Block I: Theory principles
2. Classical conditioning I Profª Paula Balea
Departamento de Psicología
Pre-associative
learning
Pre-associative learning
There is an elementary form of learning that do not involve the association
between two stimuli.
This type of learning also produces a change in behavior (sometimes long-lasting)
resulting from experience. However, this change is produced by the repeated
presentation of a single stimulus.
Because no association is involved, this learning is referred to as non-associative
or pre-associative learning.
Pre-associative learning: changes in behavior that occur as a consequence of
the repeated presentation of a stimulus. Includes two phenomena: habituation
and sensitization.
Pre-associative learning
Habituation
Habituation: progressive + = presentation
decrease of a response to a of the stimulus
stimulus that is presented
repeatedly.
Habituation of startle
It usually occurs with stimuli of low or moderate intensity (e.g., the distant sound response
[Link]
of construction work in the street) or with simple stimuli (see study by Bashinski et
al., 1985).
Adaptive function: Allows us to focus our attention on the really relevant stimuli of the environment.
Pre-associative learning
Habituation
Habituation is stimulus-
specific
the response reappears if the
stimulus is changed
Pre-associative learning
Habituation
+ = presentation of a toy
Time playing
New toy
Time playing
The phenomenon of habituation is very evident with toys; after
having played with a toy several times, the time the child
spends with the toy tends to decrease.
As habituation is stimulus-specific, if we give the child a
different toy, the (play) response is recovered.
Pre-associative learning
Sensitization
Sensitization:
Increased response to a stimulus to a stimulus that is presented
repeatedly.
+ = presentation of
the stimulus
It usually occurs with high intensity stimuli (e.g. excessively loud music from a neighbor) or complex
stimuli (see study by Bashinski et al., 1985).
Unlike habituation, sensitization is not stimulus specific.
Habituation and sensitization of visual attention in infants.
Bashinski, Werner & Rudy (1985)
Visual stimuli elicit an observation response in infants. This response can be
measured by the length of time they maintain their gaze on an object before shifting
their eyes elsewhere.
Habituation and sensitization of visual attention in infants.
Bashinski, Werner & Rudy (1985)
8 trials of 10 seconds
Sensitization
Simple stimulus
Habituation
Habituation Complex stimulus
Sensitization (initially)
The results show that the attentional response to a stimulus changes as the infant
becomes familiar with the stimulus. Moreover, the nature of the change (habituation vs.
sensitization) is determined by the characteristics of the stimulus.
Basic concepts of
Classical
Conditioning (CC)
Basic concepts
• Classical conditioning (CC):
associative learning in which a
relationship is established between two
stimuli regardless of the individual's
response.
• Adaptive function: it allows us to
predict/anticipate the occurrence of
relevant events and to react accordingly.
Basic concepts
• Unconditioned stimulus (US): biologically relevant stimulus that automatically/innately elicits a response.
• Unconditioned response (UR): an instinctive, innate, untrained response to the US.
• Conditioned stimulus (CS): an initially neutral stimulus that, after being paired with the US, acquires the
capacity to elicit the response that usually elicits the US (or a related one).
• Conditioned response (CR): response elicited by the CS after repeated and paired presentations of the CS
with the US. Its presence indicates that the individual has learned the relationship between the stimuli.
CS US
After its repeated Naturally
presentation with the produces a...
US, it ends up
producing a...
CR UR
Basic concepts
Conditioned stimulus (CS): an initially neutral stimulus that, after being
paired with the US, acquires the capacity to elicit the response that usually
elicits the US (or a related one).
Timeline of classical conditioning
Conditioned Stimulus
Unconditioned stimulus
Responses
CR UR
(appears after (is always present)
learning)
Timeline
Symbols commonly used to summarize the classical
conditioning procedures
X, Y, Z... = Conditioned Stimuli (CSs). E.g. the bell.
+ = Unconditioned Stimulus (US). E.g. the food.
– = Absence of the Unconditioned Stimulus
| = Separation between phases of an experiment/study intended to elicit different types of learning.
, = separation betwen trials within a phase
? = Indicates that this is a test trial where the ability of the CS to evoke CR is tested.
