Transcript No Slide Title

Innate Knowledge
(what an organism is born with)
Experience leads to changes in knowledge
and behavior
Learning refers to the process of adaptation
Of behavior to experience.
Memory refers to the permanent records that
Underlie this adaptation.
How experience changes an organism
Learning Perspective
EVENT
Change in
Behavior
Memory (Cognitive) Perspective
EVENT
Change in what
organism knows
Learning vs Cognitive (Memory) Perspectives
Learning focuses on simple tasks
Pavlovian Conditioning
built-in
reflex
CS
tone
food
salivation
US
UR
food
salivation
tone
new learned
association
salivation
Operant Conditioning
Animal makes random response
then accidentally presses bar
R
(response)
Re
(reward or
reinforcement)
Animal is more likely to press bar
Memory perspective focuses on complex tasks
Recall
Present a list of words (STUDY)
(wait)
Write down all the words (TEST)
Word Completion
S
M
T O
What’s the word?
Question answering
What is the red pigment that carries
oxygen in the blood called?
The Learning Perspective uses Animals as
Subjects
can control environment
belief that laws of learning apply to all
animals
The memory perspective uses humans
•we care more about humans
•language stimuli
The learning perspective takes an Associationistic
Approach
Mind is a collection of innate reflexes and learned
associations stored in the brain
Complex behavior is gradually built up out of
simple associations
Memory perspective adopts
the information processing approach
The brain is a computer
•It has software or programming called the
mind
•It has a “programming language”
Experience constantly adds to the program
It has a “central processor”
(innate knowledge)
It has a large “hard disk”
(long-term store)
NID Experiment
Learning “It’s a conditioning study!”
42
unconditioned
stimulus
(US)
NID
42
“forty-two”
unconditioned
response
(UR)
“forty-two”
conditioned
stimulus
(CS)
after learning
NID
“forty-two”
Memory “It’s a memory experiment”
cue
study NID
test NID
cued recall
item-to-be-recalled
42
?
Pavlovian Conditioning
Definitions
food
US
air in eye
shock
...
...
...
...
salivation
UR
blink
“ouch!”
US is paired with CS
CS
tone
US
food
tone
salivation
salivation
CR
conditioned
response
Sometimes the CR is like the UR
Sometimes the CR tries to compensate for US
Example:
If US is shock, CR is fear
and behavior that compensates
for pain
Forgetting Conditioning
Acquisition
Extinction
Strength
of CR
US is
paired
with CS
CS never
paired
with US
Extinction
Is it because conditioning is gone?
OR
because conditioning is inhibited?
Inhibition Hypothesis
+
CS
US
acquisition
CS
US
inhibition
extinction
Loss Hypothesis
+
CS
US
acquisition
CS
US
extinction
CS
US
Evidence Favors Inhibition Hypothesis
1.
prob.
of
CR
Spontaneous Recovery
wait a while
time
Suggests that the original learning was not lost
2. Disinhibition Effect
External Inhibition
light food
salivation
light
salivation
light
sali … TONE!
(salivation stops)
extinction
light food
salivation
light
salivation
light
sali
light
sa
light
nothing
light
TONE!
salivation starts
again
Tone inhibits the inhibition from extinction
Shows extinction is due to inhibition, not loss
3. Stimulus-compounding experiments
(Rescorla, 1979)
Experimental
Group
Control
Group
Phase 1
Tone
Shock
Tone  Fear
(same)
Phase 2
Extinction
Tone + Light - 0
Tone - 0
no shock
Light - 0
stimulus
compound
Phase 3
Light + Buzzer - Shock
Test to see
if light
is inhibited
Experimental group develops
fear more slowly
The light became a conditioned inhibitor of fear
Extinction leads to inhibition
What is learned in Pavlovian Conditioning?
S-R view
OR
S-S view
tone CS
tone CS
food  salivation
US
R
food  salivation
US
R
Which one is right?
Sensory Pre-conditioning Experiment
Phase 1
Buzzer
CS2
Phase 2
Light
...
...
Light
CS1
Food
US
Light
Phase 3 (test)
Buzzer
...
Salivation
Salivation
?
What happens?
S-R view predicts no salivation
S-S view predicts salivation
Results show salivation supporting S-S view
Sensory pre-conditioning shows S-S association is learned
Buzzer
predicts
buzzer
causes
salivation
S-S
Buzzer
predicts
buzzer
does not
cause
salivation
Light
Food
Salivation
Light
Food
S-R
Salivation
Conclude: at least some of the learning is S-S
