092408 324 Lecture

Classical Conditioning Applications Little Albert Watson and Rayner Exposed Albert to a white lab rat Child played with the rat – displaying no fear NS = Rat He then made a loud sound as Albert touched the rat CS = Rat; US = hammer strike on steel This made albert cry in the presence of the rat And other white furry items They never extinguished the response

More on fear - treatment You can also treat fears using classical conditioning Fear existed for rabbits Slow exposure at closer intervals – to the rabbit Paired with food as well An example of counterconditioning (exposure therapy – systematic desensitization) Hate!

Aversion therapy Paraphilia treatment Pair inappropriate sexual arousal with an unpleasant experience

Natural Selection What is it? Survival of the fittest? Survival of the most fit? Survival of the fit enough? Variation in species Some variation is better suited for the conditions than other variations Favorable ones are able to reproduce THE ENVIRONMENT DOES THE SELECTING The organism does not have a choice in the matter

Natural Selection and Moral Value Natural Selection is neither good or bad A given environment may select a given trait - but as the environment changes, that trait may become undesirable and the species less fit Some selected traits may have negative and positive effects

Natural Selection - examples The Blind Watchmaker The peppered moth Humans caused a change in environment Predators did the selecting Rock pocket mouse Owls did the selecting

Natural Selection, Creationism, Evolution - MYTHS http://www.newscientist.com/channel/life/dn13620 Please review all articles in this area

Natural Selection in other forms There are two more levels of natural selection One that happens on a cultural level Selection of behavior in a society – this leads to customs, etc. One that happens on an individual level You learn to do a given behavior in a given environment and not in others

How has natural selection prepared us? Reflexes A relationship between an event and a behavior They serve some function to the organism Dust in your eye – you blink (protective function) Too much tequila – you vomit (protective function) The key here is the relationship (not the behavior itself)! Occur in nearly all members of the species and appear at predictable developmental stages. They may not be useful to the organism

Reflexes – can be modified Sensitization – eliciting a response may increase the probability of the response to stimuli Loud sounds Habituation – evoking a given reflex response repeatedly will result in a reduction in the intensity or probability of that response Loud sounds

Modal Action Patterns Similar to reflexes Found in nearly all members of a species Strong genetic basis Elicited by “releasing stimuli” Different than reflexes Involve the entire organism More variability in the actual behavior Examples: Cats confronting a dog Migration in a “v” formation

General Behavior Traits Current thinking is that humans probably don’t have MAPs or Instincts We do have GBTs Wide range of stimuli elicit a wide range of responses – aversives and aggression

Evolved Modifiability Learning! What changes when learning occurs is behavior. Behavior is a reliable measure What changes behavior is experience Stimuli – anything that can be detected Learning was selected for Learning increases survival rate

Studying Learning Only been studying learning scientifically for about 100 years. Explanations must refer to physical events Identifying the events that produce something Falling bodies – greeks – “need to return to it’s natural place” We must avoid circular arguments Bob doesn’t work because he’s lazy He’s lazy cause he doesn’t work

Operational Definitions Must define behavior in terms of how to measure it Lever press Key pecking Running a maze Defining feeling and thinking in operational terms Must be observable

Measuring learning How do we measure learning? If defined incorrectly, we may not observe learning when it is happening Maze completion vs. errors Examples of ways to measure: Learning as error reduction (ebbinghaus) Learning as a change in topography (mirror art) Learning as intensity of behavior (strength, etc) Learning as speed Learning as latency (time from stimulus to response Learning as a rate Learning as fluency (errors and rate)

Measuring learning How? What do we use? Cumulative recorder Instances of behavior recorded and plotted with time to produce a rate Computers Stop watches Exam scores

Research Designs Anecdotal Evidence First or secondhand reports of personal experience Not scientifically sound but provide research questions and areas of further study Case Study A highly detailed analysis of a person or problem No ability to detect cause Descriptive Study More representative than a case study Still no experimental control Cannot detect cause – they are correlational in nature

Experiments Researcher manipulates one or more variables and measures the effect of this manipulation on other variables Independent and dependent variables Between subjects designs Comparing groups Random assignment Within subjects designs – single subject design Comparing a group to itself

Non-human research High levels of control Genetic Learning Ethical reasons Cannot do some research on humans Objections Not ethical People are not rats No practical value IACUC

