Lecture Notes: Stimulus Control and Respondent Conditioning

Class Rep Announcements

• Discord Server:
  • A Discord server has been set up for students to contact the class representatives (specifically). One of the class reps is Natalie.
  • The link will be shared soon, possibly on Friday or whenever Michelle has time.
  • Purpose: To address course-related issues and communicate concerns to course coordinators.
  • Not for special conditions, which are handled separately.
• Memory Lab:
  • The memory lab is due on Sunday. Students who haven't started should begin working on it.
• Lecture Attendance:
  • This cohort has one of the highest lecture attendance rates in a long time.
  • Students are encouraged to take breaks and treat themselves.
• Lab Access Over Holidays:
  • Psychology students can access labs during scheduled lab times over the holidays, even if there are no scheduled lab sessions.
• Study Groups:
  • There is interest in study groups for upcoming tests. Cohort forward will start those, potentially in week seven.
  • A form will be sent out to determine availability.
  • AUPSA will support the study groups with snacks, drinks, and potentially senior students.

Addressing the Class

• Gratitude for Attendance:
  • Appreciation for students attending the lecture, especially given the test tomorrow.
  • Attendance makes interactive lectures easier.

Recap of Previous Lecture: Intermittent vs. Continuous Reinforcement

• Continuous Reinforcement: Reinforcing every response.
• Intermittent Reinforcement: Reinforcing some responses according to a rule (schedule of reinforcement).
• Resistance to Extinction:
  • Intermittent reinforcement produces behavior more resistant to extinction.
  • Example: Rats intermittently reinforced (30% of the time) took longer to extinguish behavior compared to those continuously reinforced.
  • Explanation: Transition from reinforcement to extinction is harder to detect with intermittent reinforcement.

Schedules of Reinforcement

• Rules to specify how intermittent reinforcement is administered.
• Ratio Schedule: Based on the number of responses.
  • Fixed Ratio: Reinforce every nth response exactly.
  • Variable Ratio: Reinforce on average every nth response.
• Interval Schedules:
  • Fixed Interval (FI): A response is reinforced when a fixed amount of time has elapsed since the last reinforcer; a response must be made after the interval.
    • Example: FI-15 second schedule: After 15 seconds, the first response is reinforced.
    • Pause After Reinforcement: Animals pause after reinforcement, then accelerate responding until the next reinforcement.
    • Explanation: Not due to consuming the reinforcer (as variable interval schedules don't show pauses).
    • The pause is due to the predictability of the next reinforcer.
      • Animals learn they have to wait before responding.
      • Variable interval schedules don't have this predictability, so responding starts immediately. This is an example of using a scientific experiment to tease apart two possible explanations for a phenomenon.
• DRO Schedules (Differential Reinforcement of Other Behavior):
  • Distinction from fixed interval and fixed time schedules is essential.
  • Fixed Interval (FI): Response is reinforced when time has passed (must respond).
  • DRO Schedule: Every response restarts the timer (must not respond).
  • Fixed Time (FT): Reinforcer occurs after a time interval, regardless of response (doesn't matter whether you respond or not).
  • DRO schedules are effective for eliminating unwanted behavior: both negative punishment of unwanted behavior and positive reinforcement of all other behaviors.
  • Important to offer an alternative behavior to earn reinforcements.
• Fixed Time Schedules (FT):
  • Can be used to study superstition experimentally.
  • Skinner's Pigeon Experiment: Pigeons exhibited idiosyncratic behaviors before the reinforcer arrived.
  • Law of Effect: Behaviors followed by reinforcement are repeated, even if they don't cause the reinforcement.
  • Temporal Association: Close timing between response and reinforcer is key.
  • Human Superstitions:
    • Similar experiments work with humans (Wagner and Morris's clown experiment).
    • Most superstitions are about negative reinforcement (avoiding bad outcomes).
    • Irrational beliefs can be interpreted as accidental reinforcement of behaviors.

Stimulus Control of Operant Behavior

• Focus shift: From consequences to stimuli that precede behavior.
• Operant behavior is behavior controlled by its consequences. Stimulus control of operant behavior refers to how stimuli that precede the behavior affect the probability of that behavior.
• We don't always behave the same in different circumstances.
• Stimulus context indicates the consequences for responding.
• The law of effect can be refined: When a response is reinforced, it becomes more likely in the same stimulus context.
• Changing the stimulus context changes the probability of the response because different stimulus contexts signal different consequences for responding.

Examples of the Stimulus Context

• Example (Petting a Kitten): Approaching and petting a cute kitten is reinforced (pleasant experience).
  • Generalization: This behavior generalizes to other animals based on similarity.
  • Probability of petting depends on the similarity to the kitten.
    • High probability for similar cats.
    • Less probability for a possum.
    • Even less for a ferret or wolf.
  • Positive Punishment: If the cat bites or scratches, it becomes less likely I will approach it again. This would also tend to generalize.

Studying Stimulus Control

• Training a Pigeon:
  • Train a pigeon to peck a yellow key for food reinforcement (VI schedule).
  • VI schedules maintain a high, constant rate of responding.
  • Change the stimulus to an orange-colored key.
    • Prediction: Generalization will occur because orange is similar to yellow, but with less behavior.
    • If changed to a blue key the behavior would most likely drop further.

