Chapter 9: Extinction Flashcards
Extinction
- Omitting the US (classical) or reinforcer (instrumental)
- conditioned response declines
- forgetting ? Not the same; changes bc of a passage of time, not because of experience
effects of extinction procedures
- target response decreases whine no longer = in reinforcement
- after extinction burst - increase in response variability (at first)
- frustration and aggression
- depression
- resurgence (regression)
Neuringer et al. (2001)
Rats presented 2 levers and a key in operant chamber
- 3 responses in a row to get food
- during acquisition:
- group 1: had to vary response pattern
- group 2: no variation necessary
- then extinction (no reinforcement)
results:
1. rate of responding - decrease in overall rate of responding for both groups (during the reinforcement phase the control group with no variation had higher rates of responding)
2. Response variability - increase in response variability during extinction for both groups (but the highest response variability was found in the experimental group!)
Tomie et al. (1993)
methods:
- rats water deprived
- four phases:
1. alternative schedule of: - 3 min: VI-30 sec delivery of water
- 3 min: no water (ext) singled by tone (S-)
2. - no tone and tone conditions both = no water
3. - same as phase 1.
4. - no tone and tone both = water
behavioural measure: target bite bar (target biting is a sign of frustration in rats)
Result:
phase 1.
- no water presented during tone sessions. Lots of target biting during tone sessions
see slide 12
extinction and original learning
extinct does not reverse/undo original learning
- how do we know this?
1. spontanteous recovery
2. renewal
3. reinstatement
4. resurgence
extinction is not the _______ of acquisition
reversal
spontaneous recovery
when responding recovers after a period of rest, after extinction trials
Rescorla (1996) Spontaneous recovery: operant conditioning
- responses (lever press or nose poke) acquired, then extinguished
- R-rest: tested 7 days post-extinction
- R-No rest: tested shortly after extinction
- graph: R-Rest group had higher response rates always, they were really high initially then dropped off; the no rest group always had the lower responses!
renewal of original excitatory conditioning
incerased behaviour when the contextual cues that were present during extinction are changed
Bouten and King (1983) renewal of fear
fear conditioning in context A
then extinction in context A
test with context A => no fear
fear conditioning in context A
extinction training in context B
test with context A => fear!
- fear measured in conditioned suppression of lever pressing
- important: original acquisition generalizes across contexts more readily than extinction performance does!!!!
what does renewal tell us?
extinction is about learning new S- contexts rather than “un-learning”
reinstatement of conditioned excitation
rapid recovery of conditioned behaviour produced by exposures to the US
Reinstatement of fear in humans: :bar and Phelps (2005)
fear conditioning -> extinction -> US presentations -> tested with CS ( = recovery of fear)or (= no fear)
Resurgence of conditioned behaviour
appearance of an extinguished response caused by the extinction of another behaviour
techniques for enhancing extinction
- greater # of trials (decrease in spontaneous recovery)
- massing and spacing
- many alternations of extinction and rest sessions
- extinction in many diff contexts qa
- deepen extinction by using compound stimuli
Compound stimuli in extinction, Rescorla (2006)
- rats
- three stimuli: light, noise, tone
- acquisition of lever pressing in presence of stimuli, presented independently
- compound extinction phase (light w/ one of the auditory stimuli; other with auditory alone)
results:
- compound stimuli = compound extinction trials deepened the extinction of the stimulus (also reduced reinstatement and slows rate of reaquisition)
what is learned in extinction?
- SR and RO associations are not eliminated ( shown by spontaneous recovery, renewal and reinstatement)
why does extinction reduce responding is learning persists?
due to an inhibitory S-R association
- non-reinforcement of a response in the presence of a particular stimulus (S) produces an inhibitory association
extinguished effects will be highly specific to the _____ in which the response was extinguished
context!
Inhibitory S-R associations in instrumental extinction: rescoral 1993)
- Noise and Light discriminative stimuli in extinction create expectancy
- only one response available in extinction (inhibitory S-R)
- the extinction procedure produced an inhabit S-R association that was specific to a particular stimulus and response
- can’t be due to R-O or S-O associations bc changes in these would effect R1 and R2 equally
Paradoxical rewards effects
you might expect that more acquisition training = behaviours that are harder to extinguish
- NOT TRUE
overtraining extinction effect
- the more acquisition trials, the greater the expectancy of reward, and the greater the frustration when extinction is introduced
- produces more rapid extinction
- odd!
Ishida and Papini (1997)
overtraining extinction effect
titles that were trained longer showed more rapid extinction
magnitude reinforcement extinction effect
- more rapid extinction if trained w/ larger rather than smaller magnitude reinforcer
- expectancy of greater reward produces greater frustration when it is not forthcoming
Hulse (1958) magnitude reinforcement extinction effect
group reinforced 100% of the time w/ a large reward showed rapid extinction
partial reinforcement extinction effect (PREE)
- schedule in effect prior to extinction is important
- chief characteristic is whether the schedule was CRF or PRF
- much less resistance to extinction with PRF
- high and low gamblers study
Discrimination hypothesis
Jenkins (1962)
easier to notice at the start of extinction if on CRF than on PRF
jenkins:
- 2 groups of pigeons, 1st on CRF, 2nd on PRF
- both put on CRF, then immediately on extinction
- 2nd group took longer to extinguish
Frustration theory (Amsel)
- intermittent reinforcement (PRF) has rewarded and non-rewarded responses
- reward responses motivation, non-rewarded responses are frustrating
- sometimes what you expect will be a non-reinforced response produces a reinforcer
- hence, frustrated responses lead to future expectation of reward; on CRF can’t learn to respond when expecting non-reward
Sequential theory (Capaldi)
- based on what subjects learn about the memory or non-reward
- eg. you can’t remember whether or not a response was reinforced in recent past
- based on memory, a non-rewarded trial can become a cue for responding
Behavioural momentum
- analogy with Newtonian physics
- behaviour that has a lot of momentum will be hard to disrupt through manipulation
- studied using multiple schedules of reinforcement
- diff reinforcement schedules are in effect for diff stimuli presented in succession
- add disruption (e.g. extra food b/w components, extinction, novel salient stimulus, etc)
- test how responding is affected for each schedule
momentum of human behaviour: Mace et al. (1990)
VI 60, more reinforcement, more momentum
lever pressing extinguishes less when…
pressing also leads to info about food elsewhere
Fading
slowly extinguishing aspects of discriminative stimuli signalling reinforcement
- increases success rates, decreases anxiety, prompts intrinsic reinforcement
- taking the training wheels off the bike
DRO
- help alleviate frustration
- blurred line b/w extinction and negative punishment
extinction
response no longer produces reinforcement
negative punishment
response leads to withdrawal of reinforcement
