In the last section, we’ve been
talking about a variety of variables and concepts that relate to classical
conditioning. Now, we know these things work. The question then becomes, why
do these things occur. As a result of that, we have a variety of theories
that relate to classical conditioning, and why classical condition occurs in
the first place. So let’s begin this discussion by talking about stimulus
substitution theory, which we see in slide two.
Stimulus substitution theory was presented by Pavlov. What
Pavlov contended is that the presentation of a conditioned stimulus excited
some particular brain area. When the UCS followed the CS, the brain
structures responsible for processing the conditioned stimulus and the
unconditioned stimulus were active at the same time. Consequently, that led
to a new neural pathway between the two particular centers. So, as we see on
slide three, when we activated one center, the other became activated as
well. After we did this pairing over a period of time and then presented the
conditioned stimulus alone, the conditioned stimulus in essence became a
substitute for the unconditioned stimulus.
Now other people did not agree with Pavlov’s conclusions.
One of these people is Siegel and as we see beginning in slide four. Seagal
contended that the conditioned response and the unconditioned response were
different. To demonstrate that, he used morphine as an unconditioned
stimulus and analgesia as an unconditioned response. Basically what he was
looking at was electric shock and used what is called foot shock. What he
used was light or tone as a conditioned stimulus. So, we have a conditioned
stimulus (a light), we have the morphine as an analgesic, and then foot
shock. Over a period of time what you should see is that when the morphine
is stopped, the light or tone should act as an analgesic. That’s not what
happened. What he found was that as you stop the unconditioned stimulus,
there was actually an increased sensitivity to pain, not a decreased
sensitivity. He found similar studies with other types of things. For
example, he found that an unconditioned response to insulin is hypoglycemia.
When the conditioned stimulus is paired with insulin, you basically get
hyperglycemia, major difference.
An unconditioned response to alcohol is hypothermia. A
conditioned response to a conditioned stimulus paired with alcohol basically
became hyperthermia, you became too hot.
The question becomes why. What Siegel basically says, and
as we see in slide six, is that tolerance represents the conditioning of a
response that is opposite of the unconditioned drug’s effects. What we have
are environmental cues that are present during the drug administration. That
is, the CS’s that antagonize the drug’s action and ultimately result in
lower pharmacological reactions to the drug. An example would be that if
you’re taking a particular drug, and you have some kind of cue out there
that’s used as a CS prior to using that drug. When you stop that drug,
basically the opposite begins to occur. What we find says Siegel (as we see
in slide seven), is that there’s an increased response to the drug can be
induced just by changing the stimulus. For example, let’s say that we have
an addict that’s out there using heroin, and in this environment they go to
a hotel or a lobby or whatever it may be to get their heroin and use it.
Let’s say it’s a hotel room. So they go a hotel room, they get their drugs
and they use it in their hotel room and then they go to sleep or whatever.
Then they change the location. So they go from the hotel
room to say a person’s house. They then take the same amount of the drug,
same kind of drug, but in this new location they have an overdose. The
question is why. Because in essence as Seagal identifies, as we see in slide
seven, the new environment doesn’t elicit the conditioned response that was
opposite the drug’s effect. That is, these conditioned stimuli out there
basically prepare the organism for the drug. When those conditioned stimuli
are not out there, you have a problem. This particular application is seen
in slide eight. Since the cues are different as we see in slide nine, you
get the opposite effect.
Now that wasn’t the only model out there out there related
to classical conditioning. The classic model that really kind of basically
accommodates and criticisms of the other models was developed by Rescorla
and Wagner and that’s called the Rescorla - Wagner model.
As we see in slide 10, Kamin assumed (remember from last
time) that surprise determined whether conditioning occurs. Rescorla and
Wagner basically expanded on that. They said how much surprise would
determine how much conditioning occurs. If you were surprised, you should
get lots of conditioning. So what we would do is take some kind of tone,
follow it by a shock, and you would get more conditioning than if the shock
was expected. That is the next time. As we can see in slide 12, there’s an
actual model of this. What Rescorla and Wagner basically say is this. When
you pair a CS and a UCS, there is an association that is formed between
them. If you repeat that pairing, the strength of that association increases
until it stabilizes at some rate. As we can see here, there’s a curve in the
figure and it goes up and maximizes at some point.
