However, extinction procedures are only effective when matched to the function of behavior (Iwata, Pace, Cowdery et al. 1994; Richman et al. 1998). The previous example of the caregiver withholding a toy in response to SIB would be appropriate if SIB was maintained by access to the toy (i.e., social positive reinforcement). However, if SIB was actually maintained by escape or avoidance from nonpreferred demands, then withholding delivery of a toy likely would not affect SIB responding. Instead, extinction would have to take the form of escape extinction, where escape from the demands would not be allowed contingent on SIB. In general, extinction is applied in three ways, depending on behavioral function, including (1) withholding or minimizing a consequence when the behavior is maintained by social positive reinforcement, (2) by not allowing avoidance or escape as a consequence when the behavior is maintained by social negative reinforcement, and (3) by attenuating or eliminating the sensory consequence when the behavior is maintained by automatic reinforcement, often referred to as “sensory extinction” (Rincover 1978) .

Extinction rarely is utilized alone, and most often is used in conjunction with other interventions (Vollmer, Sloman et al. 2009). The inclusion of other interventions, such as functional communication and differential reinforcement procedures, aids in reducing the temporary increases in challenging behavior associated with extinction known as “extinction bursts” (Lerman and Iwata 1996). Extinction bursts can be especially problematic when treatment is for severe challenging behaviors, such as SIB or aggression, where even temporary increases in frequency or intensity may be unacceptable. In addition, the inclusion of other interventions with extinction not only targets reducing the challenging behavior, but also increasing an alternative or desired appropriate behavior, which usually is a goal of behavioral interventions (O’Neill et al. 1997). For example, Hausman et al. (2009) combined extinction with functional communication in the treatment of SIB, aggression, and property destruction in a 9-year-old female with autism, whose challenging behaviors were maintained by access to rituals . Extinction targeted eliminating the functional link between the challenging behaviors and access to rituals , while an appropriate alternative behavior (i.e., communication) was established that allowed for access to rituals .

The goal of DR is to render the child’s problem behavior inefficient so that more socially acceptable behaviors can occur. With instances where the behavior is unable to be eradicated or the prior DR procedures are not completely feasible (e.g., the behavior occurs at too high rate), differentially reinforcing low rates of the behavior (DRL) may be more appropriate initially so the child can contact the contingencies in place, until further fading can occur. Additionally, extinction may not always be possible due to the severity of the problem behavior (e.g., aggression), and as a result, it may be necessary to manipulate the parameters surrounding the schedule of reinforcement to make it more favorable for emitting acceptable behavior instead of problem behavior (Athens and Vollmer 2010; Piazza et al. 1997). Also, DR schedules can be difficult to implement due to the complexity of the treatment packages (e.g., having to monitor specific periods of time where the behavior does not occur) .

In order to effectively use FCT, the function of the challenging behavior needs to be discerned. Therefore, prior to FCT, the function of the challenging behavior should be identified via a functional behavioral assessment (Durand and Merges 2001; Tiger et al. 2008). Afterwards, an appropriate communicative response that will serve the same function as the challenging behavior needs to be identified and taught to the individual. Several factors should be taken into consideration when deciding on the communication response. For example, the individual should possess adequate motor skills to produce the communication response efficiently (Tiger et al. 2008). Response effort at producing the communication response should be low, at least during skill acquisition (Horner and Day 1991); high effort responses may be best for generalization purposes (Hernandez et al. 2007). In addition, a decision will need to be made between a topography-based system, such as verbal language or sign language, and a selection-based system, such as picture exchange. Topography-based systems are more portable, and eliminate the potential difficulty of an individual having to discriminate among various pictures, as is the case with selection-based systems. However, selection-based systems seem to facilitate acquisition of the communication response quicker, and may be better in situations where the individual lacks verbal language or has difficulty with fine motor skills (Horner and Day 1991; Tiger et al. 2008).

As an intervention, FCT rarely is used alone, and more often is part of a treatment package. Most often, FCT is paired with extinction or a punisher, in which case the functional reinforcer is delivered contingent on the communication response, while all challenging behaviors are ignored, or result in an aversive consequence (e.g., time out). Hagopian, Fisher, Sullivan, Acquisto, and LeBlanc (1998) reported that FCT alone was not effective in treating the severe challenging behaviors of 21 individuals with IDDs, and was only effective when paired with extinction or a punishment. However, in some circumstances, FCT only can be effective, usually when the schedule of reinforcement for the communication response is denser than that of challenging behavior (Kelley et al. 2002; Worsdell et al. 2000), but also when the schedules are similar (Casey and Merical 2006).

