Punishment and Negative Reinforcement - PowerPoint PPT Presentation

Loading...

PPT – Punishment and Negative Reinforcement PowerPoint presentation | free to download - id: 47ce76-NWIzZ



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Punishment and Negative Reinforcement

Description:

Punishment and Negative Reinforcement Avoidance Discrete Trial Avoidance S signals that an aversive event will occur If aversive stimulus already has begun ... – PowerPoint PPT presentation

Number of Views:33
Avg rating:3.0/5.0
Slides: 31
Provided by: CarlWL6
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Punishment and Negative Reinforcement


1
Punishment and Negative Reinforcement
2
Avoidance
3
Discrete Trial Avoidance
  • S? signals that an aversive event will occur
  • If aversive stimulus already has begun,
  • Escape response terminates stimulus
  • If aversive stimulus has yet to begin,
  • Avoidance response prevents/postpones stimulus
  • Shuttle Avoidance
  • long box with a small partition in the middle.
  • Light/tone signal s shock will occur
  • Must jump to other side to prevent/escape shock
  • Escape often is the precursor to avoidance

4
Free-Operant Avoidance
  • Also known as Sidman Avoidance
  • Named for Murray Sidman (1953)
  • No explicit stimulus serves as SD
  • Aversive stimulus is set to occur at set
    intervals
  • Response resets interval
  • No response results in presentation move to
    worse schedule
  • When highly trained avoidance responses are
    consistent and efficient

5
An Example
  • Following a response, shock set to occur at 10
    sec intervals
  • 10 s Response-Shock interval
  • Following a shock, shock set to occur at 5 sec
    intervals
  • 5 s Shock-Shock interval
  • Response resets interval back to 10 s
    response-shock interval

6
Punishment
  • A reduction in the probability of a specific
    response as a result of the immediate delivery of
    a stimulus of a stimulus for that response. The
    stimulus is designed as a punishing stimulus, the
    whole procedure is deigned as punishment
  • Azrin and Holz 1966

7
Types of Punishment
  • Positive
  • Response leads to presentation of Aversive
    stimulus, leads to reduction in behavior
  • Examples include
  • Shock, spanking, yelling
  • Negative
  • Response leads to removal of Pleasant stimulus,
    leads to reduction in behavior
  • Examples include
  • Response cost, time out

8
Terms
  • Punisher
  • Must lead to decrease in future responding
  • Aversive Stimulus
  • Something that one will work to avoid or escape
  • Differentiation not so clear
  • But neither is simply something bad

9
Very Important
  • By itself
  • Punishment only suppresses behavior
  • The punished behavior and other behavior
  • Could lead to complete suppression of most
    responses
  • Reinforcement of other behaviors aided by
    punishment of unwanted response

10
Critics of Punishment w/in Behavior Analysis
11
Arguments against Punishment
  • Thorndike
  • Small monetary losses as consequence in memory
    studies, did not significantly improve memory
  • Skinner
  • Argued on grounds of effectiveness not morality?
  • Argued that Sr and Sr- are not reciprocal
    processes
  • Punished behavior much more likely to reappear
    after punisher is removed
  • Sr- gives rise to aggressive counter attack

12
Skinner 1938
  • Food reinforced lever pressing
  • Soon, lever slap was used as punisher
  • Responding suppressed but not weakened?
  • Behavior re-emerged at pre-punishment levels
    almost immediately after EXT

