Lecture 5: Pavlovian Conditioning

Lecture 5: Pavlovian Conditioning

1. Predictive learning: the ability of an organism to predict outcomes

   • learning about relationships in the environment

2. Associations

   • In the mind of the experiementer
     • Set up a contingency in the experiment but does this mean I set up a contingency in the brain?
   • Why talk about associations?
     • Why not beliefs..reason..or memory?
       • But we believe these things because they functionally associated
       • Maybe not reasoning
       • Associations provide a biologically plausible explanation
         • brains are made of connections that can change with experience
           • Long Term Potentiation
           • Memory based on stimulus provoking a fragment response
             • Learning must be based on associations
               • Not everyone believes this!

3. Nature of associations

   • CS activates US and generates CR
     • Re-watch whole section!!!
     • But not perfect.. CR and UR sometimes opposite
     • How can we tell if the link is CS-R or CS-US?
       • Is it CS-US or CS-UR?
         • Sensitivity to US revaluation
           • Test by conditioning rats then devalutation of US
     • Stimulus Subsitution Theory
     • CS-CS links (before conditioning)
     • Second order conditioning
       • Conditioning reinforcement
         • Star Charts!
       • Some evidence for SR learning
     • CS-US links (in your imagination)
       • Dwyer et al
       • Link to OCD or false memories or phobias?

4. What do we learn?

   • Timing: the CS-US Interval
   • Attention: Latent Inhibition and Predictablity
     • Predictablity: Rats orient less to stimuli that are predicted by a signal
     • L.I: A CS that has been pre exposed produces less learning
       • Latent inhibition is a poor name!
   • Learned Irrelevence
     • Learn less about previously poor predictors

5. Prediction error (P.E)

   • Predicitive earning will correct for previous errors
   • the amount that an outcome occurs, less than the predicted outcome
   • In neuroscience the most energy efficient way of responding to environment
   • Blocking effect
     • Pre training to A blocks learning to X
     • Contingency
     • Food allergy
     • Taken as some indication of prediction error learning
   • Does P.E matter?
     • dopamine prediction error in monkeys
     • Adding prediction error gives rise to unblocking (e.g in rats)
       • Way this was done: 'feroiusly cleaver"
     • Does prediction error apply to human learning
       • fMRI evidence
         • Provides some face validity
       • People with schizophrenia show less blocking
         • Disrupted prediction error signal in psychosis
           • Psychosis related to aberrant prediction error
       • Insights into chemotherapy from learning
         • Nausea and anticipatory nausea
           • Role of scapegoats