Explaining behavior Mentalism Events happen in the environment – they in turn generate an idea and the idea causes the behavior (the organism decides what to do) Natural science approach Behavior is caused Causes precede their effect Only natural phenomena are employed parsimony

Classical Conditioning Studying gastric secretions in dogs Noticed “psychic secretions” Mental events are tied to biological outputs Experience altered the actions of the glands The gland matched the stimulus The gland became active at other stimuli

Classical Conditioning – the process Identified two types of stimuli Unconditional responses The natural reflex to a stimulus permanent Conditional responses A reflex develops through experience Not permanent

Classical Conditioning – the components A stimulus A response to that stimulus Unconditional reflex US  UR Conditional reflex NS CS  CR

Classical Conditioning - Examples Eyeblink conditioning Galvanic skin response Change in conductivity of the skin Increase conductivity when paired with shock Commercials Sex Conditioned suppression Suppression of behavior when paired with a shock Critter is responding at a given rate – we can suppress it Long presentation of the CS followed by shock Decrease in responding in the presence of that stimulus (CS) Taste aversion learning Food (cs) – poison (us) Rule out remaining illness Two groups – one receives cs and us. One receives us. Measure food eating behavior later – it suppresses.

Classical Conditioning - Trace Trace conditioning Present CS – turn it off – allow time – present US Term came from a “memory trace” As the time interval increases – conditioning decreases

Classical Conditioning - Delay Short – present the CS before the beginning of the US – leave them both on – then turn off at the same time For eyeblink – between .5-1 second delay Overlap is the key here Long – present the CS earlier and for longer – slightly weaker conditioning

Classical Conditioning - Simultaneous Not a very effective method for conditioning CS and US coincide at the exact same moments Seems to act as a different stimulus (the combination)

Classical conditioning - backward The CS comes on after the US Not very effective – if effective at all Not efficient

Contingency and Contiguity Contingency is a major key! the degree of prediction from the CS to the US effects the amount of conditioning Rescorla – p(us/cs) and p(us/no cs) Vary these probabilities using a 2 minute tone at random intervals .4 that the US would occur during a CS; .2 a US would occur during a noCS When the first was higher than the second – you would see suppression Contiguity also plays a role Interstimulus interval (ISI) – time between the termination of the CS and the onset of the US (trace) Time from onset of CS and onset of US (delay) Typically the shorter the interval the more effective the conditioning The response matters! Taste aversions can have long ISIs and still produce learning, other stimuli are not so effective.

Compound Stimuli Two or more stimuli occurring together (sound and sight - CR) then paired with a US Can test the effects of this by presenting one of the CSs alone after pairings Often you get conditioning to both But not always

Blocking Kamin – conditioned suppression procedure (css = light, tone, light tone; cr = shock) Two groups: blocking and control One stimulus seems to block conditioning to the other – no new predictability. Group Phase 1 Phase 2 Test Phase Result Blocking 1 Light Light/tone Tone Tone elicits no CR Control ------- Light/tone Tone Tone elicits CR

Overshadowing Intensity of the CS effects conditioning trials Loud CS and soft CS  US = CR Test with either CS +CS = CR -CS ≠ CR But you can then use the –CS by itself and get conditioning One seems to overshadow the other

Experience with the CS Latent inhibition Presence of a CS in the absence of the US Delays acquisition in the future Prediction is decreased

Acquisition Strength of CR Number of trials CR is increasing in strength Learn more on early trials than on later ones

Extinction Passage of time has very little effect on strength of CR Use extinction to remove or reduce the CR Present the CS without the US Remember predictability!

Spontaneous Recovery The response returns Time between extinction trials will affect this The larger the gap the larger the SR On each trial you will see less and less SR

Rapid Reacquisition Acquisition – extinction – reacquisition With each extinction/reacquisition the organism learns the CR quicker

Generalization and Discrimination Generalizability curve Basically a normal curve Response to stimuli follows a predicable pattern The further away from the original stimuls the less likely the CR will be elicited Discrimination The above allows for discrimination – to respond to one CS but not a similar NS

Classical Conditioning Apps Little Albert and fear conditioning – watson and raynor NS = White rat Response was playing CS  US (hammer on metal) CR = screaming, etc (fear) This generalized to other small white furry objects Did not extinguish this response

Other applications Paraphilia treatment Aversion therapy Pairing unpleasant stimuli with inappropriate sexual arousal Conditioned taste aversions Latent inhibition holds New foods are more likely to develop in to taste aversions Chemotherapy

More applications Drug Overdosing Compensatory conditioned response You develop CR to the environment and method of ingestion Here’s a case where the CR is different from the UR US for taking marijuana is tachycardia (fast heart rate) The CS is a friends house. You go to the house, your heart rate DROPS (CR = bradycardia) You then take the drug This is a partial explanation for tolerance Effects of a different environment can produce overdosing.