13
Avoidance - organisms responses prevent the
occurrence of an aversive stimulus. - eg moving
around a dark room - hold hands out front - feel
with feet - the avoidance response is followed by
nothing - no particular pleasure, you simply
dont get hurt - how can the absence of something
provide reinforcement - 1st avoidance experiments
by Bechterev 1913 - human subjects finger on a
metal plate warnings CS followed by shock
(US) subjects quickly lift finger off plate
upon being shocked learned to withdraw finger
upon presentation of warning. - actually not
formal Classical condit since delivery of US
depends on subjects behaviour - starting in the
30s avoidance was studied in detail - Guinea
Pigs rotating wheel - Tone CS shock
US - shock stimulated running on the wheel - CC
group shock always 2 sec after start of
tone - avoidance group if wheel moved during CS
prior to shock then no shock delivery - so the
performance of the avoidance group was superior
to the classical group - what can explain this
facilitation of performance?  
  If viewed from a classical conditioning
paradigm the avoidance trials are like
extinction! And should have attenuated the
development of the conditioned response. So if
pairings are important the results should be
opposite! Therefore the results require more
than a classical conditioning explanation.   Discr
iminated avoidance signaled avoidance - discrete
trials, each initiated by CS - the events that
occur next depend on subject - if respond before
shock, CS terminated avoidance - if respond
after shock, CS shock stay on until escape  
  - common apparatus shuttle box 2
compartments, low barrier animal shuttles back
forth between two sides of box shuttle
avoidance (animal can be shocked on either side
of box) - a variation on this is one way
avoidance where 1 side of box is always the
start side and one side the safe side so here
there is a shock compartment and a safe
compartment. This is usually learned more
rapidly than shuttle avoidance.   Two process
Theory Mowrer 47 and Miller 51 - a long
standing theory no longer reviewed as a
complete explanation of avoidance - 2 mechanisms
1) classical conditioning process activated by
pairings of the warning stimulus (CS) and the
aversive event (US) when subject fails to make
avoidance response. Through association with US,
CS comes to elicit fear. 2) Fear is an arousing
state that motivates an organism it is also
aversive, so reduction in fear can provide
negative reinforcement - since fear is elicited
by CS, termination of CS results in reduction of
fear! So, the 2nd component is instrumental
reinforcement of the avoidance response through
fear reduction. - the two processes, classical
instrumental conditioning are interdependent.
Classical conditioning must occur 1st, after this
the instrumental conditioning creates extinction
trials for classical conditioning procedure.
- This theory predicts a constant interplay
between the two processes - it explains avoidance
behaviour in terms of escape from conditioned
fear rather than prevention of shock - Thus the
instrumental response is reinforced by a tangible
event (fear reduction) rather than merely the
absence of something (aversive stimulation).   Acq
uired Drive Experiments - attempt to demonstrate
2 diff processes in avoidance 1) condition fear
to a CS with a pure classical conditioning
procedure 2) animals periodically exposed to
fear-eliciting CS and allowed to perform an
instrumental response to terminate it (thereby
reducing far) (No shocks in this phase) - If the
2 process theory is correct then subjects should
be able to learn phase 2 this is called an
acquired drive study because the drive to perform
the instrumental response (fear) is learned
through classical conditioning and is not
innate   Brown Jacobs 49 rats in shuttle box
confined to one side light/tone CS shock
control gp no checks - phase 2 subject place
on one side of box, CS on until crossed to other
side (no shocks, one-way procedure) - would rats
learn to cross if only reinforcement was turning
off CS? - two groups had similar latencies at the
start of training. - As training progressed the
shock conditioned animals learned to cross
faster - ? termination of a fear condition
stimulus is sufficient to provide reinforcement
for an instrumental report. These results
provide strong support for the two process
theory!  
  Independent measurement of Fear during
Acquisition of avoidance behaviour. - if 2
process theory correct then conditioning of
avoidance and conditioning of fear should proceed
together - however, conditional fear and
avoidance responding are not always highly
correlated - animals become less fearful as they
learn the avoidance response. - look at
conditioned suppression as a measure of amount of
fear - animals learn to bar press for food, then
a shock conditioned CS presented measure
suppression of lever press behaviour extent of
suppression is thought to be a measure of fear
elicited by the CS   Kamin et al 63 rats lever
pressing on VI schedule - CS tone ? shock US in
shuttle box for a number of groups until
successful avoidance on 1, 3, 9 or 27 consecutive
trials - then back to lever pressing tone
presented - lower values greater
disruption - suppression for gps that
successfully avoided 1, 3 9 shocks, but with
more training response suppression
declined - suggests that fear decreases during
extended avoidance training however the
decrease in fear is not accompanied by decrease
in avoidance!  
  - If fear reduction is necessary to reinforce
avoidance behaviour how can the behaviour persist
in the face of declining CS elicited fear? Some
suggestion that both the CS the background
acquire fear, but as training continues get
