Transcript Classical Conditioning

Unit 7: Learning
Learning is what makes us human.
• Adaptability
– Ability to cope with new and changing
circumstances
• Does history always repeat itself? Can we change the
course we’re on?
• Idea underlies all therapy
Learning is the heart of: Behaviorism
• Belief that behavior is learned
– genetics has little to do with it.
• “nurture-only” belief, mainstream1960s and 70s.
• Behaviorists believed learning done through
associations.
– Create habits (relating behavior to expected outcomes)
• Behaviorists want psychology to be a science:
– observable, testable and measurable.
Attributes of Learning
A. Learning = relatively permanent change in
behavior or knowledge as a result of experience
B. Distinction between learning and performance
•
Performance is an exhibited behavior
•
Learning can be inferred from performance, but performance
is not always an accurate measure of learning
C. Learning is adaptive
– Unlike most species whose behavior and bodies
have adapted to a certain environment
– humans have the capacity to learn
– Increase in speed of adaptation in comparison
• Learning and thinking (cognition) have
allowed us to adapt to all environments on
earth, without our bodies changing.
Learning IS Conditioning
•
Conditioning is the process of learned
associations
1. Classical – associate 2 stimuli to anticipated
events
2. Operant – associate a response (behavior)
and its consequence
•
Repeat the good ones, avoid the bad ones
3. Observational – learn from others’
experiences and link to our own associations
Classical Conditioning
• Ivan Pavlov: Classical Conditioning.
http://www.youtube.com/watc
h?v=nE8pFWP5QDM
Pavlov (early 20th century)
influenced Watson
• Watson is the father of behaviorism
– Little Albert experiment
“psychology should be an objective science
based on observable behavior”
• They ignored mental processes, but today we
know that consciousness is important
Definitions
1. Classical conditioning: learning that takes place
when originally neutral stimulus comes to
produce a conditioned response because of its
association with an unconditioned stimulus
•
In other words – CREATING A REFLEX
2. An unconditioned stimulus (UCS or US)
produces an unconditioned response (UCR
or UR), even in the absence of previous
training
3. A conditioned stimulus (CS) is a stimulus that has
come to elicit a conditioned response (CR) because
the organism associates the conditioned stimulus
with the unconditioned stimulus
- Initially called the “neutral stimulus”
So let’s go back to Jaws…
Pavlov’s Classical Conditioning:
Pavlovian Response or “Psychic Reflex”
• accidental psychologist: studying digestion initially
• theory links to emotion, temperament, neuroses, and language
Pavlov’s Experiment
Unconditioned vs. Conditioned Responses
• In Pavlov’s demonstration UR = CR
– Salivation
• Although UR and CR consist of same behavior,
there are subtle differences
– CR usually weaker or less intense
• Sometimes UR and CR are different but
related
– Animal given a shock, UR = pain, CR = fear of
imminent pain
Conditioned Reflex
• Classically conditioned
responses described as
reflexes
– Involuntary and automatic
Taste aversions?
Trials
• How long does it take to learn something?
Classical Conditioning Applications
1. Conditioned Fear and
Anxiety
•
•
Phobias
Irrational fear due to
classical conditioning
•
•
Fear of dentist drill
Careful though –
susceptibility of
irrational fear is mostly
based on genetics…
how?
Common Fears based on association
2. Emotional Responses
– Arousal
• Smell of first love’s cologne/perfume
3. Physiological Responses
• Sexual arousal in quails
– Conditioned to become aroused by nonsexual
stimuli
– Conditioned to elicit increased sperm release
– Fetishes for inanimate objects
• Difficult to test connections to human sexual fetishes
Acquisition: initial stage of learning
something
1. Involves repeated pairings of the CS and the
UCS/US
Acquisition Paradigms (patterns)
• What are the different ways in which the
initial learning can take place?