NCR is a commonly used intervention for treating challenging behavior for numerous reasons. First, this procedure does not require the child to emit a response (i.e., DRA) or to go for extended periods of time without emitting a problem behavior (DRO) in order to obtain reinforcement (Hagopian et al. 1994). Second, NCR can be useful for interrupting or preventing automatically maintained behavior by providing alternative sources of reinforcement (Favell et al. 1982; Roscoe et al. 1998). Third, logistically, NCR is considered to be easier to implement than other behavior strategies such as extinction or DR procedures because there is an exact schedule for delivering reinforcement and few criteria for administration. Finally, NCR may be most appropriate for individuals who engage in high rates of aggressive, disruptive, or SIB that is difficult to block or place on extinction (e.g., a child that is stronger than his/her care provider) .

There are some considerations that must be addressed when using NCR. For example, it is possible to “accidentally” strengthen the relationship between problem behavior and the delivery of reinforcement, if a problem behavior occurs shortly before the delivery of the reinforcer (i.e., adventitious reinforcement). In such situations, it may be helpful to use a combination of NCR and DRO (Vollmer et al. 1997), with reinforcement still delivered on a schedule, but only provided after a specific amount of time has passed without a problem behavior (e.g., 5 s) .

Commonly used punishment procedures vary from mild verbal reprimands to full restraint and will vary based on the behavior that needs to be decreased. Punishment selected should be functionally related to the problem behavior. For example, timeout from reinforcement would be more appropriate for a child who engages in problem behavior to obtain attention rather than a child who engages in problem behavior to escape demands. Regardless of the function, if a behavior is severe enough that it frequently causes injury (e.g., head slapping or hand biting) and/or a child is at risk for permanent damage (e.g., retinal detachment or infection), it may be necessary to stop the behavior from occurring immediately with more stringent measures (e.g., a basket hold).

Factors to consider when providing punishment include the immediacy, consistency, and magnitude with which the punisher can be delivered, as well as the history of the problem behavior and implementation of prior punishers (Lerman and Vorndran 2002). Perhaps most importantly, it is critical to assess the social validity of the procedures prior to their implementation, particularly since there has been a strong movement against the use of punishment (Carr et al. 2002). If care providers or school staff do not agree with the punishment procedures suggested, treatment integrity will most likely be low. Additionally, it is recommended to conduct a punisher assessment to determine the actual punishing effects of the procedures, and the acceptability to care providers.

The primary flaw of punishment is that it does not teach what behavior the child should be emitting. For this reason, when implementing punishment procedures it is ideal to combine the punisher with a dense schedule of alternative reinforcement in an attempt to increase the more acceptable behavior. If punishment is not implemented consistently, it may prove ineffective. Thus, it is recommended that initially it be implemented on a fixed ratio 1 schedule of reinforcement, before being faded.

Surprisingly, while the use of protective equipment and restraint procedures are fairly common in treating challenging behaviors , there is comparatively little recent published research on it, especially with young children where its use is less prevalent. This might be due to the stigma associated with protective equipment, as well as increased emphasis on identifying alternative, nonrestrictive treatments. Additionally, protective equipment and restraint procedures probably are best conceptualized as facilitators of intervention as opposed to interventions themselves. While protective devices, such as arm splints, conceptually may serve the function of extinction by restricting or attenuating the consequences of a behavior, often they are used with and to allow for the success of other interventions, such as NCR programs and enriched environments (Roscoe et al. 1998).

Much of the research on protective equipment has focused on its use in sensory extinction treatments of SIB (Rincover 1978), particularly the use of arm splints and helmets for the treatment of head hitting. Moore, Fisher, and Pennington (2004) reported, treating the SIB of a 12-year-old girl with autism by utilizing a combination of a protective helmet, rigid arm sleeves, and padded gloves that eliminated or attenuated the assumed positive sensory consequences of SIB. However, the use of restraint procedures, particularly prolonged use, can be associated with a host of negative side effects, including interference with motor development (Lovaas and Simmons 1969), inhibition of adaptive skills (Wallace et al. 1999), social stigma (Rojahn et al. 1980), and the emergence of new forms of challenging behaviors (Fisher et al. 1997). Therefore, when using protective equipment, a plan for fading, gradually reducing, and eventually eliminating the restraint is necessary (Fisher et al. 1997). In many circumstances, protective devices are used where the padding or rigidity of the device can be decreased gradually and incrementally (Fisher et al. 1997; Pace et al. 1986). Additionally, the use of protective equipment noncontingently or contingently in response to challenging behavior has been evaluated with results suggesting that noncontingent use serves a sensory extinction function (Moore et al. 2004), while contingent application seems to function as punishment (Mazaleski et al. 1994).