Your experiment In groups of two You will classically condition each person to blink their eyes to a CS (of your choosing) Need CS (preferably something that produces a stable, consistent, repeatable sound) Straw Table with books Procedure Use a trace conditioning procedure Measure eye blinks (count them) Acquisition (until the behavior stabilizes) Extinction (until the behavior stabilizes at zero) Test for spontaneous recovery Write up 3 total references – two from journal articles (no magazines) – apa 5 Introduction Method Results – include charts

Operant conditioning Eliciting vs. Emitting In classical conditioning – responses are elicited from a stimulus Think reflex In operant – the organism emits a response that is then acted upon by the environment Called operant because the behavior operates on the environment and a consequence is produced Also called instrumental conditioning

E. L. Thorndike Wanted to study behavior systematically Worked frequently with cats Hungry cat in a puzzle box Could escape by pulling a string or pressing a lever Response latency decreased, showing learning Developed the law of effect Behaviors are stamped in (led to a satisfying state of affairs) or stamped out (unsatisfying or annoying state of affairs)

Law of Effect Focuses on consequences of behavior Requires an inference to the state of the organism Satisfying Unsatisfying or annoying Skinner removed the inference and changed the type of responding that was studied (from a discrete trial to a free operant response)

Operant conditioning Behavior seems voluntary in nature Behavior is controlled by consequences Not by the stimuli that elicit them Future probability of a given behavior (or class) is affected by its consequences Used probability to get rid of the inference to an organisms state Three components A response that produces a consequence The consequence A discriminative stimulus – signals that a certain consequence is now available

Operants and behavior classes Prediction of a specific response is very difficult – so skinner talked about ‘classes’ of behavior. It’s easier to predict a class of responses than a specific one Behavior tends to be of a ‘class’ of behaviors that produce the same outcome Thorndike’s cats could have pressed the lever, sat on the lever, leaned against it, etc. Each of these behaviors produce the same outcome hence they are called a class of behavior

Reinforcement Any stimulus added or removed contingent upon a behavior that maintains or increases the strength of the behavior The reinforcement matrix Probability of behavior increases Stimulus is Added Positive reinforcement Removed Negative reinforcement

Reinforcement cont. Strength can be defined as: Rate, frequency, probability, resistance to extinction Increasing strength means that it has a measurable effect on behavior But we have to have a behavior first! Dead person’s test

Positive Reinforcement Contingently adding a stimulus following a behavior that increases the probability of that behavior occurring in the future Strengthening a behavior by adding a stimulus Score on a test reinforces studying Dog greets you is reinforced by petting The positive does not mean good – simply means added

Negative Reinforcement Contingently removing a stimulus following a behavior that increases the probability of that behavior occurring in the future Strengthens a behavior by removing a stimulus Cutting up money – increase weight loss behavior Remove headache – increase aspirin taking Negative does not mean bad!!!! Often considered escape or avoidance Usually these stimuli that you escape or remove are aversive – but that is an inference

Contingency is the key The stimuli must be added or removed CONTINGENT upon a behavior Even b will occur if and only if event A occurs.

Types of reinforcers Primary Not dependent on their association with other S r Food, water, sex, drugs Not money! Secondary Pair a stimulus with one of the above it becomes an S r Buzzer paired with water

Complex Behavior through Reinforcement Shaping Successive approximations of desired behavior Shaping lever pressing in rats Secondary sr = sound of pellet dispenser Quick, immediate Use your clicker! Chaining – longer chains can break easier than shorter ones Task analysis Forward chaining Backward chaining You can ‘break’ the chain at any link and it won’t continue

Extinction of S r responses Similar to classical Behavior followed by no consequence Produces an extinction burst You also get spontaneous recovery Resurgence As one response is extinguished – other previously reinforced responses tend to appear Asking nicely is getting extinguished. The child may then scream (Something they learned as a younger child)

Variables effecting reinforcement Contingency Like classical conditioning – as you vary the probability of reinforcement for a given behavior then the behavior varies In other words – you’re developing a sort of ‘rule’: if this, then that. Contiguity The more immediate you can reinforce the better Prevents other behavior from being reinforced