a. Trace Conditioning
–
–
–
–
–
CS is presented and terminated BEFORE presentation of the
UCS/US
Conditioning often effective when the interval BETWEEN
presentation of the CS
the UCS/US is about a half second
Fear studies; dependent on usage of hippocampus
http://www.youtube.com/watch?v=PsGjh6ul7mE
b. Delay Conditioning
– when CS is presented and continues at least until the
UCS/US is presented
– Often times paired with trace conditioning in studies
– Hippocampus-independent
– Fear expression
Fear Expression in Rats
c. Simultaneous Conditioning
– Occurs when CS and the UCS/US are presented
and terminated at the same time
– Anti-smoking ads
d. Backward conditioning
– Occurs when the UCS/US is presented before the
CS
– http://www.youtube.com/watch?v=fT6IWAIf580
e. Temporal conditioning
– Occurs when CS is at a fixed period of time
between presentations of the UCS/US
• i.e. dog starts to salivate at 7:59am because s/he is fed
at 8am everyday
Extinction
1. A procedure that leads to gradual weakening
and eventual disappearance of CR
•
Involves repeatedly presenting CS without
pairing it with UCS/US
Spontaneous Recovery
1. Occurs when previously extinguished CR suddenly
reappears after a period of training
2. Renewal effect
•
•
response will reappear if animal is returned to original
environment where acquisition took place
Proves that extinction is a suppression not an erasure
(unlearning)
•
Explains drug abuse and relapse and difficulty getting rid of
phobias permanently
Classical Conditioning II
A. Generalization
1. CR occurs to stimuli similar to
the CS
•
even though these stimuli may
have never been associated with
the UCS
2. The more similar the stimuli are
to the CS, the greater likelihood
of generalization
–
Ex: Pavlov’s dogs salivating at a
different tone than in training
John Watson and Generalization
• Little Albert
– sight of white rat (CS) was paired with loud noise
(UCS) until the CS alone produced crying and
other responses indicative of fear
• Ethical issues with Little Albert
– Never heard from again
– Failure to ensure no lasting ill effects
B. Discrimination
1. stimuli similar to the CS does not produce a
CR
2. The less similar new stimuli are to the original
CS, the greater the likelihood of
discrimination
•
What happens to generalization gradient when
an organism learns a discrimination?
How are discrimination and
generalization similar? Different?
Higher-Order Conditioning
1. A CS functions as if it were a UCS
•
Result: classical conditioning does not
depend on natural US (remember the sexy
quails?)
Higher-Order Conditioning
E. Factors that Affect Conditioning
1. Contiguity: The closer two stimuli are in space and
time, the stronger the association between them.
------------------------------------------------------------------2. “Belongingness”: The “fit” between CS and US
3. Contingency: “Information value.” The higher the
correlation between two stimuli, the stronger the
conditioned response.
4. Salience: More intense or noticeable stimuli
condition more rapidly.
1. Contiguity model
•
•
Argues conditioning will occur whenever CS and
UCS are paired
Based on Pavlov
2. CS-US belongingness: not all CS’s and US’
associable
3. Contingency model
• Argues CS must reliably predict UCS for
conditioning to occur
– based on work of Rescorla and Wagner (took indepth look at each trial of conditioning)
– Supported by phenomena like blocking (Kamin)
• Multiple CS tests, results?
Contingency Phenomena
• US pre-exposure effect: Presenting the US
repeatedly prior to CS-US trials slows
acquisition.
• CS pre-exposure effect: Presenting the CS
repeatedly prior to CS-US trials retards
acquisition. (a.k.a. Latent Inhibition)
Leon Kamin: Blocking
• US has to be “surprising” for association to
occur
– Selective attention and learning
4. Salience effects
• Overshadowing – in compound or higher
order conditioning, the more salient CS wins
One last one… d’oh
• Test devaluation: critical period when
stimulus/response has occurred too much,
creating an adverse effect
In Conclusion!!
2 models
1. Contiguity
•
Conditioning will occur when CS and UCS are paired
2. Contingency
•
CS must reliably predict UCS for conditioning to occur
(there are a lot of things that could get in the way of
that)