The use of protective equipment and restraint procedures is controversial and has come under increased scrutiny in recent years (Day et al. 2010; Matson and Boisjoli 2009). There is a clear movement to either reduce or eliminate the use of protective equipment and other restraint procedures with persons diagnosed with IDDs and ASDs (Day et al. 2010; Miller et al. 2006). However, even though there has been an emphasis on using alternative, nonrestrictive procedures to treat challenging behaviors in persons with ASDs, particularly those that reinforce appropriate alternative behaviors, the use of restrictive procedures occasionally is warranted (Matson and Boisjoli 2009). Therefore, the legitimate concerns regarding the rights and welfare of persons with challenging behaviors must be balanced with the risk these behaviors present to those individuals (Day et al. 2010). For example, the use of a protective device likely is warranted when the challenging behavior poses an immediate and imminent danger to the individual, such as self-injurious eye poking that may result in permanent eye damage or blindness.

At the center of the decision to use protective equipment and restraint procedures is determining whether their use is merely one of convenience for caretakers, for safety of the individual and others, and/or for treatment purposes (Matson and Boisjoli 2009). As a result, practitioners and researchers have sought to delineate guidelines and procedures for not only the effective, but also ethical use of restraint procedures (ABAI 2010; Matson and Boisjoli 2009). In general, restraint should only be employed with the child’s welfare as the primary goal and with the consent of the child’s caregiver(s); only the safest and least restrictive procedures that are effective should be used; restraint should be used only as a last resort and after extensive evaluation of nonrestrictive procedures has justified their use; restraint should be based on the results of a functional assessment and used in conjunction with other reinforcement-based procedures in a planned manner within the clear goals of a behavior intervention or treatment plan; and restraint procedures should be implemented by trained persons, overseen by a mental health professional, such a licensed psychologist or certified behavior analyst, and monitored closely for their effectiveness using objective data (ABAI 2010; Day et al. 2010; Matson and Boisjoli 2009).

Psychopharmacological interventions may be used to treat an entire suspected psychiatric disturbance, a specific behavior as a symptom of a psychiatric diagnosis, or a behavior occurring in the absence of a psychiatric diagnosis. For the purposes of this chapter, research will be presented for psychiatric interventions for specific behaviors regardless of the possibility of the behavior being part of a psychiatric disturbance, as such is reported elsewhere in this book. However, as previously mentioned, prior to intervention, it is of paramount importance to rule out the possibility of these challenging behaviors being a symptom of a psychiatric disorder. In these instances the therapeutic approach may differ with emphasis on treating the underlying disorder, as opposed to treating a symptom or behavior. For example, a child whose aggressive behaviors are related to an anxiety disorder likely would receive a different medication than if the aggression was related to a psychotic disorder or frustration from difficult academic demands .

Although there is a long history of and need for psychopharmacological interventions to treat challenging behaviors in children with ASDs, there are several limitations in the research literature that should be acknowledged. First, there are few well-controlled studies supporting their use (Matson and Dempsey 2008). In fact, only two medications are Federal Drug Administration (FDA) approved for symptoms related to autistic disorder, both for treatment of irritability: aripiprazole (Abilify^TM) and risperidone (Risperdal^TM) (Blankenship, Erickson, Stigler et al. 2010; Singh et al. 2011). The majority of published research reports on off-label use of medications, and more often than not these studies are “open label,” which lack appropriate controls (Matson and Dempsey 2008). Second, medication effects typically are assessed via caregiver report on a standardized measure, such as the ABC (Aman et al. 1985), which are prone to subjective bias. Studies are needed that use direct measures of behavior, which would provide better indicators of improvement. Third, several different standardized measures have been used to assess medication effects, making it difficult to draw comparisons across studies for similar medications. Finally, challenging behaviors are not always the focus of medication interventions. Many medication studies target core, global features of autism or larger psychiatric symptoms, such as mood or anxiety. While many studies report global improvements in autism and psychiatric symptoms, the impact on challenging behavior alone is not always clear .

