05 03 2012 I Recap of Tuesday A Formal models of Pavlovian Conditioning 1 Blocking 2 overshadowing II Comparator Hypothesis A Ralph Miller B Acquisition of CS US association assumed to follow simple contiguity rule 1 e g Bush Mostellar C Many learning phenomena are actually due to performance processes not acquisition processes learning performance distinction D Architecture of model 1 Subject goes through a CS US pairing and is tested with presentation of CS X direct US memory a If an association has been formed the association causes a retrieval of a memory of the US b If there were other events occurring during training comparator stimuli those stimuli will be called from memory when the CS is presented as well as the US which is being retrieved from memory i The comparator stimulus was probably also associated with the US indirect US memory c Presentation of CS X directly activates a memory of the US and also indirectly activates a memory of the US i The two memories are compared to each other and the output of this comparison determines the nature of the response performed 2 Response rule a Direct indirect excitatory CR b Indirect indirect Inhibitory CR 3 Example a Training session Tone Light Shock b Learning i Tone shock association ii Tone light association iii Light shock association c Test tone behavior d In terms of the model i Tone is presented Tone calls to mind a memory of the shock Tone calls to mind a memory of the light o Light shock association calls to mind the memory of the shock also ii Comparison process takes place If the directly retrieved memory was strong we should see a strong fear response If the indirectly retrieved memory was stronger we should see inhibition Decision is made based on the information that s compared in the animal s memory iii If the two stimuli were presented together and are about the same strength the one presented X should elicit the stronger response because it follows a direct retrieval pathway rather than indirect retrieval In a blocking group trial A is paired with the US on it s own first and creates a stronger association which then leads to a stronger retrieval in the test To combat this extinguish association with stimulus A III Cue Competition cues compete with each other A Examples 1 Blocking 2 Over shadowing a Phase 1 Ax b Phase 2 Ac Test x CR 3 Over expectation a Phase 1 A X b Phase 2 AX i Small response to X c Phase 3 Ad Text X CR 4 Conditioned Inhibition a Training A AXi A becomes a conditioned exciter ii X becomes a conditioned inhibitor b This can be thought of in context of the comparator hypothesis i Tests of X should elicit no response X calls up X A associations no X US associations Follows the bottom path on the graph and inhibits c In the Rescorla Wagner model i Negative associative strength IV Extinction of Inhibition A AXA After training Extinction of X 1 According to the Rescorla Wagner model attenuate Cl 2 According to the comparator hypothesis no effect B After training extinction of A 1 According to the Rescorla Wagner model no effect 2 According to the comparator hypothesis attenuate Cl V Attentional Models of Conditioning A Rescorla Wagner model US effectiveness determines how much is learned on a trial 1 A US that is surprising is effective 2 A US that is predicted is not effective B Attentional models look at the CS effectiveness as determining how much is learned on a trial 1 Mackintosh 1975 a Attention to a CS increases on the next trial when that CS is the best predictor of the US on the current trial attention to all other CSs will decrease 2 Pearce Hall 1980 a Attention to CS decreases on the next trial when the outcome is not surprising on current trial attention to CS increases on the next trial when the outcome is surprising 3 Neither of these accounts predict blocking in one trial they are ruled out