Thursday, May 6, 2010

The Conditioning of Paca and Simba

Twelve-week old Paca, new to the family in April of 2009. Petting and attention have always been desirable unconditioned stimuli for the kittens. 


Simba (left) and Paca (right) out on the porch in the Fall of 2009; they have learned much about the outdoors environment since this picture was taken.


(Note: All of the punishments and rewards must be administered as temporally close to the time of the action that merited them, and they must occur every time the actions happen. In addition, the punishments cannot be signaled and they must be high-intensity aversive stimuli (which, as I have defined them, they probably are). Otherwise, behavioral modification can get messy!)  (Domjan, p. 175))


The Conditioning of Paca and Simba:
A Behavioral Modification Proposal by Sam Rosenfeld




This video gives a few examples of conditioning concepts that apply to the kittens. Myself on guitar and alto saxophone.




I will propose a method by which I can train my family’s kittens, Paca and Simba, not to incessantly meow and make noise for attention every time there is some auditory or visual stimulus near the gate to their room. I will outline a functional analysis of the behavior and plan to use the principles of conditioned inhibition, extinction, and behavioral regulation to conduct this behavioral experiment. Simultaneously, I will train my family and myself to extinguish our reinforcement of the behavior using extinction, negative punishment, and positive reinforcement.
I want to go over a few concepts in learning and conditioning that apply to my kittens already before I propose the ways that I want to modify their behavior with the aforementioned methods. First, the cats most certainly know their unconditioned stimuli (UCS), or stimuli that elicit vigorous responding in the absence of prior training. (Domjan, p. 254) Some examples include petting, food, catnip, attention (visual or verbal), and play. For the behavioral modification, however, I want to focus on two of the most powerful unconditioned stimuli in their environment: my Mom and their favorite leopard-print blanket. All of the mentioned UCSes elicit unconditioned responses (UR) of excitement and/or appetitive behavior. Second, there are a number of conditioned stimuli (CS) that already play an important part in their conditioning; these are stimuli that don’t elicit conditioned responses or activate representation of unconditioned stimuli until paired with unconditioned stimuli through conditioning procedures. (Domjan, p. 245) Examples of CSes relevant to Paca and Simba are patting/hand gestures, calling their names, footsteps near the gate down to their room, and possibly even the video camera. I say “possibly” because they may be undergoing appetitive classical conditioning with the camera, and definitely undergo this conditioning with the other CSes/UCSes I mentioned. This type of conditioning was discovered by Ivan Pavlov in 1927, and is defined as a type of conditioning in which a neutral CS becomes associated with a desirable UCS; in this way, unconditioned responses can become conditioned to the CS as conditioned responses (CR). (Pavlov, 1927) A few examples of classical Pavlovian conditioning include opening a drawer (CS) becoming associated with food (UCS); patting the couch or slowly moving our hands toward the cats (CS) becomes associated with petting and attention (UCS); potentially, the camera (CS) becomes associated with attention, play, food, petting, (UCSes). In these examples, the conditioned responses would be their begging and meowing.
There are more complex conditioning procedures and effects pouncing about in that lovely playpen of the kittens, one of them being operant conditioning. This occurs when the response is defined by the effect it produces on the environment, and can involved various associations between a contextual conditioned stimulus, a response, and an outcome to that response known as a reinforcer. (Domjan, p. 249; Rescorla, p. 151) For example, Paca and Simba have learned that no one will come down the stairs to play with them unless they respond to the stimulus of footsteps near the gate by going down the steps and moving out of the way of the person coming in. In this way, the kittens have learned that they can achieve the reinforcement of play, attention, or any other UCS more often in their environment by giving this conditioned response whenever footsteps come within certain proximity of the gate (contextual CS). They have also learned through operant conditioning that meowing and begging incessantly (conditioned response) when they see or hear a family member (conditioned visual and auditory stimuli) can cause the outcomes of food, petting, and even a chance to spend time on their favorite blanket (unconditioned reinforcers). This type of operant conditioning is known as an S (R-O) association, or “three term contingency”, a higher-order relation established where S sets the occasion for a response-outcome association. (Domjan, p. 114) From our end, my family and I have an aversive unconditioned stimulus of constant meowing in the house (even though it’s cute, it makes doing anything requiring silence in our living room very difficult, as seen in the first segment of the video), and we can usually get the outcome of quiet cats if we respond with attention or feeding. In this way, the kittens have conditioned me and my family to reinforce the very behavior we are looking to extinguish! 


