configural learning Flashcards
1
Q
configural learning
A
- change how stim is represented
- sometimes stim can interact: generalisation decrement
- when you add something or take something away from a stim, the effect is not fully additive
2
Q
summation
A
- Rescorla-Wagner assumes a stim comprising two or more components (a compound stim) is equivalent to the sum of its parts
- responding to a compound stim is determined by the sum of responding to its components
3
Q
structural discrimination
A
- pigeons 6 compound stim, each comprising 2 colours, one on each side
- 3 compounds paired with food, 3 compounds with no food
- all colours associated equally with food and no food
- summation says no difference between associative strength on reinforced and nonreinforced trials
4
Q
negative patterning discriminations
A
- associative theory cannot explain tone + click –> no food (not summation here)
- if A and B have enough associative strength to elicit responding, then compound of A and B must elicit more responding, not less - violates summation principle
5
Q
problem and solution for negative patterning discriminations
A
- animals can learn discriminations that cannot be solved with Rescorla-Wagner assumption that stim compound = sum of its parts
- accurate performance, on e.g. nonlinear -ve patterning and structural discriminations, mean some adjustments to theory required
- Wagner (1971) & Rescorla (1972) suggested the unique stim account: a stim compound is not the combination of its elements
6
Q
configural cues
A
- Includes a further stim, generated only when those elements are presented together
- configural stim not very salient
- only learned about when absolutely ‘forced’
7
Q
can configural cues explain all summation failures?
A
- no
- can’t explain generalisation decrement
- when you add something to a stimulus, or take something away, the effect is not fully additive
8
Q
external inhibition
A
- condition a stim and then add a neutral cue at test
- presence of B reduces response to A
- but summation predicts response should be the same, as A is always present
9
Q
pearce theory of stim generalisation
A
- limited capacity memory buffer representing overall pattern of stimulation that is present; every stim is a configure and unique
- a compound stim not the sum of its elements
- click & light is one configural cue, tone & light as another
- Pearce - because of generalisation between the 2 unique configural cues
- needs a way to work out how much generalisation occurs between these 2 configural cues
- acknowledges configural cue contains elements of clicker and of the light
- uses this to work out how much associative strength of click+light generalises to tone+light
- generalisation depends on amount of associative strength possessed by click+light, similarity of click+light to tone+light
10
Q
how the pearce model works
what is it asking
A
- associative strength of source of generalisation?
- what are the common elements mediating generalisation?
- what % are common elements of source of generalisation?
- what % are common elements of target of generalisation?
11
Q
occasion setting
A
allow flexible use of associations
12
Q
conditional control of associations
A
- same stim can be associated with 2 diff outcomes
- which association is retrieved is conditional on the context in which the stim is presented
- allows associations to represent knowledge in a versatile way
- context appears to control access to CS –> US association
13
Q
conditional control of associative retrieval
A
- standard associative theories can’t explain this - association forms and that’s it
- normal associative learning results in an association between a CS & US
- if access to this association can be controlled by a further stim in a way that is not associative, this is occasion setting
- not associative - often evident when phenomenon cannot be explained by summation
AKA: conditional cues, occasion setters, modulators, facilitators, retrieval cues
14
Q
holland (1989)
A
- 2 groups of animals
- type of discrimination in group PP is called a positive patterning discrimination
- if in group FP, discrimination between reinforced and nonreinforced tone trials is due to summation with the weak light –> food association
- group PP shouldn’t respond differently to the tone on reinforced and nonreinforced tone trials - if the light has no strength it can’t produce summation
- associative theory cannot explain how the light controls responding
- in this experiment the light may be called a conditional cue, a modulator, a facilitatory, or am occasion setter
- tone = target CS
15
Q
holland’s and-gate theory
A
- light facilitates operation of associative link, facilitating flow of activation the tone CS to the food US
- light is not acting as a CS, it isn’t acting by summation, not a normal association
- so it won’t transfer its properties to another stim associated with food
- will not show summation
16
Q
testing the and-gate theory
A
- if light is a CS you get summation with other CSs for food - light acts on US representation
- if light is an occasion setter for the tone –> food association it won’t affect responding to the click at all
- if light is a CS you get summation with other CSs for food, because light acts on US representation, so light will boost responding to the clicker just like it boosts responding to the tone
- if light is an association setter for the tone –> food association it wont affect responding to the click at all
17
Q
negative occasion setting
A
- occasion setters can turn associations off as well as on
- R-W says extinction erases associative strength - CS left with no associative strength
- but we know spontaneous recovery happens
18
Q
bouton & nelson (1988)
A
- inhibitory association is controlled by the context - context acts as an occasion setter
- context can be many things, including time
- changing the context, reduces impact of inhibitory association & fear returns
- spontaneous recovery when context = time
19
Q
applications of occasion setting
A
- SZ
- drug addiction
- anxiety disorder
20
Q
SZ and occassion-setting
A
- often strong overlap with frontal/executive control tasks & occasion setting
- EF deficits may be occasion-setting deficits
- several cognitive symptoms of SZ can be characterised as difficulties with occasion-setting tasks
- e.g. lexical ambiguity task: ppts define word based on semantic context
- word is a CS retrieving a semantic meeaning- occasion-setter is semantic context
- ppl with SZ have trouble with lexical ambiguity task (Titone et al., 2000)
21
Q
stroop task
SZ & occasion-setting
A
- ppl with SZ have trouble e.g. Westerhausen et al. (2011)
- seeing a simple coloured shape evokes naming its features - colour and shape
- when the stim is a written word, associations are with word meaning & written word stim’s properties
- strongly biased for reading - trained to pay att to meaning
- in stroop we have to use other association and suppress tendency to name word
22
Q
ramos et al. (2002)
occassion setting & drug tolerance
A
- ethanol induces hypothermia, but after extended exposure tolerance develops
- this tolerance can become conditioned to stim signalling the ethanol
- investigated effect of extinguishing these stim - does this eliminate the tolerance effect or not?
- CSs extinguish when presented alone, occasion setters do not
- when light and injection simultaneous extinction of light removed its tolerance effects (a CS)
- when light and injection serial (good for ocassion setting) extinction of light had no effect (an occasion setter)
23
Q
inhibition in exposure therapy
A
the reason fear recovers after extinction is because extinction stems from a CS –> no fear association that is occasion-set by the context
