Bridge and Target Training: Beyond Behavior Modification
Bridge and Target training, as developed from classical and operant conditioning in the marine mammal training field, is presented as an important tool in the enlightened and humane management of animals. Two distinctive developments of this technique, the INTERMEDIATE BRIDGE AND TARGET, are presented which go beyond operant conditioning and allow better communication between human and non-human animals. Continuous rewards are bridged to 20 percent to 30 percent reinforcements on a variable schedule. Successful applications with many species are presented including horses, pigs, pigeons, dogs, and frogs. Food rewards are rarely given. Other reward possibilities are discovered or motivators can develop by asking or finding the individual animal’s desires, and these can become secondary reinforcers. Examples of accomplishments by trainer and animal teams are presented.
KEY WORD INDEX:
training, animal welfare, animal husbandry, humane
About the Author
We can find ourselves not living up to our own animal care and management standards because of various expediencies. For example, we would like to administer good, low stress health care, but may compromise that goal if our animal will not allow routine examinations, causing us to miss early health problems. If the animal does not cooperate with his own medical care, he is more likely to require anesthesia even for simple procedures, sometimes at considerable risk to the animal. Similarly, when teaching animals, if we cannot explain things to them in a way that they can understand, then we can find ourselves using physical force or bribes to make an animal conform to some wish of ours. Over the past twenty years or so, a method of animal training developed in the marine mammal community which allows us to minimize coercive, invasive, and restrictive treatment of our animal colleagues while maximizing our ability to cooperate and communicate together. This technique evolved from classical and operant conditioning, and I describe it as “Bridge and Target” (BT) training.
I find the amazing success of BT, like with the most successful behavior modification techniques, uses empathic understanding when it is at its best. The dynamics of BT are many, some of which I believe are not yet understood. It has to do with clear communication, rewards, curiosity, relationship between trainer and subject, compulsivity of some species and individuals, and much more! “Bridge and Target” (BT) training allows people to communicate with animals with startling ease and without using any physical force or constraint. Although it has been used in various degrees of sophistication throughout the marine mammal industry for a long time, it has been difficult for anyone outside the marine mammal training field to obtain a description of this technique. Most of the people actively involved in its development are oriented to entertaining the public, rather than publishing in academic journals. Also, many of the facilities where these techniques have been developed consider them proprietary, and although they may have manuals for internal use, are not interested in divulging their techniques to others.
After a rather extensive literature search to see what was available to academicians, I found nothing. Therefore, in 1990 I wrote a short manual describing the basics of how to communicate non-verbally with animals by “Bridge and Target” training (1993). It is distinct from state of the art (or science) operant conditioning in two ways. While both use a conditioned stimulus (bridge) to signal completion of a trial to the subject, “Bridge and Target” training is unique because of its use of an intermediate bridge. The intermediate bridge signals the subject that s/he is on the right track but has not completed the trial yet, like the clue “warm, warm, warm!” in the game “Hot and Cold.” The importance of this tiny addition in technique is profound. It allows us to support the subject every instant as s/he claims new territory in learning, to convert failed trials into successes, and to expand correct behavior in the face of challenge. Along with this, it allows the subject to actually process what is going on and think about it!
The second innovation of BT is the target. Although some might say this is similar to the feed-activating lever in operant conditioning paradigms, BT goes far beyond this. In Skinner boxes, animals are easily conditioned to push a lever to activate feed, or are situationally manipulated into other responses: but all in a very limited and controlled environment, usually a cage. The chances are the animals in such tightly controlled conditions will happen to push the lever as s/he investigates the food source. In the real world, in order for animals to operate any mechanism, they must first notice it in an environment rich in distractions and choices. BT helps the individual focus on a target. This can be accomplished initially by teaching the subject to touch a contact point, commonly called a target, preferably while s/he is in a controlled environment, such as a lab or training area. This focal point can then be transferred out of the lab and into the real world where we can now focus the animal’s attention in one spot, and then give information, usually via placement of one or more targets.
In BT, the act of targeting is not directly related to getting food. It is directly related to getting bridged, with occasional (less than 20 percent) food awards! There is a difference between the individual who is going after food and one who is acting in order to make contact with a target. An animal who is focused on food is not necessarily open to learning. Have you ever noticed if you follow another car to find your way somewhere, you often will not notice your path because your attention was focused on the car ahead? If you are able to plot your course with a map, you are more likely to come away with an orientation to the entire neighborhood or the entire situation. Thus, it is important to teach the trainees to focus on the targets as markers which show the path they are to take, and not as places where food is going to appear next. For this reason, animals are taught to pass by food to get to targets early in their training! Once the target is separated from eating behavior, it is possible to introduce new targets at contact points not through use of signals related to food or other reinforcers. Any part of the body can be targeted or many parts can be targeted at once. Any target can be moved relative to the animal’s body or his or her environment.
