PSYC2050

What is Learning?

There is a change, which may be invisible e.g. "behaviour potential"
The change is lasting
There is either experience and/or practice
Learning a situation is important (one situation/ context may seem more safe/dangerous than another)

What are the 2 major ways of learning?

1. Non-associative Learning
2. Associative Learning

Non-Associative Learning

Habituation- An innate response to tendencies (reflexes, taxes, instincts)
Maturation- Regular stages unaffected by practice
Fatigue- Changes due to physiological/ motivation states or evolution (disappears after break.

Habituation

Is a "getting used" to a novel stimulus. When a person attends to a novel stimulus they learn that is has no special significance and they no longer have an orienting response.

Why is habituation adaptive?

It allows us to learn that a novel stimulus is not significant and therefore you don't have to be distracted by petty events. E.g. a continuous noise in the background will at first distract, but then it won't be heard anymore.

What is Sensitisation?

Learning to intensify a response to stimuli that previously evoked little response. E.g. rubbing your arm is harmless but it will become heated and painful the longer it occurs.

Associative Learning

We measure a behaviour to infer learning, which makes it limited to observable effects.
Behaviour is caused by: The Organism, Environmental Demands, and Internal States
Small units of behaviour follow the same laws as complex behaviours.

Definition of Cognitive Psychology

The study of Mental Processes such as perceiving, attending, remembering, and reasoning. Cognitio= To know or to think.

What are the differences between Classical and Operant Conditioning?

Classical Conditioning: Ivan Pavlov, it's learning via Association.
Operant Conditioning: B.F. Skinner, it's learning via Reinforcement.

4 Elements of Pavlovian Conditioning

Unconditioned Stimulus (US)- A stimulus that elicits an unlearned response
Unconditioned Response (UR)- The unlearned response to a US
Conditioned Stimulus (CS): A stimulus to which an organism must learn to respond
Conditioned Response (CR)- The response to a CS: which is learned

With Pavlov's dogs, what other stimuli did they get a CR to?

Footsteps of the experimenter, Changing of the food dish etc.

Pavlov's Systematic Study of Learning

Started with the bell alone- no salivation
Pair the bell with food several times.
Present the bell alone again- salivation

A Typical Experiment: Habituation

An organism's decreasing response to a stimulus with repeated exposure to it. An adaptive form of learning that results from prolonged exposure to a stimulus and it is learned that the stimulus is no longer significant

A Typical Experiment: Acquisition

Pairing the US and the CS results in an increase in the CR with continuous pairing (US + CS)

A Typical Experiment: Extinction

As CS is presented multiple times without US, the CR is extinguished (CS alone).
CS becomes ambiguous.

Little Albert

Study conducted by Watson and Rayner 1920
Acquisition of Emotional Responses- Conditioned Fear
The US (Loud Noise) and the UR (Being Upset) was paired with the CS (A white Rat)
This resulted in Generalised Fear of all white, furry things.

Two Factors that Influence the Acquisition Curve

Intensity of the US- More intense= More rapid learning
Order and Timing- The CS presented before the US= Better

Interstimulus Interval (ISI)

The period of time where the CS is presented alone before the US is paired with it.
The optimal ISI is dependent on what is being learned. Humans- 0.4 to 1 secs

Delay Conditioning- SHORT

The CS is presented for a short amount of time, by itself, before being paired with the US.

Delay Conditioning- LONG

The CS is presented for a long amount of time by itself before being paired with the US.

Trace Conditioning

The CS is presented alone and then there is a period of time called the trace interval, where no stimulus is presented and then the US is presented alone. At no point do the CS and US coincide, but the CS always precedes the US. The longer the Trace interval, the weaker the conditioning.

Simultaneous Conditioning

The CS and US are presented simultaneously together and it is the worst form of conditioning.

Backward Conditioning

The US is presented before the CS. No conditioning can really happen, but occasionally there is the smallest amount of conditioning/ learning.

