Transcript Chapter 8: Stimulus Control of Behavior

Chapter 8 – Stimulus Control of Behavior
• Outline 1
– Identification and Measurement of Stimulus
Control
• Differential Responding and Stimulus Control
• Stimulus Generalization
• Stimulus Generalization Gradients as Measures of
Stimulus Control
– Stimulus and Response Factors in Stimulus
Control
•
•
•
•
Sensory Capacity and Orientation
Relative Ease of Conditioning Various Stimuli
Type of Reinforcement
Types of Instrumental Response
• The concept of behavioral control
– S (stimulus)
• Can elicit responding (R)
• Can create expectancies (O)
• “Appropriate behavior” is often controlled by cues
(stimuli) in our environment.
– Get undressed for bed
– Get undressed at school
– Kiss girlfriend/boyfriend
– Kiss random person walking on campus
• This chapter considers how stimuli come to control
our behavior.
• Identification and Measurement of
Stimulus Control
– How do we know that a behavior is under
stimulus control?
• Consider Reynolds (1961)
– Train pigeons to peck a white triangle on a
red background.
• VI schedule
– Elicited steady pecking
– Test with red key (no triangle)
– Or white triangle (no red; background is black)
• Stimulus Generalization
– How specific is stimulus control?
• Early researchers (like Pavlov) examined this question.
• Guttman and Kalish (1956)
– Train
• VI schedule
• S+ = 580nm light (yellowish orange).
– Test
• Different colors
–
–
–
–
–
–
–
–
–
520 nm (green)
540 nm
550 nm
570 nm (yellow)
580 nm (yellowish orange)
590 nm (orange)
600 nm
620 nm
640 nm (red
• Random order
• In extinction
Generalization Gradient
• The generalization gradient indicates
stimulus control by color.
– It is somewhat specific for the training
stimulus.
• But generalizes to similar colors.
• There is no gradient for the hypothetically
color blind pigeons
– Color does not control responding.
• What determines the degree of stimulus
control obtained?
• Stimulus and Response Factors
– Sensory Capacity and orientation
• Must be able to sense it
– Rats don’t do well with color stimuli
– They do very well with odor stimuli.
• Horse study from book.
– Trained to select color over gray.
» Push lever with head.
» 85% correct
» All 4 picked blue and yellow over gray
» 3 picked green over gray
» Only 1 picked red over grey
» Implies they may have poor red sensation
• Relative ease of conditioning various stimuli.
– Whether a stimulus exerts control depends on
whether it stands out from other cues in the
environment
– Children’s book
• Big pictures
• Smaller words
– Overshadowing (first noticed by Pavlov)
• CS1 – loud sound
• CS2 – dim light
• More conditioning occurred to the stronger stimulus
– Conditioning was better for CS2 if presented alone
– Overshadowed by CS1 if presented together
• Type of Reinforcement
– Aversive vs. Appetitive seems to matter
• Foree and LoLordo (1973)
– Two groups of pigeons
– Both were trained to respond when presented
with a compound discriminative stimulus
• Red light and tone
– Group 1
• Step on treadle to gain food
– Group 2
• Step on treadle to avoid shock
– Test both groups with
• Red light alone
• Tone alone
• Visual stimuli tend to gain control over
appetitive stimuli
• Auditory stimuli tend to gain control over
aversive stimuli.
• Behavior systems?
– Food RF activates feeding system?
• Rats and pigeons are more likely to find food with
vision rather than hearing.
– Shock RF activates defensive system?
• Auditory cues may be particularly adaptive for
avoiding danger.
• Outline 2
– Learning Factors in Stimulus Control
•
•
•
•
•
•
Stimulus Discrimination Training
Effects of Discrimination Training on Stimulus Control
What is Leaned in Discrimination Training?
Interactions Between S+ and S-: Peak Shift Effect
Range of Possible Discriminative Stimuli
Stimulus Equivalence Training
– Contextual Cues and Conditional Relations
• Conditioned Place Preference
• Learning Factors in Stimulus Control
– left to their own devices animals come under
stimulus control based on the stimulus and
response factors discussed above.
– But can we train animals to make finer
distinctions?
• Stimulus Discrimination Training.
– Let’s go back to the color generalization study
– Train
• VI schedule
• S+ = 580nm light (yellowish orange).
– Test
• Different colors
–
–
–
–
–
–
–
–
–
520 nm (green)
540 nm
550 nm
570 nm (yellow)
580 nm (yellowish orange)
590 nm (orange)
600 nm
620 nm
640 nm (red
• Random order
• In extinction
Generalization Gradient
• Note that the pigeons treated the 590 nm
stimulus nearly the same as the 580 nm
• Can they tell the difference?
• How could we find out?
• Train with two stimuli.
– In operant conditioning we call them S+ (Sd) and S(S∆)
• S+ responding will result in RF
• S- responding will have no effect
– Pavlovian
• CS+ (CS-US)
• CS- (CS- no US)
• For our color discrimination
– S+ = 580 nm
– S- = 590 nm
Hypothetical Result
• Effects of Discrimination on Stimulus control
– Increased stimulus control
• 1) Discrimination narrows the generalization curve
• 2) Discrimination within a dimension narrows it even more
– Makes the relevant dimension clear?
» Tone vs. loudness
• Example: Jenkins and Harrison (1962)
– Trained with tones
• 3 groups
– 1) generalization
» S+ 1000-cps tone
– 2) discrimination
» S+ 1000-cps tone
» S- no tone
– 2) within discrimination
» S+ 1000-cps tone
» S- 950 cps tone
• What is learned in Discrimination Training?
