addition, peer trainers may provide reinforcement (often in the form of a token system) to the peer tutors contingent on their use of the intervention with the peer learners in the natural environment. Peer tutors utilizing naturalistic behavioral approaches to intervention for peer learners with ASD are often taught to wait for the peer learner to initiate toward an item (e.g., reach, look at), ask the peer learner for a target behavior (e.g., label, point), provide the item when the peer learner exhibits the target behavior, and praise the peer learner after a successful interaction. Peer tutors are also taught to provide prompts as needed and take turns with the peer learner to increase opportunities for interaction (Chan et al. 2009).

Research has supported the effectiveness of these strategies, as peer tutors of various ages have been successfully taught to provide naturalistic behavioral interventions to peer learners with ASD. For example, typical preschool children were taught to use IT with three children with ASD during free play in their classroom, subsequently increasing the peer learners’ use of verbal labels and reciprocal interaction during play (McGee et al. 1992). After training, the peer tutors themselves also increased their approaches toward children with ASD. Generalization of these research findings were mixed in their study, but the teachers and peers involved in the study rated the intervention favorably. Farmer-Dougan (1994) reports successfully utilizing IT-trained adult peer tutors of adult peer learners with intellectual disability and ASD in a group home setting to increase appropriate requests during lunch. She found good generalization of the learned skills for the adult peer tutors and the adult peer learners from lunchtime to dinnertime. In addition, there was a promising increase in the overall interactions between all the residents and staff.