Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
14 Cards in this Set
- Front
- Back
|
Rules to Pavlovian conditioning
|
Latent inhibition Timing Associative Bias Blocking |
|
|
Latent Inhibition
|
•Pre-expose to CS repeatedly
•Then pair CS with US •Learning is inhibited •Animal has learned that the CS is useless, has stopped paying attention |
|
|
Timing
|
Delayed conditioning (strong) simultaneous conditioning (weak) Trace Conditioning (weak to moderate) Backwardsconditioning (weak) |
|
|
Trace Conditioning |
Optimal interstimulus interval (ISI) with less learning at delays too short or too long
|
|
|
Why is there little-to-no conditioning with simultaneous or backwards pairings?
|
CS does not predict the US!
|
|
|
Associative Bias
|
•Some associations are innately easier to make: •When tone + taste paired with poison, only taste provokes CR
•When tone + taste paired with shock, only tone provokes CR •In nature, tastes go with getting sick, sounds with getting hurt |
|
|
Blocking
|
•Once one CS fully predicts a US, another CS paired with the US will not predict the US
•Second CS is irrelevant! |
|
|
Rescorla-Wagner model of conditioning (1970s)
|
•Amount of learning that occurs with each CS-US association (trial/pairing) is driven by the discrepancy (error) between the expectation (prediction) of the US and whether or not the US actually occurred
•1st pairing of opening a cheese bag (CS = crinkly sound) + getting cheese (US), US is unexpected, a lot of learning occurs •10th pairing, US is expected, minimal to no learning occurs •Prediction error learning (aka reward prediction error): the degree to which an outcome is surprising (unexpected) modulates the amount of learning that takes place •If the US is already predicted by a CS, then it is not surprising, and no further learning occurs |
|
|
Prediction error learning
|
1.Positive (prediction) error •Novel CS precedes a US, error is large •Learning occurs
2.No (prediction) error •US is expected following the CS •No further learning occurs 3.Negative (prediction) error •US is expected following the CS, BUT no US is given •Leads to extinction |
|
|
Blocking
|
•Absence of a prediction error
•First a tone is conditioned so that Vtone(CS1) = 1.0 •A tone fully predicts a reward will be delivered (CS = US) •Next a light (CS2) is added. The formula predicts its associative strength: •ΔVLight = αβ(λ – ΣV) •ΣV = VTo n e + VLight•If we assume that αβ = 0.2 and VLight is 0 because no learning has occurred yet, then: •ΔVLight = 0.2[1.0 – (1.0 + 0)] = 0 |
|
|
Extinction
|
•During extinction, the CS is presented without the US
•This is the same as presenting a US with an intensity of 0 (λ= 0 ) •The formula predicts the associative strength during extinction: •ΔV = αβ(λ – V) but λ is now 0 (because the US is not given) •ΔV = 0.2[0 – 1] = -0.2 •The associative strength is decreasing •Use the decreased value for V (1 - 0.2) for the next trial |
|
|
Caveats to Rescorla-Wagner model
|
1.Latent inhibition is not predicted by R-W 2.Spontaneous recovery is not predicted 3.Ignores some important features such as temporal relations
|
|
|
Dopamine and Blocking
|
•VTA DA firing Conforms to R-W model of learning
•Dopamine does not fire to a blocked CS •Dopamine will fire to a tone (CS1) that predicts a pellet (US) following repeated pairings •If you now pair a light (CS2) with a pellet, dopamine will not fire to the light even with repeated pairings ΔVLight = 0.2[1.0 – (1.0 + 0)] = 0 |
|
|
Pavlovian conditioning and dopamine (DA)
|
•Use electrophysiology to record from DA cell body in the VTA
•First, DA signals to reward (US) •With repeated CS-US pairings, DA signals to CS •Absence of reward, DA is inhibited to US •DA response codes discrepancy between reward and its prediction (reward prediction error) |
