TRUE TRAINING 30 - Rewards or Lures?
In all my training, with True and other horses, I use reward as a primary technique. In general, it’s quicker, lasts longer, is more effective, creates positive interaction, and builds a bond of trust between horse and human that other training methods fail to achieve. Horse Brain, Human Brain includes several chapters about how horses learn, the strengths and weaknesses of negative reinforcement, the techniques of training by reward, and the human brain training necessary to notice reward-worthy behavior that doesn’t “pop out.” But over the past few months, I've come to realize that “training by reward” takes on very different definitions with different people. So today, I’d like to explain how rewards differ from lures in the equine brain, and why indiscriminate treats interfere with learning.
Training by reward must be distinguished from teaching by lure, which is increasingly popular nowadays. A lure is something positive that a horse knows is available and will be offered if she produces some particular behavior. She just doesn’t know what the winning behavior is going to be. In teaching by lure, trainers hold or pocket 15 or 20 treats and dole them out over a short time while the horse alters her behavior to achieve the goodies. Many people refer to this luring technique as “training by reward,” but in the language of brain science, rewards and lures have very distinct meanings. They also operate on the equine brain in different ways.
When a horse learns something new, two (or more) neural networks in her brain are activated at the same time. To really cement this new lesson, the natural brain chemical called dopamine is sometimes released into the new connection. So, if we want to enhance learning, the best brain-based technique is to be sure dopamine is being pumped into that junction at the right moment. How can we be sure of that?
By using the element of SURPRISE. The strongest dopamine release occurs when a reward is surprising. So the first, best, and most rare reward has the greatest power. Luring a horse removes the surprise. The horse knows the lures are there, and knows they will be given to her if only she can move her foot or ear just so. She has learned this. Because there is no surprise involved, luring weakens dopamine release. What we want to use instead of a known expected lure is a hidden unexpected reward… something that surprises the horse so that her brain sloshes a whole boatload of dopamine into that new connection we’ve just taught.
Many people weaken their own training power by offering a horse edible treats that are not connected to any specific desired behavior. I’m talking about the carrot that’s offered because a horse is cute or the peppermint bite she gets just because she wants one. Some people make a habit of giving treats every day at noon, at the beginning or end of every training session, or at the waning of each full moon, even though the horse has not performed any specific behavior that deserves a reward in the few seconds preceding the treat. If you must give a horse treats for inspecific behavior, just throw them in a bucket and slip it into her stall or paddock when she’s off doing something else.
There are many potential problems with indiscriminate treats, but the primary brain-based concern is that they teach the horse to expect rewards. And that expectation eliminates the surprise that strong dopamine release—and good learning—requires. Indiscriminate treats are also extremely powerful because they are edible. Strokes, scratches, verbal praise, and rest all act as excellent training rewards. But let’s face it: they’re not as good as a hunk of carrot cake! The edible nature of indiscriminate treats means we’re breaking the effect of surprise with the most powerful possible means.
So, for maximum brain-based learning, avoid lures, use mostly non-edible rewards, prohibit all indiscriminate treats, and save yummy desserts for the rare execution of an especially difficult or highly desired maneuver.