A:, B:, C:, ... = Contexts
Procedure for acquiring a CR Abbreviated (most frequent)
X+, X+, ... , X+, ... , X+
X+
These symbols are arbitrary. Sometimes the initial letters (A, B, C...) are used to denote the different CSs, or
EC1, EC2, etc. are used instead. Similarly, sometimes the terms EI1, EI2 are used to denote different EIs.
Acquisition and
generalization
of learning
Acquisition and generalization of learning
Acquisition
Development of a Conditioned Response due
to repeated presentation of an initially neutral
stimulus (later referred to as Conditioned
Stimulus), with an Unconditioned Stimulus
Trial: each time a stimulus (or two paired stimuli) are X+ X+ X+ X+ X+ X+
presented. In Pavlov's research, each time the bell is
presented followed by the food constitutes a trial.
If the CS is followed by the US, then we label them
acquisition trials, but not all trials are of this type (see
following slides).
Phobias are often acquired by this Pavlovian mechanism by which an initially neutral stimulus is
paired with an aversive one.
Watson and Rayner study (1920)
Study of Little Albert
First sessions
Sequence of
events
Sequence of
responses
Timeline
After several sessions
Sequence of
events
CS: rat
US: loud sound Sequence of
UR: crying at the sound responses
CR: crying at the rat
Timeline
Acquisition and generalization of learning
Common examples of classical conditioning in children
Phobias and irrational fears:
Potential CSs: any initially neutral stimulus. For example, the
teacher, the school, animals (spiders, dogs, etc.), specific
situations (e.g., public speaking, recess time, etc.).
US: highly unpleasant experience (verbal or physical violence,
physical pain, psychological pain, …).
If any of the CSs is associated with an aversive experience (US)
then the CS will eventually elicit a fear response (CR) similar to
the response that is naturally elicited by the aversive experience
(UR).
Acquisition and generalization of learning
Common examples of classical conditioning in children
Rejection of certain foods:
CS: any food.
US: arguments about not eating, conflicts, angry parents, ...
If unpleasant situations (US) occur whenever the food is eaten (CS), the food
itself will eventually elicit a displeasure response (CR) similar to that
naturally elicited by the unpleasant situations (UR).
Acquisition and generalization of learning
Common examples of classical conditioning in children
Conditioned taste aversion:
CS: any food.
EI: Stomach ache, vomiting…
If a food makes us feel sick once, it will be very hard for us to eat it again.
It is a very fast type of learning because our body is specially prepared to
prevent us from being poisoned or consuming spoiled food.
In “food rejection” the US is unpleasant, but in “conditioned taste aversion”
the US indicates a threat to our survival.
Acquisition and generalization of learning
When someone is attacked by a dog, the aggression is carried out by a
Generalization particular animal. However, the victim is likely to develop a fear of other dogs
CR that occurs in response to stimuli that of his breed or, usually, of any dog.
have not been subjected to conditioning.
The CR occurs because of its similarity to
the trained CS.
Appearance of CR to stimuli similar to the
trained CS, even though they have never
been paired with the US.
[Link]
Adaptive function: it allows us to respond appropriately to new stimuli without having to receive explicit training
with each one. We know that similar things have similar consequences and so if we treat them in the same way we are
likely to respond correctly. Even the same object is rarely always seen from the same angle, at the same distance, in
the same light conditions, etc. It is useful that we are able to generalize our response despite these variations.
Acquisition and generalization of learning
Generalization
In generalization procedures, the tested stimuli differ from the Moore's generalization study (1972)
trained stimuli along one dimension (color, shape, ...).
Percentage of the CR occurrence
The more similar a stimulus is to the CS, the greater the CR it elicits
(the greater the generalization); the less similar, the lower the CR
(the more the individual will discriminate).
This gradation of the ability of stimuli to elicit a CR as a function of
their similarity to the original CS gives rise to the so-called
generalization gradients.