Schedules of Reinforcement Ferster and Skinner – 1957 A pattern of SR based on numbers of responses, time and or other factors Different schedules produce different patters of responding – extremely reliable even across species

Continuous Reinforcement CRF Continuous reinforcement Great for training Not so great for the real world Extinguishes fast Basically a soda machine

Intermittent Schedules - FR Fixed ratio Sr is delivered after a fixed number of responses The post reinforcer pause Tenure, piece work, exam taking? The pause increases as the ratio increases Rapid responding otherwise Ratio strain – organism takes breaks before the SR is delivered

Intermittent Schedules - VR ON average – a given number of responses will be followed by a SR VR10 = 1,2,3,4,5,6,10,19,40 Produces a strong rate of responding No post reinforcer pause Except on thin schedules, but the pauses are smaller than FR schedules Gambling Slot machines This is essentially the addiction Strong steady rate of behavior Sports A good day in golf is essentially random You don’t know if that next shot will be the best shot of your life!

Intermittent schedules – Fixed interval After a fixed amount of time has passed the next response will be reinforced Could produce very low rates of responding Perfect timing or a clock There is a post reinforcer pause Follwed by a slow acceleration of responding rate The fixed interval scallop Waiting for a bus Reading the paper for 5 or 10 min – look up once Next minute look up once, then twice and so on

Intermittent Schedules – Variable Interval Average amount of time passes, the next response is followed by an SR 1,2,3,4,5,6,10,19,40 Steady rate of responding No PRP Checking the mail

Other schedules DRA (alternative), DRI (incompatible), DR0 (zero), DRH (behavior a minimum number of times) Fixed duration – set time then get SR Studying Fixed time Reinforcer delivered after a set amount of time – regardless of behavior Similar to welfare

Complex schedules Multiple schedules Behavior is under the influence of two ore more simple schedules, each associated with a stimulus Red light = VI10 green light = FI5 Two schedules alternate on the same key Good for studying stimulus control Mixed schedules Same as above – but no SD. Chain schedule Sr is delivered on the completion of the last in a series of schedules Red disk (FR 10) changes after the 10 th response, to yellow disk (FI 15) after the response it changes again to green on a VR 20. The last peck is then reinforced. Concurrent schedules Two ore more schedules available at a given time Great for studying choice Cooperative schedules Two organisms working on different or same schedules – but SR for either is dependent on both Total of 100R between both birds – then both receive SR Both must make at least 10 before either gets SR though

Punishment Still operant conditioning Contingency is equally as important The punishment chart: Probability of behavior decreases Stimulus is Added Positive Punishment Removed Negative Punishment

Punishment cont. Contiguity is also important The longer the delay – the less effective the punisher Intensity is also important Use a stimulus that is only as strong as you need it to be

Positive punishment Contingently adding a stimulus following a behavior that decreases the probability of that behavior occurring in the future Again – there is no good and bad here You are adding a stimulus (spank, noise, etc) It is usually a stimulus the organism finds aversive

Negative Punishment Contingently removing a stimulus following a behavior that decreases the probability of that behavior occurring in the future Response cost Things taken away – grounding

Punishment - more Difference between SR and SP Skinner argues that they are not opposite Punishment is temporary and can be thought of as suppression Often the behavior will decreases in a long term sense though Organisms habituate to some punishers

Things affecting punishment It’s a contingency! As long as the punishment contingency is in place then behavior will be reduced Remove the contingency and behavior will likely return There is no behavioral vacuum Reinforcement and punishment A punished behavior may still be receiving reinforcement So punishment depends on reinforcement – in a sense Motivation to respond Vary levels of deprivation on a food SR schedule Punish for producing the same response that produces SR The more deprivation the less effective the punisher was Availability of alternative behaviors You need to teach correct responses as well! One key – pigeons press for SR – occasionally punish. Rate of responding decreases but not to zero Two keys – left produces SR and punishment – right produces SR Learn to respond on the right key Rate of responding dropped to zero on the left key

Punishment effects and problems We use punishment because it’s reinforcing to us. Positive punishment, negative punishment usually lead to a negatively reinforcing situation for the punisher. Can be useful because it’s effects are fast. Especially if we DRO/A/I and stop reinforcing the behavior we are trying to punish Injurious behavior Sometimes has positive side effects Mentally disabled folks that were punished for self injurious behavior also seemed happier and more outgoing after the treatment

