00:00:05
hello and welcome to another episode of
00:00:10
the study of learning and behavior in
00:00:14
this particular episode we're going to
00:00:16
talk about Pavlovian learning of
00:00:19
predictive relations okay
00:00:22
now we often talk about Pavlovian
00:00:26
conditioning involving the pairing of a
00:00:29
conditioned stimulus and an
00:00:31
unconditioned stimulus and essentially
00:00:35
what we're gonna do today is to talk
00:00:37
about the many different ways in which a
00:00:41
conditioned stimulus can be paired with
00:00:43
an unconditioned stimulus so the pairing
00:00:46
is not just one thing it's kind of a
00:00:49
summary for a whole host of different
00:00:51
possible procedures and the reason it's
00:00:54
worth talking about this whole host of
00:00:57
possible procedures is that very very
00:01:01
subtle differences in a temporal
00:01:05
sequencing and overlap of condition and
00:01:08
unconditioned stimuli can make a huge
00:01:09
difference in terms of what is learned
00:01:12
how rapidly its learned and what kind of
00:01:16
behaviors you you get out of the
00:01:18
situation so if we may look at the first
00:01:22
slide this list gives you an a under
00:01:26
title order or no technical terms for
00:01:31
the various ways in which conditioned
00:01:34
and unconditioned stimuli can be paired
00:01:38
and we're gonna go look at each of these
00:01:41
specifically with respect to the timing
00:01:46
of the conditioned stimulus in relation
00:01:50
to the timing of the unconditioned
00:01:52
things and that temporal information for
00:01:58
each of these procedures is provided in
00:02:01
the next slide so let's look at the
00:02:09
short first procedure on top and all of
00:02:12
these procedures time travels from left
00:02:16
to right
00:02:17
okay and each of these pair of lines
00:02:21
represents a conditioning trial
00:02:24
so on the first conditioning trial type
00:02:30
of conditioning that's depicted is short
00:02:32
delayed conditioning and here the
00:02:35
conditioned stimulus comes on and
00:02:39
shortly thereafter the unconditioned
00:02:41
stimulus comes on and these on sets are
00:02:45
indicated by the vertical displacement
00:02:48
of the line and the conditioned stimulus
00:02:51
stays on for a while and then goes off
00:02:53
and then the unconditioned stimulus
00:02:57
similarly goes off and there is a bit of
00:03:00
overlap between the conditioned stimulus
00:03:03
and the unconditioned stimulus and this
00:03:06
by far produces the most rapid
00:03:08
acquisition of conditioned behavior and
00:03:11
you get vigorous responding so usually
00:03:15
when people are talking about csus
00:03:17
pairings they're usually talking about
00:03:18
the so called short delayed procedure
00:03:22
now this is in contrast to let's jump
00:03:29
down to the simultaneous conditioning
00:03:33
procedure that's next to the last one
00:03:36
down there now if you think about it
00:03:40
Babylonian conditioning involves the
00:03:42
establishment of an association so it's
00:03:45
an association between the CS and the US
00:03:48
and it so it's kind of the linkage
00:03:51
between these two stimuli and you might
00:03:54
think and and it's reasonable of the
00:03:59
thing that to produce the strongest
00:04:02
linkage you should present the stimuli
00:04:05
as close together as possible and
00:04:09
simultaneously if possible and that's
00:04:12
the procedure for simultaneous
00:04:13
conditioning so in simultaneous
00:04:15
conditioning the conditioned stimulus
00:04:17
and the unconditioned stimulus occur
00:04:18
simultaneously and you might think this
00:04:23
produces most rapid learning strongest
00:04:26
evidence of learning and if you thought
00:04:29
that you would be wrong
00:04:31
in fact simultaneous conditioning is
00:04:33
really problematic it doesn't produce
00:04:36
very vigorous conditioned behavior and
00:04:39
why is that that's kind of strange well
00:04:43
and this is where the concept of a
00:04:45
predictive relationship comes in the way
00:04:48
to think about these various procedures
00:04:50
is on the procedures that produce the
00:04:53
best learning our procedures in which
00:04:55
the conditioned stimulus has a strong
