Classical Conditioning F24 PDF

Of dogs Learning and bells (classical & Memory conditioning) Of Dogs and Bells (Classical Conditioning) Associative Learning The process by which events / sensations / ideas / behaviors that are linked in experience get linked together in memory Learning to associate one...

Of dogs Learning and bells (classical & Memory conditioning) Of Dogs and Bells (Classical Conditioning) Associative Learning The process by which events / sensations / ideas / behaviors that are linked in experience get linked together in memory Learning to associate one stimulus with another stimulus or a particular behavior/response  British Associationist Theory 1. Contiguity Events experienced at the same time (temporal contiguity) or in the same place (spatial contiguity) tend to be associated 2. Frequency The more often we experience these contiguous events, the stronger the association between them will become 3. Intensity The more intense a feeling the stimulus produces, the stronger the association that will be learned  Plan for today 1. Pavlov’s drooling dogs 2. Conditioning procedures 3. A matter of timing 4. Surprise! How conditioning works 5. Not all associations are equal 6. Real world applications Ivan Pavlov (1849-1936) Began his career studying physiological processes Particularly interested in understanding reflexive behavior Specific application to digestion (in dogs) A curious observation… Classical (“Pavlovian”) Conditioning When a neutral stimulus produces an automatic response after being (repeatedly) paired with a stimulus that naturally produces that response Terminology is important... Unconditioned Stimulus (US) Something that reliably produces a naturally occurring reaction in an organism (food) Unconditioned Response (UR) A reflexive reaction that is reliably produced by an unconditioned stimulus (salivation) Conditioned Stimulus (CS) A stimulus that is initially neutral and produces no reliable response in an organism (a tone) Conditioned Response (CR) A reaction that resembles an unconditioned response but is produced by a conditioned stimulus (salivation)  Classical conditioning Unconditioned Unconditioned Stimulus Response US UR  Classical conditioning US UR Food Salivation  Classical conditioning US UR Any recognizable stimulus in CS the environment Conditioned stimulus  Classical conditioning US UR Repeated presentations of US + CS CS mean that the CS reliably signals the UR  Classical conditioning US UR CS  Classical conditioning US UR CS  Classical conditioning US UR CS  Classical conditioning US UR CS  Classical conditioning US UR/CR CS Strength of Conditioned Strength of CR Response Classical conditioning  Classical conditioning The co-occurrence of the CS and the US lead to them becoming associated This means that the CS, which originally produced no response, can now inherit the response produced by the US Note: Classical Conditioning does not depend on any voluntary action of the organism. The UR/CR is automatic and unavoidable  Plan for today 1. Pavlov’s drooling dogs 2. Conditioning procedures 3. A matter of timing 4. Surprise! How conditioning works 5. Not all associations are equal 6. Real world applications A. Types of Conditioning Appetitive Conditioning: US is pleasant (food, sex, drink) Aversive Conditioning: US is unpleasant (shock, loud noise, air puff) Eyeblink Conditioning B. Measuring conditioning  Measurement of conditioning Conditioning is usually measured in 1of 2 ways: – Magnitude of response (how much drool?) – Probability of response (did they blink?) Test trials are used in which only the CS is presented, and the rate of responding is measured C. Extinguishing an old association How do we get the drool to stop? Extinction After repeatedly presenting the CS alone, the CR decreases and eventually disappears. When CS is presented alone (no US), it no longer predicts the CR.  Strength of CR 32 Is extinction just forgetting? Acquisition and Extinction of Eyeblink Conditioning NO! Extinction happens much faster than normal forgetting; often just a few trials of the CS alone to get rid of the CR Is extinction “unlearning”?  