Problems cont. Escape / avoidance The punished often try to avoid or escape the punishment This is a negatively reinforcing situation for them – escaping the punishment reinforces whatever they just did (to escape) Tuning out, lying Aggression Likely when escape is possible We attack the punisher Apathy If aversive stimuli are common then you can get a suppression of wide ranges of behavior Use SR to counteract this. Abuse Easy to abuse Starting with weak punishers and escalating Extinction bursts – reinforced leads to outlandish behavior – then punished severely. Imitation You learn to punish by being punished (observational learning)

Ethics and alternatives Is it OK to punish? Not a question for science Response prevention Works wonders with cats DRL DR0 – period of time without providing the problem response is reinforced – then change criteria DRI

Self Control Clear definition Behavior, not outcome A contract – signed and posted Public commitment Tracking behavior – baseline Stability A change technique in itself Graph it Modifying your SDs Adding Removing Delay from SD to behavior

Self control - cont Modifying your SR Identify them first (Very difficult) Make a list of possible sr Add or remove Be ready for extinction bursts Do not use something that you do or obtain often Delay from R to SR Thoughts – these serve as SDs for placeholders for SRs – very useful Using friends Contract – make them stick to it The self management problem

Educational Applications Behavior Analysis in the classroom Behavior analysis is the field in which experts in human learning apply their trade. Working in the schools is a common occurrence Misbehaving students are dealt with by reprimands and punishment but appropriate behavior is ignored Reverse this – focus on the good behavior – ignore the bad. The result was that student behavior changed significantly Instructor behavior changed too – she was reinforced for reinforcing. No child left behind act Punitive vs reinforcing

More on education Programmed instruction Teaching Machines Personalized system of instruction Used for university course Implications State your objectives Test over small amounts Provide Sr whenever possible Establish stimulus control

Clicker Training Karen Pryor and the clicker The logic of the clicker Conditioned reinforcer Immediate and salient “ Don’t shoot the Dog!” Title of her book – but it has a good point We often think the dog or the child is at ‘fault’ but we must look at the conditions they are operating under.

Learned Helplessness Dogs are restrained and given a tone-shock They are later moved to a shuttle box (two sides of a large chamber) and given a shock – naïve animals move to the side that does not produce shock. Even at the presentation of the tone. The dogs that were restrained – remained in the shock part of the box – laying down to accept shock – even with coaxing (including food) they could not get the dogs to escape. Learned industriousness Rats trained that a lever press will escape shock Other rats given inescapable shock Then exposed to a shuttle box with shock on BOTH sides. The rats that learned they could escape continued to move to the other side of the box – even though they still got shocked. The helpless ones just stayed put. Reinforce a high level of effort and persistence increases probability that they will work hard at difficult tasks for a prolonged period.

Gambling What is compulsive gambling? Is it a person issue? (“there must be something wrong with them”) How can you gamble sometimes and not get hooked while others cant? Pigeons gamble too 3s access to food for 50 key pecks (work) Then set up a gambling situation with a key that acts like a slot machine with the payout being longer access to food – first three days set up to be a lucky bird – with 15s access to food. After that they are better off working to get the right amount of food for the day. Pigeons became hooked and continued to gamble Removed the gambling component - ethics Reinstated it and the birds went back to gambling

Superstition Noncontingent reinforcement can reinforce behavior that has no connection to the reinforcer Superstition Maintained by either SR+ or SR-

Observational Learning Change in behavior due to the experience of observing a model Observation and classical conditioning Confederate and observer Confederate receives shock after a buzzer sounds (acting) – observer watches Observer later showed conditioning to the buzzer Eyelid conditioning Observer watched conditioning trials – they eventually produces a CR to the CS

Observational learning cont. Alternate explanation Maybe watching a blink is a CS for blinking Tone paired with the sight – now produces the blink Might be higher order conditioning Monkeys reared in captivity with wild parents Parents show fear of snakes – subjects do not Subjects will reach across a snake for food in the same compartment Then have them watch a parent in the presence of a snake The subjects then show fear of snakes.