Intertrial Interval (ITI)

The period of time between trials in an experiment; the longer the better.

What is the Optimal ISI for Eyelid Reflex?

300 ms= 80% to CR

What is the Optimal ISI for Taste Aversion?

30 mins= 7.5 to Aversion

2 Types of Pavlovian Conditioning

Excitatory Conditioning: CS predicts the occurrence of US.
Inhibitory Conditioning: CS predicts the absence of US.

2 types of tests for it to be an Inhibition

Retardation Test
Summation Test

Retardation Test

To make sure an inhibitor is an inhibitor.
If a CS- indicates the lack of a UR, and then another CS is paired with the CS- and there is a UR. The CS- = Inhibitor.
E.g. if you have never gotten sick from eating capsicum and then you eat capsicum and onions together (and you've never had onions before) and you get sick. The Capsicum is a CS-.

Summation Test

Presenting an excitatory CS with an inhibitor creates a plus minus effect and therefore elicits a smaller response than if the excitatory CS was presented alone or with a neutral stimulus.

Excitatory Conditioning

Cs predicts US; CS usually elicits CR.

Inhibitory Conditioning

CS predicts the absence of US, so there is no CR (This requires tests)

Extinction and Re-Acquisition

Extinction: Learning that CS no longer indicates CR.
Re-Acquisition: After extinction, Learning that the CS= CR. The process is faster.

What happens during extinction?

Forgetting: passive decay (incorrect)
Unlearning: active (correct)

Spontaneous Recovery

If the CS is reintroduced after a "break"/ extinction, then the CR can reappear.

Renewal

When extinction is context specific.
If acquisition is in context x, and extinction is in context y, then if CS is presented in context x, CR reappears.

Reinstatement

AKA: Reminder Effect.
When the US is presented alone after extinction, then the CS is presented it leads to the CR.

What does Bouton 1994 says happens to the CS after extinction?

CS becomes ambiguous.

CS preexposure: Latent inhibition

When your exposed to a stimulus before you're conditioned to it, it retards learning.

Blocking

We don't learn about a novel CS that is paired with an excitatory CS. i.e. which is already very predictive of the US

What was Kamin's (1968) experiment, how did it show blocking?

Rats were divided into control and blocking groups. In the control groups, rats saw both a light and heard a noise then got shocked. This was repeated until they developed a CR. They reacted to just the light by trying not to get shocked.
In the blocking group, rats heard a noise and then got shocked until a CR developed. Rats then heard a noise and saw a light, then got shocked. They were tested to see if they reacted to the light. They didn't= Blocking.

How does the study by Kamin (1968) prove assumptions of conditioning wrong?

1. Any stimulus can't be paired with any response as the pairing of the light and shock didn't lead to conditioning for the light.
2. The more two stimuli are paired does not mean that the association will be stronger as both the blocking and control group had the same amount of light/shock pairing.
3. Conditioning does not change trial to trial in a regular way as the rats simply ignored the light.

Superconditioning

Learning is faster if a novel stimulus is trained together with an inhibitory stimulus. i.e. which is very predictive of the absence of the US

What was Rescorla's (1971) study, how did it show supercondtioning?

A tone is played and there is no response. If the tone is then paired with a light and there is a response, i.e. shock, the rat unusually quickly associates the shock with the light= superconditioning.

Contiguity

Events occur together- in time or space.
Mere pairing of two closely timed or located events should result in learning.

Contingency

Co-variation of events, based on paired and isolated presentations= pairing things is not enough.
Conditioning changes trial to trial in a regular way.

Pavlovian Conditioning Generalisation

Other similar stimuli may also produce the CR. The more similar to the original CS, the more likely it is to elicit the CR.

Discrimination: Does Generalisation Last?

Early on in acquistion, generalisation may cause the learner to respond to a variety of stimuli. As learning continues, they can discriminate between each similar stimuli.