discriminative fear to background ? more than
fear to CS. So CS termination could continue to
be fear reducing   Asymptotic Avoidance
Performance - Two process theory makes specific
predictions about the nature of
performance - predicts that the strength of the
avoidance response will fluctuate in
cycles - extinction of conditioned fear by
successful avoidance leads to less reinforcement
for reduction of fear the avoidance response
will cease to occur in time to prevent the
US. - When shock occurs CS is paired with US
again an reinstates fear avoidance
reconditioned - so avoidance should go through
cycles of extinction reacquisition - usually
avoidance does not fluctuate in cycles
- avoidance is highly resistant to extinction ie
hundreds of trials of avoidance after shock
off - persistence of avoidance is difficult to
explain with 2 factor theory - recent empirical
findings suggest a clue if focus on the role of
response feedback stimuli - avoidance response
followed by a period free from shock, there are
feedback stimuli as a result of avoidance
- these include change in location, stimuli
from making response such as touching
manipulating lever (tactile) also
proprioceptive feedback - because there is no
shock these are correlated negatively with shock
this can lead to the development of conditioned
inhibition. - so, response cues can become
conditioned inhibitors of fear - After extensive
avoidance training subjects experience 2
conditioned stimuli on each trial 1) fear
eliciting warning stimulus (CS) followed by
2) fear inhibiting feedback cues from the
avoidance response (CS-)   Therefore asymptotic
avoidance trials should not be viewed as
extinction trials with CS alone, but as trials
where CS is followed by fear-inhibitory response
feedback cues. - Data indicates that the
presentation of fear-inhibiting stimuli following
a CS can block extinction of CS - Because the
inhibitory stimulus is a signal for the absence
of shock, occurrence of the CS- makes the absence
of the shock on avoidance trials fully expected
this protects the CS from any changes in
associative strength. alternative way to think
about this the absence of shock is attributed
to CS-, the absence of shock is attributed to
CS-, therefore reevaluation of CS unnecessary
so response feedback cues prevent CS extinction
so CS still elicits fear and motivates
avoidance.   EXTINCTION OF AVOIDANCE THROUGH
REPSONSE BLOCKING CS ALONE EXPOSURE   - avoidanc
e persists for along time - are there procedures
that can lead to rapid extinction? - flooding or
response prevention present CS in avoidance
situation, but prevent subject from making
response thus exposed to CS without being
allowed to terminate it flooded with CS
- important variable is duration of forced
exposure to CS. - Schiff, Smith Prochaska
72 - rats CS tone, one way shock avoidance
- after 10 successful avoidance trials a barrier
blocked off the safe compartment and ratio
received various lengths of CS exposure without
shock - 1, 5 or 12 blocked trials, CS for 1, 5,
10, 50 or 120 sec - Then barrier removed and
subjects tested for extinction - rat in box, CS
presented until animals crossed to safe side NO
SHOCKS subjects tested until they took 120
seconds or longer to cross over - blocked
exposure to CS facilitated extinction - effect
determined by the total deviation of exposure to
the CS - increases in the total duration of
blocked exposure resulted in more rapid
extinction.  
  - Two process view predicts flooding produces
avoidance extinction through extinction of fear
to the CS. - however conditioned suppression
experiments have found in some situations
avoidance extinguishes more rapidly than fear,
while in other situations fear extinguishes more
rapidly than avoidance. - probably extinction of
fear is only 1 factor responsible for effectives
of flooding - another variable is being
prevented from making the avoidance
response. - Katzen Kerman 74 first trained
rats to avoid shock in shuttle box then 50
extinction trials - pairs of rats 1 member
shuttle response not blocked so rat could turn
off CS by crossing over - other subject yoked
so received same duration CS but shuttle response
blocked by barrier - 3rd group control no
extinction trials during this phase all groups
then got standard extinction - control gp
greatest of crossings - fewest crossings with
subjects with blocked responses. Two groups with
identical CS exposure performed very differently
- so response blocking can facilitate extinction
so more than just Pavlovian extinction of Cs
going on  
  - Perhaps subjects learn a response that is
incompatable with the avoidance behaviour during
blocked exposure to CS this contributes to the
observed loss of avoidance responding.
  Non-Discriminated (Free operant)
avoidance - could animals learn to avoid if no
warning signal? - SIDMAN 53 (Non-Discriminated,
Free Operant or SIDMAN AVOIDANCE) shock is
scheduled without warning ie every 10 sec a
behaviour is specified an avoidance response (eg
lever press) each response prevents delivery of
shock for a set time (eg 30 sec). Animals will
learn to avoid shock under these conditions even
if no warning. - 2 intervals the S-S interval
(shock-shock) and the R-S interval
(response-shock) this is the period of safety
after a response  
  - this avoidance allows responses to occur at
any time - another diff from signaled avoidance
response there only effective if during CS here
a response anytime will reset the R-S interval
if R-S interval 30 sec - shock rescheduled 30 sec
after each response by always responding right
before R-S interval over the safety period can be
indefinitely prolonged - these studies involve
longer training animal receives more shocks