Observational Learning - operant Monkey learned to do a task to receive SR (pull a chain) An observer watched the model perform the task 5 times Observer get’s a chance to perform task Given six trials Observers benefitted Many responded correctly on 1 st trial 75% of solutions 30s 47% in 10s Failures were close to correct Did the correct action just with not enough force

Observational Learning - Operant Cats solving a problem Spin a disk to receive SR 4 Watched while one worked Two watched for the full 30 trials Two watched for the last 15 trials Observers OUTPERFORMED the models Models – 62s to solve trial 1 Observers 15 trials – 57s to solve on trial 1 Observers 30 trials – 16s to solve on trial 1

Imitation and generalized imitation When an observer duplicates the behavior of a model Imitation of a behavior may not lead to the consequence (contingency) Generalized imitation (getting SR for imitating in general) Baer and Sherman used a puppet to provide SR for imitative behavior Puppet modeled 4 behaviors (mouthing, head nodding, speaking nonsense and lever pressing) First three when imitated resulted in positive feedback from the puppet Imitation of the lever press was never SRed All imitating increased Reinforced the general tendency to imitate

Variables effecting observational learning Consequences of a model’s behavior Consistency is important Consequences of the observers behavior Learning history Have you been reinforced for observing/imitating in the past Model characteristics influence learning Attractive Competent – fellow student vs. lab assistant – learned more from the assistant Likable Prestigious

other characteristics Attention getting The eye patch – those that saw a model with an eye patch were more likely to remember The mood – those that saw a model display a mood (good or bad) were more likely to remember what the model said – compared to neutral Age Young ones don’t necessarily learn more than older Recalling what you saw (novel performances of a model)– adults were much better Survival Polar bears and the Inuit – dangle seals So observing (and imitating) can increase survivability

Generalization and stimulus control Watson and SR psychology Too simple – behavior is much more complex Not misguided – antecedent stimuli play a role We need to understand how a creature will respond to a given stimulus, similar stimuli and new stimuli Generalization gradient Train on a red light – VI 1 Now test to see if a blue light will work This produces extinction (problem) Reinforcement problem So we do probe trials (single extinction trial) mixed in the session (90% red, 10% blue)

Generalization Response generalization To a given stimulus – response may be slightly different each time Stimulus generalization To different stimuli you will get similar responses This follows a generalization gradient Semantic generalization CSs were words (presented while they were having sweet snacks) – salivation was the CR Synonyms and homophones were the test Synonyms produced stronger responding than homophones After 9/11 there were many crimes of those against Arab descent or the Muslim faith.

More generalization Training effects generalization Two sets of pigeons One group trained to respond to a lit key with a 1000mhz tone on a VI schedule One group received presence and absence training Respond to the presence of a given stimulus and not to others (1000mhz) Some trials were identical Other trials included no tone but a lit key with no SR available Extinction trials with different freq of tone present to test for generalization The birds in the latter group developed generalization to the sound Initial birds responded to all tones

Discrimination Behavior occurs in the presence of a given stimulus but not in the presence of others Light on = S+ light off = S- Inverse relation between generalization and discrimination Detecting discrimination in the generalization gradient Flat means no discrimination Steep means high levels of discrimination

Pavlovian discrimination CS+ is paired with the US CS- appears without the US present Dogs learned to respond differently to a 100bpm and 96bpm metronome That’s 4 beats a minute difference!!!

Operant discrimination training S+ and S- are types of SDs Examples Red light (S+) Blue light (S-) Bach and Stravinsky (birds learned to detect the difference between the two) Picasso and Monet – birds learned to discriminate – and to pics by the same artists that they had never seen before! Could even discriminate between other painters in the same schools.

Attending to Specific Properties S+ = triangle on red key S- = circle on green Bird 107 is attending to the red and bird 105 is attending to the triangle

Fading procedures Start with one stimuli Change properties of that stimulus until the desired discrimination is being made Pigeons have a hard time discriminating between horizontal and vertical lines but red and green are easy Start with red and green Superimpose vertical on red and horizontal on green and slowly remove the color Now the bird can discriminate

Errorless discrimination Shape a pigeon to respond on a red key for food 3 min periods Alternated with dark key as S- (only 5 seconds) Duration was gradually lengthened Then from dark to dim to brighter and so on until the green key is as bright as the red At end red is S+ Green produces extinction Only 10 pecks on the green key for ALL birds If we just trained the discrimination without fading then birds would have made 100s of responses

Matching to sample Select from two or more alternatives (comparison stimuli) the stimulus that matches the standard (sample) Comparison stimuli One is S+ one or more is S- Illuminate light (red / green) then go dark (need an observing response) Two or more selection keys – all lit at the same time, one red, one green S+ is the key that matched the sample Can teach them to select the opposite Colors and positions change from trial to trial Correction procedures – no switching if an error was made This gets rid of ‘side preferences’

Equivalence relations Reflexivity – A=A Symmetry – A=B then B=A Transitivity – A=B and B=C then A=C

Stimulus control Discrimination training brings behavior under the influence of SDs We learn when to engage in a given behavior We learn when a given behavior is unlikely to receive SR The SD in the 3term contingency is there because of this effect.