00:04:59
predictive relationship with respect to
00:05:01
the US now how can you assess these
00:05:05
predictive relationships well let's
00:05:06
think about an actual situation that
00:05:09
situation that many of you have no doubt
00:05:12
encountered now driving along the road
00:05:15
particularly in a rural road you come
00:05:19
across a railroad crossing right and a
00:05:22
railroad crossing is indicated by their
00:05:24
cross sign and there are flashing lights
00:05:27
if the train is about to come
00:05:30
and there's also warning bells that ring
00:05:34
when lights flash so they're trying to
00:05:37
get your attention and trying to get
00:05:40
your attention so that you're don't
00:05:43
drive across the track and get hit by
00:05:45
the train
00:05:46
okay now under do you ordinarily if
00:05:52
you're driving along and and the train
00:05:59
signal comes on do you immediately stop
00:06:03
well I don't know about you but often
00:06:09
what I see and sometimes what I do when
00:06:13
it'll Comicon these railroad warning
00:06:17
signals I gotta look quickly left and
00:06:20
right and if I don't see a train I got
00:06:22
it and try to get across the tracks so
00:06:27
in that so ordinarily the railroad
00:06:30
signals are not programmed in a way that
00:06:33
makes the signal a good predictor of the
00:06:38
train if the railroad signal was
00:06:44
according to a short delay procedure
00:06:46
that means like this railroad signals
00:06:50
start to flash and boom two or three
00:06:52
seconds later the Train is there so if
00:06:56
you haven't stopped when the signals
00:06:59
come on near the Train it's gonna hit
00:07:01
you that would be a short delay
00:07:03
procedure and if the railroad crossings
00:07:06
were programmed that way you would lose
00:07:10
a lot of people in traffic in railroad
00:07:13
accidents but the word would get around
00:07:18
that boy if you see those things
00:07:20
flashing better not cross the track
00:07:22
because the train is gonna be there
00:07:24
right away
00:07:25
now consider a simultaneous conditioning
00:07:29
procedure and a simultaneous procedure
00:07:33
there warning lights and warning bell at
00:07:38
a railroad crossing starts when the
00:07:40
Train is actually at the crossing if the
00:07:44
warning bell it comes on at the same
00:07:48
time that you see the Train does the
00:07:51
warning bell give is the warning bell
00:07:54
necessary for you to predict a train no
00:07:57
you can just respond on the basis of
00:08:00
seeing the train so on the within a
00:08:04
simultaneous pairing kind of situation
00:08:07
the conditioned stimulus is irrelevant
00:08:10
yeah if what we're basically trying to
00:08:13
predict and adjust to and respond to
00:08:17
appropriately is the unconditioned
00:08:19
stimulus which in this example is the
00:08:22
train so if the warning stimulus in a
00:08:24
train are there at the same time that
00:08:26
morning stimulus is irrelevant
00:08:28
and that's why simultaneous conditioning
00:08:30
typically doesn't produce very strong
00:08:33
conditioned responses our conditioned
00:08:35
response was meant to be anticipate
00:08:37
anticipatory their reflections or of the
00:08:40
prediction that the US will occur and in
00:08:43
a simultaneous case then the conditioned
00:08:46
stimulus doesn't provide any predictive
00:08:49
information it's useless and so you
00:08:51
don't you come you don't respond to me
00:08:54
so
00:08:55
the simultaneous case let's move up in
00:08:59
the diagram to the long delay
00:09:02
conditioning procedure this is in fact
00:09:05
what is typically employed at railroad
00:09:09
crossings the warning signal comes on
00:09:12
and it stays on a while before the chain
00:09:16
shows up so under these circumstances is
00:09:24
the onset of the signal indicative that
00:09:29
you're about to see the train well no
00:09:33
not really the onset of the signal tells
00:09:36
you that the train will not be there for
00:09:39
a while and so with a long delay
00:09:47
conditioning procedure what you see
00:09:53
subjects have repeated experience with
00:09:57
this kind of appearing is that the
00:09:59
condition response will be delayed until
00:10:03
