Lecture 8 - Theories of Conditioning Flashcards
Outline Pavlov’s Temporal Contiguity Theory
Closeness time form associations
Evidence ice-cream and chemotherapy taste aversion isn’t necessary. Over-ride conditioning
Outline Smith and Roll 1967 study that conditioning taste aversion over long delays
Rodents x-rays exposure. Artificially induce sickness.
Animals make associations even over many hours. Reduce preference for drink with saccharin
Outline Kamin’s Blocking Experiment
Inspired Rescorla Wagner
In blocking have number successive trials learning and associating a pair, then later add in another component e.g. light. Create compound = no learning light.
Info value as well as closeness in time important determinant learning
Who inspired Rescorla Wagner Model
Kamin’s Blocking Experiment
Outline Rescorla-Wagner Model
Focus surprisingness US.
Limitation determined how overall outcome could be rated by saliency
Learning gradually increases
Evaluation of Rescorla-Wagner Model
Cant measure particular trial, but ONLY series trials
Don’t have currency to measure all different intensities
Effect of CS salience on learning
CS intensity relative to background
Depressed acquisition function resembles that seen with a CI
Does the Rescorla-Wagner 1972 Model predict extinction?
YES
By trial 2 - smaller change in associative strength
Outline conditioning with compound stimuli in Rescorla Wagner Model 1972
US surprisingness now depend on how well this event is predicted by all available stimuli.
Cue competition = looking all available cues influencing associations
Outline Conditioned Inhibition
Summing associative strengths and know what it looks like when going through extinction
Unpack what is occurring
Light = maintain excitation
Noise = inhibitor = negative associative strength
Inhibitor is a safety signal keeping original learning in tact preventing it from extinguishing individuals
When tested together appeared extinguished
How to test for Conditioned Inhibition
Train excitator and compare outcomes
Transfer Test - take inhibitor and present it with another CS predicting something interest and see if it dampens down behaviour like established inhibitor - summation test
How can inhibitors be treated?
Emotional
Foot shock acts inhibitor/safety - approach and make lever presses get to it
Inhibitor positive - treated something nasty
What is the take home message from the Rescorla-Wagner 1972 Theory Model
Amount learning any one trial depends how surprising US is
Outline Wagner 1981
Modification include CS factors. Moved away saliency CS fixed. CS must also be surprising.
Animals react novel stimuli with orienting response or repeated presentation or declines habituations.
Associability stimulus decline latent inhibition
Outline rat orienting towards light CS as evidence Wagner 1981
Light may or may not lead food deliveries.
Reaction light gradually diminish if nothing happens.
Sometimes food = some reinforcement. Orientation light maintains = interests stays up as have to focus to know outcome
What is the main difference between UK and USA Rescorla-Wagner Model
Americans focused on outcome
UK focus on information value of signal like Wagner 1981
Outline Mackintosh 1975
Animals pay attention/condition stimuli established good predictors significant events
Discrimination learning - acquired distinctiveness of cues
Cues in discrimination learning more effective once chance process As associative strength CS and US increases, associability same CS other outcomes increase
Associative strength poor predictors should decline
Who influenced Mackintosh 1975
Lawrence 1950studies on discrimination learning
Acquired distinctiveness of cues
Associability is not fixed factor dependent on intensity.
Outline Mackintosh-Turner Design Test Phase 1
Already have established predictor shock. Later have compound presentation where 2 stimuli leads shock. But type shock hanged. Would expect some unblocking and some learn. Not redundant
Outline Mackintosh-Turner Design Test Phase 2
Added stage light but shock level remained unchanged unlike phase 1
Associability poor predictors declined
Introducing light as poor predictor as nothing has changed
Suppression ratio = LESS LEARNING
Conclusions from Mackintosh-Turner Design Test
Good predictor yield better learning in future.
Bad predictor = decline in learning
BUT orientation maintained when not good predictor as stimuli partial reinforcer. Inconsistent schedule maintains attention
Outline Pearce-Hall 1980
Opposes Mackintosh
Animals not waste time processing events with known consequences
Stimuli established consequences, automatic mode and cannot enter new associations
Stimuli unknown consequences processed controlled mode and enter new associations
When we are not sure = pay attention
Outline Hall and Pearce experimental design 2
Set stimulus as predictor. Compared novel stimulus see if its better or worse being later predictor.
Tone followed by shock. Phase 2: same tone followed larger shock. Compared this to light followed by shock, then a tone followed by larger shock= different predictor
Why did they use a different predictor in Hall and Pearce experimental design 2?
If stimulus set up as predictor as Mackintosh thinks, then conditioning be stronger, suppression ratios be lower with group as tone as predictor
Outline the results of Hall and Pearce experimental design 2
Group who did not have tone set up as predictor earlier (2) better conditioning and predictor.
Opposes Mackintosh
Evaluation of Hall and Pearce experimental design 2
Assumes they successfully set up tone as a predictor in stage 1
Phase 1 training wasn’t very effect .45.
Counter-argument for Evaluation of Hall and Pearce experimental design 2
Hall and Pearce 1982 reported similar experiment in which phase 1 training more effective
