Flashcards in Lecture 3: Learning (Curtis) Deck (53):
an enduring change in the MECHANISMS OF BEHAVIOR involving specific stimuli and/or responses that results from prior experience with similar stimuli and responses
-> defined in terms of a change in the mechanisms of behavior rather than a change in the behavior itself
mechanisms of behavior
underlying machinery that makes behavior happen --> physical machinery in the CNS
the startle response
a defensive response of evolutionary importance
ex: a sudden jump and tensing of the muscles
a DECREASE IN RESPONSE as a consequence of repeated exposure to a stimulus even though the stimulus remains the same
ex: a mare no longer threatening people when they walk by her baby and nothing bad happens
two important things to remember about habituation:
the initial response (mare being threatening toward people when near her baby) is an INNATE repsonse.
the decrease in response (mare no longer being threatening to people) is LEARNED
what goes on neurologically with habituation?
response fails to occur because changes in the CNS block the relay or sensory neural impulses to the motor neurons
the habituated response is restored by exposure to a strong extraneous stimulus paired with the stimulus to which the animal has habituated
ex: a human actually does threaten the foal, then the mare returns to threatening passerby again
an increase in responsiveness produced by repeated stimulation.
more intense stimuli produce greater increases in responsiveness and with more intense stimuli the sensitization effects persist longer
similarities between sensitization and habituation
both reflect how an organism ends up sorting out what stimuli to ignore and what stimuli to respond to
both processes involve learning about just one stimulus
simple mechanism whereby an organism learns about RELATIONSHIPS BETWEEN STIMULI and alters its behaviors accordingly
unconditioned stimulus (US)
elicits a particular response without the necessity of prior training
response that occurs without prior conditioning
does not elicit a particular response initially, but comes to do so as a result of becoming associated with a US
response that comes to be made to the conditioned stimulus as a result of classical conditioning
the number of pairings needed between a conditioned stimulus and a neutral stimulus depends on:
1. intensity of stimulus
2. relevance of neutral stimulus
when only ONE SPECIFIC STIMULUS elicits the response
response elicited by a diverse but related group of stimuli
taste aversion learning - what is it and what are its 2 unique properties?
a robust form of classical conditioning
ex: an animal eats something --> gets sick. source of the CS and US is the same.
1. strong taste aversions can be learned with just one pairing
2. learning occurs even if animal does not get sick until several hours after exposure
if animal is repeatedly exposed to a CS without further pairing with the US, then the animal response will eventually cease
ex: if you stop pairing the whistle with the fish, the whistle wont have the same meaning anymore
what is the difference between extinction and habituation?
habituation is loss of an INNATE response
extinction is loss of a LEARNED response
in both situations, the stimulus intensity remains the same but the response decreases
operant conditioning (instrumental conditioning)
behavior that is affected by the immediate consequences. the behavior occurs because it was previously instrumental in producing certain responses. ie goal directed behavior
ex: a dog has learned that the action of sitting when a person says "sit" results in food.
a good event
increases the probability of the behavior occurring in the future
decreases the probability of the behavior occurring in the future
positive operant conditioning
a controlling stimulus is ADDED to environment
negative operant conditioning
a controlling stimulus is REMOVED from the environment
describe the way operant conditioning works
there is an eliciting stimulus: a hand signal, command, etc. --> gives the animal information about the immediate consequences of a given behavior --> response: behavior performed --> controlling stimulus: food reward
adding something to the environment immediately subsequent to the behavior to increase the frequency of behavior.
"if you do X, good things will happen, so keep doing X."
removing something from the environment to increase the frequency of behavior
ex: car begins beeping annoyingly -> you put on your seatbelt -> car stops beeping. this increases the likelihood of you putting on your seatbelt in the future
adding something to the environment to decrease the frequency of a behavior.
probability that a behavior will decrease as a consequence of a controlling stimulus occurring immediately subsequent to behavior.
ex: you do not put on your seatbelt -> car starts beeping -> beeping is being added to the environment to reduce the frequency of you NOT putting on your seatbelt
"if you do X, bad things happen, so stop doing X."
ex: a bark collar
important point to remember about negative reinforcement and positive punishment
you generally can't have negative reinforcement without positive punishment first.
negative reinforcement = removing something from the environment to increase the frequency of behavior
positive punishment = adding something to the environment to decrease the frequency of a behavior
in our seatbelt example, the beeping was added to the environment to decrease our behavior of NOT wearing our seatbelt (positive punishment). when desired behavior was achieved (wearing a seatbelt) the beeping was removed from the environment (negative reinforcement)
for punishment to be effective, 3 conditions must be met:
1. immediate - within seconds of behavior
2. consistent - occur every time
3. appropriate for individual - interrupt behavior, but not cause fear
which punishment do you start with?
removing something from the environment to decrease the frequency of a behavior
referred to as a "time out" - if you do X, good things don't happen, so stop doing X.
prompting and fading
initially getting the animal to do a desired behavior by touching the animal and giving an eliciting stimulus - gradually fade the intensity of the prompt
rewarding the "first step" of a behavior and gradually reward further and further into progression of the desired behavior - all done without touching the animal.
- often done for things that aren't likely to be done naturally.
ex: teaching a dog to lie down. give treat for each step in reaching desired position, gradually require animal to engage in behavior that is more and more like final behavior.
schedules of reinforcement -
determines how and when the response will be followed by a reinforcer
what conditions must be met for schedules of reinforcement?
- occurrence of a certain number of responses
- passage of time
- presence of certain stimuli
- occurrence of other responses
schedule of reinforcement: ratio
depends on number of responses animal has performed
schedule of reinforcement: continuous
animal is reinforced for each behavior
MOST EFFECTIVE when learning a NEW behavior
schedule of reinforcement: fixed ratio
gets a treat after a certain number of responses
schedule of reinforcement: variable
a different number of responses is required for the delivery of each reward.
if on a steadily increasing variable ratio schedule, the behavior is likely to become very persistent
schedule of reinforcement: interval
depends on the amount of time that has passed
a response is reinforced only if it occurs after a constant and set amount of time after last reward was given.
responses are reinforced if they occur after a variable interval since the delivery of the previous reinforcer
what affects how fast and how well learning occurs?
a neutral stimulus (clicker, "good," whistle, etc.) that is paired with food via classical conditioning. eventually comes to have close to same value as food, must still be followed with the food to prevent extinction.
interference with the learning of new instrumental responses as a result of exposure to inescapable and unavoidable aversive stimulation
- animal learns there is no consequence to its behavior and stops trying.
- applies to dogs that have experienced chronic, inescapable abuse
flooding and problems with flooding
deliberate exposure of the animal to a stimulus until the response extinguishes or the animal habituates.
the problem: once session begins, exposure to the stimulus must continue until the response ceases. therefore animals with strong fears may injure themselves.
when is it best to use intermittent reinforcement?
good at teaching behaviors you want, REALLY good at keeping behaviors you dont want from happening
expose animal to a low-level stimulus and increase the intensity of the stimulus while maintaining the same response