Transcript PowerPoint slides

Chapter 3
Classical Conditioning
Ivan Petrovich Pavlov
(1849-1936)
• Classical (Pavlovian)
conditioning
• Medical physiologist
• Digestion
• Human/animal
differences
• Conditioned reflexes
en.wikipedia.org/wiki/File:Ivan_Pavlov_(Nobel).png
en.wikipedia.org/wiki/File:One_of_Pavlov%27s_dogs.jpg
Terminology
• Unconditional stimulus (US)
– Stimulus that elicits the innate reflex (e.g., food)
• Unconditional response (UR)
– Reflex action that occurs in response to US (e.g.,
salivation)
• Conditional stimulus (CS)
– Any stimulus that doesn’t originally elicit the UR
(e.g., bell)
• Conditional response (CR)
– The action elicited by the CS (e.g., salivation)
Conditioning and Awareness
• Awareness of conditioning not required for
learning
Innate
•
•
•
•
US-UR is an innate stimulus-behaviour
“Reflex”
Hardwired
Stereotypic pattern of behaviour
Example: Bell and Food
Later Trials
First Few Trials
CS
CS
US
US
UR
CR UR
Time
CS = bell
US = food
UR = salivation
CR = salivation
Processes
• Acquisition
– Acquiring a CR
– E.g., pair CS with US
• Extinction
– Reduce/eliminate a CR
– E.g., present CS without US
Measuring Conditioning
• Sometimes difficult to measure CR
– e.g., if CS & US close together, CR & UR can
overlap
• Test trial (probe trial)
– Give CS alone
• Intensity
– Does CR intensity increase with experience?
Example: Eyeblink Conditioning
• Airpuff on eye
• Blink
• UR vs. CR eyeblinks
CS (tone)
US (airpuff)
CR (blink)
UR (blink)
– UR blink faster than CR blink
Example: Taste Aversion
•
•
•
•
Very strong
Very persistent
Usually conditioned after one presentation
Experiment
–
–
–
–
Rats fed novel food (CS)
Injected with lithium chloride (US)
Choice: novel food or regular food
Chose regular food
Higher-Order Conditioning
• CSs and USs can be associated (First-order)
• CSs can be associated with other CSs
• Second-order conditioning
First-Order Conditioning
first-order
tone (CS1)
salivation (CR)
food (US)
Second-Order Conditioning
second-order
light (CS2)
tone (CS1)
salivation (CR)
Risk of extinction?
tone (CS1)
salivation (CR)
food (US)
CS+ and CS• CS+ (excitatory CS)
– CS predicts occurrence of US
– Activates behaviour related to US
• CS- (inhibitory CS)
– CS predicts non-occurrence of US
– Suppresses behaviour related to US
PAVLOV’S PROCEDURE
Trial Type A
Trial Type B
CS+
CSUS
• Randomize trial type presentation
NEGATIVE CONTINGENCY PROTOCOL
CSUS
• Context cues serve as CS+
Testing for CS• CS- produces absence
of CR
• No CR
– You’ve produced CS– Haven’t learned
anything
• How to measure
nothing…
• Summation test
– Measure CR with CS+
– Compound stimulus of CS+
& CS-; measure CR
• Retardation of acquisition
– Trained CS- and novel
stimulus; pair both with
novel US for same number
of trials
– Measure CR for both
– Prior learning of CSinhibits learning new
association
Short Delay Conditioning
• Strongest and most rapid
• Simple autonomic responses: 5-30 seconds
• Quick skeletal responses: 0.5 seconds
CS
US
or
Long Delay Conditioning
• Other distracting stimuli?
• Timing estimation required
CS
US
or
Trace Conditioning
• From “memory trace”
• Must remember CS
• Other stimuli interfere
trace interval
CS
US
Simultaneous Conditioning
• Weaker than short delay
• CS can’t signal onset of US
– Not predictive
CS
US
Backward Conditioning
• Ignores order; US comes first
• CS has no predictiveness
• Might become CSCS
US
Influences in Classical
Conditioning
CS-US Contiguity
• Closeness together in time and/or space
• Usually, more learning if greater contiguity
between CS & US
• Type of conditioning may influence this
• e.g., eyeblink vs. taste aversion
CS-US Contingency
•
•
•
•
If-then situation
X iff Y
Consistency of pairing CS and US
Greater contingency, greater learning
Stimulus Features
• Nature of stimulus affects its conditioning
ability
• Intensity
• Novelty
Compound Stimuli
• Two+ simple CSs presented at the same time
• Paired with US
Overshadowing
• Salience
• Exclusive regulation of CR by most salient
CS in compound stimuli
Latent Inhibition
•
•
•
•
Repeatedly present neutral stimulus (N)
Pair N with US
Harder to condition N as CS
CS- or habituation
Blocking
•
•
•
•
CS1 -- US
CS1 and novel stimulus (CS2) with US
CS1 --> CR
CS2 --> no or very weak CR
Textbook Error: p. 77
• “But suppose we eat two foods, one spicy
and the other bland. If we then become sick,
thanks to blocking we are likely to develop
an aversion to the spicy food -- even though
it may have been the bland food that caused
our illness.”