An overview of my living room: the kittens stay downstairs, but like to come up to the gate and meow/make plenty of noise when we're trying to do non-cat related things! Original illustration April 2010

The kittens stay downstairs (in the lower portion of this picture), and come up the stairs to the gate whenever they see or hear anyone in the living room or the computer room (upper portion of the picture). Original illustration April 2010

Identifying the Reinforcers for the Meows

An important part of behavior modification is figuring out which unconditioned stimuli serve as the most powerful unconditioned reinforcers for behavior. Functional analysis is the procedure of identifying and isolating the reinforcers that decrease certain behaviors by putting the subject in a situation where a specific behavior will occur, presenting the stimulus of interest and observing the increase or decrease in response; this has helped us to discover the kittens’ most significant reinforcers. From our observations, the reinforcer that is likely responsible for the meowing behavior could be a leopard-print blanket to which the kittens have a Harlow-esque contact comfort attachment; in addition, the kittens treat my Mom as a secure base that they stick close to when she has the blanket on, further displaying similarities to the Rhesis monkeys in Harlow’s attachment studies. (Harlow & Zimmerman, 1959) We can play with the kittens for hours after they have been begging at the gate, and then feed them when we’re ready to go, but relatively soon after we leave they will start up their howling again. This discounts play, food, and our presence as the reinforcers for their behaviors. 
However, when my Mom puts this particular blanket on her lap, the kittens will vigorously engage with the blanket in a way that closely resembles nursing and will discontinue their whining after she has comforted them for a significant period of time. Interestingly, they do not seem to engage in the contact comfort behavior as much if my Mom is not the one holding the blanket, which may mean that she is the discriminative stimulus that governs this behavior. Stimulus discrimination causes differential responding in the presence of two or more stimuli. (Domjan, p. 208) So, without the stimulus/secure base of my mother acting in conjunction with this specific blanket stimulus and a calm environment, the kittens cannot be soothed.  In contrast with the stimulus discrimination they display with the attachment behavior to my mother and the blanket, the kittens actively display stimulus generalization when they are at the top of the stairs; any type of stimulus that can be heard or seen from the stairs that they associated with the family’s presence will cause them to engage in begging behavior. This leads me to conclude that the strongest unconditioned reinforcer available for behavioral modification of the kittens is the leopard-print blanket calmly held by my Mom. 

Behavioral Modification for the Whole Family!