Generalization Gradients

• Graphs showing the amount of responding as a function of a stimulus dimension (e.g., color).
  • Training Stimulus: S+ (positive discriminative stimulus), signals reinforcement. Pecking is reinforced.
• Decremental Gradient:
  • Shape: Peaked, with most responding at the training value.
  • As you move away from the original trained stimulus context, the less responding you should see. The less it should generalize.
• Incremental Gradient:
  • If pigeons received electric shocks for pecking the yellow key they would no longer pick it.
  • This tendency not to pick it would generalize, leading it to peck an orange key more than yellow, and so forth.

Terminology Review

• Decremental Gradient: Most responses at the original training stimulus (S+).
• Responding generalizes to similar stimuli.
• Amount of generalization depends on similarity.
• Stimulus Control: The extent to which stimuli that precede or accompany operant behavior affect the rate or probability of that behavior.
  • Less generalization = more stimulus control.
• Discrimination: Behaving differently in the presence of different stimuli, change the context and observe whether you get different behavior.
• Generalization and discrimination are opposites.
• Stimulus control means lots of discrimination, not much generalization.

Real Data: Guttman and Kalish (1956)

• First to demonstrate generalization gradients empirically.
• Technological Reason: No reinforcement during testing with new colors.
• If reinforces are given during testing, the animal will ignore the colors because they will get reinforced regardless of what the color is.
• Tests must be in extinction to see the effect of previous training without further reinforcement.
• Partial Reinforcement Effect: Use intermittent reinforcement to extend behavior during extinction.
  • Trained picking yellow key on a variable interval schedule, then tested with other colors.
• Four Groups of Pigeons: Trained with different S+ (different colors of key).
• Trained at four different training values.
• If they had only trained one group, it could be explained why they preferred a color from previous training. It was important they train different groups at different training values.
• When they were trained at certain nanometers and certain colors, that is what they picked at most. When the stimulus changed, responding decreased.
• Large stimulus control (steep gradients): Relatively small color changes produced big decreases in behavior. Pigeons are pretty good at this.
• Pigeons have better color vision so they have good at determining if the colors are different.

Summary of Stimulus Control of Operant Behaviour

• Nothing is learned in isolation. We always learn things within a particular stimulus context.
• Change the context, change the behavior in a systematic way.
• Stimulus Control: Less generalization, more discrimination.
• ABC of Behavior:
  • Antecedents (stimuli before behavior) set the occasion for responding.
  • Behavior (responses) followed by consequences (reinforcers).
• Example: Actively train discrimination where responses to a red key are reinforced, but not to a green key.
  • Stimulus signals the consequence for responding.

Interesting implication of Stimulus Control

• Learning is likely to improve by recalling material within the same stimulus context.
• Experiment: Godin and Badley (underwater learning).
  • Tested whether learning a list of words underwater or on land affected recall.
  • Words were recalled better if they were being tested in the same context that they were learned.
  • Experiment failed to demonstrate this.
• Training generalisation: It is important to train for generalisation in applied behaviour analysis.
• The way different contexts are distinguished needs to be trained.

Respondent Conditioning (Classical Conditioning) : Another way Stimuli Affect Behaviour that isn't Signaling the Consequences of Responding.

• Stimulus control of operant behaviour: consequences are signalled by stimuli.
• Respondent conditioning: based on the association between two stimuli.
• Pavlov’s dogs

Definitions for Pavlov's Dog Experiment

• Unconditioned Stimulus: A stimulus that automatically elicits an unconditioned response.
  • Example: Meat for a dog elicits salivation.
• Unconditioned Response: Example: salivation
• Previously Neutral Stimulus: Might be paired with the unconditioned stimulus to reliably predict it.
  • Pavlov would ring a bell before giving the dog food.
  • Ring the bell before the food.
  • The Bell did not elicit salivation but elicited salivation once they realised food was coming.
• Conditioned Stimulus: Neutral stimulus becomes a Conditioned Stimulus

Details About Pavlov's Dog Experiment

• US (Unconditioned Stimulus): food, which reliably elicits an Unconditioned Response of salivation.
• Ring a bell just before presenting the food.
  • Initially, bell does nothing.
• Bell had become a Conditioned Stimulus(CS) that elicits salivation (Conditioned Response = CR).
• Conditioned stimuli elicit conditioned responses, while unconditioned stimuli elicit unconditioned responses.
• The Bell is NOT signaling. It is a different mechanism here.
• Learned behaviour is when an animals learn to predict future events.
  • In Pavlovian conditioning, classical conditioning, the bell predicts that food is coming, the dog can use the bell to predict the future.
  • Pause after reinforcement is something animals uses to predict the future for something as simple as food.

Short Video Demonstrations from other Modules

• Classical conditioning can be strong enough to make you sick by influencing your immune system.
• Classical conditioning role in phobias.

Video Demonstration 1: Immunity Suppression Example

• Philip Zimbardo
• People can unintentionally condition the rat's immune response.