In Bridge and Target training, the target is presented to the animal, showing the individual exactly what s/he must do to be successful. Trial and error is eliminated. Once the animal has learned to contact a target, for example with his or her muzzle, s/he will contact the target whenever and wherever it is presented. Even if a target is presented across a room or a field, the subject will understand that if s/he wants to be bridged, s/he must travel to contact the target with his or her muzzle. This is why it is so easy to teach animals to “come” by this technique, and why it is so easy to correct run-away animals. The animal learns by following a moving target, we can define any motion or behavior. Targets can be presented in a series in order to form certain behaviors, complex behaviors, positions or even be used to communicate concepts like, “wait,” in a process analogous to “connecting the dots.”
To summarize, the introduction of a fully formed target concept 1) transmits the specific requirements for success to the animal, eliminating trial and error,
2) focuses the individual’s attention to the place where the trainer will deliver the next information, in a mode that prepares him or her for learning (versus eating),
3) expands the flexibility of what we can discuss with the trainee because we can extend targeting to body parts that are not directly related to eating (so neither attention nor activities are limited to eating related activities). Moreover, animals quickly learn their way around this system so we can show them the desired end point. Then they will work with us to get there, sometimes improving the methods and results.
THE FRUITS OF TRAINING
Examples of applications of “Bridge and Target” training with various types of animals in various settings are listed for your consideration.
We routinely set up management programs for animals to allow them to assist in their own care. Retrieval of food and water dishes, toys, bedding, trash, and nets helps to free the trainers time for other meaningful interaction adds a sense of EMPOWERMENT for the individual in training. Incidentally, the subjects are usually also willing to deposit trash in a sack or bin. Subjects can learn to carry things, retrieve, including one another and things that are out-of-reach for the trainer.
Animals can also learn to stand quietly on a scale to be weighed, to separate out of a group when requested, to have specific places to facilitate feedings in groups and to allow one another to eat unmolested. They also learn not to challenge doors, gates, or food deliverers. I used to swim (laden with fish) among a group of adult male and female California Sea Lions, and was treated very politely and respectfully by all. I found it took only a few minutes work to
teach the animals their own names, and their own personal “station symbols” (rather like place cards at a formal dinner). This process greatly relieved feeding-time stress in group housing situations.
Animals who have been conditioned to return to a signal or call are more likely to be recovered if they get out. We recovered several lovebirds who were part of a colony which had received only rudimentary bridge and target training. Recovery time was about 15 minutes. The birds came to the target signal of the verbal cue, “Here,” and a finger signal on outstretched arm and hand. The birds had been bridged to a target of touching beak to finger, and they came flying from the treetops. A client in California reported a pony recently trained to respond to BT had charged a gate and was on a busy boulevard where he had a line of exasperated drivers stopped. She chased the pony, commanding him to come, calling his name to no avail. When she remembered to call him to the target, she reported the pony spun on a dime, and RAN back to her. He stood to be haltered to return home, to the standing ovation of some frustrated drivers! Other dramatic results reported to me include: 1 — Bob Cook and Jeanne Frost (personal communication, May 1991) regarding separate incidences of calling large dogs off chasing stock.
2 — Susan Conway (personal communication, March 1993) regarding calling a pig away from a confrontation with another pig when both came together accidently. Susan reports being over 100 feet away when she heard the noise and called one of the pigs to her. Even animals with a long history of running away when called can often be corrected in 5-20 minutes. The usual technique takes three steps listed below: 1. Bridge the vocalization, such as saying “here,” or the word “x” by giving food WHILE the animal is doing what you want. There are three trials here.
2. Isolate the verbal signal by saying the word and THEN present food or other reinforcement most preferred by the individual. There are three trials here.
3. This is the 7th trial, where you simply give the verbal cue along with the target, for example two fingers presented in clear, crisp delivery. Usually no food or reinforcement are needed here, although a variable schedule of reinforcement can be successfully used up to about 33 percent of the time. Last year, I gave a seminar at which our hosts commented that all the applications with horses sounded interesting, but if we could get their confirmed runaway dog, Emma, to come when called would be very convincing. A few minutes later, Emma dashed out the door as we were entering, affording the opportunity to address this challenge. In about 15-20 minutes she was coming at top speed whenever called. The next day during the seminar, I introduced Emma as a star pupil. Then we released her to have the run of the stable. About an hour later, I reminded the audience about Emma and proceeded to call her to the target, even though she was nowhere in sight. After about thirty seconds, no Emma. I called once more, in case she might not have heard. Thirty seconds later, still no Emma. I conceded that she might be out of range, or not interested in coming when I was interrupted by my assistant, who drew our attention out a huge, second story window, where we saw little Emma, ears flying in the wind, still about a half-block away, running to get to the target! When trained to come when called, the individuals can be given more freedom. Catastrophes such as becoming lost in a hurricane can be prevented and much more can result from this simple training.
