Judith A. Dygdon, PhD PSYC 580 Part IV 1 Introduction to Classical Conditioning Here we begin our study of the learning of elicited behavior. Remember that we have defined an elicited behavior as a behavior that occurs in response to an antecedent stimulus. We often use the model S ---- R to capture elicited behavior. We have talked a bit about natural, or unlearned or untrained, elicited behavior. Now we will begin to discuss how new elicited behaviors develop, and existing ones change, through an organism’s experience with the environment. Keep in mind that we will assume that principles of learning apply across species, unless we have reason to believe that that is not the case. For this reason, I use the term “organism” instead of “person” when I am speaking generally about a subject of learning. Understand “organism” to mean and rat, pigeon, dog, person, etc. Classical Conditioning is the name we give to the model we use to organize and describe the ways in which environmental experiences influence elicited behavior. Classical conditioning is also known, synonymously as, Respondent Conditioning and Pavlovian Conditioning. Classical conditioning captures the process through which a neutral stimulus comes to elicit a response. Classical conditioning captures the process through which a neutral stimulus (in other words, a stimulus that is meaningless to the learning organism) comes to take on meaning or value for the learning organism. A more elaborate definition of neutral stimulus than the one I offered above is a stimulus which elicits from the organism no response, or simply an orienting response, which quickly habituates. Classical conditioning was first clearly described in public writing by Ivan Pavlov (though we now know that others were studying the same phenomena at the same time) in Russia in the early 1900's. (His main publications on this material Judith A. Dygdon, PhD PSYC 580 Part IV 2 appeared in the 1920's.) Please note that it is appropriate to talk about classical conditioning being “discovered” but not “invented”. Classical conditioning is a natural process through which organisms are equipped to learn from their environments. Like so many other natural processes, once we understand it, we can harness its power and use it to do good things (unfortunately, bad things as well). BASIC MODEL ---In each “conditioning” trial: UCS ----- UCR | | | CS If I were to put words to this model, I would “read” it as follows. “In any given trial, a CS is presented with a UCS. A UCR naturally follows the UCS.” OK – time for definitions: Conditioning trial = one presentation of the chosen CS and UCS to the learning organism. UCS (some authors use US) = Unconditioned Stimulus or Unconditional Stimulus - A stimulus that elicits a response without training. A stimulus that naturally elicits a response. UCR (some authors use UR) = Unconditioned Response or Unconditional Response - The response that follows a UCS. Judith A. Dygdon, PhD PSYC 580 Part IV A response that follows a stimulus without training. In the preceding three sections you read, we talked about natural, elicited behaviors. This would be one of those. CS = Conditioned Stimulus or Conditional Stimulus Before training, a neutral stimulus. After training, it elicits a response it did not elicit before. I could add something to the diagram I presented above – After many presentations of: UCS ----- UCR ----- CR | | | CS This is observed: CS If I were to put words to this diagram, I would “read” it as follows. “After sufficient presentations of the CS and the UCS, a response (CR) is elicited by the CS.” One more definition: CR = Conditioned Response or Conditional Response - After training, a response elicited by the CS. A very interesting note on language. Note that I give you alternative first words 3 Judith A. Dygdon, PhD PSYC 580 Part IV 4 for each of the four terms in the classical conditioning model. For example, I call a CR either a conditioned response or a conditional response. What’s with the different endings? This is a problem that comes from sloppy translation. Recall my earlier discussion of Pavlov. He was the first to publish a description of classical conditioning. He was Russian and wrote in Russian, but his work was quickly translated into English for eager, English speaking psychologists. The terms he used were translated into the –ed ending forms and Conditioned Response, Unconditioned Stimulus, etc. quickly became established in the English psychological lexicon. Years later it was realized that, had Pavlov’s work been translated more carefully, the result would have been terms with –al endings. I don’t expect you to see this now, but, as you will see, the –al endings are more accurate. What this means is, today the –ed endings are used because they are well established, but many people would like to see a change to the –al endings because they’re more accurate. You will come to see why the –al endings are more accurate as we continue to talk about this material. For the time being, expect to see both. Let’s look at some examples of respondent conditioning. First, let me point out