Konorski’s (1948) second-order conditioning experiment
Phase 1
light . . . food  salivation
Phase 2
buzzer . . . light  salivation
buzzer
salivation
This is second-order conditioning
Phase 3
light . . . shock  leg withdrawal
Test phase
buzzer . . .
What happens?
salivation or leg withdrawal
S-S view
light
buzzer
S-S
food
salivation
S-S
buzzer
shock
leg withdrawal
Buzzer should lead to leg withdrawal
S-R view
food
light
salivation
buzzer
shock
leg withdrawal
Buzzer should lead to salivation
Conclude:
Both S-S and S-R learning occur
Conclude:
Both S-S and S-R learning occur
But why didn’t leg withdrawal occur in
Konorski’s experiment?
Phase 1
light
S-S
food
salivation
When do you get conditioning?
Pavlov
“If CS and US occur
at around the same time”
Temporal Contiguity View
Modern View (Rescorla)
“If the CS predicts whether
the US will happen”
Contingency View
Contingency
Perfect positive contingency
lightning
no lightning
thunder
20
no thunder
0
0
345
prob (thunderlightning) = 1.0
prob (thunderno lightning) = 0
Strong negative contingency
stars
sun
1
no sun
200
no stars
200
60
Contiguity without Contingency
bird
no
bird
airplane
10
20
no plane
20
40
prob.
(birdplane) = .33
prob.
(birdno plane) = .33
bird and
plane are paired
A quick test for contingency
no
Stim 2 Stim 2
Stim 1
a
b
No Stim 1
c
d
a·d > c·b
then positive
a·d = c·b
zero contingency
a·d < c·b
then negative
You can have a positive contingency even when
pairing is the least frequent possibility
Example: can you learn that
and “cat” are associated?
hear
“cat” no “cat”
see
100
900
1,000
no
200
9,800
10,000
prob (“cat”
prob (“cat”no
) = .10
) = .02
positive contingency
Contingency and Conditioning
Standard Experiment
tone
time shock
tone tone
shock shock
tone
shock
shock no shock
tone
4
0
no tone
0
3
perfect positive
contingency
tone leads to fear
Unpaired Experiment
shock no shock
tone
0
4
no tone
4
0
perfect negative
contingency
tone becomes a
conditioned inhibitor
of fear
Random Pairing
T
T
S S
T T
S
S
shock no shock
tone
2
2
no tone
2
2
zero
contingency
No conditioning occurs even though tone and shock
are occasionally paired
Partial positive contingency
shock no shock
tone
2
1
no tone
1
2
conditioning
occurs
Conclude: contingency, not contiguity matters
Random Pairing Experiment
shock no shock
tone
20
20
no tone
20
20
zero
contingency
no conditioning
occurs
Shows that there must be some contingency
between CS and US to get conditioning.
Contiguity is not enough.
Fear vs Anxiety
•if tone predicts shock then animal becomes
afraid after tone
(like a phobia)
•if tone does not predict shock (random pairing)
animal ignores the tone and experiences
something like anxiety (unfocused fear)
Blocking Effect
phase 1
tone
shock
16 times
phase 2
tone+light
shock
8 times
phase 3
light alone
NO FEAR IS ELICITED!
Shows that contingency alone doesn’t produce
conditioning
Get conditioning
when
(a) CS predicts US
AND
(b) CS tells something we didn’t
already know
Explaining the Blocking Effect
trial 1
CS
tone
US
shock
•CS is surprising
US is surprising
•Process CS and US
•Create association between CS & US
trial 2
tone
stock
trial 3
trial 8
CS  memory of  fear US
tone
shock
shock
Shock is not surprising any more so no
additional strengthening of association occurs
trial 3
trial 8
CS  memory of shock  CR fear
tone
US-shock
Strength of
CS-US Association
Tone fully activates memory of US shock, so
when shock comes it is not at all surprising.
No additional learning occurs.
no more
learning
rapid
learning

1

2

3
 
4 5

6

7

8
Phase 2
trial 9
CS1 CS2
tone + light
memory
of shock
fear
The light is surprising, so it is processed. But the
tone completely predicts the shock
US
shock
So the US shock is not processed, so no
association is formed between light and shock.
Explains blocking effect
General Conclusion
Stimuli are associated when they provide
information that the organism doesn’t already
know.
Modern view of conditioning
and the blocking effect
(1) Form association between stimuli only if they
are actively processed or “rehearsed” together.
(2) Stimuli are processed only if they are
unexpected
(3) As conditioning proceeds both the CS and US
become less surprising. So they are processed less
and, hence, less additional learning occurs.