Sensory Preconditioing
• Pair two neutral stimuli repeatedly
• Pair one with US repeatedly until CR
produced
• Test other stimulus
• CR produced
Number of CS-US Pairings
asymptote
CR Strength
• Acquisition curve
• Non-linear
• Asymptote
Conditioning Trials
Intertrial Interval
• ITI
• Time between each CS-US pairing (i.e.,
between trials)
• Generally, around 30 seconds effective
Extinction of CR
Extinction
•
•
•
•
•
CS without US --> Extinction
Weakening and stopping of CR
Not forgetting
A type of conditioning
CS paired with absence of US
Spontaneous Recovery
•
•
•
•
After extinction, let time pass
Present CS again (no US)
Temporary, small return of CR
Shows extinction is not forgetting
Relearning/Reacquisition Effect
• Extinguish CR
• Recondition with CS-US pairing
• Fewer trials required
Putting it Together
Strength of CR
Acquisition
CS&US
Extinction
CS alone
Trials/Time
Spontaneous
Reacquisition
Recovery
CS alone
CS&US
Theories of Classical
Conditioning
Associationism, Stimulus Substitution,
Preparedness, Rescorla-Wagner
Associationism
• Linking together of:
– Events
– Memories
– Actions and consequences
• Contiguity, similarity, contrast
• Central to study of learning and behaviour
Ebbinghaus’ Memory
Experiments
• 1880s
• Nonsense syllables
– E.g., ZOG, PAF, TOB
• One subject
• Recite from memory
• Savings
– E.g., if 10 trials initially, then
after a delay 3 more trials,
savings = (10-3)/10 = 7/10 =
70%
Major Findings
• List length
• Effects of repetition
100
Percent Savings
– Overlearning
Forgetting Curve
75
50
• Effects of time
25
• Role of contiguity
• Backwards associations
20min 1hr 8.8hr 1day 2days 6days 31days
Time between study and relearning
Classical Conditioning
• Innate US-UR reflex pathway
• CS is associated with the US
• Through the associative process, CR is
produced
Stimulus Substitution Theory
•
•
•
•
•
Pavlov
CR and UR produced by same neural region
CS takes on properties of US
Substitution
CR should be the same as UR
Example: Sign Tracking
• Response not required
• US often food
• Stimulus (CS) indicates
US availability
• Subject “tracks” the
sign more and more
• CS takes on properties
of US
• Pigeon autoshaping
• Longbox autoshaping
F
= CS
F
= US
Biological Predispositions
Burns &
Domjan
(2000)
Timberlake &
Grant (1975)
Problems with SST
• CS not a complete substitute for US
– e.g., eyeblink differences
– Magnitudes
• CSs produce different responses
– Omissions and additions
• Compensatory conditional responses
Preparatory Response Theory
• Learn responses that prepare organism for
US occurrence
• Sometimes CR same as UR, sometimes
different
Example: Drug Tolerance
•
•
•
•
•
Neurophysiological dependencies
Siegel (1975)
Contextual stimuli act as CSs
Compensatory CR
Morphine
Contextual Stimuli Theory
• Rats on hotplate
• Between-groups study
• Independent variables:
– Morphine or placebo
– Location of injection (Home or Injection room)
• Dependent variable: time to lift feet
Results
• Control (placebo): 13 sec.
• Exp. Gr. 1 (morphine): 24 sec. (day 1) to 13
sec. (day 4)
– Injection room gives contextual cues
– Compensatory CR
• Exp. Gr. 2 (morphine):
– Day 1-3 injection room: 24 --> 13 sec. latency
– Day 4 home room: 28 sec. latency
Interpretation
•
•
•
•
•
•
US: Morphine
UR: Pain reduction
CS: Injection room
CR: Pain sensitivity
CS prepares rats for morphine injection
Body homeostasis
Rescorla-Wagner Theory
• Contiguity account
• Associative strength
• CS acquires limited amount of associative
strength on any one trial
Three Factors in Theory
• Maximum associative strength
• Difference between current and maximum
strength
• Number of additional CSs
Rescorla-Wagner Equation
DVn = c( - Vn-1)
DVn: change in associative strength for CS on
one trial
c: represents salience of CS and US; a constant
(0.0-1.0)
: maximum associative strength (magnitude
of UR)
Vn-1: associative strength already accrued by CS
Acquisition Phase
• Example: set c = 0.25,  = 10.0
• Vn-1 starts at 0.0
– For the first trial Vn-1 = V1-1 = V0
– For the second trial Vn-1 = V2-1 = V1
• First CS-US pairing:
DVn = c( - Vn-1)
DV1 = 0.25(10.0 - 0.0)
= 2.5
• Second CS-US pairing:
DV2 = 0.25(10.0 - 2.5)
= 1.88
• Total associative strength Vn (or “VTotal”) after two
trials:
V1 + V2 = 2.5 + 1.88 = 4.38
• Third CS-US pairing:
DV3 = 0.25(10.0 - 4.38)
= 1.41
Trial
0
1
2
3
4
5
6
7
8
9
10
DVn Vn (VTotal)
0.00 0.00
2.50 2.50
1.88 4.38
1.41 5.79
1.05 6.84
0.79 7.63
0.59 8.22
0.45 8.67
0.33 9.00
0.25 9.25
0.19 9.44
Associative Strength (Vn)
Acquisition Phase

DV3
DV2
DV1
Trials
Extinction
• Example:
– Set c = 0.25,  = 0.0
• After first extinction trial:
DVn = c ( - Vn-1)
= 0.25(0.0 - 10.0)
= -2.5
Associative Strength (Vi)
Extinction
DV1
DV2
DV3
Trials
 = 0.0
Blocking
• Learned CS blocks subsequent CSs
• Example
– CS = tone, novel CS = light
– c = 0.25,  = 10.0
– Completed 8 trials with just tone, V8 = 9.0
 DVn = 0.25(10.0 - 9.0)
• Only 1 unit of associative strength left to
split between the tone and the light
• Ultimately, Vtone=9.5 and Vlight=0.5