even after extensive training animals dont avoid
all shocks - procedures sensitive to individual
differences - rate of responding is controlled
by the value of the S-S and R-S intervals - the
more frequently shocks occur without responding
(S-S interval) the more likely the animal is to
learn avoidance - increasing the periods of
safety (R-S interval) also promotes avoidance
behaviour - the relative values of S-S and R-S
are also important (ie R-S cant be shorter than
S-S) - this task challenges the two process
theory because there is no explicit CS to elicit
conditioned fear and it is not clear how
avoidance reduce- attempt to use 2-process theory
S-S and R-S are short and fixed, ? they are
predictable - suggests that animals learn to
respond to the passage of time as a signal for
shock. Temporal conditioning - assume the
passage of time after last shock (in S-S
interval) or last response (in R-S interval)
becomes conditioned to elicit fear. Since the
timing starts over after each occurrence of the
avoidance response the response removes the fear
eliciting temporal cues so time cues take on
the role of the CS - this predicts subjects will
do most of their responding at end of R-S
interval when fear is high. - Results obtained
are consistent with this prediction - BUT many
animals successfully avoid a great many shocks
without distributing their responses in this
manner - also avoidance behaviour has been
successfully conditioned with variable interval
S-S an R-S interval - so perhaps 2 process
theory is not a good or complete explanation of
non-discriminated avoidance.  
  Delay - time interval between response and
punisher - increasing the interval results in
less suppression of behaviour   Schedules of
Punishment - same as schedules of reinforcement
have large affect on suppression of
behaviour - so higher the ratio, lower the
suppression  
  Effects of punishment on schedules of Positive
Reinforcement - FI or VI schedule of positive
reinforcement punishment produces a decrease in
overall responding but no change in response
pattern VI stable rate, FI scallop
pattern - FR punishment increases length of
post-reinforcement pause but has little effect on
ratio run - shock delivery early in ratio run
increases the post-reinforcement pause more than
shock delivered near the completion of the ratio.
In addition, punishment has less effect on
instrumental responses that produce more frequent
positive reinforcement.   Availability of
Alternative Responses - in many experiments the
punished response is the only one to get positive
reinforcement such as food - by decreasing rate
of responding subject also decreases the food it
gets therefore conflict between suppressing
behaviour to avoid punishment and responding to
obtain positive reinforcement - the availability
of an alternative source of reinforcement greatly
increases the suppression of responding produced
by punishment. - Herman Azrin 64 adult
(humans) males - response levers - each VI to
produce a cigarette after behaviour occurring
at a stable rate responses on 1 lever resulted in
an obnoxious noise for 1 gp only 1 lever
available during this phase, in another both
levers available 1 punished - when punished
response was the only way to obtain cigarettes
punishment produced moderate suppression - when
the alternative lever was available, responding
on the punished lever ceased completely - thus
availability of an alternative response for
obtaining positive reinforcement greatly
increased the suppressive effects of
punishment.  
  Effects of a Discriminative Stimulus   Discrimin
ative Punishment if responding is punished in
the presence of a discriminative stimulus but is
not punished when the stimulus is absent here
the suppressive effects of punishment will become
limited to the presence of the discriminative
stimulus.   Dinsmoor 52 rats, food at VI 2
min, after training 5 min punishment alternated
with 5 min no punishment. During punishment the
lights in the chamber were turned off and each
lever press resulted in shock. During safe
period the lights were on no shocks delivered
food was delivered during all times - rats
quickly learned to restrict responding to the
safe periods when the lights were off
responding was suppressed.   Real World
Discriminative Stimuli - strict teacher or
parent - police cars   Stimulus
Control - similarity of situations and
generalization of punishment - Honig Sliuka
(64) punishment of key-pecking in pigeons
during training key light varied 490-610
nm - pecking reinforced to all wavelengths once
pecking was equal on all key wavelengths they
selectively punished pecking to 550 nm with
shock  
  - punishment was effective from the outset
- got generalization gradient centered around
the punished wavelength as training continued
the generalization gradient sharpened
considerably. - if subjects find that responding
to similar situations is safe then responding
only suppressed in a situation where it is
actually punished. - ie teaching child not to
cross street need it to generalize!   Punishment
to signal Positive reinforcement - punishment
does not always decrease behaviour sometimes
people seem to seek out punishment - punishment
may become a signal or discriminative stimulus
for the availability of positive
reinforcement - if this occurs punishment will
increase rather than decrease responding. - pigeon
s peck key for food reinforcement VI schedule
then when trained each response punished with
mild shock sufficient to reduce response rate to
50 - then periods when punishment was in effect
were alternated with periods of no punishment
the pecking response was only reinforced with
food during the punishment periods. No other
signal to tell birds about punishment or food.