– Example
• S+ (light) S- (tone)
• 1) learn about S+ alone
– Respond during light
– Learn nothing about tone
• 2) learn about S- alone
– Suppress responding during tone
• Learn nothing about light
• 3) learn about both (Spence’s Theory)
– Respond during light
– Suppress responding during tone
• Spence’s Theory of Discrimination
Learning
– The S+ becomes excitatory
• Signals RF
– The S- becomes inhibitory
• Signals lack of RF
• How do we test this?
– Responding to S+ and not responding during
S- is not enough
• Any of the 3 theories predict this
• The peak shift can be considered
evidence for Spence’s view.
• Range of Possible Discriminative Stimuli
– Many kinds of stimuli have been examined
• Simple
– Color
– Tone
• Complex
– Number
– Time of day
– Kind of music
» Carp
» Blues vs. Classical
» John Lee Hooker vs. Bach
• Types of Stimuli continued
– Artists
• Monet vs. Picasso
– Internal
• Hunger
• Drugs
• Drugs
– Cocaine = left lever
– Saline = right lever
• Antagonist?
• Other drugs?
– Amphetamine?
– Caffeine?
• Stimulus Equivalence Training
– We have seen that discrimination can
sharpen stimulus control
• Treat similar stimuli differently
– Can we produce the opposite effect.
• Train animals to treat very different stimuli
similarly?
• Honey and Hall (1989)
• Phase 1
Group 1
Noise = Food
Clicker = Food
Group 2
Noise = nothing
Clicker = Food
• Phase 2
Noise= foot shock
Noise = Foot shock
• Test
Clicker
Clicker
• Which group is more afraid of the clicker?
– Group 1
• Seems a common outcome causes the rats to treat the stimuli similarly.
• Common Coding – a typical equivalence experiment
– Based on Urcuioli, Zentall, Jackson-Smith, and Steirn (1989)
•
•
•
•
•
Phase 1 (Many-to-One Matching-to- Sample)
R  R+GV  R+GG  G+RH  G+R-
• Phase 2 (new comparisons)
• R  B+Y• G Y+ B• Test (does learning transfer)?
• V  B Y?
• H  Y B?
•
Does Equivalence training cause stimuli to become harder to
discriminate?
– Based on Kaiser, Sherburne, Steirn, and Zentall (1997)
•
•
•
•
•
Train
R  R+GV  R+GG  G+RH  G+R-
•
•
•
•
•
Test (discrimination)
Consistent
Inconsistent
R+
R+
V+
VGGHH+
•
Which Group learns faster?
– pecks S+ 90% of the time
• Sidman
– True equivalence must demonstrate three
concepts
• 1) Reflexivity (sameness)
– If A = A, B = B, C = C, and so on.
• 2) Symmetry (bidirectional equivalence)
– If A = B then B = A
• 3) Transitivity (transfer equivalence across
stimuli)
– If A = B and B = C then A = C
• Spoken and written speech involves these
three aspects of equivalence.
• Reflexivity (sameness)
– Apple = Apple
– Orange = Orange
=
=
• Symmetry (bidirectional equivalence)
– A(object) = B (word)
– B (word) = A (object)
•
= Apple
• Apple =
• Transitivity.
– If A (object) = B (spoken word)
– And B (spoken word) = C (written word)
– Then A (object) = C (written word)
• If
=
• And
= Apple
• Then
= Apple
• Some have argued stimulus equivalence
is a human trait
– Requires language
• People with good verbal skills can form equivalent
relationships easily.
• Those without have much more difficulty.
• Animals?
• Reflexivity (sameness)
– Pick the thing that looks the same
• Based on Zentall and Hogan (1978)
• Train (Identity matching-to-sample)
– R R+G– G G+R-
• Test (with “novel” stimuli)
–BBY
–YYB
• Symmetry (bidirectional equivalence)
– Based on Zentall, Sherburne, and Steirn (1992)
• If red = food, then food = red
• If green = no food, then no food = green
• Train (differential outcomes procedure)
– R  R+G- (food)
– G  G+R- (noFood) do over until correct to move on
• Test
– Food  R G?
– No Food  G R?
• Transitivity
– Based on Steirn, Jackson-Smith and Zentall (1991)
• Phase 1 (Differential Outcomes)
– R  R+ G- (food)
– G  G+ R- (no food)
• Phase 2
– food  B+ W– no food  W+ B-
• Test
– R  B W?
– G  B W?
• Train (Differential Outcomes)
• Phase 1
– R  R+ G- (food)
– G  G+ R- (no food)
• Phase 2
– food  B+ W– no food  W+ B-
• Test
– R  B W?
– G  B W?
• Contextual Cues and Conditional
Relations
– Conditioned Place Preference
• Inject Rat with drug and confine to one
side of chamber
• Test later (next day)
– Drug free
– Which side do they prefer?
• Heroin – Good sick?
• Two groups of Thirsty Rats
– Morphine Group
• Inject with morphine
• Place in Side 1 with Sacch.
– Control
• Inject with Saline
• Place in Side1 with Sacch.
• Test
– open access
– two bottles in each side
• Sacch vs. Water
• Results
– Morphine Group?
• More time in Side 1
• avoid Sacch.
– Control?
• Equal time each side
• prefer Sacch.