1. Reconditioning occurs more rapidly for the specific CS-US pair NO! 2. CR may reappear if you wait a while after a period of extinction (spontaneous recovery) 3. The CR may reoccur in a new environment (renewal) Strength of CR So what’s going on? Pavlov’s hypothesis: extinction leads to the creation of additional inhibitory connections between CS and UR During acquisition (CS + US) excitatory connection develops between CS and UR because CS predicts US Pavlov’s hypothesis: extinction leads to the creation of additional inhibitory connections between CS and UR During extinction (CS alone), new inhibitory connection develops between CS and UR because now CS predicts the absence of the US Pavlov’s hypothesis: extinction leads to the creation of additional inhibitory connections between CS and UR Thus, now there are 2 different pathways that have been created: 1. CS → US → drool 2. CS → nothing → inhibit drool  Evidence: Conditioned inhibition Acquisition Phase + Test Phase + D. Higher-order conditioning a CS can become a US! First-order conditioning Second-order conditioning + E. Generalization & Discrimination Generalization Generalization Conditioning with one stimulus tends to spread to similar ones (e.g., tones of a similar frequency) Discrimination Training Discrimination Conditioning will NOT spread (discrimination) when training includes similar stimuli that do not predict the US  Plan for today 1. Pavlov’s drooling dogs 2. Conditioning procedures 3. A matter of timing 4. Surprise! How conditioning works 5. Not all associations are equal 6. Real world applications  Delay Conditioning Inter-stimulus interval (ISI) Inter-trial interval (ITI) CS US Time  Short delay CS conditioning US Long delay CS conditioning US Trace Conditioning CS US Simultaneous CS Conditioning US Backwards Conditioning CS US Strength of Conditioned Response An ISI of about 450 ms. (just under half a second) is typically optimal Inter-stimulus interval (ISI) in milliseconds  Classical conditioning interim summary 1. US can be pleasant (appetitive), like food, or unpleasant (aversive), like an air puff to the eye 2. Can be measured based on the magnitude or probability of CR 3. CR can be extinguished by presenting the CS in isolation (extinction) 4. Extinction works through the development of a new inhibitory connection between CS and UR 5. A highly learned CS can become a US for a new stimulus (higher-order conditioning) 6. The CR can spread to stimuli that are similar to the CS (generalization), but adding similar stimuli that are not paired with the US during acquisition will reduce this (discrimination)  Plan for today 1. Pavlov’s drooling dogs 2. Conditioning procedures 3. A matter of timing 4. Surprise! How conditioning works 5. Not all associations are equal 6. Real world applications British Associationist Theory 1. Contiguity Events experienced at the same time (temporal contiguity) or in the same place (spatial contiguity) tend to be associated 2. Frequency The more often we experience these contiguous events, the stronger the association between them will become 3. Intensity The more intense a feeling the stimulus produces, the stronger the association that will be learned Is contiguity sufficient for conditioning?  Trial # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Tone (CS) CS + US = 4 CS ONLY = 4 US ONLY = 0 Neither =8 Shock(US) Rescorla & Wagner (1972)  Trial # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Tone (CS) CS + US = 4 CS ONLY = 4 US ONLY = 4 Neither = 4 Shock(US) Rescorla & Wagner (1972)  Which condition leads to classical conditioning? Trial # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 CS + US = 4 Tone (CS) CS ONLY = 4 US ONLY = 0 Shock(US) Neither = 8 CS + US = 4 Tone (CS) CS ONLY = 4 US ONLY =4 Shock (US) Neither = 4 Rescorla & Wagner (1972) There’s more to learning than just contiguity (co-occurrence) alone: It’s about expectation or prediction! Thus contingency is critical. Is contingency sufficient for conditioning?  Blocking Phase 1 Phase 2 Test Experimental A US A+B US B A+B US B Control Kamin, 1969  Blocking 90% 80% % of time responding 70% Strength of CR: 60% 50% 40% 30% 20% 10% 0% B Control B Experimental A +B B Experimental A US A +B US B US Control Kamin, 1969  Blocking Conditioning only occurs if CS is both a useful (informative) and non-redundant predictor of the US Challenges the idea that associations are learned independently, and that contingency is sufficient for conditioning! Associations learned depend on previously learned associations So the secret of learning is…  Key Ideas 1. If there is no surprise (violation of expectation), then learning will not occur The CS has to provide new predictive information 2. So, connections between CS and UR are modified due to prediction error: the difference between what the organism expects/predicts will happen and what actually does happen 3. Error-driven learning of this type is a general and powerful way the brain can learn and modify connections  Plan for today 1. Pavlov’s drooling dogs 2. Conditioning procedures 3. A matter of timing 4. Surprise! How conditioning works 5. Not all associations are equal 6. Real world applications British Associationist Theory 1. Contiguity Events experienced at the same