The only way to tell if food would be delivered
was to see if punished. - higher rate of pecking
occurred during punishment periods than during
safe periods. - punishment became an SD for food
reinforcement   Self Punitive Behaviour - ie
making punishment an SD for positive
reinforcement - self punitive behaviour can also
result from prior escape training called
Vicious-Circle behaviour rats given escape
training in which entire runway electrified
have to run length of it to reach safe goal box
after subjects have learned this an extinction
procedure begun control no shock experimental
group final third of runway is electrified
Thus experimental groups encounter shock if they
run during extinction, but control groups receive
no punishment. - punishment increases resistance
to extinction of the running response in the
experimental subjects - instead of suppressing
behaviour, punishment of conditioned escape
behaviour facilitates responding. - some innate
defensive responses appear to be relatively
resistant to punishment - Punishment also leads
to a reallocation of responding suppression of
a punished response leads to increase in
expression of other responses. - there are few
systematic theories of punishment - Conditional
Emotional Response theory - Estes 44 based on
the observation that a conditioned stimulus that
has been paired with shock will suppress the
performance of food reinforced instrumental
behaviour - condition to lever press associate
CS with shock - see effects of CS presentation on
lever pressing - interpretation was that CS
produces competing responses incompatible with
lever pressing - Estes suggests that stimuli
experienced just prior to punishment serve the
function of signaling punishment so if lever
press right before shock then stimuli associated
with lever pressing become associated with shock.
These cues will acquire conditioned aversive
properties will elicit conditioned emotional
responses that are incompatible with the punished
behaviour. - can explain why more intense/longer
shocks produce more response suppression more
vigorous conditioned emotional responses why is
contingency important?   Stimuli associated with
response are more closely related to performance
of behaviour so conditioned emotional response
likely to interfere with responding   Estes 69
re thought problem new account of much of
conditioned suppression paraphrased in
motivational terms shock conditioned stimuli
disrupt food reinforced responding by evoking
an emotional or motivational state incompatible
with the motivation to maintain the food
reinforced behaviour - shock conditioned stimulus
inhibits the motivation to respond based on
positive reinforcement   Avoidance Theory of
Punishment   Dinsmoor 54, 77 Stimuli that set
the occasion for the instrumental response become
conditioned by the aversive stimulus when the
response is punished so stimuli acquire
aversive properties - subjects learn to escape
from the conditioned aversive stimuli related to
the punished response by engaging in a behaviour
incompatible with the punished response
performance of the alternate activity results in
suppression of the punished behaviour so dont
weaken the punished response rather strengthen
competing responses.   One strength of this
theory is that both avoidance and punishment can
be understood in the same theoretical
framework - problem is that all of the
difficulties in explaining avoidance now occur
for punishment as well - another difficulty is
that the stimuli that acquire conditioned
aversive properties are not under the direct
control of the experimenter so are difficult to
investigate   Punishment and negative law of
effect - Thorndike (1911) punishment opposite
to positive reinf. - Thorndike (1932) failed
to find supporting evidence but other researchers
have retained notion - Premack positive
reinforcement when able to engage in valued
activity is contingent on prior performance of
lower value activity - punishment reverses this
a low valued activity (shock) occurs contingent
on the performance of a higher-valued behaviour
- undergoing shock can punish lever
pressing - running wheel with drinking tube
rats thirsty so drinking more likely than
running drinking was followed by forced running
drinking was suppressed! - the same contingency
can produce opposite outcomes depending on the
relative value of running and drinking when
running more likely (rats not thirsty) then
running positively reinforced drinking! So same
thing could be positive reinforcer or punishment
depending on the state of the organism - only
difference is in punishment animal is forced to
engage in lower-valued activity   Punishment in
Humans - evidence suggests many similarities
between effects of punishment on animals and
effects on humans - Bucher Lovaas 68
electric shock to treat self destructive
behaviour in autistic children - autistic
children stereotyped behaviour, some engage in
terrible self destructive behaviour - eg John
physical damage so severe had to be restrained
for 24 hrs a day! how to eliminate the
behaviour one way might be to reinforce other
behaviours incompatible with self injury ignore
self-injury John so severe this was not an
option - used punishment instead taken to
specific room restraints removed shock each
time he hit himself - during 15 baseline sessions
no punishment average of 250 self-hits per
session - when punishment introduced behaviour
disappeared almost immediately there were a
number of experimenters  
14
Counterarguments
  • Almost NO evidence that Sr- is more effective
    than Sr
  • Rewarded behavior also is likely to ext when
    reinforcer is removed
  • Avoidance is one of the most resistent to ext
    schedules