that, in this “lecture”, I will be using examples of excitatory classical conditioning. This is training that “tells” the organism “something important” is going to appear. There does exist inhibitory classical conditioning, which tells the organism that something important will not appear. We’ll introduce that label when we need it. Judith A. Dygdon, PhD PSYC 580 Part IV 5 Pavlov’s dogs and work with simple responses Who hasn’t heard of Pavlov and his dogs! The experiences he provided his dogs could be captured by the following diagram: Conditioning trials: UCS meat powder ----- UCR salivation ----- CR salivation | | | CS tone Test trial: CS tone Expressed in words, Pavlov identified a stimulus that naturally elicited a response from his laboratory dogs. That stimulus was meat powder and the dogs naturally salivated when it was placed in their mouths. Pavlov also identified a neutral stimulus for his dogs – a simple tone. The first few times he played the tone for his dogs, they oriented to it, but that response quickly habituated. Next, for each dog, Pavlov played the tone, then put meat powder in the dog’s mouth. Of course, after the meat powder went in, salivation happened. Pavlov gave the dogs this pairing many times. Curiously and reliably, after many pairings, the dogs salivated, but not just when the meat powder went into their mouths, they salivated after the tone, but Judith A. Dygdon, PhD PSYC 580 Part IV 6 before the meat powder. The previously neutral tone came to elicit a salivation response. Note that the dogs did not salivate willy-nilly. Early on, they salivated specifically to the meat powder. Later, they came to salivate to the tone, even when it wasn’t followed by meat powder. This response to the tone was new for the dogs, and hence we say that learning occurred. We call the new response (salivation to the tone) a CR or conditional response. Another example: I’ve got one! I love the smell of gasoline. I have for as long as I can remember. I have come to see, however, that most people do not find that smell appealing. This led me on a search for the explanation of why I find the smell of gasoline pleasant. Here is what I think happened. When I was a kid, gasoline was quite inexpensive (I know I am dating myself with this story.) It was so inexpensive, that it was not unusual for people to make a special trip just to “gas up the car”. I do recall that this scene played many times in my experience. My father, in preparation for one of these “gas up the car” trips, would say to me: “Judy, want to come to the gas station with me Whoa! Of course my answer was yes. I’d get to spend time with my father. Just the two of us – no little sister! Do you see classical conditioning trials here that might explain my appetitive (i.e., good) emotional response to the smell of gasoline? Judith A. Dygdon, PhD PSYC 580 Part IV 7 In my example, and in the Pavlov example, the CR looks much like the UCR. We used to believe that that was the rule but an interesting series of studies taught us otherwise. This the story of these very clever experiments which are referred to as the Sign Tracking or Autoshaping studies.. Pigeons were put in standard laboratory cages built specifically for pigeons. Each cage was big enough to enclose the bird. Each had a food delivery system so that we could present an appetitive (i.e., good, or pleasant to the organism) stimulus to the pigeon. Each had a plastic, depressible key that could be illuminated from behind. Here is what the experimenters did: They lit up the key a few times. At first the birds oriented to it, but this stopped quickly. The key light, obviously, was neutral and meets our requirements for a CS. Next, the experimenters illuminated the key and before the light went out, the put food (a UCS) in the birds' dish. When training started, the birds did nothing when the key light came on, but as soon as the food appeared they scurried over to it and pecked at the food. These trials were repeated many times. A picture of the design of these trials appears below. Conditioning trails: UCS ----UCR food delivered to dish go to dish in feeding area of cage and eat the food | | | CS key lights up Judith A. Dygdon, PhD PSYC 580 Part IV 8 in different area of cage As the conditioning trials went on, the experimenters expected to see a CR develop, and they expected to see a CR that was identical to the UCR – that is, they expected that as training wore on, the birds would scurry over to the dish and wait for their food. Well, a CR did develop, but not the one that was expected. As the training wore on, the birds started to run over to the disk and peck at it! Did a CR develop? Absolutely. Before training the illuminated key elicited only an orienting response that quickly habituated. Now it elicits pecking and hanging around! Was the CR the same as the UCR? Absolutely not. What did these experiments (and others like them, since) teach us? The CR won’t always be identical to the UCR. We know define classical conditioning as Classical conditioning captures the process through which a neutral stimulus comes to elicit a response it did not