time (temporal contiguity) or in the same place (spatial contiguity) tend to be associated 2. Frequency The more often we experience these contiguous events, the stronger the association between them will become 3. Intensity The more intense a feeling the stimulus produces, the stronger the association that will be learned  Principles of Stimuli & Responses 1. Intensity of the US matters (for example, shock conditioning is faster than eyeblink conditioning) 2. Intensity of the CS matters (salient cues are learned faster) 3. Past experiences can modify the intensity of the CS  Latent Inhibition Phase 1 Phase 2 Test A Experimental A US B A US B Control Shohamy et al., 2000  Latent Inhibition A US Control % of time responding Strength of CR: A A US Experimental 74  Principles of Stimuli & Responses 1. Intensity of the US matters (for example, shock conditioning is faster than eyeblink conditioning) 2. Intensity of the CS matters (salient cues are learned faster) 3. Past experiences can modify the intensity of the CS 4. Species-specific differences (humans and pigeons respond more to visual stimuli than auditory ones; less so for rats) 5. CS-US-CR compatibility: Different USs are more easily conditioned to (associated with) certain CSs  Experience two possible CSs Rats drink saccharine-flavored water. Whenever they make contact with the tube, they activate a bright light and noisy buzzer.  Then only one of two possible US-UR pairs: Some rats are shocked Some rats are exposed (physical pain) to X-rays (inducing nausea) Which association(s) will be learned? Test: Rats given a choice between: – Saccharine-flavored water (no light/buzzer) – Non-flavored water attached to the light and buzzer. DV: Which will the rat avoid? Which association(s) will be learned? Saccharin- Light and flavor Buzzer Shock- pain X-ray- nausea Which association(s) will be learned? Saccharin- Light and flavor Buzzer Shock- pain X-ray- nausea  CS US Increased human Skin Conductance Response (SCR). CR: Moisture (sweat) on the skin controlled by the sympathetic nervous system Fear responses are learned faster, harder to extinguish when shock is paired with ancestral dangers than ancestrally innocuous things (e.g., spiders vs. butterflies)  Plan for today 1. Pavlov’s drooling dogs 2. Conditioning procedures 3. A matter of timing 4. Surprise! How conditioning works 5. Not all associations are equal 6. Real world applications  Applications Advertising Understanding drug abuse – Environmental CSs (e.g., appearance and smell of drug- taking location) produce CR drug craving in anticipation of US drug: expectation can elicit a preparatory (or compensatory) response – Environments can trigger relapse Explaining and treating fear and anxiety John B. Watson (1878-1958) Before conditioning: No fear of rat Conditioning trial: + = CS: Rat US: Noise UR: Fear Before conditioning: No fear of rat After conditioning: CS: Rat CR: Fear Generalization of Fear Phobias  Phobias Phobia: overwhelming, irrational fear of an object, place, or situation. Best explained as arising through a classical conditioning procedure.  Phobias Stimulus in a phobia often something that was threatening to human ancestors (e.g. snakes, heights, closed in spaces) This suggests we have a biological predisposition to fear certain things Not everyone who has a fear-evoking experience develops a phobia Can we use Classical Conditioning to eliminate phobias?  Exposure Therapy – Exposing the patient to the feared stimulus in a safe environment. Like normal extinction trials. Counterconditioning – Conditioning the patient to have a pleasant feeling and reaction toward something that they once feared or disliked. Accomplished by pairing the feared stimulus with something good (appetitive) so that it predicts good things Systematic Desensitization: – Present successive approximations of CS while patient stays relaxed. Eventually CS no longer elicits reaction Initial state FEAR Desired state CALM How to get there? CALM How to get there? CALM After lots of practice… CALM  Take Home Points 1. Classical conditioning is a simple, yet incredibly powerful learning capacity that is highly conserved across species 2. Critically, classical conditioning allows us to predict and prepare for future events by building expectations for what will happen 3. Raw co-occurrence of CS and US (contiguity) is not sufficient: CS must provide non-redundant and predictive information 4. So, we only learn where there is some prediction error (a violation of our expectations). Error-driven learning is a general process that may occur throughout the brain 5. We can harness classical conditioning to influence others and to treat psychological problems