There are a few ways to condition the kittens to stop begging for attention by using theories that emphasize controlling strong unconditioned reinforcements such as Mom + Blanket. We could try extinction, in other words simply refuse to reinforce in any way the behavior we want to fizzle out, but this method alone is going to be ineffective. One major reason is the kittens have learned the behavior through partial reinforcement, meaning that we have given them the UCS only some of the times that they meow and only some of the times we go near the gate (CS), which makes the cats more persistent and extinction more difficult than extinguishing behaviors with perfect CS/US contingencies. (Jenkins, 1962; Theios, 1962) Another major reason it might not work is because my mom and I have a particularly strong attachment to the cats, which translates into both a desire to satisfy their begging and an aversion to prolonged meowing. We almost always fall into a positive reinforcement trap where we’ll give attentional positive reinforcement after the cats perform the behavior we dislike, which reinforces that behavior. In turn, this creates a negative reinforcement trap where we feel the only way to remove the aversive stimulus is to respond.  (Wierson & Forehand, 1994)
One effective behavioral modification procedure we could use in conjunction with extinction is conditioned inhibition. This type of procedure, originally explored by Pavlov with canines, involves conditioning suppression of a behavior by using a CS to signal the absence of the UCS.  (Domjan, p. 75; Pavlov, 1927) We already know that the kittens associate footsteps near the stairs with UCSes of all sorts. What if we walked up to the gate when they were meowing too much and put a large sheet over the gate? It may sound silly, but the kittens would learn that while the sheet is over the gate, they would not receive any type of reinforcement for their begging. This is a form of negative punishment, because it involves taking away attention, a desirable UCS; this also employs stimulus discrimination. Once they have been quiet for some operationally defined period of time, 5 minutes for example, we could remove the sheet (end the punishment) and positively reinforce the desired silent behavior with a short Mom + Blanket time, ex. 5 minutes.
There are still other procedures that could be used to extinguish the begging behavior, including the behavioral regulation approach.  (Timberlake, 1984) This approach recommends taking the organism’s preferred distribution of activities, a point labeled on a graph as the behavioral bliss point, and creating a schedule line to represent how the time the subjects must spend doing the less probable behavior before they are allowed to spend time doing the more probable behavior. Behavioral regulation procedure could be adopted for the cats: by graphing the time spent quiet on one axis and the time spent with Mom + Blanket on another axis, graphing the behavioral bliss point, and creating a schedule line that does not go through that point, we could create an instrumental contingency that would boost the amount of time the kittens are not meowing while controlling the delivery of desired unconditioned reinforcement. Time with Mom and the blanket becomes contingent upon time spent not whining. We can easily control the availability of substitutes because, after all, there aren’t any substitutes for this blanket or for my Mom; additionally, the aversiveness of being quiet is not so strong that the kittens won’t be able to obey the schedule line.
Conditioning the family and myself not to interfere with the kitty conditioning shouldn’t be too difficult; all we need is a simple set of positive reinforcement and negative punishment rules. For example, every time we reinforce the kittens when they have been quiet for five minutes, we can reward ourselves by having some desired unconditioned stimulus (ex. Five chocolate covered raisins). Conversely, if anyone reinforces the kittens when they are begging or meowing, that person loses internet priviliges for a significant amount of time (ex. Four hours). It may be very hard for me and my Mom to perform the conditioned inhibition procedure of covering up the gate, but we need to stick to it to make the stimulus an effective signal and to keep punishment consistent; therefore, we may need a stronger reinforcer for employing these methods than chocolate raisins (ex. Needlepoint sewing for my mom, playing music for myself). We may also need strong punishments than losing internet priviliges because we are likely to sacrifice internet time for cat time; instead of losing internet priviliges, my mom could lose the privilige of reading her favorite books for a day, and I could lose the privilige of listening/playing music for a day.
Using knowledge of conditioning procedures, I hope to have great success in modifying the behavior of my favorite kittens in the world, Paca and Simba, as well as the behaviors of my family so that the process moves quickly and smoothly. I hope you enjoyed the project, and thanks again for reading! 


~ Special thanks to Mom and Dad for the great footage and images of the kittens


Works Cited

Domjan, M. (2005). The Essentials of Conditioning and Learning (Third Edition ed.). (V. Knight, Ed.) Austin, Texas, USA: Wadsworth CENGAGE Learning.

Harlow, H. F.; Zimmerman, R. R. (1959). Affectional Responses in the Infant Monkey. Science , 130 (3373), 421-32.

Jenkins, H. M. (1962). Resistance to extinction when partial reinforcement is followed by regular reinforcement. Journal of Experimental Psychology (64), 441-50.

Pavlov, I. (1927). Conditioned Reflexes: An investigation of the Physiological Activity of the Cerebral Cortex. (G. Anrep, Ed., & G. Anrep, Trans.) London: Oxford University Press.

Rescorla, R. (1988). Pavlovian Conditioning: It's Not What You Think It Is. American Psychologist , 43 (3), 151-60.

Theios, J. (1962). The partial reinforcement effect sustained through blocks of continuous reinforcement. Journal of Experimental Psychology (64), 1-6.

Timberlake, W. (1984). Behavioral Regulation and Learned Performance: Some Misapprehensions and Disagreements. Journal of the Experimental Analysis of Behavior , 41 (3), 355-75.

Wierson, M.; Forehand, R. (1994). Parent Behavioral Training for Child Noncompliance: Rationale, Concepts, and Effectiveness. Current Directions in Psychological Science , III (5), 146-50.