15
Key features of effective punishment
  • Intensity
  • Response-dependency and schedule
  • Immediacy
  • Consistency
  • Availability of alternative behavior
  • Co-occurrence with reinforcement
  • Existence of conditioned punishers

16
Intensity
  • Start off at the highest acceptable severity
  • More humane in long run?
  • Starting with mild shock and increasing intensity
    works very poorly!
  • However, starting intense and then reducing
    intensity
  • suppression continues resulting in more
    suppression than the mild shock would have
    produced on its own.

17
Avoidance - organisms responses prevent the
occurrence of an aversive stimulus. - eg moving
around a dark room - hold hands out front - feel
with feet - the avoidance response is followed by
nothing - no particular pleasure, you simply
dont get hurt - how can the absence of something
provide reinforcement - 1st avoidance experiments
by Bechterev 1913 - human subjects finger on a
metal plate warnings CS followed by shock
(US) subjects quickly lift finger off plate
upon being shocked learned to withdraw finger
upon presentation of warning. - actually not
formal Classical condit since delivery of US
depends on subjects behaviour - starting in the
30s avoidance was studied in detail - Guinea
Pigs rotating wheel - Tone CS shock
US - shock stimulated running on the wheel - CC
group shock always 2 sec after start of
tone - avoidance group if wheel moved during CS
prior to shock then no shock delivery - so the
performance of the avoidance group was superior
to the classical group - what can explain this
facilitation of performance?  
  If viewed from a classical conditioning
paradigm the avoidance trials are like
extinction! And should have attenuated the
development of the conditioned response. So if
pairings are important the results should be
opposite! Therefore the results require more
than a classical conditioning explanation.   Discr
iminated avoidance signaled avoidance - discrete
trials, each initiated by CS - the events that
occur next depend on subject - if respond before
shock, CS terminated avoidance - if respond
after shock, CS shock stay on until escape  
  - common apparatus shuttle box 2
compartments, low barrier animal shuttles back
forth between two sides of box shuttle
avoidance (animal can be shocked on either side
of box) - a variation on this is one way
avoidance where 1 side of box is always the
start side and one side the safe side so here
there is a shock compartment and a safe
compartment. This is usually learned more
rapidly than shuttle avoidance.   Two process
Theory Mowrer 47 and Miller 51 - a long
standing theory no longer reviewed as a
complete explanation of avoidance - 2 mechanisms
1) classical conditioning process activated by
pairings of the warning stimulus (CS) and the
aversive event (US) when subject fails to make
avoidance response. Through association with US,
CS comes to elicit fear. 2) Fear is an arousing
state that motivates an organism it is also
aversive, so reduction in fear can provide
negative reinforcement - since fear is elicited
by CS, termination of CS results in reduction of
fear! So, the 2nd component is instrumental
reinforcement of the avoidance response through
fear reduction. - the two processes, classical
instrumental conditioning are interdependent.
Classical conditioning must occur 1st, after this
the instrumental conditioning creates extinction
trials for classical conditioning procedure.
- This theory predicts a constant interplay
between the two processes - it explains avoidance
behaviour in terms of escape from conditioned
fear rather than prevention of shock - Thus the
instrumental response is reinforced by a tangible
event (fear reduction) rather than merely the
absence of something (aversive stimulation).   Acq
uired Drive Experiments - attempt to demonstrate
2 diff processes in avoidance 1) condition fear
to a CS with a pure classical conditioning
procedure 2) animals periodically exposed to
fear-eliciting CS and allowed to perform an
instrumental response to terminate it (thereby
reducing far) (No shocks in this phase) - If the
2 process theory is correct then subjects should
be able to learn phase 2 this is called an
acquired drive study because the drive to perform
the instrumental response (fear) is learned
through classical conditioning and is not
innate   Brown Jacobs 49 rats in shuttle box
confined to one side light/tone CS shock
control gp no checks - phase 2 subject place
on one side of box, CS on until crossed to other
side (no shocks, one-way procedure) - would rats
learn to cross if only reinforcement was turning
off CS? - two groups had similar latencies at the
start of training. - As training progressed the
shock conditioned animals learned to cross