just before the train arrives just
00:10:06
before the US arrives so in their
00:10:13
example of a railroad crossing if you
00:10:16
hear their warning stimulus come on you
00:10:19
first see the flashing light and hear
00:10:21
the bell that tells you the train is not
00:10:24
gonna be there for a while and so that's
00:10:27
why most drivers that I see as soon as
00:10:31
the warning signal comes on they
00:10:35
accelerate to try to get across the
00:10:37
track of course it's not a good idea to
00:10:41
do that your car may malfunction and
00:10:45
there are examples of busses and things
00:10:48
that get stuck on the track and get hit
00:10:51
by the train but and and you can't be
00:10:55
sure what the warning delay is and so
00:11:01
it's best not to take a risk with that
00:11:05
but Pavlov actually studied both
00:11:09
the short delayed procedure and a long
00:11:11
delay conditioning procedure he observed
00:11:16
rapid development and condition
00:11:18
salivation with the short delayed
00:11:19
procedure the long delayed procedure the
00:11:22
subjects initially responded when the CS
00:11:26
game on
00:11:27
but then the salivation started to occur
00:11:34
later and later during the conditioned
00:11:37
stimulus so that it occurred just before
00:11:42
the food was presented any salivation
00:11:46
that occurred early during me in an
00:11:48
interval was inhibited and Pavlov called
00:11:53
this phenomenon the inhibition of delay
00:11:57
you clearly get inhibition of delay in
00:12:00
practical situations for sure okay so
00:12:04
that's the long delay procedure let's
00:12:06
move up to the trace conditioning
00:12:09
procedure this is also a procedure that
00:12:12
Pavlov study and it's so kind of
00:12:16
interesting here well we sort of liked a
00:12:19
short delay procedure we start the
00:12:22
conditioned stimulus shortly before the
00:12:25
unconditioned stimulus but here we turn
00:12:27
off off this the warning stimulus for a
00:12:32
while before the u.s. is presented so
00:12:36
there is a gap there is a temporal gap
00:12:40
between CS and the US and the gap can be
00:12:48
as little as a half a second and having
00:12:52
that gap in there makes a huge
00:12:55
difference if you have no gap you get
00:12:59
rapid learning and that's kind of like a
00:13:01
short delay procedure you stick in a
00:13:03
half a second gap and all of a sudden
00:13:06
learning slows down a great deal and
00:13:11
when do you suppose the subject risk
00:13:14
makes the conditioned response well what
00:13:17
is the best predictor of the
00:13:19
unconditioned stimulus under these
00:13:22
circumstances
00:13:23
imagine if the wall at a railroad
00:13:25
crossing the warning lights and and Bell
00:13:29
came on and then it ended and movement
00:13:32
few seconds after that the train comes
00:13:35
by you know the best predictor of the
00:13:41
train is the end of the warnings
00:13:44
stimulus and so you actually get
00:13:47
conditioned responses during the trace
00:13:50
interval and Pavlov hypothesized that
00:13:59
you're actually learning to respond to
00:14:03
the lingering neural trace of the CS
00:14:07
which has been turned off so trace
00:14:11
conditioning it's kind of an interesting
00:14:14
interesting phenomena and it actually
00:14:17
has been the subject of attention by
00:14:23
numerous neuroscientists who are trying
00:14:25
to understand how organisms make use of
00:14:31
time and form associations that require
00:14:39
bridging gaps of time we're actually
00:14:42
bridging a gap of time so trace
00:14:45
conditioning procedure it's kind of
00:14:47
complicated and rather interesting okay
00:14:50
let's move down to the last procedure on
00:14:54
this slide
00:14:55
in this last procedure notice that the
00:14:58
order of the conditioned stimulus and
00:15:00
the unconditioned stimulus have been
00:15:02
reversed and so that's why this is
00:15:04
called backwards backward conditioned
00:15:07
there is this order to see us in us our
00:15:10
backwards in in there
00:15:12
they're uncondition sing us comes on
00:15:15
first and it goes off and shortly
00:15:18
thereafter or right afterwards the
00:15:21
conditions thing this comes on and goes
00:15:23
no so under these circumstances what
