Learning 1: Introduction to Cognition - PDF

INTRO TO COGNITION: LEARNING 1: ASYNCH Dr Nils Muhlert Today we will… ¨ Find out more about classical & operational learning ¤ Findout why buffaloes walk on roads ¤ Understand how we can develop associations Habituation ¨ Decline in tendency to respond to stimuli once it become...

INTRO TO COGNITION: LEARNING 1: ASYNCH Dr Nils Muhlert Today we will… ¨ Find out more about classical & operational learning ¤ Findout why buffaloes walk on roads ¤ Understand how we can develop associations Habituation ¨ Decline in tendency to respond to stimuli once it becomes familiar ¤ 1st instance of stimuli: startle response ¤ 2nd instance: reduced startle ¤ 3rd instance: hardly any ¤ 4th instance: ignored ¨ Ignoring the familiar allows us to concentrate on more important activities! ¨ Has to rely on memory Classical conditioning ¨ In habituation, event recognised as familiar, but nothing new is learned. decrease in response to stimuli ¨ Learning = associations ¤ Either strengthened… ¤ …or weakening of existing associations ¨ Ivan Pavlov (1849-1936) ¤ Dogs salivate when: n Meat placed in mouth n Sight of meat n Sight of dish n Sight of person who brings meat n Sound of that person’s footsteps Reflexes ¨ Unconditioned: ¤ Meat placed in mouth – salivation ¤ No learning needed! ¤ (meat = unconditioned stimulus - US) ¤ (Saliva = unconditioned response - UR) ¨ Conditioned ¤ Sound of bell/ person’s footsteps - salivation ¤ Learning needed ¤ Bell = conditioned stimulus – CS ¤ Saliva = conditioned response - CR Varieties of conditioned responses ¨ By pairing CS with US, we can make: ¤ Crabs retract their eyes (Feinman et al, 1990) ¤ Fish thrash ¤ Octopuses change colour (Hanlon et al, 1996) ¨ In humans, training CS with US can: ¤ Cause increase in galvanic skin response (US loud noise) ¤ Reflexive eye blink (US puff of air on eye) ¤ Link words to sexual arousal (Hoffman et al. 2002) Instrumental conditioning ¨ Edward L. Thorndike (1874 – 1949) ¤ The response depends on the behaviour. n Reinforcement ¤ E.g. meat would only come IF the dog salivates. BF Skinner ¨ 1904-1990; Psychologist, followed from Thorndike ¨ Insisted on sharp distinction from conditional processing. ¤ Actionsaren’t just reflexes – they’re voluntary (chosen by animal) ¤ Operants – action operates on environment to bring about change linked to consequence. Law of effect ¨ Responses that are rewarded = Strengthened ¨ Responses that are not rewarded (or punished) = weakened ¨ Strength of response is adjusted according to that response’s consequences. ¨ Animals show learning curves – faster responses following reinforcement Today we’ve learned ¨ Classical conditioning works by pairing US with CS ¤ Not dependent on the animal responding in a certain way ¨ Instrumental (or operant) conditioning uses law of effect. ¤ Uses the animal’s response to weaken or strengthen the association. INTRO TO COGNITION: LEARNING 1 Dr Nils Muhlert Today we will… ¨ Find out more about classical & operational learning ¤ Find out why buffaloes walk on roads ¤ Learn how heroin addicts increase tolerance through classical conditioning ¨ Introduce Tolman’s ‘cognitive map’ theory ¤ Watch chimpanzees make tools In the asynch material you… ¨ Found out about habituation ¨ Learned about classical conditioning ¨ (meat = unconditioned stimulus - US) ¨ (Saliva = unconditioned response - UR) ¨ Bell = conditioned stimulus – CS ¨ Saliva = conditioned response - CR Another Classical Conditioning example ¨ Clip from “The Office” In this clip, the unconditioned stimulus was A. The computer beep B. Eating the Altoid C. Salivation/Fresh breath D. Asking for a mint E. None of the above In this clip, the conditioned stimulus was A. The computer beep B. Eating the Altoid C. Salivation/fresh breath D. Asking for a mint E. None of the above Acquisition of CR Acquisition of CR (2nd order) ¨ Pavlov Extinction ¨ Wolf likes his dinner  Spontaneous recovery & Reconditioning ¨ After a break, wolf tries his luck again Prob of CR 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Trials Generalisation ¨ Wolf sees similar dining places Discrimination ¨ Over time, wolf realises that tool shack doesn’t stock cats ¨ In early trials, wolf generalizes to tool shack (CS+) ¨ Later, wolf learns to discriminate between cat shack & tool shack (CS-), no more salivation ¨ Wolf learns tool shack = ‘no food’ = an inhibitor Conditioned fear ¨ CR can be much more complex than salivating ¤ E.g. Conditioned emotions ¨ Conditioned emotional response procedure ¤ Estes & Skinner, 1941; Kamin, 1965 ¨ Fear response is superimposed on ongoing activity ¤ Initially,wolf ignores the footsteps, salivates. ¤ In later trials, wolf starts to suppress salivation. Relevance of conditioned fear ¨ May be similar to human behaviour ¨ Does this explain phobias? ¨ What about fear of non-conditioned associations? ¨ Some fears are ‘biological’ (Poulton & Menzies, 