Rescorla- Wagner Model

The level of conditioning is a result of an internal comparison between the expected and actual strength of the US.
Expectation is based on prior experience with the US- previous trials.
Strength of the US is fixed, e.g. Mild Shock.

When does the Rescorla-Wagner Model suggest that the CR becomes stronger?

It assumes that a CR gets stronger if the CS and US pairing is SURPRISING.

What is the explanation by the Rescorla-Wagner Model for blocking in Kamin's (1968) study?

The light doesn't change how surprised the rat is at being shocked, the rat ignores the light because it already expects the shock to happen because of the noise.

What is the explanation by the Rescorla-Wagner Model for superconditiong in Kamin's (1968) study?

If light is not followed by a shock it will be associated with the absence of shock. However, the rat will be surprised it got shocked after a light and tone was presented, therefore making the learning process faster.

What does the Rescorla-Wagner Model have difficulty explaining?

It has difficulty explaining Configuring and CS Pre-Exposure.

What does the Rescorla Wagner Model assume about Classical Conditioning?

It assumes that classical conditioning is about information processing rather than purely associations.

What did Mackintosh say about Predictiveness?

If a stimulus is better correlated with reinforcement than the other stimuli that accompany it, then attention to the stimulus will increase and attention to the other stimuli will decrease.

Predictive Error (Pearce & Hall, 1980)

Stimuli that accurately predict the events that follow them will receive little attention, whereas those that inaccurately predict these events will be paid considerable attention.

Taste Aversion

Unlike regular classical conditioning, it only takes one single association to acquire an aversion. The association can be delayed for up to 24 hours.

Why is Taste Aversion different to other Classical Conditioning?

It takes a while to feel sick after eating food
It's important to know which foods are bad as soon as possible (only needing one association)

Garcia Effect- Preparedness

Some associations are learned faster than others.

How does classical conditioning affect your tolerance to heroin?

When taking heroin, there are usually similar environmental cues. The endocrine system associates these cues with taking heroin and ramps up the tolerance response. If these cues are not present, the body isn't as prepared to tolerate the heroin, making overdose more likely.

What is the Heroin Overdose Experiment?

They gave heroin to rats in the same or different room, or a control where they just had heroin. The mortality rate for the same room when taking heroin was 32.4% which was less than if the heroin was taken in a different room, 64.3%. Context Conditioning affects tolerance.

Systematic Desensitisation

In clinical psychology, when the US is presented without the CS.

Why do Phobias persist?

People avoid the situations where they see the phobic object, but that way they can't use systematic desensitisation.

Fear Hierarchy

Participant is given relaxation training.
They are gradually exposed to the things higher on the fear hierarchy as they relax.
After 3.5 years, 70% of people systematically desensitised were no longer fearful of flying.

What is the difference between Operant and Pavlovian Conditioning?

Pavlovian conditioning is involuntary--reflexive-- the stimulus precedes the response, while the response comes after the stimulus, and the role of the learner is passive. Whereas, Operant conditioning is usually voluntary, the stimulus is presented after the desired response, and the response happens before the stimulus, the role of the learner is active.

Operant Conditioning

Actions that result in a reward tend to be repeated or become more frequent.
Actions that result in punishment tend to be avoided or become less frequent.

An Example of Early Life Operant Conditioning

"If you are a good girl/ boy, you can have an ice-cream".
"If you are a bad girl/boy, you can get a smack".

Thorndike's Methods

Cats were put inside a puzzle box. They could escape by pulling on a string, stepping on a platform, and turning a latch on the door. Cats became quicker at this with experience.

What are the 2 options that will get an animal to do something.

1. Wait.
2. Reinforce any behaviour that could lead to desired behaviour.

Shaping

Reinforcement of any behaviour that leads/ or is similar to the target behaviour.

Superstitious Behaviour

If a pigeon or other being is randomly rewarded, they will start to assume that the things they do are the cause of the rewards, even if there is no actual causation.