faster - ? termination of a fear condition
stimulus is sufficient to provide reinforcement
for an instrumental report. These results
provide strong support for the two process
theory!  
  Independent measurement of Fear during
Acquisition of avoidance behaviour. - if 2
process theory correct then conditioning of
avoidance and conditioning of fear should proceed
together - however, conditional fear and
avoidance responding are not always highly
correlated - animals become less fearful as they
learn the avoidance response. - look at
conditioned suppression as a measure of amount of
fear - animals learn to bar press for food, then
a shock conditioned CS presented measure
suppression of lever press behaviour extent of
suppression is thought to be a measure of fear
elicited by the CS   Kamin et al 63 rats lever
pressing on VI schedule - CS tone ? shock US in
shuttle box for a number of groups until
successful avoidance on 1, 3, 9 or 27 consecutive
trials - then back to lever pressing tone
presented - lower values greater
disruption - suppression for gps that
successfully avoided 1, 3 9 shocks, but with
more training response suppression
declined - suggests that fear decreases during
extended avoidance training however the
decrease in fear is not accompanied by decrease
in avoidance!  
  - If fear reduction is necessary to reinforce
avoidance behaviour how can the behaviour persist
in the face of declining CS elicited fear? Some
suggestion that both the CS the background
acquire fear, but as training continues get
discriminative fear to background ? more than
fear to CS. So CS termination could continue to
be fear reducing   Asymptotic Avoidance
Performance - Two process theory makes specific
predictions about the nature of
performance - predicts that the strength of the
avoidance response will fluctuate in
cycles - extinction of conditioned fear by
successful avoidance leads to less reinforcement
for reduction of fear the avoidance response
will cease to occur in time to prevent the
US. - When shock occurs CS is paired with US
again an reinstates fear avoidance
reconditioned - so avoidance should go through
cycles of extinction reacquisition - usually
avoidance does not fluctuate in cycles
- avoidance is highly resistant to extinction ie
hundreds of trials of avoidance after shock
off - persistence of avoidance is difficult to
explain with 2 factor theory - recent empirical
findings suggest a clue if focus on the role of
response feedback stimuli - avoidance response
followed by a period free from shock, there are
feedback stimuli as a result of avoidance
- these include change in location, stimuli
from making response such as touching
manipulating lever (tactile) also
proprioceptive feedback - because there is no
shock these are correlated negatively with shock
this can lead to the development of conditioned
inhibition. - so, response cues can become
conditioned inhibitors of fear - After extensive
avoidance training subjects experience 2
conditioned stimuli on each trial 1) fear
eliciting warning stimulus (CS) followed by
2) fear inhibiting feedback cues from the
avoidance response (CS-)   Therefore asymptotic
avoidance trials should not be viewed as
extinction trials with CS alone, but as trials
where CS is followed by fear-inhibitory response
feedback cues. - Data indicates that the
presentation of fear-inhibiting stimuli following
a CS can block extinction of CS - Because the
inhibitory stimulus is a signal for the absence
of shock, occurrence of the CS- makes the absence
of the shock on avoidance trials fully expected
this protects the CS from any changes in
associative strength. alternative way to think
about this the absence of shock is attributed
to CS-, the absence of shock is attributed to
CS-, therefore reevaluation of CS unnecessary
so response feedback cues prevent CS extinction
so CS still elicits fear and motivates
avoidance.   EXTINCTION OF AVOIDANCE THROUGH
REPSONSE BLOCKING CS ALONE EXPOSURE   - avoidanc
e persists for along time - are there procedures
that can lead to rapid extinction? - flooding or
response prevention present CS in avoidance
situation, but prevent subject from making
response thus exposed to CS without being
allowed to terminate it flooded with CS
- important variable is duration of forced
exposure to CS. - Schiff, Smith Prochaska
72 - rats CS tone, one way shock avoidance
- after 10 successful avoidance trials a barrier
blocked off the safe compartment and ratio
received various lengths of CS exposure without
shock - 1, 5 or 12 blocked trials, CS for 1, 5,
10, 50 or 120 sec - Then barrier removed and
subjects tested for extinction - rat in box, CS
presented until animals crossed to safe side NO
SHOCKS subjects tested until they took 120
seconds or longer to cross over - blocked
