Extinction Flashcards
Extinction as a procedure?
Classic conditoining: CS - US learning first, after that CS - No us
Operant: na S (RO), S (R/)
Experience of no reinforcement after presenting CS
Extinction as effect:
CR - or R decreases
Is an adaptive process!
Traditional: unlearning R-W model
Opposite of acquisition
> but this is wrong
Is actually learning new relationship
Spontaneous recovery
Passage of time in itself can create the return of extinct response.
Often partial return of CS - not as frequent as before
Extinction suppresses conditioned behavior - it does not erase it
= classical conditioning + operant conditioning
Renewal
Return of CS after leaving the context of extinction
Contextual cues that were there during extinction are changed.
> ABA = renewal
A shift in context only disrupts extinction process, it does not disrupt acquisition processes.
Bouton - hypothesis renewal
The relationship between CS-US
and when response should occur will be mediated through the context.
> it is a context controlled INHIBITORY CS-US association
The extinct CS becomes ambiguous. It can ether signal that the US comes or does not come.
Context will resolve this ambiguity
Reinstatement
Return of CR > because of exposure of US alone
But! : has to happen in the same context as US was exposed
Sometimes > reinstatement because of another US similar to the original US
what is
Sensitivity to post-extinction US/ devaluation
Starting point: test in which the post conditioning - US devaluation effect is visible ( “S-S”)
Question: if extinction does not erase the CS-US relationship - > US devaluation should also work after extinction!!!
PROBLEM: after extinction you do not have any more observable behavior
“Tric”: increase CS↑ through association CS with a different US!
what is the result of post extinction devaluation?
CR level in general lower in group Extinction than control
However… since the CS-US relationship is still intact…
In both groups same strength of US devaluation!
(Dev < NotDev)
what are Increasing Pavlovian extinction performance methods:
- Increase the amount of extinction trials(?)
2 . longer ITI’s during extinction
> reduces spontaneous recovery and renewal
- Extinction in multiple contexts
> reduces “renewal”
> “uitzondering-op-de-regel” wordt “nieuwe regel”
- “Reminder cues” for extinction-performance
explicit extra cue during extinction and test
> reduces “renewal” and spontaneous recovery “
what is prime extinction?
through presenting CS before the extinction happens. Long term memory facilitation.
Not only consolidation consolidatie (originally learned) also reconsolidation retrieving informatione
reconsolidation: storing new information
CS first after that extinction
Less spontaneous recovery, renewal, reïnstatement
What is learned during extinction? - role of inhibitory S-R association
Rescorla:
Frustration of absence positive expected outcome > creates
Inhibitory association between present stimuli and conditioned behavior CR ( inhibitory S-R) that will suppress the conditioned behavior = extinction
Paradoxical effects of reinforcement
If expectation of US is higher, then during extinction phase bigger frustration and quicker extinction!
Over training and extinction:
After asymptote of training is reached still additional trial are executed.
Responding declines more rapidly - extinction - when overtraining has happenend!
Strong US effect and extinction
Extinction > responding declines faster if expected reward was bigger
Partial v.s. Continuous US and extinction
Extinction > happens faster with continuous reinforcement than partial (intermittent) reinforcement
Discrimination Hypothesis for continuous US & extinction
Easier to detect transition from reinforcement phase to extinction phase during continuous reinforcement than partial.
Discriminating between the two phases easier.
However: this theory is not correct!!
JENKINS & THEIOS
FRUSTRATION THEORY
AMSEL
Partial reinforcement: learning to kept doing response even when there is no US after CS. so even if US does not appear, you will be used to still giving response
Continuous reinforcement: you learn that US always comes after CS so if CS is not there you will quickly stop.
SEQUENTIAL THEORY
CAPALDI
Memory of the previous trials. Even if US did not occur. Memory of not reinforced trial will trigger to give response during partial reinforcement