How is Superstitious Behaviour seen in Golfers (Lee et al., 2011)?

Golfers were told they were playing with a putter belonging to a famous golfer are significantly more accurate than golfers who were not told this.

Chaining

Acquiring a behaviour is easier if done in smaller bits and pieces. This can be done forwards or backwards. It is best to start with the last behaviour in the chain and work backwards.

Three Term Contingency

The stimuli.
The response.
The stimuli that follows the response- the reinforcer.

Reinforcer

Increases a Behaviour

Punisher

Decreases a Behaviour

Positive Reinforcement

Adds something to increase a behaviour. E.g. Rewarding a child with a lollipop for good behaviour.

Negative Reinforcement

Removes something to increase behaviour. E.g. Cleaning the house because your mum was nice to you.

Positive Punishment

Adds something to decrease behaviour. E.g. Smacking a child for bad behaviour.

Negative Punishment

Removes something to decrease behaviour. E.g. Taking away a toy because the child was naughty.

Out of punishment and reinforcement, which is more effective?

Reinforcement is more effective than punishment.

Problems with Punishment

It isn't as permanent as reinforcement.
Individuals are more motivated when they're being reinforced.
Punishment reduces trust and increases aggression.

How to Punish Effectively

No Escape.
As intense as possible- with limits.
Continuous schedule.
No delay.
Over a short period.
No subsequent reinforcement.

What side effects should be watched when using Punishment?

Changes in other behaviours.
Aggression.
Fear.
Modelling of Violence.
Learned Helplessness.

Schedules of Reinforcement: Continuous

Reinforce each response.

Schedules of Reinforcement: Intermittent/ Partial PRF

Reinforce only some responses.

Schedules of Reinforcement

Fixed Ratio: every nth time
Variable Ratio: On average, every nth time
Ratio Schedules= Most Efficient
Fixed Interval: First after n seconds- scallop effect
Variable Interval: On average, first after n seconds

Reward Variables

DRIVE: Reinforcement depends on how much the organism wants the reinforcer. E.g. hungry vs sated organism.
SIZE: Animals in a Skinner box learn faster if they get more food, but diminishing returns.
DELAY:

Stimulus Control

The behaviour happens when the stimulus is present and doesn't happen when the stimulus is absent.

Stimulus Discrimination

The tendency to only respond to the stimuli that has been reinforced.

Stimulus Generalisation

The general tendency to respond to similar stimuli that have been reinforced.

Drive Reduction Theory

Reinforcers maintain physiological homoeostasis.
Drives need to be satisfied.
Stimulation needs to be reduced.

Behaviour Regulation

Behavioural Homoeostasis- not food (stimulus), but eating (behaviour) is reinforcing.

Premack Principle

A high probability behaviour can reinforce a low probability behaviour.

How does Brown, Spencer, & Swift (2002) use the Premack's Principle?

Their patient refused to eat all buy a few specific foods.
Low probability behaviour: eating new foods.
High probability behaviour: eating favourite foods.
He was told that if he ate a small amount of new foods he could eat some of his favourite foods- the boy began to eat more new foods.

Escape Learning

Emit a response that terminates an aversive consequence- negative reinforcement.

Avoidance Learning

Emit a response to prevent the occurrence of an aversive consequence altogether.

Avoidance Learning: The Two Factor Model

Emit a response to prevent the occurrence of an aversive consequence altogether.
On way Avoidance.
Present CS (tone) followed by US (shock)
Initial trials feature escape, then avoidance takes over.

One Way Avoidance (Avoidance Learning)

Easy: faster with intense aversive stimuli, different compartments and the subject spending long time in a safe compartment.

Two Way Avoidance (Avoidance Learning)

Difficult: faster with weak aversive stimuli and similar compartments. E.g. phobia- avoid stimulus, reduces fear= negative reinforcement.