As a result of these limitations, psychiatrists are challenged with basing their understanding of medication utility on a less than optimal body of research and more often on case study reports, and sometimes must refer to reported results and clinical trials of medications used in the general population for similar symptoms to guide their decisions. To complicate matters, children with ASDs often present with multiple challenging behaviors which may require multiple medications for intervention (Carlson et al. 2006). Delineating the effects (both positive and adverse) of one or more medications on one or more challenging behaviors make the job of the psychiatrist even more difficult .

The most common medication classes used with ASD populations include psychostimulants, antidepressants, antipsychotics, mood stabilizers, and alpha agonists (Blankenship, Erickson and McDougle 2010; Handen and Lubetsky 2005; Matson and Dempsey 2008; Singh et al. 2011). As a comprehensive review is beyond the scope of this chapter, the following will present only a brief overview of the various medications used to treat challenging behaviors of children with ASDs. For more comprehensive, detailed reviews, the reader is directed to the following resources: Carlson et al. 2006; Handen and Lubetsky 2005; Matson and Dempsey 2008; Singh et al. 2011; Sweeney et al. 1998 .

SSRIs increasingly are used to treat repetitive behaviors in children with ASDs. Fluoxetine (Prozac^TM), fluvoxamine (Luvox^TM), sertraline (Zoloft^TM), citalopram (Celexa^TM), and venlafaxine (Effexor^TM) all have been reported in the treatment of children with ASDs, and results have been promising. Fluoxetine and fluvoxamine both have been reported to reduce repetitive, compulsive, and ritualistic behaviors (Peral et al. 1999; Posey et al. 2006). Fewer studies of SSRIs have reported positive effects on SIB. One report documented reductions in SIB following the use of fluvoxamine (Yokoyama et al. 2002), but its limited benefit may be offset by concerns over side effects (McDougle et al. 2000). Paroxetine also has been reported to reduce SIB (Posey, Litwiller et al. 1999; Snead et al. 1994). However, in many of these studies, reductions in stereotypy and SIB were modest and not all subjects responded positively to the medications (King et al. 2009). In many instances, adverse side effects were reported, including agitation, anxiety, insomnia, hyperactivity, restlessness, and appetite loss (Cook et al. 1992; Peral et al. 1999) .

One other medication that has received significant attention for treatment of SIB is naltrexone (Trexan^TM), an opiate antagonist. Naltrexone has been used under the hypothesis that SIB is maintained by automatic positive reinforcement in the form of the release of endogenous opiates (Sweeney et al. 1998). The action of naltrexone is to block the effect of the endogenous opiates, therefore removing the reinforcing consequence of SIB. In a review of the literature that included non-ASD subjects and adults, Symons, Thompson, and Rodriguez (2004) reported that 80 % of subjects evidence reductions in SIB following use of naltrexone, with 47 % demonstrating reductions as high as 50 % from baseline rates. Case studies of subjects with ASD have reported similar findings (Chabane et al. 2000), but these have not been supported by clinical trials (Campbell et al. 1993) .

More recently, atypical, or second generation, antipsychotics have become increasingly used for externalizing behaviors because of their demonstrated effectiveness and better adverse effect profiles. Included in this class of medications are risperidone, aripiprazole, quetapine (Seroquel^TM), olanzapine (Zyprexa^TM), ziprasidone (Geodon^TM), and clozapine (Clozaril^TM). Collectively, all of these medications appear useful in the treatment of externalizing behaviors in children with ASDs, but the research support often is inconsistent with as many as half or more of participants in some studies either showing no response to the medication, or experiencing adverse side effects leading to discontinuation of the medication (Handen and Lubetsky 2005; Matson and Dempsey 2008; Singh et al. 2011) . Risperidone and aripiprazole have the distinction of being the only medications in this class which are FDA approved for the treatment of irritability and aggression in children with autistic disorder, and have been the most researched. The evidence appears strongest for risperidone with numerous studies, including case reports, open label studies, and double-blind, placebo-controlled studies demonstrating significant reductions in irritability, aggression, tantrums, and other destructive and disruptive behaviors (Demb 1996; Findling et al. 1997; Shea et al. 2004). In a two-part study by the Research Units on Pediatric Psychopharmacology Autism Network (McCracken et al. 2002; RUPPAN 2005), children with ASDs taking risperidone showed significant improvement in measures of irritability, tantrums, and aggression over a placebo-control group. The improvements for the risperidone group persisted for 6 months, and relapse of symptoms was observed with the implementation of a placebo phase following the risperidone phase. Risperidone also has been used successfully to treat irritability and aggression in children with ASD as young as 2 years of age (Boon-Yashidi et al. 2002; Posey, Walsh et al. 1999). Few severe side effects have been reported in these studies, with the most common side effects being weight gain and sedation (Singh et al. 2011).