00:15:27
does the condition stimulus allow you to
00:15:31
predict it doesn't allow you to predict
00:15:34
the us because the US has already
00:15:36
happened what the CS allows you to
00:15:40
predict is that you're gonna get an
00:15:44
inter trial interval nothing is going to
00:15:46
happen for a while until the next
00:15:48
conditioning truck in this particular
00:15:52
case the train would come by and then
00:15:56
the lights and warning bells would come
00:15:59
on on the DOE circumstances the white
00:16:04
lights and bail would tell you that the
00:16:08
train is gone and you're not going to
00:16:10
see it again for a while and so in a
00:16:13
sense they're the conditioned stimulus
00:16:18
signals the absence of the unconditioned
00:16:22
stimulus and that leads to a special
00:16:24
kind of learning that's called
00:16:26
inhibitory conditioning which we'll talk
00:16:30
about in one of the future episodes here
00:16:34
so um here are all the different ways in
00:16:39
which condition are not going to send
00:16:41
stimuli can be paired and you can see
00:16:44
you get very different results depending
00:16:45
on what the precise temporal sequencing
00:16:49
is and I think this is really
00:16:53
interesting not just because it tells us
00:16:58
more about circumstances under which you
00:17:01
respond but more importantly if we may
00:17:04
go to the next slide the the real
00:17:07
take-home message for that that I take
00:17:12
home the reason I think these
00:17:15
differences in procedures are so
00:17:18
interesting and important is that they
00:17:22
reflect the organisms interpretation
00:17:26
readout of the temporal structure of the
00:17:29
environment you know events in our
00:17:32
environment occur in time some things
00:17:35
occur first and others later and others
00:17:38
later yet and so forth and so there's
00:17:41
kind of a temporal sequence of things
00:17:42
and conditioning procedures and the kind
00:17:47
of condition behavior that develops is a
00:17:52
reflection of how we order events in
00:17:56
time okay
00:17:59
if we order events in time far apart
00:18:03
then we get behavior more akin to
00:18:06
inhibition of delay if we order events
00:18:09
closer together we get behavior more
00:18:11
akin to what's produced by short delayed
00:18:14
conditioning another way to think about
00:18:19
these things is that different results
00:18:23
of various conditioning procedures
00:18:25
reflect the organisms readout of the
00:18:29
causal structure of the environment
00:18:31
every one of the things we're always
00:18:32
trying to figure out is what causes what
00:18:37
what he meant results or is responsible
00:18:41
for the causing something else and
00:18:44
sometimes the cause and consequence are
00:18:49
pretty close together you know you you
00:18:51
got your hot stove and you burn yourself
00:18:54
and that occurs just right away other
00:18:57
times that causal events are pretty far
00:19:00
apart you know you start boiling water
00:19:03
on a stove and it's not going to boil
00:19:05
for a few minutes
00:19:07
there's a you former relationship with
00:19:10
someone else
00:19:11
and there are aspects of that
00:19:13
relationship that will not be evident
00:19:15
sometimes for months or years afterwards
00:19:18
my favorite one is is pregnancy you know
00:19:22
certain events cause pregnancy and I
00:19:28
result in the birth of a newborn those
00:19:32
events are nine months apart for human
00:19:35
beings
00:19:36
so the causal interpretation of what
00:19:41
causes what and what is the causal
00:19:43
texture of the environment requires
00:19:46
paying attention to the order of events
00:19:49
and the timing of those events and so
00:19:51
forth and I think Pavlovian conditioning
00:19:53
reef reflects how we interpret the
00:19:58
causal structure of the environment well
00:20:04
that's my story for today and I hope
00:20:06
you've found it interesting and even if
00:20:09
you did not find an interesting I hope
00:20:12
you'll pay close attention to it because
00:20:15
these temporal details make a big
00:20:18
difference in terms of how learning
00:20:21
proceeds thanks a lot and I have a good
00:20:25
day
00:20:33
you
