Use of a Token Economy to Treat Vocal Stereotypy In a Boy with Autism

Background: A core feature of autism is repetitive, nonfunctional behavior. A common form of this includes loud vocalizations referred to as vocal stereotypy. Although vocal stereotypy usually does not cause harm to the individual or others, it can be impairing to social, academic, and adaptive functioning. Vocal stereotypy can result in social isolation and placement in a more restrictive academic setting. Consequently, children with autism who engage in loud vocal stereotypy are often prevented from participating in regular education activities.  Vocal stereotypy may be automatically reinforced by the sensations that it generates in the individual. Previous research on the treatment of automatically maintained vocal stereotypy has shown that it can be treated.  However, many of the treatment options may not be feasible in the everyday environment. For example, research has shown that vocal stereotypy reduces when individuals engage in a preferred activity, such as listening to music, and subsequently lose access to the activity contingent on vocal stereotypy. However, this intervention is entails allowing constant engagement in preferred activities leaving no time for academic and daily living tasks. Thus, there is a need for data demonstrating successful treatment of vocal stereotypy that will easily transition to the everyday environment.

Objectives: The purpose was to examine the effects of a treatment package used to treat automatically reinforced vocal stereotypy in a child with autism without interfering with academic instruction or other daily activities.

Methods: Carl, a 12 year-old male with autism, exhibited loud vocal stereotypy which resulted in a more restricted classroom placement than his level of functioning required.  The use of a Token Economy plus Response Cost (TE+RC) arranged to allow for inclusion of academic instruction was implemented.  Access to the computer was identified as the reinforcer. During baseline, attempts to access the computer were blocked and there were no programmed consequences for vocal stereotypy. During treatment, tokens were earned for the absence of screaming for a specified time period.  Additionally, tokens were lost contingent upon vocal stereotypy.  The required number of tokens to access reinforcement, the length of the reinforcement interval, and the length of the specified time with no screaming were all gradually increased simultaneously.  The terminal goal was 30 min of academic instruction required to access 5 min of reinforcement.

Results: During baseline Carl engaged in high and stable rates of vocal stereotypy (M = 10.3 RPM).  During TE+RC, vocal stereotypy dropped to zero levels (M = 0.5 RPM).  When baseline was repeated, rates increased (M = 2.1 RPM).  When TE+RC phase was repeated, rates of vocal stereotypy remained at near zero levels. The token requirement increased from 1 to 15, the length of the DRO interval increased from 10 s per token to 116 s per token resulting in a 29 min work interval.  The reinforcement interval increased from 1 minute to 5 minutes. 

Conclusions: The token economy package was successful at reducing vocal stereotypy while allowing for everyday activities to continue.