exposure to CS facilitated extinction - effect
determined by the total deviation of exposure to
the CS - increases in the total duration of
blocked exposure resulted in more rapid
extinction.  
  - Two process view predicts flooding produces
avoidance extinction through extinction of fear
to the CS. - however conditioned suppression
experiments have found in some situations
avoidance extinguishes more rapidly than fear,
while in other situations fear extinguishes more
rapidly than avoidance. - probably extinction of
fear is only 1 factor responsible for effectives
of flooding - another variable is being
prevented from making the avoidance
response. - Katzen Kerman 74 first trained
rats to avoid shock in shuttle box then 50
extinction trials - pairs of rats 1 member
shuttle response not blocked so rat could turn
off CS by crossing over - other subject yoked
so received same duration CS but shuttle response
blocked by barrier - 3rd group control no
extinction trials during this phase all groups
then got standard extinction - control gp
greatest of crossings - fewest crossings with
subjects with blocked responses. Two groups with
identical CS exposure performed very differently
- so response blocking can facilitate extinction
so more than just Pavlovian extinction of Cs
going on  
  - Perhaps subjects learn a response that is
incompatable with the avoidance behaviour during
blocked exposure to CS this contributes to the
observed loss of avoidance responding.
  Non-Discriminated (Free operant)
avoidance - could animals learn to avoid if no
warning signal? - SIDMAN 53 (Non-Discriminated,
Free Operant or SIDMAN AVOIDANCE) shock is
scheduled without warning ie every 10 sec a
behaviour is specified an avoidance response (eg
lever press) each response prevents delivery of
shock for a set time (eg 30 sec). Animals will
learn to avoid shock under these conditions even
if no warning. - 2 intervals the S-S interval
(shock-shock) and the R-S interval
(response-shock) this is the period of safety
after a response  
  - this avoidance allows responses to occur at
any time - another diff from signaled avoidance
response there only effective if during CS here
a response anytime will reset the R-S interval
if R-S interval 30 sec - shock rescheduled 30 sec
after each response by always responding right
before R-S interval over the safety period can be
indefinitely prolonged - these studies involve
longer training animal receives more shocks
even after extensive training animals dont avoid
all shocks - procedures sensitive to individual
differences - rate of responding is controlled
by the value of the S-S and R-S intervals - the
more frequently shocks occur without responding
(S-S interval) the more likely the animal is to
learn avoidance - increasing the periods of
safety (R-S interval) also promotes avoidance
behaviour - the relative values of S-S and R-S
are also important (ie R-S cant be shorter than
S-S) - this task challenges the two process
theory because there is no explicit CS to elicit
conditioned fear and it is not clear how
avoidance reduce- attempt to use 2-process theory
S-S and R-S are short and fixed, ? they are
predictable - suggests that animals learn to
respond to the passage of time as a signal for
shock. Temporal conditioning - assume the
passage of time after last shock (in S-S
interval) or last response (in R-S interval)
becomes conditioned to elicit fear. Since the
timing starts over after each occurrence of the
avoidance response the response removes the fear
eliciting temporal cues so time cues take on
the role of the CS - this predicts subjects will
do most of their responding at end of R-S
interval when fear is high. - Results obtained
are consistent with this prediction - BUT many
animals successfully avoid a great many shocks
without distributing their responses in this
manner - also avoidance behaviour has been
successfully conditioned with variable interval
S-S an R-S interval - so perhaps 2 process
theory is not a good or complete explanation of
non-discriminated avoidance.  
 
 
18
Response Dependency
  • Contingency
  • although response independent aversive
    stimulation can result in some general
    suppression of behavior, significantly more
    suppression of behavior occurs if the aversive
    stimulus is produced by the instrumental response
  • rats VI 1 min food 3 groups
  • never shocked
  • brief shock every 2 min unrelated to bar press
  • brief shock about every 2 min contingent on lever
    press

19
(No Transcript)
20
(No Transcript)
21
Availability of Alternative Responses
  • Often, punished response is the only one to get
    positive reinforcement
  • By decreasing responding, subject also decreases
    reward earned
  • conflict b
About PowerShow.com