CT FEAT: Families Helping Children Achieve Their Full Potential
  Print This Article    

Self-Stimulatory and Stereotyped Behavior in Autism; A Review of Behavior Analytic Perspectives

Roberta Daversa
Southern Connecticut State University
December 2001 Abstract

One of the distinguishing features of many children with autism is their frequent engagement in repetitive, apparently non-functional behaviors.  These have been historically been referred to as stereotypic behaviors.  Currently behavior analysts and others recognize other inappropriate behaviors directed at producing self-sensory stimulation  and have broadened the area of concern to include all of these “self-stimulatory” behaviors, some of which are more subtle to detect than traditionally defined stereotopy.   This class of behaviors is problematic in all of its forms because of the obstacles to learning they present and their socially stigmatizing nature.   Viewing these inappropriate behaviors as variations of normal functioning is effective in understanding possible functionality of these behaviors.  This in turn offers important insight into the treatment of such behavior excesses.  Recent research on the motivational components of antecedent conditions can also prove an important tool in autism treatment.

Many children with autism engage in apparently non-purposeful, repetitive sequences of fixed behaviors including movement patterns  (hand-flapping, body-rocking, spinning or flipping of objects) and repetitive vocal sequences (echolalia).   The presence of stereotypical movements has been a key feature of the syndrome of autism since Kanner originally described it in the 1940’s.  These stereotyped behaviors are characterized mostly by their cyclical and rhythmic nature (Berkson, 1983) and have been documented and studied in other human populations such as nursing home patients  (Cohen-Mansfield, 1991) and drug addicts (Lewis, Baumeister and Mailman, 1987), and extensively in animal research.  Along with stereotypies, behaviorists generally include other non-rhythmic behaviors, such as atypical body postures and staring into lights, within a broader category of human behaviors referred to as self-stimulatory behaviors.   These behaviors are produced and received by the performer, and are not always pathological in nature.  Some examples of normal self-stimulatory behaviors are dribbling a basketball and playing solitaire (Newsom  & Lovaas, 1987).    However, because of their often excessive and behaviorally unusual nature in children with autism, self-stimulatory behaviors can be socially stigmatizing, thereby precluding significant inclusion in natural social, occupational and educational environments.   These behaviors pose significant obstacles to learning and left untreated they can easily become the dominant behavior in later years of life.

Research and treatment of problem behaviors in people with developmental disabilities has become an area of prolific study, with numerous articles addressing assessment and intervention  (Cuvo, 1997).  This paper will attempt to provide the reader with an overview of behavior research on the effect of self-stimulatory behaviors on learning and the acquisition and maintenance of such behaviors among children with autism from a functional perspective of their consequences and in terms of their motivational antecedents.

Self-Stimulation and Learning

Early research in the area of self-stimulation and learning in children with autism included a study by Lovaas,  Litrownik, and Mann, 1971) which documented significant increase in response latencies when children were engaged in self-stimulatory behavior.  A few years earlier, in 1968, Risley had reported an increase in certain appropriate behaviors when self-stimulatory behaviors were punished (Koegel & Covert, 1972).  These studies suggested to other researchers the possibility that skill acquisition could be enhanced through a reduction of self-stimulatory behaviors.  In the early 1970’s Koegel and Covert conducted experiments that attempted to define a functional analysis of the relationship of discrimination learning and self-stimulatory behavior in children with severe autism.    They examined (1) whether the children they studied failed to acquire discrimination learning tasks when engaged in self-stimulatory behaviors, (2) whether the suppression of such behaviors resulted in an increase in learning of the desired behaviors and (3) whether there was a spontaneous reduction of self-stimulation  (without external suppression) during skill acquisition.  In one experiment that they conducted, subjects were trained to respond by pressing a bar for a reinforcer when a light and tone were presented (SD) and to refrain from responding during the negative stimulus (SD) interval (absence of light and tone).  Training was first provided without suppression of self-stimulatory behavior for a significant number of trials, and then during a period in which self-stimulatory behaviors were suppressed (by the experimenter sharply saying “No!” and when necessary, delivering a “brisk” slap on the hands to interrupt the self-stimulation).    Koegel and Covert documented that skill acquisition was not achieved during the “no suppression” period and that the target discrimination task was learned during the “suppression” phase.  The data revealed a clear, almost perfectly inverse relationship in the rate of correct responses to the rate of self-stimulation.    