Another aspect of general management that is facilitated by training is leading animals to use environmental features, whether it be an artificial honey tree, the entire yard (including changing established trails), slides, squeegee doors (to shed water from the coat of an otter entering his or her den) or using scratching brushes.
MEDICAL CARE AND SCIENTIFIC RESEARCH
Much of the justification of training for medical management and research hinges on the issues of stress and anesthesia. I have found that stress reactions are a leading cause of sickness and death, either from compromised immune response or activation of fight/flight resulting in injury, and even failure to thrive. Besides altering the immune function, stress can alter other aspects of body chemistry, skewing research results. Consider the effects of measuring the basal metabolic rate of a stressed animal in light of research indicating that a stressed animal may consume more oxygen than the same animal during peak exercise (Ray, et al., 1980). This research, which used individuals willing to be tested, included a profile of the respiratory rates of whales during sleep, jumping, high speed swimming, and that of the female shortly after her mate died. After her mate died, she lay in the water submerged much of the time and her respiration rate was as great or greater than her previous peak respiratory rate.
At the University of Maryland, we trained pigs to stand voluntarily while blood was collected from the vena cava, about one inch from the pig’s heart, via a 5-inch needle. The biggest problem reported by Benny Erez was that once trained, the pigs competed to be first, so he had to take a little extra time to teach each her name, and to only come when called. The total training time for this behavior was approximately 1 hour per animal, accomplished in sessions of 5 minutes or less, scattered over 2.5 weeks.
Many of my clients have Vietnamese Potbellied Pigs who are routinely anesthetized for trivial procedures such as vaccines, blood samples, and hoof trimming. These pigs voluntarily cooperate with these procedures, without the need of anesthesia, reducing stress, risk, and expense. It can take as little as one day to prepare pigs to be vaccinated without restraint, about 3-7 sessions for taking rectal temperatures, about a week for hoof trimming, and about two for blood sampling. In situations where the individual leaves his home territory to get medical treatment, s/he must also be taught about the new environment and people.
At the National Zoo in Washington, D.C., we trained all the animals in the “Beaver Valley” for general husbandry and medical management. We also collected volunteered blood samples from the grey seals. We brushed the wolves and treated their ears with insect repellent and salve, collected a hair sample from the paw of a polar bear, examined a broken bear canine without anesthesia, brushed the teeth and examined the mouths of the seals and sea lions, and collected grey seal urine samples. We also did minute physical examinations on both types of pinnipeds, and treated eyes, ears, and skin topically.
Veterinarians are not required to participate in the training, although the animal must learn to tolerate a number of people playing the role of the vet so that the animal learns to regard the vet as just another one of many people who are preoccupied with animal body functions.
PERCEPTION MODIFICATION OR COUNTER-CONDITIONING
In “Bridge and Target” training, the third thing the animals are taught to do is to relax on cue, a key step in preparing them for changing their attitudes about things. In order to teach the animal to relax on cue, the trainer says “easy,” “x,” or any other cue word while massaging the animal’s muscle groups. After a short time, simply hearing the word cue helps the trainee relax. The phrases counter-conditioning, desensitization, acclimation, habituation and perception modification refer to various aspects of the same process. “Perception modification” (Cover, 1994) is the description of the process by which we can change the way an animal looks at a given event. For examples, in less than 20 minutes trainers can significantly calm frightened or aggressive animals. We can teach individuals to enjoy bath time, clipping, or spraying, and even to relax on cue. This process can be so low-keyed that the subject is oftentimes unaware of the challenge being redefined.
Perception Training is a means of giving animals coping skills and changing their perceptions of stressful or disliked events. As mentioned above, Benny Erez trained market hogs to stand voluntarily while blood samples were taken. Since these pigs all learned to want to be first, they were certainly regarding the collection of blood differently than before! Hence, this one procedure gave three benefits. Judging from the overt signs of stress, such as running, squealing, and straining, we reduced the stress these pigs experienced as a result of being restrained or poked with a needle, and, had pigs “requesting” to be bled. If stress was reduced, the quality of blood sample was improved, whether for research or diagnosis. Finally, as reported by the humans involved in the procedures, we reduced their stress considerably.
