Lecture 3 Flashcards
1
Q
What is an excitatory presentation?
A
- Two things separately activate things in the animals mind, under conditions, there’s a new association formed as it allows the presentation of the tone to excite a memory of food
- Under circumstances, presentation of tone can stop memory of food from becoming active = inhibitory learning
2
Q
When does learning take place?
A
- On a given trial to the extent that there is a large discrepancy between what is known/expected and what happens
- When discrepancy is large = much is learned, when small = little is added to what is known
3
Q
How does associative strength increase?
A
- Large discrepancy due to shock, but as the animal expects it, the discrepancy becomes smaller and less surprise so less learned association
- Curve of learning: characteristically negatively-accelerated learning curve which changes in associative strength become progressively smaller as training continues = does go up by by smaller increments until it reaches an asymptote
- Delta V = change in associative strength, Lambda = what actually happens, Sigma V = what animal expects to occur over all trials
- This difference is then multiplied by the product of the fixed learning rate parameters for the light and shock
4
Q
What is an example of a learning curve?
A
- Required to learn what sprays killed alien bugs
- Learning function is curved but reached asymptote in different intervals
5
Q
What happens to the associative strength in a compound that occurs after the shock?
A
- Blocking group have learnt about light, so learning curve of compound e.g noise and light stays at light level which is 100 = as they expect shock as a result of the light
- Associative strength of the compound as a whole then goes up to 100
- In the blocking group, the associative strength of the noise stays at 0, but compound conditioning where light has not been presented alone, the associative strength goes up to 50 because each component adds up to 50, or another number depending on the salience (expectation is that it increases, but there are two compounds here)
6
Q
How to generate inhibitory learning?
A
- Tone followed by food, Tone+light followed by nothing
- Animals come to show conditioned responding when the tone is presented alone (goes to get food), but not when tone accompanied by light
- When light is presented alone, it does not appear to have any tendency to provoke a response = learning about the light is behaviourally silent
7
Q
What is the summation test?
A
- G1 have classic training: tone followed by food and then tone and light lead to nothing
- G2: control: light followed by nothing
- Training a transfer where both G1&2 get noise followed by food
- G1: light reduces conditioned response to the noise but it does not so in group 2
- THEREFORE: animals learn about the light interfering with what they know about stimuli that predict food
8
Q
What is the retardation test?
A
- G1: tone followed by food and tone+light followed by nothing, G2: light followed by nothing, G3: nothing at all
- G1-3 light followed by food
- It is harder to established conditioned response to the light in G1 than in G2 and 3, it is also harder to establish conditioned responding to the light in group 2 than 3
- G1 has learnt something about the light that makes it difficult for the light to acquire the ability to elicit conditioned responding
- Even that light draws attention
9
Q
What is the inhibitory link account?
A
As tone can excite food, inhibitory association stops memory of food from becoming active when presenting light = put both together results in a lesser conditioned response
10
Q
What is the no US account?
A
- Tone activates memory of food
- Light activates memory of no food = could be an emotional response
- Light stops animal responding due to structural inhibitory link between no food and food e.g think of a banana and no banana at the same time
- Results in two different conditioned response
11
Q
How do we apply the model to extinction?
A
- Going from light followed by food to light followed by nothing
- Assuming that the asymptote has been reached during training, there is no US presented during extinction so lambda = 0
- Extinction trials result in VL decreasing to 0 according to a negatively accelerated function
- Change of strength is negative, and because the light starts out with no associative strength = becomes negative
12
Q
What are the explanation of the results of Summation and retardation tests?
A
- S: the light’s negative association strength combines the excitatory strength of the noise to produce a net value of 0 (adding different numbers together with different signs = results in a smaller positive value)
- R: Following stage 1 = when the light is paired with food it takes longer for the light to move from having negative value
13
Q
Is this account accurate?
A
- When you present the conditioned inhibitor alone, its inhibitory strength should diminish: the associative strength of the light will move from -100 to -80 until Vlight= 0
- The prediction that the inhibitory properties of a stimulus will be subject to extinction was tested and failed to find any evidence that the inhibitor extinguish when it is repeatedly presented alone
- Accept the No US account instead of inhibitory link
14
Q
What are the strengths of the Wagner Model?
A
- Precise: making ordinal predictions about the effects of manipulation at a trial-by-trial event
- Considerable explanatory power
- Makes novel and counterintuitive predictions some of which have been confirmed
15
Q
What is superconditioning?
A
- IF increases in associative strength are dependent on surprise = should be making learning easier and quicker if you increase the discrepancy
- STUDY: Inhibitory group - noise established as inhibitor for shock, uncorrelated group =noise presented uncorrelated with shock leaving it with zero associative strength
- Both groups receive noise+light leading to shock trial
- Light represented alone and elicits more responding in group inhibitory rather in group uncorrelated
16
Q
What is overexpectation?
A
- When 2+ stimuli presented with US, animal learns predictive value of each stimulus
- If total amount of conditioning exceeds US, animal may have a weaker response as it expects more than it receives
17
Q
What are some weaknesses of the model?
A
- Overshadowing is not predicted to occur but is observed following a single conditioning trial
- Models account of conditioned inhibition is flawed
- Suggestion that alpha value of a stimulus does not vary as a function of experience to a stimulus retards subsequent learning involving that stimulus