In a related experiment, Koegel and Covert recorded baseline rates of self-stimulatory behavior during a pre-training period.   Discrimination training was then initiated  (without suppression) and was followed by another “free” period without training.  The results showed a high rate of self-stimulatory behaviors in phase 1 (absence of discrimination training) with children engaging in these inappropriate behaviors for a majority of time in the pre-training period.  This was followed by marked decrease in self-stimulation during training (again displaying an inverse relationship between the variables, as correct responses increased, stereotypy decreased), with a return to somewhat higher levels of self-stimulation in the period immediately following discrimination training (but significantly lower levels than in the pre-trial period).    Koegel and Covert concluded that self-stimulatory behavior is incompatible with successful learning.  Some children can spontaneously reduce self-stimulatory behavior without external suppression during discrimination training and we can expect acquisition of the skill in these cases.   For others, external suppression of self-stimulatory behaviors is a prerequisite for learning new skills (Koegel & Covert, 1972).  Later work by Koegel found that suppression of self-stimulatory behavior produced an increase in appropriate play without the use of external reinforcers, but that regression to base stereotypy rate returned when suppression was discontinued.

Behavioral Functions of Self-Stimulating Behavior

While one feature of stereotypic behavior is that it appears to be non-adaptive and purposeless, behavior analytic research suggests that stereotypic and other self-stimulatory activities serve to provide and regulate sensory input from the environment.  It has been proposed that these behaviors are operant responses that serve as an effective block during times of excessive or aversive environmental stimulation and can provide sensory input during periods of under-stimulation.   Lovaas, Newsom and Hickman proposed a theory of “perceptual reinforcement” in which they relate self-stimulatory behavior to the innate need for stimulation among all organisms and that the excesses in these behaviors among children with autism is a consequence of the absence of appropriate alternative behaviors for accessing such stimulation.

 (Lovaas, et al., 1987).

Our position is that everyone, handicapped or not, both requires and can be rewarded by certain amounts and kinds of stimulation, as indicated in the literature on sensory deprivation and sensory/perceptual reinforcement.  The major difference between normal and handicapped persons lies in how the stimulation is obtained: Normal individuals obtain their stimulation primarily through normal verbal, social, work and recreation/leisure behaviors, whereas retarded and autistic persons obtain theirs through simpler behaviors, many of which are repetitive and stereotyped  (Newsom & Lovaas, 1987).

These operant responses are maintained by the automatically reinforcing perceptual consequences (Newsom & Lovaas, 1987).  The consequences of self-stimulatory behaviors are so highly reinforcing that this type of behavior is often chosen over other primary reinforcers   (Lovaas et al., 1971).

It has been found that the removal of a specific sensory consequence causes a reduction of self-stimulation associated with that particular consequence (Rincover, Cook, Peoples and Packard, 1979).  Rincover and colleagues studied the role of sensory reinforcement in maintaining self-stimulation, and whether this preference for specific sensory stimulation could be used to facilitate appropriate toy play.   Four children with autism were observed while engaging in stereotypic behavior and the behaviors were categorized by various sensory consequences they produced. Self-stimulatory behaviors (and the specific sensory consequences produced) were defined individually for each subject.   For example they observed that one child, Reggie, continually chose to spin a plate on a hard surface, and appeared to lean in to listen to the sound of the plate as it whirled around.  It was determined that the sensory pay-off for this behavior was auditory.  In the first part of the study, sensory extinction procedures were designed for each child, corresponding to the type of sensory consequences identified with their behaviors.  For Reggie, carpeting was installed on the flat surfaced he used for plate spinning.    The carpeting did not interfere with the spinning, but no sound was audible from the spinning plate during this part of the experiment.  Similar sensory extinction procedures were designed and implemented for each child, based on their specific stereotypy.   Extinction of the self-stimulatory behavior occurred in all the children when the sensory reinforcer was removed.   In phase 2, Rincover et al. found that the defined sensory reinforcers could be used to train appropriate play with toys that imitated the preferred sensory reinforcer (e.g. a music box and an autoharp for Reggie to provide auditory feedback) and that these behavior changes remained intact for several months without external suppression of self-stimulation or the use of external reinforcers.