The training process can help build trusting, mutually respectful relationships between people and animals, and it is the means of forming a closer understanding of our animal colleagues through direct communication. Because we have become efficient at transmitting information to animals, that part of the art of training has become trivial. Motivation is now the real challenge. A key part of managing motivation is to persuade the animal to reveal himself to you, so that you understand what he wants (during that particular encounter). The most direct way to get this kind of information is to simply ASK THE ANIMAL WHAT S/HE WANTS. Animals have already been taught to tell us things, for example honeybees have told us what kind of electromagnetic field is present, whether a cow wants food or a date with a bull (Varner et al, 1988), what their body parts are called (with pigs and horses), and what kind of food they prefer (with dogs). Although, like people, animals do not always choose what is healthiest for them, they certainly can tell us what they prefer, and this can be an important avenue for testing animal welfare provisions from now on.
Just as we can learn other languages, most non-human animals can learn a fair amount of human language, whether or not they can speak it. Animals are routinely taught concepts such as wait, in, out, over, under, remember-this-list-of-items, go find another animal (or human), help another animal (or human). They can learn cues in different modes, such as sight and/or sound, and in different languages. It only takes about 3-7 trials to switch a cue on a behavior, say from one language to another, or from a word cue to a gesture cue. Animals are routinely taught to respond to gestures as subtle as knee jerks (Romer, 1989) or forehead tension as in the case of “Clever Hans, the horse.” Most of the trained animals I have known understood well over 200 cues and/or words, and were on the way to learning others before I stopped keeping count. I have been working with a pig who can identify 8 colors, the numbers 1-22, the alphabet, 20 picture flash cards, and a number of words. We are now in the process of expanding her concepts of how the names/pictures/words/letters are interrelated. I suspect this individual could learn a considerable amount of English, and possibly to read even though English is a very unsuitable language for speaking with animals because it is so capricious and illogical. English with the same letters representing many different sounds or no sounds at all, with words that can sound alike but mean different things, and with redundant, overlapping letters, such as c, k, and s.
Chickens and pigeons learn to assist in getting on their costumes and equipment, voluntarily putting wings through sleeve holes. These birds, often maligned as stupid, learn basic obedience, can “heel,” come when called, learn to enjoy being bathed, learn to hop into the bath on cue, to perform a complex series of events without human intervention. In a wild west show with chickens, we had no fighting and no fraternization amongst our cast of mixed-sex adult chickens, and the chickens would practice their “parts” on their own time, (as do many show animals). It surprises many that chickens learn as quickly as dolphins, monkeys, bears and pigs. . . they just do not do all the same things. Chickens are better at some things and worse at others. In Japan, I saw fish trained to swim through colored hoops in a specific order, even when the hoops were mixed up, and fish trained to pull a blind down in their exhibit when the aquarium was closing. I have seen frogs orient to wave pressures, electric fish identify electromagnetic fields, pigeons notify troops of ambushes along roads in Vietnam or search for downed pilots in the ocean, dogs detect screwworms and termites, pigs search for drugs, seals find black boxes from oceanic air crashes, and iguanas stay on command.
A DIRECTION FOR FUTURE RESEARCH
An important contribution to animal welfare science would be to teach animals the concept of pain related to body parts, so that we can ask them “where does it hurt?” Cross species communication is the new frontier.
Cover, K. (1994). Changing bad times to good. Potbellied Pigs Magazine.
Cover, K. & Zeligs, J. The Syn Alia Series on Animal Training. Bridge and target training made easy. Author: Sarasota, FL.
Ray, Roger D., et al. (1980). Social organization and synchrony in a pair of killer whales. Proceedings of the International Marine Animal Trainers Conference. Symposium conducted in Key Biscayne, Florida, October to November.
Romer, S. (1989). An educational pinniped performance using subtle stimulus control and discrete visual bridging. International Marine Animal Trainers Conference Proceedings. Symposium conducted in Chicago, IL, November, pp 185-9.
Varner, M. , Cover, K. & Clingerman, K. (1988). Use of operant conditioning for identification of estrus in Holstein heifers. Proceedings, American Association for the Advancement of Science, Annual Meeting. Symposium conducted in Boston, MA on February 12-15.
Syn Alia Animal Training Systems
Kayce Cover earned her Bachelor of Science degree from the University of Maryland in 1990. She has been training domestic and exotic animals for research, public education, and entertainment for over 20 years. Cover has worked at the University of California at San Diego, Scripps Institutions of Oceanography, Mystic Marinelife Aquarium, the National Zoo and the University of Maryland at College Park and now heads Syn Alia Animal Training Systems, a private consulting firm. Cover has a training and certification program in Bridge and Target training, and has written a manual explaining how and why this method works. Cover’s current projects include teaching a pig to read English, developing a program for training dressage horses without burnout, and working with handicapped children via Bridge and Target.
Copyright 2004 by Syn Alia Training Systems Â®