Learning, Attention & Memory Flashcards
Learning
enduring change in behaviour, resulting from experience
•learning associations
Conditioning
process environmental stimuli + behavioural responses become connected
–Classical/Pavlovian: learn 2 types of events go together
–Operant/Instrumental: learn behaviour leads to particular outcome
Classical Conditioning
neutral stimulus elicit a reflexive response because associated with stimulus already produces that response
neutral: no reactions, no reflexive response
• pair it with something that produces response
Unconditioned stimulus (US) + Unconditioned response (UR)
US: automatic have reaction to, automatically elicits response
UR: salivating, automatic response
Conditioned stimulus (CS) + Conditioned response (CR)
CS: neutral becomes paired with US
CR: looks similar to UR, reaction to CS, but not identical
Classical Conditioning
US (food) produces UR (salivation)
neutral produces no salivation response
US presented just after neutral repeatedly, US continues to produce UR
neutral stimulus alone produces CR
Acquisition
gradual formation of an association between the conditioned + unconditioned stimuli
•Best when US presented right after CS
Extinction
process in which the conditioned response is weakened when the conditioned stimulus is repeatedly presented without the unconditioned stimulus
take pairing away, eventually response extinguished, now learning tone doesn’t = food anymore
Spontaneous recovery
process in which a previously extinguished response re-emerges following presentation of the conditioned stimulus
we wait, present CS again, produces some CR still
Stimulus generalization
stimuli that are similar but not identical to the conditioned stimulus produce the conditioned response
not a specific stimulus, but also any identical stimulus
• Not just the tone, but any tone
Stimulus discrimination
differentiation between two similar stimuli when only one of them is consistently associated with the unconditioned stimulus
know 1 tone leads to food, but other doesn’t
Second-order conditioning
something is consistently paired with the conditioned stimulus, without the unconditioned stimulus, and leads to a conditioned response
associating some other stimulus with CS, then this other stimulus starts eliciting CR
Phobias
acquired fears that are out of proportion to the real threat of the object or situation
•little albert: came to associate all white things with terrified response
•paired white rat + loud noises
•amygdala crucial for fear conditioning
Classical Conditioning: Later Developments
- Not all CS-CR pairings are the same!
* Some associations are easier to learn than others
Conditioned food aversion
Associating a particular food with an unpleasant outcome (i.e., illness). Can be formed in one trial, even if the illness doesn’t occur right away.
•even if food isn’t the thing that makes you sick
Biological preparedness
animals are genetically programmed to fear some objects more than others
we’re programmed to fear particular objects that makes sense we are afraid of them. They pose much more of a realistic threat
– E.g., phobias about snakes and heights are more common than phobias about squirrels and staplers
Prediction
In order for learning to take place, the CS must
accurately predict the US
Rescorla-Wagner model
cognitive model of classical conditioning which states that the strength of the CS-US association is determined by the extent to which the US is unexpected or surprising
– Because this leads to greater effort by the animal to understand why the US appeared
when US unexpected/surprising, it makes us wonder what caused it
•More unexpected, more encouragement to look for cues
Operant Conditioning
learning process in which the consequences of an action determine the likelihood that it will be performed in the future
Thorndike’s law of effect
behaviour that leads to a “satisfying state of affairs” will more likely occur again, and any behaviour that leads to an “annoying state of affairs” will less likely recur.
Operant Conditioning
-cat in the box
if cat hits lever, door opens, get food
•Eventually accidentally press the lever
•After a while cat learns what to do to get out of box
Reinforcer
stimulus that occurs after a response and increases the likelihood that the response will be repeated
–Primary: satisfy biological need
secondary: money that associate with rewards
Shaping
reinforcing behaviours that are increasingly similar to the desired behaviour
–Reinforcing successive approximations eventually produces the desired behaviour by teaching the animal to discriminate which behaviour is being reinforced
•We can’t just wait around for them to accidentaly do it
•Everytime animal gets a little bit closer to the trick, we reward them
Positive + Negative reinforcement
• Positive reinforcement: Increases the probability of a behaviour being repeated by the administration of a (positive, rewarding) stimulus
adding positive stimulus
• Negative reinforcement: Increases the probability of a behaviour being repeated by the removal of a (negative, aversive) stimulus
taking away negative stimulus
Positive + Negative punishment
• Positive punishment: Decreases the probability of a behaviour being repeated by the administration of a (negative, punishing) stimulus
by adding negative stimulus
• Negative punishment: Decreases the probability of a behaviour being repeated by the
removal of a (positive, pleasurable) stimulus
removal positive stimulus
Continuous reinforcement
results in fast learning,but also fast extinguishing as soon as the reinforcement is removed
when association is there everytime, learn faster
• When take away reinforcer, so it’s easily extinguished
Partial reinforcement
behaviour is reinforced intermittently
some of the time
•Variable ratio most effective (slot machine)
Types of Partial reinforcement
– Ratio schedule: Based on number of times the behaviour
occurs (e.g., you get paid for every 10 toys you build) # of times behaviour occurs
– Interval schedule: Based on the passage of time (e.g., you get paid for every hour of work)Rewarding based on time
– Fixed schedule: Reinforcement is consistent and predictable, always the same
– Variable schedule: Reinforcement is variable and unpredictable (occurs after a varying number of behaviours, or a varying amount of time) unpredictable
Biological Constraints
Animals have a difficult time learning behaviours that are incompatible with innate, adaptive behaviours
– Conditioning is most effective when the association between the behavioural response and the reinforcement is similar to the animal’s built-in predispositions
hard to get animal to do something if it’s opposite something it does naturally
Latent learning
Learning that occurs without reinforcement
• Edward Tolman argued that reinforcement had more impact on performance than acquisition/learning
– Showed that rats could learn a path through a maze without reinforcement (latent learning), but not reveal their learning until it is reinforced (red line below)
animals learn without reinforcements
•Once you add reinforcement, they perform better + faster
•Evidence shows that it has more to do with performance
Insight learning
occurs without reinforcement
aha moments, no need reinforcement
Observational Learning
behaviours are acquired or modified following exposure to others performing the behaviour
• E.g., Bobo Doll Experiment – Vicarious reinforcement
•See others engaging in behaviour, whether they get reward/punished
•Vicarious: seeing others being reinforced/punished for behaviour
Superstitious Behaviour
tendency to associate chance events with other events that occur together in time can lead us to associate chance events unrelated to reinforcements or punishments
Biological Basis of Learning
• The experience of pleasure usually results from activation of dopamine neurons in the nucleus accumbens
– More dopamine is released under deprived conditions (e.g., getting food when you are starving is more rewarding then getting food when you are not starving)
• More dopamine!more reinforcement
• More deprived = more dopamine release