Applications Concept formation We don’t have ‘concepts’ per say any more than we have extinction You demonstrate conceptual behavior You must generalize within classes You must discriminate between classes Pigeons and trees Peck a disk when a picture is displaying a tree They then demonstrated the concept of tree Pigeons and humans Peck a disk when a picture is displaying a human Alone, groups, diff sizes, shapes, ages, sexes, clothing or no clothing, full view, partially hidden, etc. Another where they pecked where humans were in the picture Got this overwhelmingly right – even without SR for that part of the task

Forgetting and Remembering The metaphor Is a reflection of our times Wax tablet, blank slate, filing cabinet, computer, etc All of these posit a place called memory (where memories are stored) Where is the memory itself? It doesn’t seem to be in one location The hippocampus is involved in forming long term memories – but they are not ‘stored’ there. The changed organism The physiological process of the neurons firing is what we call memory – but its not a thing that is moved around, etc Instead of positing a structure – suggest change.

Forgetting in learning To study memory in learning we are really studying “the study of memory is properly the study of the factors that affect our performance … after a period of no practice.” Malone Durability of learned behavior Deterioration in learned behavior following a period without practice. Do you remember = can you describe Have you forgotten = can you still do

Measuring learning Retention interval Period of time is allowed to pass without the performance of said behavior – then test to see if it persists This gap is the retention interval Free recall Perform previously learned behavior after a retention interval

Measuring forgetting Prompted recall Much like wheel of fortune Add cues until the item is remembered Why is this better than free recall? Captures the generalization gradient Relearning Reinstate original procedure Measure how fast the response is acquired compared to trail 1 Savings is the measure ebbinghaus Recognition Identify previously learned material Delayed match to sample Present sample – delay time to response. You are measuring remembering.

Measuring forgetting Extinction method Acquisition phase – then a delay to extinction phase Faster extinction should occur in the delay conditions Gradient degradation Flattening the generalization gradient Known phenomenon comparison Fi schedule 24hr break, 24 day break – rats in the later condition didn’t show scallop, at first

Time as the variable Many of the above seem to be accounted for by time But time is a hypothetical construct We are really measuring change So the reasons for forgetting is not simply time – but some change in the organism

sleep Sleep helps sometimes! In reality, it just doesn’t work against you. Lists of nonsense syllables Followed by varying lengths of sleep Compare groups – other activities The sleep groups did better Done with inactivity as well Immobilized cockroaches

More on remembering Overlearning The more you learn something the less likely you are to forget it. Prior experience Chess masters Chess masters compared to a chess club Arranged pieces on the board – study for 5 seconds 90% accurate for the masters – 40% for the chess players for reproductions Now arrange randomly No better than ordinary players Rules out the superior memory argument

Stimulus control and remembering Cue dependent forgetting Cue present at learning is not present at test time Rats learning to run a white or black maze One minute retention interval Half of each condition was ran in the other Performance decreased for the switch group Humans Learned lists of words standing or sitting Recall was better in the condition you learned in State dependent learning Organisms state serves as a stimulus Hunger, hormones, drunk, etc

APA 5 Cover page Running head; header; page numbers Centered Introduction General statement about the issue Leads to your study Method A detailed description of the procedure you used How many trials, what conditions, etc. Results Your findings – include a graph. Graph should have 3 parts. Summarize this graph in text – in terms of the three parts. Discussion A synopsis of your study in terms of the findings and other literature. Did you expect to get what you got? What problems did you have? Why might have you obtained the results you did? References Follow APA 5 formatting rules Builder, B. (2008). How to build a house. Construction Journal for Little Kids, 123(4) , 176-188 Builder, B. (2008). Building Bob Style. London: Norton.

092408 324 Lecture

More Related Content

What's hot (20)

Viewers also liked (7)

Similar to 092408 324 Lecture (20)

More from Ryan Sain (20)

092408 324 Lecture