Overview of the Issue

The process by which horses learn (equine learning theory) is not routinely taught in equine veterinary schools. As such, veterinarians graduate with a limited knowledge of equine learning theory and handle horses using traditional methods of restraint. They also tend to explain the behaviour of the horse according to that used by the general horse owning community. Ultimately there is a high prevalence of injuries sustained by veterinarians working with horses, but this may be reduced through a deeper understanding of equine learning theory.

Objectives of the Presentation

This presentation aims to present to the delegate an understanding of the mental capacities of horses and learning theory. This knowledge can be applied to horses under their care to reduce any unwanted behaviours, such as kicking out, exhibited by the horse, and to make working with the more difficult equine patient more time efficient.

The Mental Capacities of Horses

We are often guilty of overestimating the intelligence of the horses we work with. Often equine veterinarians will comment that the horse ‘knows’ what he is doing and is making a purposeful decision to misbehave, however the behaviour can often be explained using equine learning theory—see below. Behaviours should never be attributed to higher processes of learning or understanding when they can be easily be explained by more simple processes. Key points to remember include:

• Horses have a relatively poor short-term memory. They may not associate punishment with the unwanted behaviour after even a few seconds.
• Horses have an excellent long-term memory. Memories in horses are triggered and so a horse may have a fear response triggered by a specific veterinarian or procedure they previously found fearful, even if they were not exposed to the veterinarian/procedure for many years in between.
• Horses learn behaviours in the context which they experience them in. For example, a horse may associate certain areas, such as veterinary examination boxes, with certain behaviours, such as pulling away, when it would not normally offer these behaviours elsewhere.
• Horses are not capable of problem solving but instead learn even apparently complex behaviours through trial and error.
• Horses cannot learn new behaviours through observational learning. For example, they cannot develop cribbing by copying another horse.

Classical Conditioning

Classical conditioning—Making associations between two previously unrelated cues; increases the Predictability of the environment for the horse.

Most of us will remember the story of Pavlov’s dogs whereby he rang a bell before feeding his dogs meat. He then discovered if he rang the bell when no meat was present the dogs would still salivate in anticipation of being fed. This is the phenomenon of classical conditioning, when the horse learns that one cue predicts what will happen next. How quickly the association is made depends on how frequently the two events occur, how consistently one predicts another and how close together they occur.

Examples

• An older horse approaches a young horse with teeth bared and ears back. The youngster stands still and gets bitten. Next time the older horse approaches with teeth bared and ears back the youngster can predict what will happen next (a bite) and can move out the way to prevent being bitten.
• In well-trained horses the rider will give a postural cue before using their hand/leg. The horse will quickly associate the two events and react to the postural aid. The rider does not then need to use the leg/hand and the aids appear ‘invisible’
• How many horses do you know that whinny when they hear the latch slide open on the feed room door, or that become anxious when you raise the jugular vein prior to venipuncture?
• Clicker training relies on the fact that the ‘click’ predicts a reward will follow

Operant Conditioning

Operant conditioning—Making associations between a stimulus and a response; increases controllability for the horse.

Operant conditioning allows the horse to learn how its behaviour can alter the environment. It is divided into reinforcement training—increasing the likelihood a behaviour will be repeated in the future; and punishment—decreasing the likelihood a behaviour will be repeated in the future.

Positive Reinforcement

The addition of something pleasant after the desired behaviour.

This makes the horse more likely to repeat the behaviour in the future. Common examples would be food or a scratch on the withers.

The timing of positive reinforcement is vital—it needs to be as close as possible to the desired behaviour. For this reason, secondary positive reinforcement is frequently used. For example, in clicker training the ‘click’ predicts the food reward and so the click can be timed to coincide with the desired behaviour.

Examples

• Giving a horse a small food reward every time they are caught makes them very easy to catch.
• Giving a horse a quick scratch on the wither when they stand still in an examination room, makes this behaviour more likely to be repeated.

Negative Reinforcement

The removal of an aversive stimulus after a desired behaviour.

