Transcript PowerPoint slides

Classical Conditioning
Ivan Pavlov
•
•
•
•
Medical physiologist
Digestion
Reflexes
Fistula
en.wikipedia.org/wiki/File:Ivan_Pavlov_(Nobel).png
en.wikipedia.org/wiki/File:One_of_Pavlov%27s_dogs.jpg
Stimuli & Responses
• Unconditional
– Stimuli and responses whose properties
are not dependent upon prior training
• Conditional (i.e., “dependent”)
– Stimuli and responses whose properties
occur only after training
• US, UR, CS, CR
Classical Learning Features
• Reflex
• Subject’s behaviour does not cause
delivery of US
• Association of stimuli
Temporal Arrangement
Short Delay
Long Delay
Trace
Simultaneous
Backwards
Terms
• Conditioning trial
– Each CS-US pairing
• Intertrial interval
– Time from end of one trial to start of next
trial
• Inter stimulus interval
– Time from the start of the first stimulus in
pair to second stimulus in trial
Interstimulus
interval
CS
US
Conditioning
trial
Intertrial
interval
Measuring Learning
• Test (probe) trial
– Present CS by itself (no US)
• Magnitude
– How much CR occurs
• Probability
– How often CS produces CR
• Latency
– How soon CR occurs after CS
Pseudoconditioning
• Increase in response due to just the US
• Sensitization
Controls
• Random control
– Exp. Gr.: CS-US pairings
– Control Gr.: same number of CS & US, but
randomized
• Explicitly unpaired control
– Exp. Gr.: CS-US pairings
– Control Gr.: same number of CS & US, but
presented far enough apart (even separate
sessions) to prevent association
CS Types
• Excitatory CS (CS+)
– CS predicts the occurrence of US
– Activates behaviour related to US
• Inhibitory CS (CS-)
– CS predicts the non-occurrence of US
– Suppresses behaviour related to US
Inhibitory Conditioning
• Why predict non-event?
• Unpredictable aversive events more
stressful
• Craske et al. (1995) measured general
anxiety in subjects with panic disorder
– Predictable and unpredictable attacks
– Before and after anxiety ratings
Results
General Anxiety
Unpredictable
Predictable
Before
After
• Similar anxiety before
• Post attack anxiety significantly lower if
attack predictable and higher if attack
unpredictable
• Ability to predict aversive event also
allows prediction of lack of aversive
event
• Application: stress management
techniques
– Can’t eliminate all stressors
– Introduce periods of predictable “safety”
– Reduces overall stress
Producing CS• Can produce CS- for either appetitive or
aversive US
• Most research done with aversives
• Inhibitory conditioning (and inhibitory
control of behaviour) only if there is also
an excitatory context for US delivery
– Can’t have CS- without CS+
– But, can have CS+ without CS-
Pavlov’s Protocol
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…
Techniques
• Bidirectional response
– Utilizes opposing responses
– Do one with CS+, opposite with CS-
• 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 CS- inhibits learning new
association
Backwards Conditioning
• Inconsistent results across studies
– Little learning at all, CS-, CS+
• Keith-Lucas & Guttman (1975)
– Backward conditioning and biological
plausibility
– Predator attacks prey
•
•
•
•
Antelope grazing
Lion attacks
Antelope clawed, but escapes
Pain (US); proceeds sight of lion (CS)
Experimental Procedure
• Rats fed sugar pellets
• Give one-time electric shock (US)
• Lights go out
– 1, 5, 10, or 40 seconds
• Toy hedgehog added (CS)
• Observe rat one day later
Model
• Sugar pellets = grazing
• Shock = pain of attack
• Hedgehog = lion
Control Groups
• Saw hedgehog, but no shock
• Shocked, but didn’t see hedgehog
Results
• Backward conditioning not seen in
controls
• In 1, 5, and 10 sec delay groups, got
backward conditioning
– Avoid hedgehog
– Don’t eat much when hedgehog present
• Fear induced by hedgehog is CR
Conclusion
• Biologically relevant CSs can cause
backward conditioning
Emotional Conditioning
•
•
•
•
Wide range of emotional responses
Emotions universal
Positive and negative
Emotional response to stimulus
reflexive
• Conditioned Emotional Responses
(CERs)
John Broadus Watson
• Hard-line Behaviorism
• British Empiricism
(nurture over nature)
• Early work with rats
• Shift to infant research
• Opposed
Introspectionism and
Freudian theories
Conditioning of Fear
• Watson & Raynor (1920)
• Albert B.