elicit before conditioning trials. (Why did this particular response emerge in the birds? We think it is due, in part, to their innate feeding repertoire. The food elicits approach and pecking. The illuminated key provided a new, and earlier, target for approach and pecking.) Judith A. Dygdon, PhD PSYC 580 Part IV 9 Human Emotional Responses I’ll conclude this week’s online presentation with a statement about why you should care about classical conditioning. We believe that acquired emotional responses are learned through classical conditioning processes. Lots of the issues that come to clinical attention are problematic emotional responses. We will talk specifically about these after we finish our online review but consider, as a coming attraction, the case of simple phobias. Consider, for example, a dog phobia. Classical conditioning raises the hypothesis that a problematic dog fear has developed through sufficient experiences that fit the following template: Conditioning trials: UCS bite ----- UCR pain/fear (negatively valenced arousal) | | | CS dog Eventually: CS dog ----- CR fear Judith A. Dygdon, PhD PSYC 580 Part IV 10 Now, the UCS need not be a bite. That was just an easy example. The UCS need not even be delivered by the dog. It would just need to be a traumatic experience that co-occurs with dog. How long will it take to produce a conditioned response? Varies with stimuli, with responses, and across individuals. While conditioning can be accomplished with one trial, conditioned responding becomes stronger with repeated trials. Once built, are we stuck with these fears forever? Will all traumatic pairings result in fears? --- We will explore questions like these later. *** Oh, and one more thing -- I had you read an article by Gray. Here's why: This article was not Gray’s first writing on the topic, but perhaps his clearest. In it, he summarizes work that crosses the behavioral/neurological divide. This work pretty convincingly argues that we can see in humans (and of course, in other organisms) differentiation in neurological structures that support our responding to (and hence, learning from) environmental events. Our best interpretation of the data available so far is that there are three neurological systems that support our responding: The Behavioral Approach System (BAS), the Behavioral Inhibition System (BIS), and the Fight/Flight system (FFS). You will study more about the specific neurological structures involved in these in biological bases courses, but these concepts are important to us in learning because of what they have to say about our interaction with environmental events. Let me highlight some important features of these concepts: Judith A. Dygdon, PhD PSYC 580 Part IV 11 BAS -- structural system that supports responding to appetitive stimuli and appetive emotional responding, like pleasure. This appears to include responding to innately meaningful appetitive stimuli (i.e., pleasureable UCSs) and learned appetitive stimuli (i.e., pleasureable CSs). As with all other biological based characteristics (consider the weird parrallel to "height"), we expect variation across individuals. Some people will be more sensitive to appetitive stimuli than others, but we all have the capacity to respond to appetitive stimuli (just as some individuals are taller than others, but we all have the characteristic of "height"). FFS -- structural system that supports responding to natural aversive stimuli, and some aversive emotional responding, like fear. Yes, I typed "natural." Unlike the case with BAS, Gray (and others) believe that we see responding to natural aversive stimuli, and emotions like "fear" supported by its own system. As with all other biological based characteristics we expect variation across individuals. Some people will be more sensitive to natural aversive stimuli than others, but we all have the capacity to respond to natural aversive stimuli. BIS -- structural system that supports responding to novel stimuli and to learned aversive stimuli, and accompanying emotional responding, like alertness and anxiety. Yes, I typed "learned." Unlike the case with BAS, Gray (and others) believe that we see responding to learned aversive stimuli, and reponses like anxiety and alertness, supported by its own system. Note that Gray believes our responding to novel stimuli is organized in this system as well. As with all other biologically based characteristics we expect variation across individuals. Some people will be more sensitive to novel and to learned aversive stimuli than others, but we all have the capacity to respond to novel and to learned aversive stimuli. Note that, taken as a whole, this literature categorizes responding into responding to appetitive and aversive stimuli, but posits three neurological pathways to support this. As this literature has worked its way into clinical psychological practice, people seemed to have "lumped" FFS and BIS responding into one group (hence the BIS/BAS self-report instrument that the Gable et al. articles discuss) but please keep in mind that the writing on the neuro literature posits three systems. Gray (with McNaughton) in 2000 modified this model, still defining BAS as the “home” of responding to appetitive stimuli, but broadening FFS (renamed Flight/Fight/Freeze System) to cover responding to natural and learned aversives, Judith A. Dygdon, PhD PSYC 580 Part IV 12 with BIS managing responding to conflictual situations involving both appetitive and aversive activation. This is a continually developing arena, here is the point for us: It is interesting that our studies of the behavioral systems (i.e., the study of "learning") have made discoveries that parallel these biological findings. In classical conditioning we talk about aversive and appetitive UCSs and learning that comes from pairings with them. In operant conditioning, we will see that behaviors are built through the influences of reinforcement (through appetitive consequent events) and punishment (through aversive consequent events). So cool. By the way, some people are fond of saying that learning theory / conditioning models ignore emotion. Actually, emotion is critical to both classical and operant conditioning. If you don’t understand this statement, ask me about it. Judith A. Dygdon, PhD PSYC 580 Part VI 1 Once learned, a CR can spread – Here’s how. 1.) Stimulus Generalization = Once a CS has come to elicit a response (i.e., a CR), other stimuli like it will elicit similar responses. The more another stimulus is like the one involved in conditioning, the more likely it is to elicit a response. For example, we can train a rat to be afraid of a tone of a particular pitch (say 2 kHz) by pairing it with shock, many times. UCS ----- UCR shock fear | CS ----- CR 2 kHz fear tone If we do not teach this rat otherwise (explicitly) his new CR to the 2kHz tone will spread to tones of other pitches. This is a very reliable phenomenon. So reliable, that we often present our predictions in graph form and give that graph a special name. Look at graph # 1 “PDF #2” to see a picture of an expected generalization gradient for a rat with this experience. Experience will influence what this generalization gradient looks like for different individuals and for different stimuli. Continuing with the example from above, if we explicitly teach this rat that the 2 kHz tone predicts shock, but also present other tones and never pair these with shock, his generalization will be much more specific. UCS ----- UCR shock fear | CS ----- CR 2 kHz fear tone and no shock | CS all other tone pitches Its gradient would likely look like graph #2 in “PDF #2”. Judith A. Dygdon, PhD PSYC 580 Part VI 2 2.) Higher Order Respondent Conditioning = Once a CS has come to elicit a response it can be used much like a UCS is in the training of responding to other, novel stimuli. Here is a rat illustration. This is First Order – UCS shock | | CS1 tone ----UCR pain/fear ----- CR1 fear ----- CR1 fear ----- CR2 fear ----- CR2 fear ----- CR3 fear This is Second Order – CS1 tone | | CS2 light This is Third Order -CS2 light | | CS3 click Remember hypothetical Bob and his fear of dogs? Well once Bob has learned to fear dogs by experiencing “dog” and coincident trauma, he is likely to learn to fear other things in the following way. First Order Conditioning – Judith A. Dygdon, PhD PSYC 580 Part VI UCS ----- UCR bite pain/fear | | | CS1 dog and eventually CS1 dog ----- CR1 pain/fear Higher Order Conditioning (2nd order) -CS1 ----- CR1 dog pain/fear | | | CS2 new neighbor and eventually CS2 ----new neighbor CR2 fear Higher Order Conditioning (3rd order) CS1 ----- CR1 new neighbor fear | | | CS2 new neighbor’s custom motorcycle 3 Judith A. Dygdon, PhD and eventually PSYC 580 Part VI 4 CS2 ----- CR2 new neighbor’s fear custom motorcycle Why do you think I made a point to qualify “neighbor” with “new” and “motorcycle” with “custom”? This won’t go on forever, because every higher order conditioning trial is also a __________ trial. And one more thing – Stimulus Arrangements So far, I haven’t said anything about WHEN the UCS “comes on” and WHEN the CS “comes on”. In fact, a variety of possibilities for these arrangements exists. Researchers have study the impact that different arrangements in the presentation of these two stimuli have on the resulting CR. Because of this, names have been attached to many of the arrangements studies so far. You need to know the names because you’ll read about them, and you might need them to describe classical conditioning based interventions you use. We have several stimulus arrangements that, collectively, are called FORWARD PAIRINGS. They are called FORWARD because the CS comes on before the UCS. We talk about three specific forward pairings. They are: 1.) Short-delayed conditioning in which the UCS comes on shortly after the start of the CS (For example, in teaching a rat that a tone (CS) means food (UCS) is coming, we might start the tone and 2 seconds later present the food.) 2.) Long-delayed conditioning in which the UCS comes on a long time after the start of CS (In the rat example above, the tone might be sounding for 30 seconds before the food is presented.) NOTES on the first two types: The meanings of “short” and “long” are relative, not absolute. The CS and UCS will overlap for some period of time. Judith A. Dygdon, PhD PSYC 580 Part VI 5 3.) Trace conditioning in which the CS is terminated some time before UCS begins. Be definition, in trace conditioning procedures, the CS and UCS never overlap. The CS always ends before the UCS comes on. (In the rat example, the tone comes on, stays on for 10 seconds, then goes off for 10 minutes, and then food is presented. The study of FORWARD PAIRING stimulus arrangements has given rise to some vocabulary that you need to know.  