Extinction of Conditioned Behaviour (Chapter 10)
Con’t
Enhancing Extinction: Implications for Exposure Therapy
How do we enhacne exteinction?
We want the extinction procedure to maintain its longevity over time
More trials = stronger suppression
Increasing the intertrial interval produces more enduring suppression (e.g., spacing it out, dealing with it more on your own and warding off against spontaneous recovery)
Multiple contexts — extinction trials/procedures in multiple contexts means you’re more likely to generalize that extinction procedure across contexts
Take an extinction reminder cue with you — think of AA: they celebrate your achievements in sobriety with tokens — they are reminders of the process. Recovering alcoholics will often keep their tokens with them.
COmpoundin extnicton stimuli — if you do mutliple extinction sessions simultaneously, the process will take longer, but it will be deeper learning: for smokers, it is sometimes suggested to quit smoking while you quit something else (overexpectation will wekaen the strength of predictors)
Rescorla (2006)
3 discriminative stimuli; light, tone, or noise which are all predicitve of the reward
All underwent extinction individually, and in a similar way
Light and noise continue extinction procedure together
Initially, two cues recover extinction
As it’s carried out across trials in the absence of the reward, the compound stimuli
6 days break between compound trials and test trials
individual extinction of tone Y was learned, but wasn’t as resistant to extinction because of its indiviudal extinction procedure
S-O learning remains intact. What about R-O learning?
Test the specificity of reinstatement (Ostlund and Balleine, 2007)
Maintained memory of which lever did what after extncition, which is R-O association
What is Learned Extinction?
If S-O and R-O associations remain intact after extinction, then what produces the response decrement?
Inhibitory S-O learning takes place.
Paradoxical Extinction Effects
Overtraining Extinction Effect
More extensive reinforcement training leads to more rapid extinction
Produced by using rapid continuous reinforcement
Frustration is greater during non-reinforcement
Learning instrumental behaviour super well → you know quicker when it’s not working compared to not having learnt it so well
Partial Reinforcement Extinction Effect
Greater persistence in instrumental responding in extinction after partial (or intermittent) reinforcement training than after continuous reinforcement training
Extinction is slower, and frustration is less
E.g.,: something has been partially reinforced on a schedule
Example: throwing tantrum worked once, so bratty kid continues because they know it worked once (pays off occasionally, which is enough to keep persistence going)
Discrimination Hypothesis:
Learned Persistence:
Frustration Theory (A. Amsel):
Sequential Theory (E.J. Capaldi):