ASD Intervention Study on Reinforcement Schedules

• Some children with autism spectrum disorders (ASD) display persistent errors.
• These errors are not responsive to commonly used prompting or error-correction strategies.
• One possible reason is that the behavior-consequence relations are not readily discriminable.
• The study increased the discriminability of the behavior-consequence relations in conditional-discrimination acquisition tasks for 3 children with ASD using schedule manipulations and a unique visual display.
• The visual display was designed to increase the saliency of the differences between consequences in effect for correct responding and for errors.
• A multiple baseline design across participants showed that correct responding increased for all participants after the intervention.
• After 1 or more exposures to the intervention, correct responding persisted to varying degrees across participants when the differential reinforcement baseline was reintroduced to assess maintenance.
• Increasing the saliency of behavior-consequence relations may help increase correct responding in children with ASD who exhibit persistent errors.
• Key words: acquisition, autism, conditional discrimination, discrete-trial training, discriminated operant, error correction, response cost, schedule discrimination, second-order schedule

Behavior-Acquisition Programs

• The central goal of almost all behavior-acquisition programs is to establish one or more discriminated operants.
• The discriminated operant has traditionally been considered the fundamental analytic unit for the scientific study of behavior.
• It consists of one or more antecedent stimuli that occasion a specific response, followed by the reinforcing consequence produced by the response.
• Discriminated operants always occur in contexts in which other responses produce other reinforcers, even when the experimenter arranges only a single operant contingency.
• The fundamental analytic unit should be the concurrent discriminated operant.

Discrete-Trial Training

• A typical discrete-trial training arrangement to teach conditional responding involves presenting a conditional discriminative stimulus (SD), and two or more response options are available.
• Each option is correlated with a different schedule, conforming to the concept of a concurrent discriminated operant.
• The concurrent schedules typically programmed during training are reinforcement (e.g., fixed-ratio [FR] 1) for the response corresponding to the conditional SD (or “correct” response) and extinction for any other responses (an “error”).
• It is somewhat of a misnomer to refer to responses that do not produce reinforcement as "errors."
• Skinner's dictum is that “The subject is always right implies that the trainer is responsible for errors, or instances of misarranged stimulus control.
• Instances of misarranged stimulus control can result from confusability at any point in the three-term contingency.

Confusability

• Errors may occur because alternative SDs are highly similar and easily confused.
• Errors may occur because responses are similar and easily confused.
• Errors may occur because reinforcement schedules are similar and easily confused.
• Confusability can occur at the level of the antecedent-behavior relation or the behavior-consequence relation.
• Consider a fixed-interval (FI) schedule in which the behavior-reinforcer relation involves a one-to-one correspondence between a single response and immediate delivery of the reinforcer.
• The FI that must precede that single response is often not highly discriminable.
• Delivery of the reinforcer increases resistance to extinction for all of the responses that lead up to that single response.

Spillover Effects

• These spillover effects could easily be eliminated by introducing an SD that effectively signals the completion of the FI.
• Most acquisition training procedures used with individuals with autism spectrum disorder (ASD) focus on reducing confusability among antecedents, behaviors, or antecedent-behavior relations, usually through the introduction and subsequent fading of controlling prompts.
• A controlling prompt is an antecedent stimulus that reliably occasions the correct response.
• An example involves teaching a child receptive identification skills using a progressive prompt delay.
• The trainer presents the conditional SD and immediately models the correct response, then progressively lengthens the delay between the conditional SD and the prompt.
• Similarly, controlling prompts have been used to teach correct response differentiation between similar responses, such as dance steps.
• Investigators have used a wide variety of other prompt and prompt-fading procedures to promote skill acquisition among children with ASD and developmental disabilities.
• Therapists should consider the discriminability of both antecedent-behavior relations and behavior-consequence relations when they attempt to analyze and modify discriminated operant responding in applied settings.

Saliency for Correct Responses and Errors

• Interventions that increase the saliency of the behavior-consequence relations for correct responses and errors should address error patterns.
• One set of training methods targets behavior-consequence relations to reduce errors during discrete-trial training, known as error correction.
• Research on error-correction procedures has produced inconsistent and idiosyncratic findings.
• An alternative approach to reducing persistent errors displayed by individuals with ASD is altering the training arrangement in ways that increase the saliency and discriminability of the behavior-consequence relations.
• Ways to increase likelihood that participants easily discriminate that only correct responses produce reinforcement and that error patterns produce near-zero obtained rates of reinforcement or an unfavorable consequence.
• The components were introduced as a package because each component was designed to support the central purpose of increasing the saliency of the concurrent schedules in effect for correct responses and errors so that reinforcement of correct responses did not produce spillover effects and also strengthen errors.