2002) Relations between CR and UR ¨ The CR and UR seem similar ¤ When it hears the bell, the dog salivates (CR) ¤ When meat is placed in its mouth, it salivates (UR) ¨ But they’re subtly different! ¤ Meat in mouth = more salivation, richer in enzymes ¨ For fear conditioning ¤ UR – exposure to taser – wolf jumps and howls, HR increases. ¤ CR – footsteps – wolf freezes, tenses muscles, HR slows = fearful anticipation Conditioning & drug use ¨ Diabetes ¤ Injectinsulin, to deplete blood sugar levels. ¤ After many injections, sight of needle starts to trigger early physiological response = INCREASE in blood sugar levels. ¤ Body is striving for homeostasis - CR prepares for US n Compensatory reaction Conditioning & drug use ¨ Gerevich reported case of KJ, 26 year old male ¤ First seen in rehab with wife, 1997 ¤ Reported to take 1g heroin/day ¤ Died of overdose, Jan 9th 1999 ¨ Day before, took 0.5g heroin with wife at home ¨ Following day, 0.5g in subway station toilet. ¤ Confirmed by blood & urine samples at pathology ¤ Lack of body’s normal anticipation/ conditioning? Gerevich et al., 2005 Harm Reduction Journal Instrumental conditioning ¨ Edward L. Thorndike (1874 – 1949) ¤ The response depends on the behaviour. n Reinforcement ¤ E.g. meat would only come IF the dog salivates. BF Skinner ¨ 1904-1990; Psychologist, followed from Thorndike ¨ Insisted on sharp distinction from conditional processing. ¤ Actionsaren’t just reflexes – they’re voluntary (chosen by animal) ¤ Operants – action operates on environment to bring about change linked to consequence. Operant conditioning ¨ Skinner and his operant conditioning box Law of effect ¨ Responses that are rewarded = Strengthened ¨ Responses that are not rewarded (or punished) = weakened ¨ Strength of response is adjusted according to that response’s consequences. ¨ Animals show learning curves – faster responses following reinforcement Instrumental conditioning in action Cognitive learning ¨ Did the dog learn about Darwin through classical or operant conditioning? ¨ Edward Tolman (1886-1959) Cognitive views on conditioning ¨ In classical conditioning procedures, animal does not respond to CS in same way as US ¤ subtlydifferent ¤ Sometimes very different (e.g. fear learning) n CS becomes signal that US will follow, animal makes preparations. n Signal needs to occur within close timeframe in classical conditioning, so as to learn… Contiguity (Pavlov) ¨ For wolf, footsteps need to occur just before taser, in order to learn association frequency of conditioning Relative -.5 0.5 1 1.5 Stimulus Interval (seconds) Contingency ¨ Many signals were present throughout wolf’s event ¤ Cat meowing, crab claws clacking, etc ¨ Only loud footsteps allowed prediction of taser. ¨ Stimulus needs to be not only contiguous with target event. ¤ Also contingent – reliably occur prior to event Contingency effects ¨ Rescorla, 1967, 1988 ¤ Rats exposed to tone (CS) & shock (US) ¤ Tone not always presented prior to shock ¤ In 2 groups of rats, tone indicated 40% chance of shock (CS-US pairing). ¨ Some rats - additional shocks without warning (tone is little use, rats learned to ignore) ¨ Other rats – with no tone, much lower chance of shock ¤ Tone is slightly useful – these rats showed conditioning What about when no contingency occurs 1. Wolf hears footsteps – but not always followed by taser. ¤ 50% probability ¨ In this situation, animals generally fearful BUT ¤ Absence of footsteps is now inhibitor of fear (wolf is relaxed). 2. Wolf gets tasered 50% of time, without warning ¤ 50% probability ¨ Wolf must be afraid all the time Schwartz et al., 2002: Psychology of learning & Behavior How do animals learn ¨ The answer may be surprising… ¤ Surprise ¨ Probabilities are updated on trial by trial basis ¤ Ifa result is unexpected, then wolf changes his sensitivity to a situation. The Blocking Effect ¨ Leon Kamin, 1968 1. 2. 3. 4. Blocking in humans – temp in celcius & Fahrenheit Cognitive views on instrumental conditioning ¨ Law of effect ¤ Responses strengthened by pos reinforcers ¤ Responses weakened by punishment ¨ Latent learning (Tolman & Honzik, 1930) ¤ Acquisition of new knowledge without change in behav 10 days exploring …then test Latent learning (2) ¨ Harkness & Marouda (1985) ¤ Desert ants leave nest, look around for food ¤ After finding food – heads straight back to nest ¤ Food can be as far as 50 m away (entrance to nest 1 mm) ¤ Ant uses movement velocity (speed, direction), to compute & update position. ¤ Indicates learning, but not by conditioning!! This week we’ve learned ¨ Classical conditioning works by pairing US with CS ¤ Can be 2nd order or more. ¤ Allows generalisation & discrimination, can be used to condition fear responses. ¨ Instrumental (or operant) conditioning uses law of effect. ¨ Cognitive map theory accounts for non-result based learning. Reading for this lecture ¨ Skinner BF (1976) About Behaviorism. New York: Vintage Books