Lots of people get negative reinforcement confused with punishment. Try and think of it from the mathematical sense of the word, negative means removal or subtraction of something and reinforcement means to increase the likelihood the horse will repeat the behaviour. A better way to remember this is often removal reinforcement. Pressure is applied to the horse which motivates it to remove the pressure, provided the pressure is released as soon as the desired response is offered the horse will offer that response next time the pressure is applied. An important thing to remember is that the pressure should never cause pain or induce fear as these are both detrimental to learning.

Examples

• If a horse is too hot, they seek shade; if they are thirsty, they drink; if a fly irritates them, they swish their tail. Each time, the horse alters their behaviour to remove the aversive stimulus.
• If you put pressure on the horses halter (pressure), they should walk forwards and the pressure be released.
• If the horse feels pressure from the halter when tied up and pulls back, breaking the rope, this releases the pressure and the horse is more likely to repeat this behaviour in the future.

Positive Punishment

The application of an aversive stimulus after an unwanted behaviour has occurred.

This makes the behaviour less likely to be repeated in the future. Again try to understand this in the mathematical sense: Positive means the application or addition of something and punishment tells us the behaviour is unlikely to be repeated again in the future. A good example of punishment would be an electric fence.

The act of the horse pushing on the fence results in an aversive stimulus (electric shock), this results in the horse being less likely to push on the fence (the unwanted behaviour).

Negative Punishment

The removal of a pleasant stimulus after an unwanted behaviour has occurred.

Again this makes the behaviour less likely to be repeated in the future and again think of the words in the mathematical sense: negative means the subtraction of something and punishment tells us it is to make a behaviour less likely to occur again in the future. An example would be not giving the horse its feed if it was badly behaved in a ridden session before—we can obviously see this is a very ineffective technique (as the horse will not associate its behaviour with the consequence) but amazingly it is still widely used by horse owners.

The Problems with Punishment

• It lowers the motivation of the horse to trial new responses in training. Punishment tells the horse what not to do but does not tell it what it should be doing instead. Therefore, when punishment is used frequently in training, the horse becomes reluctant to offer new responses, in case they are wrong and are reprimanded for them.
• The horse can become desensitized to the punishing stimulus. If the punishing stimulus is not enough to suppress the behaviour the horse will gradually become habituated to increasingly powerful and painful punishing stimuli with obvious welfare consequences.
• Timing—To suppress a behaviour the punishment needs to occur at the same time as the behaviour occurs. If it occurs afterwards there is a good chance the horse will associate the punishment with the immediate body reaction/ posture of the person, not with the unwanted behaviour. Therefore the behaviour is not likely to be suppressed and the horse will be more confused and anxious in training
• The horse may have an extreme reaction to the stimulus. This is negatively reinforced if the reaction stops/ delays the punishment meaning the horse is more likely to offer extreme reactions in the future.
• It creates powerful fear associations with the person/whip. Fear is learned from a single response and never completely forgotten. We always want to avoid fear in training.

Summary Including 5 Key “Take Home” Points

1.  Horses have limited ways in which they can learn new behaviours that are based on operant conditioning.

2.  Horses are excellent at making associations between two events.

3.  Reinforcement of a behaviour means it is more likely to repeated.

4.  Punishment of a behaviour means that it is less likely to be repeated.

5.  When we talk about positive or negative from a behaviour perspective, we are indicating whether something is being added to or subtracted from the scenario.

Summary

Horses have limited ways in which they can learn and we should not assume an unwanted behaviour is being exhibited because the horse is naughty or has a desire to upset us in some way. By understanding the processes by which horses learn, we can not only understand why unwanted behaviours are offered but also understand how to prevent them being repeated in the future.

References

1.  Pearson, G. (2015) Practical application of equine learning theory, part 1. In Practice. 2015;37:5 251–254 doi:10.1136/inp.h2046.

2.  Pearson, G (2015) Practical application of equine learning theory, part 2. In Practice. 2015;37:6 286–292 doi:10.1136/inp.h2483.

3.  McGreevy, P and McLean, A. (2010) Equitation Science. Published by Wiley-Blackwell, West Sussex, U.K. IBSN 978-1-4051-8905-7