Other studies have identified the function of stereotypy as a method of social interaction, i.e. to influence the behavior of others in order to either access or escape stimulation or instructional requests.  Self-stimulatory behaviors can be maintained in the individual with autism through negative reinforcement by reducing or escaping from unwanted social interactions.    These studies suggest that stereotypy and other self-stimulatory behaviors serve multiple functions for the child with autism (Kennedy, Meyer, Knowles, and Shulka, 2000). 

Recent research on the multiple functions of stereotypic responses includes two studies by Kennedy and colleagues (Kennedy et al., 2000).  In one, a multi-element design was utilized to determine the rate of stereotypic responding under four separate conditions (attention, demand, no attention and recreation).  During the attention condition, the experimenter and the subject engaged in separate activities, with the researcher providing 5 seconds of attention to the student when stereotypy occurred.    During the demand condition, the child was given an instructional request every 15 seconds, with praise for correct responses and full physical prompts for non-compliance.  Task demands were suspended for 30 seconds immediately following stereotypic activity (to assess the impact of negative reinforcement in the form of escape from environmental demands).  During the no-attention condition the child received no social interaction or assigned activities, assessing the rate of stereotypic behaviors in the absence of environmental stimulation (and presumably for sensory reinforcement).  During the recreational condition, the child was given various preferred activities and praised every 15 seconds in the absence of stereotypy (control condition to minimize stereotypic behavior).  The results confirmed that stereotypic responses served multiple functions: to access positive (attention) and negative (escape) reinforcer functions within   a social context, and to provide sensory stimulation (consistent with perceptual reinforcement research). 

In a second experiment, multiple reinforcer functions of stereotypy that were identified in the preceding experiment were paired with functional communication training (FCT) to teach separate alternate behaviors for each function that was previously determined to influence the occurrence of stereotypic responses (social attention, escape from instructional demand and sensory stimulation).  Hand raising was taught to acquire adult attention, signing “break” was taught to escape instructional demands, and signing “more” was taught to access preferred sensory stimuli.  Kennedy et al. documented significant decreases in stereotypic behaviors following the FCT training.    Their research supports past studies on the functional components of self-stimulatory behaviors  (perceptual reinforcers and social interactions).  In addition, their work offers important implications for the design of treatment programs for self-stimulatory behaviors, with considerable emphasis on the need to provide alternative behaviors in the treatment/education considerations for children with autism.

The Influence of Motivational Antecedent Events on Problem Behaviors

In its simplest form, Skinner’s three -term contingency of operant conditioning includes the context of the behavior (antecedent), the response (behavior) and the stimuli that follow the response (consequences).   Since then, behaviorists have recorded and manipulated antecedent events in recognition of their effects on behavior, with an emphasis on the discriminative properties of antecedent stimuli (those that signal that reinforcement is accessible).

Stimulus control and the basic response-reinforcement relationship have been extensively addressed in both arenas of application and research of behavior science.  Additionally, behavior analytic researchers have also noted the influence of motivational antecedent conditions.   In 1982, Michael proposed the concept of the establishing operation (EO) as a label for the motivational features of antecedent events, thus distinguishing these motivational influences from the discriminative properties of antecedents.   The concept of the EO has facilitated the study of antecedent effects, and the manipulation of motivational aspects that influence behavior.    Recent research on the effects of EOs can be organized into three categories: demonstration of the influence of EOs on behavior, manipulation of EOs in behavioral assessments, and manipulation of EOs to change behavior  (Iwata, Smith and Michael, 2000).

