Transcript Learning
Learning Learning is a change in behavior as a result of past experience Even though our principal interest is in how humans learn, you’ll notice that a great deal of learning research is conducted using animal subjects instead Electrical Activity of the Neuron • Less Complexity •Greater Control Over Variables •Generality of Principles Classical Conditioning (Pavlovian Conditioning) Ivan Pavlov (1849 – 1936) Famous Russian Physiologist who won a Nobel Prize in 1904 for his work on the biochemical analysis of digestive secretions of the stomach Electrical of the Neuron HisActivity later notoriety stems from the widely-known studies he performed on conditioning of the salivary response But his actual discovery of conditioned reflexes occurred in a different context much earlier Classical Conditioning (Pavlovian Conditioning) Procedure : 1. Dog transported from colony room to test cubicle and chained to a hitch post 2. Food was prepared on a table 10’ away from subject 3. Dog was fed and 10 minute waiting period allowed Electrical Activity of the Neuron 4. Fistula clamp was opened to allow for collection of digestive secretions which occurred in response to the arrival of food in the stomach A mistake in the procedure leads to a discovery! The fistula clamp was not closed properly allowing for leakage. Secretions were observed taking place prior to feeding. Pavlov termed these “Psychic Secretions” Classical Conditioning (Pavlovian Conditioning) Electrical Activity of the Neuron To more conveniently study “psychic secretions”, a new procedure was devised involving the salivary reflex The Classical Conditioning Paradigm Stage 1 (Before Conditioning) UCS UCR (Food) (Salivation) US (UCS) - Unconditioned Stimulus A stimulus which reliably and without an prior experience reflexively causes a response to occur (the Unconditioned Response – UCR or UR) Some of the Most Frequently Studied Classically Conditioned Responses Autonomic Activity dry food or acid in mouth salivation electric shock electrodermal activity (GSR) electric shock or loud noise cardiac acceleration electric shock or loud noise sudden respiratory inspiration change in illumination electric shock pupillary reflex food in viscera gastro-intestinal secretion insulin injection hypoglycemia Some of the Most Frequently Studied Classically Conditioned Responses Skeletal Activity nipple on cheek or lips rooting, sucking loud noise airpuff to cornea suborbital electric shock blow to the eye eyeblink or nictitating membrane response food in the mouth mastication electric shock heat pin prick flexion, limb withdrawal Some of the Most Frequently Studied Classically Conditioned Responses Skeletal Activity nipple on cheek or lips rooting, sucking loud noise airpuff to cornea suborbital electric shock blow to the eye eyeblink or nictitating membrane response food in the mouth mastication electric shock heat pin prick flexion, limb withdrawal Brain Activity illumination EEG occipital alpha rhythm The Classical Conditioning Paradigm Stage 1 (Before Conditioning) UCS UCR (Food) (Salivation) CS (Bell) A neutral stimulus – that is, one that does not initially cause the UCR to occur. This stimulus might be referred to as the to-be-conditioned stimulus (CS) The Classical Conditioning Paradigm Stage 2 (During Conditioning) CS UCS UCR (Bell) (Food) (Salivation) During conditioning we repeatedly pair, close together in time, the originally neutral stimulus with the UCS Temporal Contiguity is the key facet of combining the stimuli for conditioning to in fact occur The Classical Conditioning Paradigm Stage 3 (After Conditioning) CS (Bell) CR (Food) (Salivation) After conditioning we present the originally neutral stimulus by itself and observe a conditioned response (CR) We refer to the process that has occurred as “Acquisition” since a new response is being acquired to the originally neutral stimulus The CR and the UCR are not identical responses. The CR is always a fractional component of the UCR. The Fractional Anticipatory Nature of Classical Conditioning Classical Conditioning represents a learned connection between environmental events that are likely to happen together. We become more effective in our response to events if we prepare for them before they occur. A dog can more efficiently eat and swallow food if he salivates in anticipation of feeding. Example of driving away in your car after class is over. Preparing your keys in advance is anticipatory of a situation that you are soon to encounter. The response is fractional as you can not insert and turn the key until the lock is encountered. Organisms that prepare for events before they occur are likely to be more effective in their response. Stages of Learning Stimulus Generalization Stimulus Generalization Decrement The degree of response to test stimulus is gradually reduced as stimulus differs more and more from the original Discrimination Discrimination is said to occur when the subject responds differently in the presence of one stimulus versus another. Conditioned Emotional Response Originally neutral stimuli can take on aversive characteristics as a result of being paired with primary aversive stimuli. Watson and Raynor (1920) explored the development of a conditioned emotional response in the infant “Little Albert” Instrumental Conditioning This form of learning is called “Instrumental” because the subject desires his environment to change in some way. But in order for this to occur, his own behavior will be instrumental in bringing about the desired change. For example, I’d like my environment to include a hot steaming pizza in front of me right now. But for this to become true, I must act in a certain way. My behavior will be instrumental in causing this to occur. An alternate terminology for this for of learning is to call it “Operant Conditioning” because the behavior operates upon the environment, changing it in some way. Instrumental Conditioning Distinguished From Classical Conditioning Classical conditioning involves an association between stimuli. The subject comes to learn that one stimulus is contingent (dependent) upon the occurrence of another. E.G., Pavlov’s Dogs learned that food delivery was contingent upon the ringing of the bell Instrumental conditioning involves an association between a response and a stimulus. The subject comes to learn that a stimulus is contingent (dependent) upon the emission of a response. E.G., A dog learns that a doggy treat is contingent upon posturing into the “begging” position Instrumental Conditioning Instrumental Conditioning was first defined and studied by E. L. Thorndike The Law of Effect (Thorndike) The likelihood and strength of a behavior are determined by the consequences it produces. If a behavior is followed by a positive state of affairs, the response becomes stronger (reward training). The term “Reinforcement” is used to refer to any stimulus which has the capability to strengthen behavior If a behavior is followed by a negative state of affairs, the response becomes weaker (punishment). Such negative stimuli are referred to as being “Aversive”. Administration of negative and positive stimuli creates four distinct Instrumental Conditioning Training Paradigms. The Asymmetrical Law of Effect The general belief amongst psychologists that reward and punishment differ significantly in their effectiveness in modifying behavior Punishment suppresses behavior but its effects are diffuse (Suppression affects all behaviors, not just the punished behavior). Punishment produces disruptive negative emotionality. Punished behaviors show a large spontaneous recovery. Instrumental Conditioning Paradigms Nature of the Stimulus Target Response Causes Application of the Stimulus Target Response Prevents or Causes Removal of the Stimulus Positive Negative Positive Reinforcement Punishment (Appetitive Conditioning) (Aversive Conditioning) Negative Reinforcement Omission (Escape or Avoidance) The Cumulative Response Record Variables Influencing Instrumental Conditioning Amount of Reinforced Practice Learning Curve Negatively Accelerated Asymptote Variables Influencing Instrumental Conditioning Amount of Reinforced Practice Amount of Reinforcement Large Magnitude of Reward Small Magnitude of Reward Variables Influencing Instrumental Conditioning Amount of Reinforced Practice Amount of Reinforcement Quality of Reward Delay of Reward Reinforcement Schedule Schedules of Reinforcement CRF (Continuous Reinforcement) – Each response is followed by reward. p (reward) = 1.0 EXT (Extinction) – Each response followed by no reward p (reward) = 0.0 PR (Partial Reward) – Responses are followed by reward with a probability greater than 0 but less than 1. 50% PR schedule would mean subjects would be reinforced on a random ½ of their trials The Partial Reinforcement Effect (PRE) CRF PR