Condic. Test
Subjects: rabbits 400 Hz
CS = Tone 1200 Hz 800 Hz
US = shock to the cheek 1200Hz 1200 Hz (CS)
UR = eye blinking to the shock 1600 Hz Frequency (Hz) of the test tones
CR = eye blinking to the tone 2000 Hz
[Link]
Acquisition and generalization of learning
The strongest responses
Generalization come from stimulation of the
Drops of saliva in 30 seconds
areas closest to the thigh
Pavlov demonstrated the generalization gradient by
attaching small vibrators to various parts of a dog's
body.
After conditioning salivation to stimulation in the
thigh, he stimulated other areas.
The closer the stimulated point was to the thigh, the
stronger the conditioned response (Pavlov, 1927).
Rear leg Pelvis Shoulder
Front paw
thigh Trunk Front leg
Stimulated body part
Acquisition and generalization of learning
Discrimination
It is the opposite phenomenon of generalization.
It is the ability to respond differentially to the trained stimulus.
We say that there is a lot of discrimination when the individual
responds to the trained stimulus, but hardly exhibits any response to
Percentage of the CR occurrence
similar stimuli.
If a dog salivates in the presence of a 1200 Hz tone but not in the
presence of a 5 Hz tone, then we say that it is able to discriminate
between them.
Discrimination is visible in the generalization gradient itself. For
example, the rabit blinks to a 1200Hz tone but much less to a 400 Hz
tone (it discriminates, to some extent, between the two).
But discriminations can be trained deliberately with discriminative
training. Frequency (Hz) of the test tones
Acquisition and generalization of learning
Discrimination
Discriminative training.
A CS1 is associated with the US, while another similar CS2 is not
associated with the EI.
The organism learns to "discriminate" between stimuli, responding
only to the one associated with the US, even if the CS is very
similar.
If a rabbit blinks in a generalized way to
different tones, and we want him to blink
only to a 1000 Hz tone, then we can use the Tone 670 Hz - no Food
following training: Tone 1000 HZ - Food
Tone 1500 Hz - no Food
Group B: simple conditioning to the 1000 Hz tone
Group A: discriminative training
Acquisition and generalization of learning
Discrimination
Discrimination training allows us to determine the differential threshold (the minimum
difference we are able to detect in a stimulus) of an individual or a species.
• By presenting an animal with a progressively more similar CS+ and CS-, we can
determine its ability to differentiate between stimuli of different frequencies,
wavelengths (colors), etc.
• If the animal responds to the CS+, but not to the CS-, then we know that it is capable
of detecting the difference; if it salivates to both, then its senses do not allow it to
distinguish between them.
Discrimination studies have been used to determine the degree to which patients with
autism spectrum disorders (ASD) are able to discriminate emotions. Some studies
indicate that their discriminative ability is impaired (e.g., they are unable to differentiate a
neutral face from a mild expression of anger or joy).
Acquisition and generalization of learning
Generalization and discrimination
Generalization and discrimination are
two sides of the same coin. To say that
there is too much generalization is the
same as saying that there is too little
discrimination.
Depending on how much the organism
responds to stimuli similar to the
trained one, different gradients of Flat gradient Bell-shaped gradient Sharp gradient
generalization can be distinguished. A lot of generalization Some generalization Very little generalization
Very little discrimination Some discrimination High level of
discrimination
Extinction and
related
phenomena
Extinction and related phenomena
Extinction
Reduction of a CR due to repeated
presentations of the CS in the absence of the
US.
Extinction trial: trial in which the CS is
presented without being accompanied by the
US with which it was previously associated.
It has a clear parallelism with therapies involving X+ X+ X+ X- X- X- X- X- X-
X+ X+ X+
exposure to feared stimuli (phobias, obsessive
compulsive disorder...). It is the experimental analogue
of these therapies.
Acquisition and extinction procedure
X+ | X-
Extinction and related phenomena
Then... Three phenomena indicate that:
Does extinction eliminate excitatory learning? Or… • The original learning (EC-EI) is not eliminated
does the individual "unlearn" the CS-US association? by extinction. In other words...
The answer is NO • Extinction learning (CS - noUS) is labile,
unstable...
After extinction, the CR recovers relatively easily if one of three possible situations occurs.
Depending on the cause (the situation that generates the recovery), the phenomenon is given a
different name.