Michael defines the concept as “an environmental event, operation or stimulus condition that affects an organism by momentarily altering (a) the reinforcing effectiveness of other events and (b) the frequency of occurrence of that part of the organism’s repertoire relevant to those events as consequences . . .” Thus the EO serves two functions:  it alters the effectiveness of response consequences (reinforcers and punishment), and it changes the frequency of the related behaviors.  For an example Michael uses food deprivation.   It is an EO in that it is an environmental condition that both increases   the reinforcer value of food and increases behaviors directed at obtaining such reinforcers.    He also coined the term abolishing operation to refer to those EOs that have the opposite effect on reinforcers and behaviors.  An example of an abolishing operation is food satiation (Michael, 1993).

The relationship of EOs to problem behavior is a relatively new topic in behavior analysis.   Several common EOs have been identified in the motivation of problem behavior, including attention deprivation, deprivation of tangible items, deprivation of stimulation, etc.    It has therefore been suggested that application of this concept to the treatment of problem behaviors among people with developmental disabilities could benefit by focusing on modifying the environmental challenges that these people experience, rather than on the manipulation of contingencies (McGill, 1999).   

Clearly the study of motivational aspects of antecedent conditions is an area worthy of further investigation.  It offers potential to increase efficacy of treatment programs for self-stimulatory behaviors, especially when combined with current understanding of the functional components of these behaviors and the consequences of neglect in this area of autism treatment.


Berkson, G.  (1983).  Repetitive stereotyped behaviors.  American Journal of Mental Deficiency,  88,  239-246.

Cuvo, A. J. (1997)  Review of Behavior  Analysis In Developmental Disabilities 1968-1995 (3rd Ed.)  edited by Iwata et al. Journal of Applied Behavior Analysis,  30,  595-598.

Iwata, B.A., Smith, R., & Michael, J.  (2000).  Current research on the influence of establishing operations on behavior in applied settings.  Journal of Applied Behavior Analysis,  33,  441-418.

Kennedy, C. H., Meyer, K., Knowles, T. & Shukla, S. (2000).  Analyzing the multiple functions of stereotypical behavior for students with autism: implications for assessment and treatment.    Journal of Applied Behavior Analysis,  33,  559-571.

Koegel, R. L., & Covert, A. (1972). The relationship of self-stimulation to learning in autistic children.   Journal of Applied Behavior Analysis,   5,  381-387.

Lewis, M. H., Baumeister, A. A., & Mailman, R. B. (1987). A neurobiological alternative to the perceptual reinforcement hypothesis of stereotyped behavior: A commentary on Self-stimulatory behavior and perceptual reinforcement. Journal of Applied Behavior Analysis,  20, 253-258.

Lovaas,  O. I., Litrownik, A., & Mann, R. (1971).   Response latencies to auditory stimuli in autistic children engaged in self stimulatory behavior.    Behavior Research and  Therapy,   2,  39-49.

Lovaas, O. I., Newsom, C., & Hickman, C. (1987). Self-stimulatory behavior and perceptual reinforcement.  Journal of Applied Behavior Analysis,  20,  45-68.

McGill, P.  (1999).  Establishing operations: Implications for the assessment, treatment, and prevention of problem behavior.  Journal of Applied Behavior Analysis,   32,  393-418.

Michael,  J. (1993).   Establishing  operations.  The Behavior Analyst, 16, 191-206.

Newsom, C., & Lovaas, O. I. (1987). A neurobiological nonalternative: Rejoinder to Lewis, Baumeister, and Mailman.  Journal of Applied Behavior Analysis,  20,  259-262.

Rincover, A., Cook, R., Peoples, A., & Packard, D. (1979). Sensory extinction and sensory reinforcement principles for programming multiple adaptive behavior change. Journal of Applied Behavior Analysis,  12,  221-233.

Back to Top



CT FEAT, P.O. Box 370352, West Hartford, CT 06137-0352 Phone (860) 571-3888
Copyright 2010 CT FEAT
CT FEAT    Contact Us