Classical Conditoning Flashcards
Pavlov experiment
CS-US
CR-UR
Pavlov theory of temporal contiguity
Main conclusions drawn from over 7000 conditioning experiments –
(1) The CS and US must come into temporal contiguity
(2) They must do so repeatedly
Eyeblink conditioning experiment in rabbits
Conditioned responses increase with more trial blocks
Associations learned in classical conditioning
Bell (CS) = Food (US)
Salivate (CR) Salivate (UR)
Auto shaping in pigeons
Peck rate increases along the 20 trial blocks
Steadily
What factors determine the rate of learning?
The events to be associated need to be paired in time so that they occur together
Two ideas will be associated if they occur closely together in time, if we eat the apple it might taste crisp and sweet.. (Hume, 1700s)
‘Cells that fire together, wire together’ (Hebb, 1949)
Second order conditioning and sensory pre conditioning
Second order conditioning
Phase 1 - L-Food
Phase 2- T-L
Test- T?
Sensory preconditioning
Phase 1 T-L
Phase 2 L-Food
Test T?
Pavlov 1927
Theory of temporal contiguity
Pavlov reported that conditioning could occur with delays of several minutes, provided the CS continued to act until the onset of the US (‘delay conditioning’)
But if the CS comes to an end some time before the onset of the US (‘trace conditioning’), it is much harder to form CRs
Pavlov showed learning with trace intervals up to 2-3 min
Temporal contiguity…
Generally improves learning
Delay conditioning
CS before US
Trace conditioning
CS and then US much after
Simultaneous conditioning
At the same time
Backward conditioning
US before CS
Different ways to present CS and US in time
Delay conditioning
Trace conditioning
Simultaneous conditing
Backward conditioning
Bernstein experiment on taste aversion in learning
Exception to the contiguity rule
Something special about taste CS and illness US
Mapletoff – novel ice cream of maple and walnut (better control ‘Hawaiian’)
Participants children undergoing chemotherapy, ice cream 1 hour before session in Bernstein’s 1978 experiment
Pavlov theory of temporal contiguity
Studies of trace conditioning show that temporal contiguity between CS and US is generally important
But studies of taste aversion show that temporal contiguity is not necessary for successful associative learning
Does the role of temporal contiguity depend on the importance of other factors too?
Effects of stimulus intensity
Overshadowing
Overshadowing or cue competition effect on conditioned suppression in rats (Mackintosh, 1976)
Latent inhibition
LI
Stimulus pre-exposure (in phase 1) reduces later learning (in phase 2) when the stimulus is now relevant
LI can apply to both excitatory and inhibitory stimuli, i.e. LI of a potential inhibitor
Conditioned inhibition
Inhibitory learning
Phase 1: CS -> US
Phase 2: [CI + CS] -> no US
When the conditioned inhibitor (CI) is present, the CS is not reinforced
Effects of an exciter with an inhibition or another exciter
One exciter - conditioned responding is normal
One exciter + one inhibitory - less than normal
Two excitors- more conditioned responding
Functional significance
It makes sense to restrict learning to events that are likely to indicate the causes of important outcomes
Timing of events: temporal contiguity is generally necessary but not sufficient
Some potential signals don’t stand out from the background (overshadowing)
Pre-exposure can render stimuli ‘irrelevant’ (latent inhibition)
Significance of some events may be qualified by other stimuli (inhibitors, other excitors)
Some cues are redundant (blocking)
Some stimuli are poorly correlated with outcomes (relative validity)
Design and results of a relative validity experiment
The ‘correlated’ group has a better predictor (so X is redundant), whereas for the ‘uncorrelated’ group stimulus X is no worse than A or B at predicting US or no US, hence attention is maintained to X as its consequences are unknown and some conditioning to X is demonstrated (because half the time the US follows)..
Kamins unblocking experiment
Design and results
Inbuilt control for overshadowing, which in theory may also reduce conditioning to the light
Kamins blocking experiments
Inbuilt control for overshadowing, which in theory may also reduce conditioning to the light
Rascorlas experiments
CS-US contingency
Positive contingency
Zero contingency
Negative contingency
Rascorlas experiments
CS-US contingency