Inter-stimulus interval (ISI) = Time lapsed between the start of the CS and the start of the UCS.  Inter-trial interval (ITI) = Time lapsed between the end of the UCS and the beginning of the CS.  Trace interval = Time lapsed between the end of the CS and the start of the UCS. Will all forward pairings result CRs? It depends on the stimuli used and the organism. Recall the rat example I gave you above. We have data that support the expectation that the short delay and long delay examples should produce CRs (indicated by approach to the food dish). I don’t expect the trace example to produce CRs in rats. However, we do have lots of data to support the effectiveness of trace conditioning arrangements in human beings. The arena from which this support comes is that of taste aversion. It looks as though humans (and probably other organisms) are ready to learn to avoid (and dislike) novel tastes that have been even coincidentally linked to illness (even if that illness does not begin until hours or days after the new taste has been experienced. Researchers have also explored “other than forward” stimulus arrangements and have found that, under many circumstances, other non-forward pairings can produce CRs. Here are the non-forward names we use so far: 1.) Simultaneous conditioning in which the UCS and CS start and end at the same time. Judith A. Dygdon, PhD PSYC 580 Part VI 6 2.) Backward conditioning in which the CS comes on as UCS ends. (I believe soon we will be talking about short delay backward, long delay backward, and trace backward stimulus arrangements.) 3.) Temporal conditioning in which some like to say “there is no CS” but the UCS comes on and off at regular intervals. The CS can be seen to be the passage of an interval of time. (For graphic representations of all of these stimulus arrangements, see graph #3 in “PDF #2”.) The study of the impact of stimulus arrangements on whether or not a CR is produce and what that CR might look like has produced some fascinating results. Why might the timing of when the CS appears relative to the UCS affect the kind of CR developed? Consider that UCSs elicit UCRs that are (often) emotional responses. In humans, we believe that emotional responses follow the path depicted in the Standard Pattern of Affective Dynamics (see its graph in “PDF #2”). For example, presenting the CS before, or just before (i. e., short delay) the UCS might be linked to the beginning of SPAD, but a stimulus that ends well into the “life” of the UCR’s SPAD (i.e., long delay) might trigger a very different (even opposite) emotional response! Judith A. Dygdon, PhD PSYC 580 Part V 1 Extinction and its Limits When we closed our talk in the previous segment, I posed a couple of questions. One of them was “Once built, are we stuck with these fears forever?” We had just finished talking about the fact that, when we encounter pathological human fear and anxiety responses in clinical situations, we first (and often very successfully) look to the classical conditioning model to explain how these reactions developed. Why classical conditioning? Because these fears (like a fear of dogs) are responses to antecedent stimuli (in this case, dogs) and could be simply elicited behaviors. Because not all humans are afraid of dogs, it is likely that when we encounter an example of this fear, it is an acquired or learned elicited response. Classical conditioning explains how elicited responses are learned. Why is that important? Because if we know the kinds of experiences that built the fear then we know something about the kinds of corrective experiences (i.e., psychotherapy) necessary to correct/undo/treat the fear. Time for a disclaimer – I am talking about fear as a simple human example of a classically conditioned response. I am doing this because we encounter fears and anxieties all the time in clinical work. However, classical conditioning explains how new emotional reactions are established (NOT JUST FEAR). It also explains how appetitive responses (i.e., likes) are acquired. Remember that. You might need it. OK – now, back to the question. Once built, are we stuck with these fears? The answer is “no” (well, mostly no). Once experience creates a new emotional response through classical conditioning, it can make that response go away through process called extinction. This is how extinction works: After a CS has come to elicit a response (like “Bob” learned to fear dogs from being bitten many times), present the CS without the kind of UCS that gave it meaning (in this case that would be presenting dog without