Lecture 9: Chaining and Differential Reinforcement

increasing behaviour:

positive reinforcement, shaping, fading, chaining (today), stimulus discrimination training, and schedules of reinforcement

Decreasing behaviour :

operant extinction, punishment (next week), antecedent control procedures (final week), and differential reinforcement procedures (today)

Behavioural chains

• Behavioural Chain aka stimulus-response chain

  • A sequence of discriminative stimuli (SDs)and responses (Rs) in which:

    • Each response except the last produces the SD for the next response

    • The last R is followed by the reinforcer.

• Each SD in a behavioural chain is:

  • A cue for the next response

  • A conditioned reinforcer for the previous response (other than the first SD)

Stimulus Response Chain

• Each response produces a cue for the next response

• SD 1 —> R1 —> SD 2 —> R2 —> SD 3 —> R3…—> S+

• S+ = reinforcer at the end

• Stimulus-response connections are “links” in the chain

• If any response is not produced by SD before it, the chain will be broken

• “A chain is only as strong as its weakest link”

Chaining, Fading, and Shaping Compared

• All are gradual change procedures

  • Gradually, through a series of steps, all produce new behaviour

• Shaping steps:

  • Reinforcement of closer and closer approximations

• Fading steps

  • Reinforcing the final desired response in the presence of closer and closer approximations to the final desired stimulus for that response

• Chaining steps

  • Reinforcing more and more of the specific stimulus-response links

Chaining Methods

• Total Task Presentation

• Backward Chaining

• Forward Chaining

Total Task Presentation

• Attempt all steps of the chain at once

• Continue until every step is mastered

• Assistance provided as needed

• Reinforcement provided upon completing last step

Backward Chaining

• Last step first

• Then, next-to-the last step is taught and linked with last step

• On each trial, last step in the chain is reinforced

Forward Chaining

• Initial step first

• Then, first and second… etc.

• On each trial, last step that was developed is reinforced

Effective Behavioural Chaining

• Do a task analysis: Identify the components of the final sequence

  • Task analysis – process of breaking a task down into smaller steps or component responses to facilitate training

  • Components should be selected so that there is a clear stimulus signaling the completion of each component

  • Review each of the controlling stimuli (SDs) for each of the responses in the sequence

    • If similar stimuli control different responses, there is more chance for error and confusion

Task analysis

Process of breaking a task down into smaller steps or component responses to facilitate training

Influencing the Effectiveness of Behavioural Chaining

• Consider strategies for independent use of prompts by learners

  • Written task analysis

  • Picture prompts

  • Reciting self-instructions

Guideline for Effective Application of Behavioural Chaining

1. Do task analysis

2. Consider strategies for independent use of prompts by learners

3. If necessary, do a preliminary modeling trial

4. Decide on method and teach the units in the proper sequence

5. To expedite learning, use fading procedures to decrease extra help that may be needed by client to perform some of the steps

6. If using backward or forward chaining, make sure that on each trial the learner performs the entire set of steps learned up to that point

7. Early in training, use ample reinforcement; decrease gradually

8. Make sure that reinforcer used is effective

Differential Reinforcement of Low Rates (DRL)

• Reinforcing only low rates of responding

• Behaviour will occur at low rate

• Limited Responding DRL

  • Maximum number of allowable responses during an interval

  • Useful when:

    • Some of the behaviour is tolerable

    • Less is better

• Spaced-Responding DRL

  • Behaviour does not occur during an interval, after interval passed, behaviour must occur for reinforcement

  • Behaviour spaced over time

  • Want to reduce a desirable behaviour (a little is desirable)

Differential Reinforcement of Zero Responding (DRO)

• Reinforcer presented only if response does not occur within an interval

  • If response occurs, timing of interval starts again (resets)

  • aka Differential reinforcement of other responding – reinforcement for any other behaviour

  • Length of interval increased gradually until:

    • Behaviour occurring very rarely or not at all

    • A minimum amount of reinforcement is given for nonoccurrence

Differential Reinforcement of Incompatible Responding (DRI)

• Similar to DRO, but specifies an incompatible response

  • Undesired response is not possible when engaged in the incompatible response

  • Reinforce incompatible response

Differential Reinforcement of Alternate Behaviour (DRA)

• Extinction of undesirable behaviour

• Reinforcement of other, topographically different behaviour (not necessarily incompatible one)

guidelines for effective use of DR procedures

• Decide which type of differential reinforcement procedure will be used

• Decide which reinforcer to use

• Record baseline data

• If possible, inform the individual of the procedure to be used

• Follow the selected procedure