Problem 4 Flashcards
Classical conditioning/
Pavlovian Conditioning
Refers to a form of learning in which one stimulus predicts an upcoming event
ex.: doorbell predicts the delivery of food
Unconditioned stimulus
US
Refers to a stimulus that evokes a response naturally
–> occur unconditionally without prior training
ex.: FOOD –> evokes hunger
Unconditioned response
UR
Refers to a natural response to the natural occurring stimulus (US)
–> occur unconditionally without prior training
ex.: food –> HUNGER
Conditioned stimulus
CS
Refers to a cue that is paired with an US and comes to elicit a conditioned response (CR)
ex.: BELL –> food
Conditioned response
CR
Refers to a trained response to a CS in anticipation of the US
ex.: SALIVATION –> in anticipation of food
Appetitive conditioning
Refers to conditioning in which the US is a POSITIVE event
–> learning to predict something that satisfies a desire or appetite
ex.: food, sex
Aversive conditioning
Refers to conditioning in which the US is a NEGATIVE event
–> learning to avoid or minimize the consequence of an expected aversive event
ex.: shocks etc
Sexual conditioning in male Quails
Experiment
Male quails were conditioned to approach + remain near a light (CS) which was associated with access through a door to a sexually receptive female (US)
Odor Conditioning in flies
Experiment
Flies were first put in a container with an odor were they were shocked, then in a container with another odor where they weren’t shocked
–> they were then placed in a container with both odors to the left and to the right
Eyeblink conditioning
Presenting a tone which ultimately predicts an air puff to the eye.
This elicits an anticipatory defensive response by blinking the eyes before the arrival of the air puff
–> aversive conditioning
In which way is the eye blink being an UR different from being a CR ?
The learned CR takes place BEFORE the onset of the US, therefore protecting the eye from the air puff
–> the UR would take place AFTER the arrival of the US, as one wouldn’t be expecting the US
Conditioned compensatory responses
Refers to an automatic response that the body experiences + that is opposite of the effects of the drug to reach a state of HOMEOSTASIS
- -> elicited by situational cues (CS)
- -> partially mediate
a) tolerance
b) withdrawal distress
c) relapse
Tolerance
Refers to a decrease in reaction to a drug, so that larger doses are required to achieve the same effect
Extinction
Refers to the process of reducing a learned response to a stimulus by ceasing to pair the stimulus with a reward/punishment
–> even though one might not respond to the CS, the learned response isn’t gone, just unexpressed
ex.: stop delivering food when bell is rung
What is evidence for the fact that the original CR isn’t lost during extinction ?
- If a long time passes before new testing, one will react with the old CR when presented with CS
- Previously extinguished CS is learned more rapidly than a novel CS
Compound conditioning
Refers to conditions when 2 cues occur simultaneously in a conditioning experiment, and therefore compete with each other
Overshadowing
Effect that occurs when a more salient cue within a compound acquires more association, and is therefore more strongly conditioned due to
a) salience
b) temporal priority
–> assuming cues are both valuable in info
Blocking effect
Classical conditioning occurs only when a cue is both a useful + nonredundant predictor of the future
ex. : if I have Doris to predict the stock market correctly, I don’t need Hermann to do the same
- -> I will therefore ignore him, no learning
Which characteristics does a stimulus have to have in order to become associated with a US ?
a) reliable
b) useful
c) bonredundant
Rescorla-Wagner model
States that changes in the CS-US associations on a trial are driven by the Prediction error
expects that CS-US association increases proportional to the degree that the US is surprising
–> the larger the error, the greater the learning
FORMULA:
Occurrence of US – Expectation of US based on CS
Prediction Error
Refers to the difference between what was predicted and what actually occurred
Error correction learning
Method in which the errors on each trial lead to small changes in performance that seek to reduce the error on the next trial