anything traumatic happening) over and over and over again. Eventually, Bob will no longer show fear when presented with dog. At this point, you might be asking “how long will it take to extinguish a learned response?” The answer to that question is related to the question “how long will it take to build a learned response”. The building of a CR depends upon the strength of the UCS used and the noticeability of the CS. The strong the UCS Judith A. Dygdon, PhD PSYC 580 Part V and the more noticeable the CS, the fewer “pairings” we will need to build a response. The converse is true of extinction. The stronger the UCS used in “building” the more extinction trials we will need in the “undoing”. So, we call this process through which we get rid of learned elicited responses, EXTINCTION. Accept the following as a picture of extinction. Note that I am using big font/boldface to capture the fact that we are presenting this eliciting stimulus over and over again, with nothing “emotionally meaningful” happening. CS ----- CR Eventually, CS ----- and later still, CS ----- CR 2 Judith A. Dygdon, PhD PSYC 580 Part V 3 Extinction should look a lot, to you, like HABITUATION, which we discussed last week. We could have depicted habituation as S ----- R Eventually, S ----- R and later still, S ----- A question for you to consider: What’s the difference between extinction and habituation? I said earlier that a having a hypothesis for how a clinical problem developed gives us guidance in designing a treatment to ameliorate the problem. All of our empirically supported experiential (as opposed to drug) treatments for problematic fears are based on classical extinction. Some of those interventions are: • Systematic desensitization • Flooding • In vivo graduated exposure Judith A. Dygdon, PhD PSYC 580 Part V 4 Once a response is extinguished, is it gone for good? Hint: If you tracked a “comparison” I made above, you might have a good guess at an answer to this question. The perhaps sad truth is that once something is part of one’s learning history, it appears to be there for good (assuming the relevant brain structures /structure functioning remains intact). How do we know that? Because extinguished responses can come back. Here are some of the ways in which extinguished responses will return/ 1. Spontaneous Recovery = After extinction, if the organism is away from the CS for awhile, and then the CS is presented again, the CR is likely to be elicited again Simple human example: After we successfully treat Bob’s fear of dogs, if he does not encounter “dog” for a period of time, he might show fear when he next encounters a dog. Does my description of the spontaneous recovery of extinguished responses remind you of anything? 2. Renewal = After extinction, if the organism encounters the CS in a new context, or in the company of stimuli, other than those present during extinction, the CR is likely to be elicited again. Simple human example = After successful in vivo exposure of fear of dogs in his therapist’s office, Bob encounters a dog on the street and feels afraid (and mutters to self: “This psychology stuff is bunk. I still hate dogs.”) 3. Reinstatement = After extinction, if the organism encounters the UCS like the one used in training, alone (i.e., not even paired with the CS!), the CR is likely to be elicited again. 4. Reacquisition = This one is perhaps the least mysterious of the lot – After successful extinction training, the CR will be reacquired through new CS+UCS pairings in a much shorter time than was required for the organism to learn the CR originally. Judith A. Dygdon, PhD PSYC 580 Part V 5 Though we typically appeal to extinction based models in therapies to “get rid of” unwanted emotional responses, it is possible to approach the problem from another, classical conditioning based perspective. To illustrate this, let me appeal an animal example of fear training. We can expose a rat to many trials of the following: UCS ----shock | | CS tone UCR fear After sufficient trials, we will observed the following: CS tone ----- CR fear As I mentioned in a previous lecture, in a case like this, because the CS signals that “something is going to occur” (i.e., the shock) we will more specifically call the CS an excitatory CS or, alternatively, a CS+. We can give this rat some additional classical conditioning trials that will provide him with considerable relief. Even though we allow the CS+ to continue to predict shock, we will also provide the following “compound CS” trials: no shock ----| | CSA and CSB tone light After sufficient trials like this, we will find that when we present the tone, the rat continues to show fear, but when we present the tone with the light – no fear! When built in this way, we call the light an inhibitory CS, or a CS-. It predicts that, although otherwise expected, “something will not occur” (i.e., the shock). Judith A. Dygdon, PhD PSYC 580 Part V When inhibitory CSs “tell” the organism that they are safe from something scary, they are often referred to as “safety signals”. In this case CS light ----- CR relief 6
Purchase answer to see full attachment