– Mother a wet nurse at Harriet Lane Home
(attached to Johns Hopkins University)
– Albert first assessed at about 8 months
– Emotionally stable, healthy
Method
• Present white rat
– No fear
• Present white rat and bang metal bar
– Produces CER of fear, avoidance,
withdrawl
•
•
•
•
US = noise, UR = startle
CS = rat CR = fear
CER generalizes to other furry objects
Video
• Study went for several months
• Intended to reverse CER conditioning,
but Albert B’s mother ended her job at
hospital
• Research led directly to Mary CoverJones’ counter-conditioning with Peter
What Happened to Albert
• Beck, Levinson & Irons (2009)
• Historical detective work
• Albert B.’s mother probably Arvilla Irons Merritte
– Douglas Merritte, born 9 March 1919
• Arvilla Merritte left Johns Hopkins
• Worked as assistant for ill wife of farmer
• Douglas Merritte died 10 May 1925, probably
from meningitis
Name
• Why Albert B.?
– Ethical concerns with confidentiality not
firmly established
– Watson may have played “name games”
– His sons William and James
– His name from John Albert Broadus,
Baptist minister… Albert B.
What Happened to Watson
• Affair with Rosalie Raynor, his grad student
• Divorce, fired, resigned as president of APA
• Worked for J. Walter Thompson advertising
agency; vice-president within two years
• Ponds cold ream, Maxwell House coffee
• Published books and articles on childcare
– Psychological care of infant and child (1928)
– Criticized by many modern child
experts/advocates, but not any more extreme than
other childcare texts of the time
– Strongly advocated against spanking and corporal
punishment
Nonhuman Studies of Fear
•
•
•
•
•
Typically use shock as US
Rats freeze
SSDS
Conditioned suppression ratio
Train operant response; train CS+ for
aversive US, test suppression of operant
response in presence and absence of CS
• Suppression video
Suppression Ratio
Suppression Ratio =
CS Responding
CS Responding + pre-CS Responding
• 0 if behaviour entirely suppressed
• 0.5 if no suppression
Pre-CS
CS
Calculation
S.R.
25
25
25/(25+25)=25/50
0.5
25
0
0/(0+25)=0/25
0
25
15
15/(15+25)=15/40
0.375
Sign Tracking
• Also now called autoshaping (Brown &
Jenkins (1968)
• 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
Biological Predispositions
Burns &
Domjan
(2000)
Timberlake &
Grant (1975)
Taste Aversion
•
•
•
•
•
US = stimulus producing illness
UR = illness, nausea, etc.
CS = (generally) novel taste
CR = nausea
Long delay or trace conditioning
Theory of Interest
• Contiguity
• Equipotentiality Premise
– Pavlov
– Doesn’t matter what you use as CS
– Any stimulus can be conditioned to any US
Initial Studies
•
•
•
•
Garcia & Koelling (1966)
Garcia, Ervin & Koelling (1966)
Difficulty getting published
Finally, accepted in Psychonomic
Science
Results
• US = poison, CS = novel flavour
• Delay between CS and US 5 - 22
minutes produced very strong CR
• Weaker, but significant CR (avoidance
of flavour) with up to 24 hour ISI!
• Violation of contiguity
Results
US=shock
Water Consumed
Water Consumed
US=X-ray
Pre-cond.
Post-cond.
Flavoured
water
“Bright-noisy
water
Pre-cond.
Post-cond.
• Violation of equipotentiality
• Some CS-US combinations more easily
learned
• Biological predispositions
Scientific Pardigms
• If evidence contradicts fundamental
premise…reject the evidence
• But… sometimes unexpected results
are correct
Eye Blink
• US = air puff UR = blink
• CS = noise, light, vibration, etc. CR =
blink
• Straight-forward classical conditioning
• Vehicle for examining neurobiology of
learning and memory
Brain Circuitry
Cerebral cortex
Mossy
fibres
CS
Pontine
nuclei
Tone
CS
Auditory
nuclei
CS
Interpositus
nucleus
CR
Red nucleus
CR
US
Climbing
fibres
Inferior olive
US
Trigeminal
nucleus
reflex
Cranial
UR paths
motor nuclei
Reticular
formation
Eyeblink
UR & CR
Corneal
air puff
US
Eyeblink Exercise
• We can do science to it!
• Yes, real science with nothing more
than a turkey baster, a pencil, and
paper.