The science behind behavior modification (Proceedings)

Learning theory is not just theory...it is the scientific knowledge developed from the learning principles of habituation, classical conditioning and operant conditioning. This includes systematic desensitization and counter conditioning as well as types of punishment.

Classical conditioning

We may wish to establish new responses by counterconditioning new desirable responses or extinguishing existing undesirable responses. New aversive or positive emotional responses are created by pairing the relevant stimulus with an aversive stimulus (loud noise), or a positive appetitive stimulus (food). Undesirable classically conditioned responses are extinguished by presenting the conditioned stimulus repeatedly without the unconditioned stimulus.

Operant conditioning

You can establish new responses by reinforcing (e.g. food) desirable behaviors (operants), or extinguish undesirable behaviors (such as attention-seeking behavior) by allowing no more reinforcement. Extinction is an active process that occurs when a response is made but reinforcement is withheld. Depending upon the response and conditioning procedure (see later), just a few or hundreds of responses may be made before extinction occurs. Forgetting is a passive process by which a response declines as a function of time. Extinguishing a learned response is much more difficult for behaviors reinforced on an intermittent schedule than on a continuous reinforcement schedule.

Clicker training

A form of operant conditioning where a bridging stimulus is used to deliver the reward (reinforcement) immediately after the desired response. It also involves classical conditioning, as the click signals that a reward is coming, therefore the animal has a positive emotional reaction.

Reinforcers

Reinforcement: The application of a positive stimulus or removal of an aversive stimulus that INCREASES the probability of a behavior recurring.

     • Positive reinforcement: Increasing the probability of a behavior reoccurring by ADDING SOMETHING PLEASURABLE. Positive means "adding something"

          - Examples: food treats to reward a "sit" command, praise for "coming" when called

     • Negative reinforcement: Increasing the probability of a behavior reoccurring by REMOVING SOMETHING AVERSIVE. Negative means "removing something"

     • Negative reinforcement is NOT punishment

          - Examples: removal of pressure of a bit in horse's mouth when horse stops, removal of pressure of dog head collar when dog stops pulling

     • Primary reinforcer: Naturally or biologically reinforcing (food, water)

     • Secondary reinforcer: Previously neutral stimuli that have been paired with primary and take on qualities of the primary reinforcers.

          - Examples: "good dog" paired with food, "click" from clicker paired with food, can opener for cat food.

Reinforcement schedules

     • Continuous reinforcement: Reinforce EACH AND EVERY response

          - Necessary for shaping. Also necessary for early steps of training, switching to intermittent schedules.

          - Intermittent reinforcement: **Much more resistant to extinction compared to continuous reinforcement** — this is why slot machines are so popular!

     • Ratio

          - Administration of reinforcement after specific number of responses (fixed ratio) or after approximate number of responses (variable ratio—most effective).

          - Interval

          - Administration of reinforcement after specific amount of time (fixed interval) or after approximate amount of time (variable interval).

Shaping behavior

Utilizes gradual successive approximation in teaching responses that do not yet exist at the operant level.

Example: Teaching a dog to open the fridge and retrieve a beverage. First teach a dog how to touch an object on command, then teach it to put that object into its mouth, then teach it (separately) to touch a handle, then teach it to grab a rope attached to the handle, then teach it to pull the rope, then teach it to grab the handle alone (without rope), then teach it to grab an object that is in the fridge, then teach it to bring that object back to you, then teach it to close the door, then teach it to grab the object THEN close the door, then put it all together!!! (obviously this description is much simpler than it actually is).

Flooding

This is a type of Habituation. This unintentionally occurs frequently in animals in nature and our homes. This is when animals are habituated to situations causing anxiety and fear, where the full-strength stimulus is presented repeatedly (does not produce aversive consequences).

Flooding can be a risky procedure! If not done properly, it can worsen the fearful response. Flooding is easiest in young animals. Stimulus at full strength is presented repeatedly until animal shows no anxiety or fear. In growing up, young animals are habituated to (flooded with) many potentially anxiety-evoking stimuli such as motorized appliances, auto rides, strange noises, and strange people resulting in habituation. Flooding, like desensitization, results in a previously anxiety-evoking stimulus producing a neutral emotional response.

Rarely do we recommend this procedure for behavior modification.

Systematic desensitization

     Therapeutic application

In adult dogs, this technique is useful in treating the aversive emotional reaction associated with innate and acquired anxieties or fears. The procedure is commonly used in treating separation anxiety and fears of loud noises. The technique involves habituation of an innate fear or extinction of classically conditioned responses.

Procedure

In adult animals, the fully intense stimulus often evokes a full-blown aversive emotional reaction (with visceral consequences) which makes it almost impossible to extinguish or habituate because the stimulus is, in fact, followed by aversive consequences. Desensitization involves presentation of a gradient of increasing intensity of the stimulus. The stimulus eventually loses its ability to produce an adverse emotional reaction. The gradient is distance or intensity. To accomplish desensitization it is necessary to identify the emotion-evoking stimulus (stimuli), determine the starting intensity which evokes no response or a weak response, and conduct trials at increasing steps of intensity. Other aspects of desensitization are covered in Sections on anxiety and fears. Desensitization may also be used to reduce excitatory reactions such as exuberant barking to door bell ringing. An example of the use of systematic desensitization is given below.

