L6 - The Content of Condtioning Flashcards
Why must experimenters be careful in designing conditioning experiments designed to be purely Pavlovian or instrumental?
The association the animal learns may not be the one the experimenter intends. Crossover between classical and istrumental.
What are the four types of outcome that can occur when PavlovianxInstrumental design is crossed with Pavlovian/Instrumental learning?
LEARN BRACKETS BITS
- Pavlovian design, Pavlovian learning (pure classical conditioning)
- Pavlovian design, “instrumental” learning (Superstition)
- Instrumental design, instrumental learning (pure instrumental conditioning
- Instrumental design, Pavlovian learning (implicit Pavlovian learning)
What test can be used to determine if a response is instrumental or Pavlovian?
The omission test can be used.
How does the omission test work?
The CR must now prevent the US from occurring. A true instrumental response will extinguish quickly, because it is within the animal’s control. A Pavlovian response will not pass the omission test - the animal will continue producing it reflexively, because it is outside conscious control.`
What paradigm is commonly used to demonstrate that a response which appears instrumental can actually be Pavlovian?
Autoshaping in pigeons. Pigeons will learn to peck at a light that predicts food. However, this behaviour fails the omission test, indicating it is a Pavlovian reflex.
Auto-shaping in pigeons
The pigeon will start pecking at the light in expectation to being delivered food.
Long-box in pigeons
Long box, with light at one end and food delivered at other end. The pigeon still feels a need to run to the light when it comes on and peck at it, and then bolt to the other end to receive the food.
Omission in pigeons
If the pigeon pecks the light that will cancel the food on that trial. The pigeon can’t stop pecking, so this behaviour fails the omission test (if this was a truly instrumental response the pigeon should have no trouble omitting the response)
Do Pavlovian CRs and Instrumental CRs pass or fail the omission test?
Pavlovian CRs FAILS omission test - subjects can’t ‘control’ their response.
Instrumental CRs PASS the omission test.
What feature of conditioned responses allowed the CS-UR theory to persist?
In many cases, the CR is equivalent to the UR. Thus it appears that a simple S-R link between the CS and UR is being formed.
List three phenomena that can’t be explained by S-R learning.
- conditioning occurs even when responses are prevented during training
- devaluation of the US affects responding to the CS. Responding depends on value of US at time of testing, not initial training
- mediated conditioning. punishing a CS can reduce responding to the US
Why can it be difficult to disentangle S-S and S-R learning? According to Holland
- A single stimuli can consist of multiple features.
- Conditioning probably happens on multiple levels.
An animal probably acquires multiple S-S connections with regard to several features of a CS, and S-R connection with regard to its behaviour in response to that stimulus
Although CR’s are useful indexes of conditioning, what is their limitation?
CR’s don’t indicate the content of what is learnt.
Give an example of different CS’ to the same US producing different CR’s.
Rats given shocks will react differently depending on the CS.
- if the CS is diffuse, like a noise, rats will freeze
- if the CS is localised, like a probe, rats will withdraw from it or attempt to bury it
This is because it is not about the sensory qualities of the US, motor response
How did Rescorla show that a Pavlovian CS can influence instrumental behaviour?
- in the first phase, Rescorla classically conditioned CS1 (sucrose) and CS2 (food) to two different stimuli, US1 and US2.
- the second, instrumental phase, Rescorla trained rats to pull a chain for US1 and press a lever for US2
- in the third phase, Rescorla presented CS’ 1 and 2. CS 1 increased chain pulling, and CS 2 increased lever pressing - if you play c1 they then pull the chain more, reminds them of the action.
