4.2 Contingency & Conditioning Flashcards
Is TC sufficient or necessary for learning?
No - the CS and US is not required to be co-occurring together in time for us to learn something about it
TC is not necessary for learning eg. Food aversion
Give examples of TC being insufficient for learning
TC has to be present and no learning happen for it be considered insufficient
Overshadowing and blocking are examples of TC being insufficient
Define contingency
The relationship between 2 events
What did the results of Rescorla’s truly random control show?
Due to the extra shocks that occur when CS isn’t present, we don’t get conditioning
Suggests the occurrence of the CS must be contingent on the presence of the US
Causality driving conditioning
What does contingency imply?
Contingency implies causality
Conditioning is an attempt of animals to understand causal relationships in the world
When the CS does not “cause” the US = no causality, no contingency & no learning
When does conditioning occur in terms of contingencies?
Conditioning occurs when there is a +ve contingency
We care about what happens when the CS is present and when it is absent
What is it called when the probabilities of CS and US occurring are equal?
Zero contingency
Who came up with a way of systematically investigating contingency?
Rescorla
How did Rescorla (1968) show how contingency affects conditioning?
Rats on PRft schedule - noise followed by shock on 40% of CS presentations
CS and US do not always occur together
Systematically varied the rate of US occurrence in the ABSENCE of the CS
In instances where no CS, there were 4 groups:
0 group = no CS and no US (No shock when CS is absent)
- 1 group = no CS, US 10% of time
- 2 group = no CS, US 20% of time
- 4 group = no CS, US 40% of time (4 more shocks when CS is absent)
What does a suppression ratio of 0 imply?
You have learned a lot
What gives the best learning in terms of the probability equation?
The largest difference between the 2 sides of the probability equation gives the best learning
Which group in Rescorla’s 1968 experiment learnt the most about contingency?
0 group learnt the most about the contingency - experienced no shocks when CS is absent so learnt the causal relationship between CS and US
Which group in Rescorla’s 1968 experiment learnt the least about contingency?
0.4 group learnt little because US occurred without CS so causal relationship is less clear - do not stop pressing the lever as no conditioned anxiety experienced
Does it matter how many shocks are presented?
No - the absolute value of the shocks is not what leads you to being afraid
Depends on what you are specifically afraid of i.e. the CS
Describe the experiment that shows what happens if probability of US in presence of CS is less than probability of US in the absence of CS.
The setup is that whenever there is a noise, there is never a shock that appears at the same time
Only in the absence of noise do you get a shock
Explicit -ve contingency that is being set up
Under the conditions of an explicit -ve contingency being set up, do we still get learning?
Yes - inhibitory conditioning occurs
The animal learns a safety signal - the tone means safety as no shock will occur
This contrasts excitatory conditioning
How did Gallistel view conditioning?
The process of conditioning is use seeking out information
Conditioning = computational process where trying to work when/if important event occurs
Only going to learn if there is worthwhile info there to be learnt about
If the CS carries no meaning, there is no learning
CS that carry the most info are the ones we learn the most about
Reference Shannon’s information theory when explaining conditioning
When we seek out info, it is all about the reduction of uncertainty in the receiver - it is not about how much info is conveyed by the signaller
Exchange of info between signaller and receiver
What is important is what happens on the receiver end and whether there is reduction in uncertainty or not
What happens when the CS provides useful info and how do we quantify this?
The animal will learn that it is a useful predictor
We can quantify how useful a CS is through observing how the animal responds
There will be no responding to useless CS’s
Give an example of meaningful CS’s in terms of catching trains.
Example: catching train to uni during peak time
Trains come every 5 mins, the train announcement doesn’t carry much info because train comes every 5 mins so don’t listen
Trains less frequent, listen out for train announcement to know when the next train is - this CS (train announcement) provides you with a lot of info
What is important for conditioning?
Ratio of the “interval between US’s: CS duration”
It reduces the amount of uncertainty by a proportionate amount
