Theory of Learning 1 Flashcards
Mahoney & Ayres, 1976
Rats; a 4-second tone was paired with a 4-second shock
Name the 4 types of conditioning that was used:
Delay
Trace
Simultaneous
Backward conditioning
What are the 2 types of forwards conditioning?
Delay conditioning
Trace conditioning
Which type of conditioning is this:
The CS precedes the US (tone predicts the shock)
Delay conditioning
(a type of forwards conditioning)
Which type of conditioning is this:
The CS precedes the US (tone predicts the shock after a short delay)
Trace conditioning
(a type of forwards conditioning, it has a trace interval)
Which type of conditioning is this:
The CS and the US are conditioned simultaneously (at the same time/ together)
Simultaneous conditioning
(eg. tone for 4 seconds and shock for 4 seconds played at the same time)
Which type of conditioning is this:
The US (shock) comes before the CS (tone)
Backward conditioning
Mahoney & Ayres, 1976
Shock experiment: suppression of licking (lick latency)
Do rats learn to associate more with forward conditioning compared to others (sim/ back)?
Rats learnt most when the tone was before the shock=
Forward conditions (Delay/trace)
Simultaneous learning group: learnt a little
Backward conditioning group: did not learn much
What is the best type of conditioning for associative learning?
A forward relationship between the CS and US
(forward conditioning)
What is necessary for causation?
Correlation
Coincidence is not enough to infer causality
(clue in the name):
1. when two or more thingshappen at the same time;
2. chance or luck
This is not enough that events are paired, must also be correlated
Sometimes pairing are not sufficient enough to form an association (correlated)
– Not enough that events are paired, must also be correlated
must occur together more than they occur separately
Are pairings or correlation more important for learning an association?
Rescorla (1968) Three groups of rats given 5 tones and some shocks.
Which group had electric shocks which were not paired with tones?
Group 3 negative – shock NEVER paired with tone
NEGATIVE tone negatively correlated with shock
Are pairings or correlation more important for learning an association?
Rescorla (1968) Three groups of rats given 5 tones and some shocks.
Which group had electric shocks paired with tones + were not paired with tones (sometimes)?
Group 2 zero – 2 tone> shock pairings, plus extra shocks
ZERO tone uncorrelated with shock
Are pairings or correlation more important for learning an association?
Rescorla (1968) Three groups of rats given 5 tones and some shocks.
Which group always had electric shocks paired with tones?
Group 1 positive – 2 tone> shock pairings, no more shocks
POSITIVE tone positively correlated with shock
Are pairings or correlation more important for learning an association?
Rescorla (1968) Three groups of rats given 5 tones and some shocks.
What was the main difference between these 2 groups?
What were the findings?
If all that mattered was pairings, then both groups should learn the same association as they had the same amount of pairings.
However, if correlation matters (contingency) then the positive group should learn to associate more due to the shock and tone being positively correlated.
Findings showed:
The positive group
Learned more associations than zero /negative control group
Conclusion: Pairings are not everything, Correlation is important
Are pairings or correlation more important for learning an association?
Rescorla (1968) Three groups of rats given 5 tones and some shocks.
Which group may have shown inhibition (less scaredness) toward the sound of the tone as it was never paired with the shock?
Group 3
NEGATIVE tone negatively correlated with shock
Rats have evolved to selectively learn about correlated events by using the idea of?
Surprise
Pairings only produce learning when US (shock) is surprising.
You first need pairing of a CS (tone) and US (shock)
But also need that US (shock) to be surprising.
Which effect is this known as?
The blocking effect
Kamin (1969): Blocking
Measuring how much fear they had of the light: conditioning to light
Same number of light > shock pairings in both groups
In Group 1 light conditioned with a pretrained noise N
shock not surprising
In Group 2 light conditioned with a novel noise N
shock not predicted, and is surprising
Findings showed which group had less learning?
Showed pairings only produce learning when the US is surprising as the rats responded to less food pallets (due to fear) in group 2
Less learning in Group 1, where shock not surprising
Rescorla & Wagner theory:
theory describes how much association strength increases on each trial (i.e. CS>US pairing)
What is the equation?
The crucial observation, that pairings only produce learning when the US is surprising, is captured by which theory?
Rescorla & Wagner theory
Which theory describes how much association strength increases on each trial (i.e. CS>US pairing)?
Rescorla & Wagner theory
Rescorla & Wagner theory:
theory describes how much association strength increases on each trial (i.e. CS>US pairing)
What do these symbols stand for?
V is strength of association between CS & US
∆V (delta V) is change in V (associative strength) after each pairing
eg. pairings of light and food: V= zero in the beginning, but as trial pairing continues V becomes low/ medium/ high
Rescorla & Wagner theory:
theory describes how much association strength increases on each trial (i.e. CS>US pairing)
What do these symbols stand for?
Alpha refers to salience of CS - its intrinsic perceptual intensity
eg. if light is strong you will notice it more than dimmer light
Beta refers to salience of US - its intrinsic perceptual intensity
eg. food: high b for peanut butter but low b for normal pellets
Rescorla & Wagner theory:
theory describes how much association strength increases on each trial (i.e. CS>US pairing)
Which symbol stands for this representation of the size of US (food)?
high = more food amount (reflection of magnitude)
Rescorla & Wagner theory:
theory describes how much association strength increases on each trial (i.e. CS>US pairing)
What do these symbols stand for?
How much the US (Shock) is already predicted
The sum E of the associative strengths Vs of everything else that is present on that trial
Rescorla & Wagner theory:
theory describes how much association strength increases on each trial (i.e. CS>US pairing)
What do these symbols stand for?
( lambda - SV ) is how surprising the US is if US is very large (N big) it’s more surprising. If US is well predicted (SV big) it’s less surprising
What 3 things does learning depend on?
Salience of the CS (high/ low tone)
Salience of US (nice/ okay food)
How surprising the US is
Equation:
When does learning stop?
Each US can only support so much learning
Learning stops when US no longer surprising
Exam question:
Trial 1= no associations learned
Trial 2= some associations learned
Learning will stop as trials continue when associative strength is equal to lambda (content of bracket is 0)
During the process of learning, learning reaches an asymptote.
This is the maximum associative strength it can reach for food US.
How do we increase the amount of lambda?
Increase with the same light intensity (CS)
But increase with double the amount of food (US)
Rat will increase its learning/ lambda/ (higher level of asymptote)
Conditioning two CSs at once: Overshadowing
Which condition would learning be the least impactful?
Condition light with no noise
Condition light with a little noise
Condition light with a loud noise
Condition light with a loud noise
-there is only so much learning one animal can do before eg. it starts to get scared
-the more salient the 2nd stimulus, the more it can take from the light
competition for strength= overshaddowing between light and noise
-if noise is more salient than light, a louder noise will create faster learning than a lower light level
-high alpha = higher conditioning
Rescorla & Wagner theory:
theory describes how much association strength increases on each trial (i.e. CS>US pairing)
If you increase the salience of light (CS) what happens?
alpha gets bigger
although maximum amount of learning is unchanged, it happens faster
(same with beta)
Overshadowing and blocking: Competition for strength
What happens to the associative strength in each condition?
In control group light gets all the associative strength
In overshadowing group strength is divided between light and noise
In blocking group noise gets all strength in stage 1, so none left for light!