Aripiprazole also has been demonstrated to reduce irritability and aggression in children with ASDs (Owen et al. 2009; Stigler et al. 2004). Stigler et al. (2009) reported that 88 % of subjects with ASDs ages 5–17 years demonstrated improved symptoms on measures of irritability, aggression, and tantrums in a 14 week, open label study of aripiprazole. However, other reports have only documented improvement with aripiprazole in one third to half of subjects, and poorer outcomes for children diagnosed with ASDs, prompting calls for more well-controlled studies to support its use (Masi et al. 2009; Valicenti-McDermott and Demb 2006). Most studies report fewer and usually milder adverse side effects, such as agitation and sleepiness, with aripiprazole (Masi et al. 2009; Singh et al. 2011; Valicenti-McDermott and Demb 2006); however, more serious side effects, such as tremor, akathisia, and facial dyskinesia, have been noted (Marcus et al. 2009; Valicenti-McDermott and Demb 2006) .

In addition to antipsychotics, medications used to treat mood disorders in the general population, such as anticonvulsants and antidepressants, have been used to treat irritability and aggression in persons with ASD. Antidepressants would appear to be effective, low-risk options for children with ASDs; surprisingly very few studies have evaluated their use in this group. Clomipramine, fluoxetine, fluvoxamine, sertraline, and paroxetine all have been reported to reduce irritability, anger, tantrums, and aggression in children with ASDs (e.g., Gordon et al. 1993; Cook et al. 1992; Yokoyama et al. 2002; Steingard et al. 1997; Posey, Litwiller et al. 1999). While the results appear promising, additional well-controlled studies are needed to evaluate their short and long-term benefit. For example, in one study, Steingard et al. (1997) found reduced aggression for eight of nine subjects using sertraline, but these effects lasted only 3–7 months for three subjects.

Anticonvulsants are commonly used to treat mood instability for persons diagnosed with bipolar disorder. As such, they have been evaluated for mood symptoms, such as agitation and aggression, with children with ASDs. Valproic acid (Depakote^TM) is one of the more common anticonvulsants and has produced mixed results. Hollander et al. (2010) reported a 62.5 % positive response rate for irritability on the CGI compared with placebo using divalproex sodium. Helling et al. (2005), though, did not find clinical improvement on the irritability subscale of ABC for 30 subjects ages 6–20 years. Other anticonvulsants, including carbamazepine (Tegretol^TM), lamotrigine (Lamictal^TM), topiramate (Topomax^TM), oxcarbazepine (Trileptal^TM), and levetiracetam (Keppra^TM) have little research supporting their use in treating externalizing behaviors of children with ASD, and are associated with numerous adverse side effects (Belsito et al. 2001; Handen and Lubetsky 2005) .

Although not a core aspect of ASDs, challenging behaviors, including stereotypies, self-injury, aggression, and disruptive/destructive behaviors frequently are observed and likely have their origins in early childhood, but often are overlooked or dismissed as transient until they cause significant problems, such as injury to the individual or others, or damage to property. As a result, there is a growing emphasis on early identification and intervention of these behaviors for children at risk for or exhibiting early signs of ASDs, especially since these behaviors often persist into adulthood, interfere with skill development, and are associated with several adverse outcomes for the individual. Given their severe nature, children presenting with challenging behaviors should be evaluated as early as possible by both a behavioral specialist (e.g., psychologist or certified behavior analyst) and psychiatrist, both of whom should have expertise in assessing and treating children with ASDs. Behavioral and psychiatric assessment techniques have evolved in recent years that are sensitive in identifying the nature and function of these behaviors, which leads to more successful treatments. Furthermore, behavioral and psychiatric interventions are available, which have proven efficacious in the treatment of these behaviors in young children. However, more, well-controlled studies to substantiate the use of several medications with this population are needed. In addition, there is a need for research that evaluates combined and coordinated behavioral and psychiatric interventions, which findings from studies with others disorders (e.g., ADHD, depression) support as the most effective treatments.