Counterconditioning

      Therapeutic application

Used to supplement desensitization to replace an aversive emotional response (fear or anxiety) with a positive (or appetitive) emotional response that is incompatible with the aversive state. Counterconditioning gives better resolution to problems than just a neutral reaction from desensitization alone.

Procedure

Usually used in combination with desensitization. For treating fears (aversive emotional reactions), pair a very mild form of the fear-evoking stimulus with food and/or affection to condition a positive emotional response that is incompatible with the internal aversive reaction. The positive internal state overcomes the aversive internal state if the aversive state is only mildly aroused. Deprivation of food for 12 or even 24 hours may be advisable in beginning counterconditioning. As the animal becomes progressively desensitized to the low-level stimulus, the counterconditioning is then carried out with the stimulus at progressively higher levels.

In the case of flooding in young animals, counterconditioning is usually built into the desensitization process. For example, the desensitization to children usually is followed immediately by positive interactions with the children.

Example - In dogs, to treat anxiety associated with having nails clipped, present the dog with food treats when nails are very mildly clipped. An example using fear of children is presented below.

Punishment

     Principles of punishment

Defined as the application of an aversive stimulus or removal of a rewarding (reinforcing) stimulus (stimuli) contingent upon the performance of a particular behavior, and that decreases the probability of that behavior occurring in the future.

Punishment is not the same as negative reinforcement Negative reinforcement is the removal of an aversive stimulus to increase probability of a behavior.

Rules of punishment

     1. Provide opportunity for the animal to perform correct behavior!

     2. Reduce the motivation for the behavior (i.e. castrate marking male dogs)

     3. Stage misbehavior if possible (i.e. set up the steak on top of the counter)

     4. Punishment should be intense enough to stop behavior, but not too severe to cause anxiety/fear

     5. Needs to be short delay between behavior and punishment, 1-3 seconds

     6. Don't gradually increase punishment, as the animal can habituate to it

     7. The punishment should be dependent on the behavior, not the punisher

     8. The behavior needs to be punished every time

Types of punishment

     1. Interactive punishment.

          A. Examples which are commonly referred to as administering "discipline" include: hitting, slapping, grabbing by scruff of neck, yelling, and correction with choke chain or control halter.

          B. Although not recommended, it is sometimes effective in punishing specific acts of misbehavior when the animal is caught in the act.

          C. Interactive punishment can produce an aversion to the person delivering the punishment and can make a fear-motivated problem worse.

          D. Interactive punishment in dogs usually involves the punisher exerting dominance over the dog and the dog being submissive. However, in many cases, it may also lead to an escalation of the dog's aggression. With cats, which are not predisposed to take such a subordinate role, interactive punishment may evoke aggression towards the punisher.

     2. Remote punishment

          A. The aversive stimulus is automatically triggered by the misbehavior or triggered by the punisher without the animal knowing it.

          B. Examples are use of: shock collars, bark collars, invisible fences, mousetraps, water sprayers, air horns, electronic sirens, and home security device.

          C. Remote punishment is delivered so as to dissociate the person delivering the punishment from the punishment. Can be used with common misbehaviors that are directed to only one or two locations. Avoids problem of producing aversion to person who is responsible for delivering punishment. Remote punishment is actually a form of classical conditioning to link an internal aversive emotional reaction with the object to which misbehavior is directed.

     3. Social punishment

          A. Involves temporarily abandoning or socially isolating the animal.

          B. Companion animals, especially dogs, do not like people to abandon them or having to leave a fun place like a walk in the park. Thus one can punish dogs by terminating pleasant social situations.

          C. Examples: walking away, leaving the house, going home from the park when misbehavior occurs.

          D. A bridging stimulus, such as a kazoo, can be used to precede the social withdrawal. If the bridging stimulus is sounded immediately after the misbehavior occurs and prior to social withdrawal, it takes on the aversive properties of social withdrawal and communicates the exact behavior that is being punished. Assume at least six trials will be needed to establish the connection between the bridging stimulus and termination of the outing or departure of owners. Bridging stimuli involving low-pitched, long vocalizations have an inhibiting effect on behavior, and are therefore better than stimuli characterized by short, high-pitched, rapidly repeated notes which speed up or initiate activity.

          E. The "time out" procedure of placing a misbehaving dog in a back room is also a form of social punishment. The dog is handled prior to isolation, so one would not expect time out to be as effective as walking out on the dog. If time out is used, a bridging stimulus will be valuable in signaling the impending isolation by bridging the time gap between the misbehavior at the actual experience of isolation.

          F. In order for the "time out" to be effective, the location needs to be a room with little social significance, ie bathrooms, laundry rooms, etc. It should NOT be the dog's crate ,(if it is to be confined there for any other reason) the back yard or other rooms where the dog is likely to spend "fun" time.