1. Spontaneous recovery
2. Reinstatement
3. Renewal
Extinction and related phenomena
Spontaneous recovery
Extinction produces a decrease in the CR, but this
effect dissipates over time. If a rest period is
introduced after extinction training, a recovery of the
response is observed.
Once the acquisition and extinction of X has been X+ X-
completed, the simple passage of time produces a X?
spontaneous recovery of the extinguished CR.
Spontaneous recovery procedure
Acquisition Extinction Test
Spontaneous recovery: Reappearance of an
time
X+, X+, ... , X+ | X- , X-, ... X- | | X?
extinguised CR that occurs by the mere passage of
time. Or simply
X+ | X- | time | X?
Extinction and related phenomena
Spontaneous recovery
Recovery is often only
partial, the CR does not
return with the same
intensity it had before
extinction.
Spontaneous recovery is the
experimental analogue of
relapse.
Rescorla’s spontaneous recovery study (2004)
Subjects: rats.
CS1 - US1: noise – small sugar ball
CS2 - US2: light - solid food
CR: entries to the food dispenser.
CS1
Time elapsing between extinction and
test:
CS2
• Test with CS1: 8 days.
• Test with CS2: 0 days.
Extinction and related phenomena
Reinstatement
Definition: Recovery of an extinguished CR due to the presentation an
unsignaled US (without the CS) after the extinction phase.
Reinstatement procedure
Acquisition Extinction Isolated presentation of US Test
X+ X- + X?
Reinstatement study by Vansteenwegen et al., (2006).
Participants: humans; EC: blue square on the screen; EI: intense noise (1 sec.)
RC: skin conductivity (increases when we are afraid).
Reinstatement procedure
Acquisition Extinction Isolated presentation of IE Test
X+ X- + X?
Conductivity
Fear CR recovered after
presenting the US in an
unsignaled manner.
Extinction and related phenomena
Renewal of response
Definition: Recovery of an extinguished CR due to a context change between
the extinction and the test phases. In other words, recovery of an extinguished
CR when the CS is presented out of the extinction context.
3 types of response renewal procedures
Acquisition Extinction Test
ABA Renewal A: X+ B: X- A: X?
AAB Renewal A: X+ A: X- B: X?
ABC Renovation A: X+ B: X- C: X?
The letters represent the context in which each of the phases takes place. 1st letter = acquisition context; 2nd
letter: extinction context; 3rd letter: test context.
Of the three procedures (ABA, AAB and ABC), the ABA procedure produces the highest recovery.
Extinction and related phenomena
Renewal of response
In the experimental setting, contextual changes may involve:
• Changes in shape
• Changes in odor
• Changes in color or patterns on the walls (stripes, dots...)
In the school setting they may involve:
• Different classes
• Different schools
• Different educational environments (family vs. school)
• ...
Extinction and related phenomena
Implications of CR recovery phenomena:
CR recovery phenomena that occur in the laboratory are replicated in daily life. This implies that
maladaptive responses (e.g., exaggerated or irrational fear responses) tend to reappear.
• A student with a phobia of public speaking will experience fear of that situation again after some
time has passed since the end of exposure therapy (spontaneous recovery).
• A student who feared balloons because once one popped, but then had this fear eliminated after
playing with several balloons that did not pop, will regain the fear if a loud noise (different from a
balloon popping) occurs (reinstatement).
• A student who fears going to school after having been in school A, but who has lost that fear after
a change of school (school B), will again fear the situation of going to school if he/she returns to
the original school (ABA renewal).
• Etc.
Extinction and related phenomena
Conclusion on the three phenomena of CR recovery after extinction:
During extinction, the association between CS-US is not eliminated, the original
learning is latent and susceptible to be manifested:
• With the passage of time.
• If the individual is re-exposed to the US.
• If the CS occurs in a context different from that in which extinction took place.
All this indicates that 'relapse’, or at least the presence of fluctuations in the individual's
life with periods in which symptoms worsen, is almost a certainty. However, as
mentioned, after successful treatment or therapy, what was once an incapacitating fear
may end up being a simple discomfort.