We tie that in with slide 13. So if we look at both
figures of slide 12 and slide 13. If we look at trial one which occurs very
early in the system, you get lots of associative strength. At time two, you
get a little bit less, until ultimately the value approaches some stabilized
level.
Rescorla and Wagner put this to a mathematical model,
where they represented the strength of each trial. On each particular trial,
the Vmax was the asymptotic value of V. That is, the peak of where it was.
V, (the associative strength) increases on each trial until it gets to V
max. And the ultimate amount of conditioning on any particular trial depends
upon the difference between V (associative strength) and the maximum amount
of associative strength.
So the model expresses four basic ideas which we see in
slide 14.
Number one, there’s a maximum amount of associative
strength that can develop between a CS and UCS, and it’s determined by the
UCS. As we’ve seen earlier, different unconditioned stimuli have different
associative strengths.
Number two. The associative strength increases with each
learning trial. However, the amount gained is affected by the amount of
previous conditioning. So as we saw in the figure, more associative strength
will occur during early trials than in later trials. You get a lot early and
you get smaller and smaller as time goes on.
Number three, the rate of the conditioning is dependent
upon the particular conditioned stimuli and unconditioned stimuli that are
presented. So as we see in slide 15, some stimuli develop very rapid amounts
of associative strength while others do it at much slower rates.
Also, some unconditioned stimuli produce a lot more rapid
learning than other conditioned stimuli. The classic example is shock or
pain produces very rapid conditioning.
And finally the level of conditioning is dependent upon
number the amount of learning prior to the conditioning of the stimulus and
the amount of conditioning associated with the unconditioned stimulus.
Thus as we see on slide 16, an unconditioned stimulus can
only support a certain amount of conditioning regardless of how many
different stimuli are associated with it. When several stimuli are presented
together such as a light and a tone, the stimuli in essence must share that
associative strength.
Now ultimately as we see in slide 17, the model has
different constructs that are similar to the ones that were described
earlier. That is, the strength of the association, (V) between the
conditioned stimulus and unconditioned stimulus in essence control the
amount of conditioning that’s displayed.
And there’s a maximum level (lambda), that determines the
strength an association can reach. And finally, the rate of learning (alpha)
determines how rapidly it can be reached.
So ultimately we can put all of this into a formula and
that’s what you see in slide 18. There’s a couple of different versions
depending on which book that you have. C ( the first one) is basically the
combination of alpha and beta. Alpha being the association of the
conditioned stimulus, and Beta referring to the intensity of the
unconditioned stimulus. The max again, (lambda) is determined solely by the
amount of conditioning with the UCS. That defines the maximum level.
Finally, the change in the associative strength is basically the change in
Vn) is the change out of Vn on that particular trial. So as you can see in
slide 19, you can put that into a mathematical formula and you can determine
how much associative strength that you will have on any particular trial.
would encourage you to walk through the numbers, walk through them very
slowly so you can understand how they work.
Now there’s some problems with the Rescorla - Wagner
model. Number one, the associative strength represents the strength of a
theoretical association, it’s not real. You can see the response rate but
you can’t directly observe the association presumed to produce the
particular behavior. So for example if an associative strength was 10, how
many drops of saliva would you expect to see. Well, you really don’t know
for sure because you need to know the values of alpha and beta before you
can predict it. That becomes very problematic, so what Rescoral and Wagner
did was they used arbitrary values to test a particular model. Basically
what the results showed was that the model predicts a learning curve of the
same shapes. You can’t predict the exact number of saliva drops but you can
predict whether salivation will increase or decrease on a particular trial.
So, in essence what the model does is gives you a good idea about what is
going to happen with classical conditioning upon each particular trial and
how that can be used in other cases. In addition to that, it explains other
variables such as blocking extremely well.
In the next section, we will kind of review some
applications of classical conditioning, so until that time you have yourself
a great day and we will talk to you soon.
Back