Transcript Stimulus Control

Operant Conditioning – Chapter 8
Theories of Learning
October 14, 2005
Class #23
Discriminative Stimulus (SD )

A stimulus in the presence of which responses
are reinforced and in the absence of which
they are not reinforced…
Basically, its a signal that a response will be
followed by a reinforcer
 It signals the availability of the reinforcer thus
increasing the probability it will occur
 The behavior then is considered to be under
stimulus control
• See next slide…

Stimulus Control


Having stimulus control means that the probability
of the behavior varies depending upon the stimuli
present
Our behavior often comes under stimulus control
without our awareness

A person that contributes to charity generously
while in church, but watches every penny spent
while at work
SD  R  SR

??  ??  ??

Stimulus Control: Generalization
Generalization is when responses to one
stimulus occur to other, usually similar stimuli
 Generally, as the training and test stimuli
become more different responding will
decline, producing what is called a
generalization gradient

Generalization Gradient
350

250

200
150
100

50
0
63
0
61
0
59
0
57
0
55
0
53
0
0
51
Responses
300
Guttman & Kalish
(1956)
Wavelength (nanometers)
Pigeons were
reinforced for
pecking a 580 nm lit
key (orange-yellow)
on a VI schedule
A test session was
then given where
many different
colored key lights
were presented in
extinction
Stimulus Generalization as a
Measure of Stimulus Control
Responses
400
350
300
250
200
150
Training S+
100
50
0
500
520
540
560
580
600
620
640
Wavelength (nanometers)
Pigeons were trained to peck in the presence of a colored light of 580 nm
wavelength and then tested in the presence of other colors.
Guttman & Kalish (1956)
Stimulus Generalization: Applied


Rincover & Koegel (1975) were attempting to train autistic children to
imitate
 One child, Joey, learned to touch his head when the Experimenter
said “do this” and touched his own head; unfortunately, it took 300
trials
 A generalization test in a different location and with a different
experimenter, resulted in 0% correct responses
 It turned out that the first Experimenter had frequently held Joey’s
hands in his lap before each trial began
 Another generalization test was conducted and this time the new
experimenter held Joey’s hands first, then said “do this” & touched
his head: resulted in 60% correct responding
 A third test was conducted and the experimenter just held Joey’s
hands in his lap: 100% correct responding
What gained stimulus control???
Stimulus Control: Discrimination

A better procedure for training Joey would
have been to use discrimination training…
Discrimination training involves presenting at
least 2 stimuli but reinforcing only one of them
 Discrimination is differential responding to
multiple stimuli
 A discriminative stimulus for extinction (S∆) is a
stimulus that signals the absence of
reinforcement

Stimulus Control: Discrimination

Azrin & Hayes (1984) used discrimination training to
increase males’ sensitivity to cues signaling females’
level of interest in males
 Subjects were shown 24 film clips of couples
conversing; half were positive instances of female
interest and half were negative instances
 Subjects were reinforced for correctly identifying
whether each slide involved female interest or not
 The male subjects increased their correct responses
by over 50% as a result of the training
Stimulus Control:
The Peak Shift Effect

The peak shift effect is when the peak of a
generalization gradient shifts away from the S∆
following discrimination training
Hanson (1959)

Procedure

Several groups of pigeons pecking different
colored key lights
• The control group was trained with just one
stimulus, a 550 nm key light, reinforced on a VI
schedule
• The experimental groups were given successive
discrimination training with 2 key lights, the 550
nm light was always the SD, and different groups
had a different wavelength as the S∆
450
400
350
300
250
200
150
100
50
0
Wavelength (nanometers)
62
0
60
0
58
0
56
0
54
0
52
0
Control
Exper
50
0
48
0
Responses
Hanson (1959): The
experimental group had a 590 S∆
“Mr. Shallow” example…

“Mr. Shallow” identifies women on just one
quality – introversion vs. extroversion

He has just broken up with Jackie who is an
introvert (he said she’s boring)
• Jackie is the discriminative stimulus for
extinction (S∆) here

Earlier in his life, “Mr. Shallow” had an
exciting relationship with Liz who was a pure
extrovert
• Liz is the Discriminative Stimulus (SD ) here
“Mr. Shallow” example…

“Mr. Shallow” moves to a new city and starts
hitting the singles bars seeking a new partner

According to the peak shift, what type of
woman will he be trying to “hook up”
with???
Stimulus Control:
Multiple Schedules

A multiple schedule is a complex schedule of
reinforcement

It is like a chained schedule except that each link is
reinforced

A multiple schedule consists of 2 or more simple
schedules presented in sequence, each resulting in
reinforcement and each having a distinctive SD
Stimulus Control: Behavioral Contrast


Behavioral contrast is a change in a reinforcer’s
effectiveness caused by previous experience with
other reinforcers
The change may be due to the schedule of
reinforcement or the magnitude of the reinforcer


Positive contrast occurs when the previous reinforcer
was less valued (current reinforcer’s effectiveness is
increased)
Negative contrast occurs when the previous reinforcer
was valued more (current reinforcer’s effectiveness is
decreased)
Running Speed (ft/sec)
Crespi (1942)
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
256-16 Pellets
16-16 Pellets
1 - 16 Pellets
2 4 6 8 10 12 14 16 18 20 2 4 6 8
Trials
Stimulus Control:
Behavioral Contrast

If presented a signal that indicates a change in
rate or magnitude of reinforcement is about to
begin, anticipatory contrast will often occur

Anticipatory contrast is a change in behavior
(increase or decrease) in inverse anticipation
of an upcoming change in reinforcement
(decrease or increase)
Seems to be the case with us to right
from the start…
¢
Kobre and Lipsitt (1972)

Participants
• Newborns

Procedure
• Artificial nipple that either gave reinforcer of
sucrose or water

Results
• Support contrast effect
Stimulus Control: Fading &
Errorless Discrimination Learning

Errorless Discrimination Learning involves
setting up the discrimination training
procedure in such a way that the organism
learns the discrimination without making
mistakes
Do all shaping (pretraining) with the SD only,
which is always associated with reinforcement
 Fade in the S∆ gradually (start with a weak
version and work up to the final strength of the
stimulus)

Stimulus Control: Fading & Errorless
Discrimination Learning

There are fewer adverse side effects (e.g.,
frustration) during learning when using an
errorless discrimination procedure
 However, organisms are less flexible when the
discrimination changes if they were trained with
an errorless procedure versus the standard
approach
 Therefore, you should only use the procedure
for factual information where the response
requirements are unlikely to change
Haupt et al. (1975)
Participant
 9 year-old girl with math difficulties
 Procedure
 Correct answers were gradually faded out

Stimulus Control:
Additional Applications

Training animals for public performance

Targeting
Eliminating problem behaviors in animals
 Training in and/or eliminating human
behaviors

Improving study habits
 Reducing insomnia

Credits:

www.radford.edu/~pjackson/ExtinctIC.ppt