Methods to test the effectiveness of reinforcement

• Placing an edible reinforcer in one of three clear containers contingent on each correct response;
• Delivering the three accumulated edible reinforcers to the participant after each of the three containers held a reinforcer;
• Removing all accumulated reinforcers from the clear containers as a response-cost contingency for errors.
• The combination of the second-order schedule for correct responses and the response cost for errors meant that strings of 3 consecutive correct responses were required to produce reinforcement.
• The goal of the containers was to increase the saliency of the accumulation and loss of reinforcers for correct responses and errors, respectively.
• The combination of the second-order schedule and the response-cost contingency greatly reduced the probability that a response pattern that included errors intermixed with inadvertent correct responses would be followed by reinforcement.

Visual Display Study

• The visual display and schedules used were informed by token-reinforcement and response-cost procedures that used light displays to signal token accumulation and loss with pigeons.
• The study included three boys diagnosed with ASD, participating in an early-intervention program.
• The treatment package was evaluated using a multiple baseline design across participants.
• After the package resulted in relatively high and stable levels of correct responding, a reversal to baseline (baseline maintenance) was conducted to determine if correct responding would maintain at high levels after the treatment was withdrawn.
• Before the experiment, a paired-choice preference assessment was conducted to identify a set of high-preference snack items for use as reinforcers.
• Differential reinforcement of correct responses was initiated in new baselines.
• During baseline, the therapist presented the conditional SD at the start of each trial. Participants were given 5 seconds to respond, and the therapist delivered praise and a high-preference snack item contingent on each correct response.

Modifications

• During treatment, trials were conducted as in baseline except:
  • Three small clear containers were positioned in a horizontal array near the therapist.
  • The therapist presented the conditional SD at the start of each trial.
  • Participants were given 5s to respond, and the therapist delivered high-preference snack items according to a FR 3 FR 1 schedule.
  • Contingent on the first correct response, the therapist delivered praise and placed an edible reinforcer in the left-most clear container.
  • Contingent on the second consecutive correct response, the therapist added an edible reinforcer in the middle container.
  • Contingent on the third correct response, the therapist placed an edible reinforcer in the right-most container, and then delivered all three snack items to the participant immediately.
  • Contingent on an error, the therapist emptied collected snack items from the clear containers and then presented the next trial.
  • If containers were empty when a participant made an error, the therapist proceeded to the next trial.

Percentage of Correct Trials Results

• The percentages of correct responses during the baseline, treatment, and baseline-maintenance phases for the three participants were recorded.
• Zane displayed moderate levels of correct responding during the initial differential reinforcement baseline, and 100% of his errors were errors of commission(saying the wrong response).
• Bard displayed moderate levels of correct responding during the initial differential reinforcement baseline, and 92% of his errors were errors of commission.
• Don (bottom) displayed moderate levels of correct responding during the initial differential reinforcement baseline, and 67% of his errors were errors of commission.
• When the containers were arranged in a horizontal array, correct responses increased and stabilized at nearly perfect levels.
• When treatment was withdrawn, each participant's responding remained consistent.
• For each participant, the intervention initially produced a decrease in the rate and amount of reinforcement delivered.
• Each correct response produced reinforcement in baseline (100%).
• For all participants in the first few sessions they all emitted reinforcement in less than half the total responses.
• These data suggest that the response-cost contingency, the visual display, or both, may have played roles in the initial increases in correct responding.

Hedge's G Statistic

• Therefore, levels of correct responding were higher than during baseline.
• Levels of correct responding during the last five data points from the final baseline-maintenance phase were higher than during baseline.
• The change in the level of correct responding from the initial baseline to the initial treatment phase was significant.
• The current study evaluated an intervention to increase the saliency (or discriminability) of the concurrent schedules on conditional discrimination tasks for three children with ASD who had emitted persistent errors when similar tasks were exposed to a variety of prompting, differential reinforcement, and error-correction strategies.
• These treatment effects were maintained when the treatment was withdrawn.
• The long term maintenance suggests that that increasing the likelihood that only correct behaviors produce (FR 3 FR 1) and errors (response cost) during treatment for participant's discrimination of concurrent responses.

Discrimination

• Another procedure to increase discrimination speed accuracy is by emphasizing behaviors relating specifically to correct answers known as the differential outcomes procedure.
• Correlates each stimulus-response that occurs will determine unique reinforcer.
• Children are more likely to correctly respond when they feel correctly responding will "give me solder"

Limitations

• Limitations of the intervention process is that the number of discriminations that can occur are limited to the level of reinforcement at an equivalent value.
• The study's discriminability intervention was applied successfully with two, four and nine stimulation stimuli successfully.
• Developing new ways will allow us to separate relation problems between those that occur without issues with reinforcement methods to help target issues with relations specifically.
• Additional experiments in the future will study how the individual pieces of the current study effects and how these occur. This can be done be separating or limiting a patients access to said visual elements.
• The study does not implement an error-correction procedure although previously studies has shown that all three all individuals need this.
• The current study showed 3FR 1 produced more effective results and produced a behavior relation with signals when the conditions have been met.

Study on clear containers, stimulus presentation, and reinforcement schedules to improve correct responding with ASD children. The study also considered persistence, generalization, and maintenance of intervention effects. Results of intervention show enhanced saliency of behavior–consequence relations.