Horses are Smart!

Published in Vetmed March 2012

According to E.B. Hanggi, at the Equine Research Foundation, Aptos, California, USA, horses are smarter than we think! They can manage ordinary daily cognitive tasks and mental challenges quite well.
In nature, food and water sources are of inconsistent quality and variable distribution. Their predators change location and habits. At the same time they must deal with interactions in the herd and need to learn and remember the identities and roles of other individuals in the herd.
Life for a domesticated horse is even more bewildering. In addition to dealing with natural situations, they often live in unsuitable environments, need to suppress instinctual behaviours and learn tasks that are not natural behaviours. They also need to co-exist with humans, who from a horses' viewpoint must be very strange indeed.
Research into the nature of cognition and perception in horses is gradually providing some fascinating insights. 
 
Horses learn through habituation throughout their lives. In a new environment, horses are very reactive and pay attention to any stimulus perceived by any of their senses. However, when certain stimuli are insignificant and are repeated regularly, the response diminishes. Habituation is a simple form of learning, but none the less important, as it allows non-vital information to be filtered out subconsciously, enabling the horse to focus on what is significant.
 
When a horse is particularly sensitive or fearful, either by its nature or through learned experience, it can be desensitized. Rough handling of its ears, or a bit banging against its teeth may make a horse extremely sensitive to being bridled, for example. This leads to head shyness. Such a horse needs to be retrained by approaching its head in increments and retreating when necessary until the horse willingly accepts normal, gentle bridling. 
 
Horses can become habituated and desensitized to almost overwhelming stressors if it is done correctly. A good trainer exposes horses in a positive manner to all sorts of sights, sounds and contacts, for instance by leaving him unrestrained in a safe corral and having humans behaving strangely around him, playing with bouncing beach balls, multicoloured balloons, umbrellas, plastic tarps, pylons, rattles etc.  This makes the horse much more capable of handling novel events calmly. If desensitization is not done correctly, it has the potential to make a horse more fearful and sensitive, for instance by removing an alarming stimulus too soon.
 
Pavlovian conditioning does not only work for dogs.
A horse also learns effortlessly when initially unimportant stimuli or events are regularly coupled with stimuli that initiate some sort of response. This form of conditioning is used by a trainer when he places a word onto a behaviour. The word "trot" would mean nothing to an untrained horse, but if it is regularly paired with the flick of a whip (associated with invoking a flight response or pain) immediately before the horse changes gait on an upward transition, the horse soon learns to respond to the verbal cue, without the use of a whip.
The use of the word "good", when correct behaviour is carried out, followed by a food treat, informs the horse that a food reinforcer is forthcoming.
To many a veterinarians dismay, horses often learn that the sight of a syringe is associated with pain. The unconditioned horse will then attempt to escape if he has not been trained to accept such handling. 
Stabled horses will learn to associate certain sounds with feeding - eg doors opening, hay being poured into a bucket, or will respond to visual cues , such as the arrival of the caretaker. They will then display sometimes unwelcome anticipatory behaviours, such as vocalizing, pawing or kicking stall doors. When reinforced, these may become conditioned behaviours.  For instance, when a horse, upon hearing grain being poured, starts kicking his door, a person may feed the horse hurriedly in an attempt to get him to settle down. This positively reinforces the undesirable behaviour and if this scenario repeats itself, the horse has effectively trained the human to feed him quickly.  
  
Operant conditioning, in contrast to classic or Pavlovian conditioning, occurs when an animal manipulates its environment to obtain reinforcement, which can be positive, when he receives something he desires, or negative - such as removing something unpleasant. Both types strengthen the relationship between stimulus and sought after response, so that when the same stimulus is given again, there is an increased chance that the horse will repeat the action.
 
In the early stages of learning the meaning of a new stimulus, the horse may respond randomly and hit upon the desired response by accident. Reinforcement of the response at the correct moment will cause the animal to repeat the behaviour. Horses excel at this type of learning, especially when positive reinforcement is given.
  
Unfortunately negative reinforcement is most often used in training. Horses are typically trained to perform actions to avoid something aversive. Under saddle they move forward when leg pressure is applied by the rider, on the ground they yield their hindquarters when pressure is applied to the flank, they back up when pressure is applied to the bridge of their nose, and they enter a trailer to avoid whips, or pressure from ropes. A good trainer refines this by reinforcing the correct and ignoring the incorrect. In this way only the slightest pressure will eventually produce the desired response, making the human/horse partnership appear effortless.
 
Research has been done with respect to the use of positive reinforcement principles to facilitate trailer loading of horses. It is not uncommon for a horse to be reluctant to be loaded onto a trailer. It can be very time consuming and hazardous for horse and handler. A horse may resist by rearing, pulling back, pawing, kicking or even falling over. If the owner gives up and fails to load the horse, such behaviour is reinforced. Traditional loading methods are based on negative reinforcement, often with some punishment for an uncooperative horse as well. Researchers Ferguson and Rosales-Rui found that , with positive reinforcement and target training, horses learned to load willingly and improper behaviours ceased. These effects also generalized to other novel situations. Horses that have learned to load through positive reinforcement do so eagerly, sometimes requiring only a verbal or visual signal, even when unrestrained. Positive reinforcement can also be targeted at lifting feet for hoof care, groundwork, grooming and veterinary handling.
 
Discrimination learning in horses has been researched since the 1930's. In discrimination tasks, horses must learn that one stimulus, and not another will result in positive reinforcement. That stimulus then begins to control behaviour, so that the horse acts in a specific manner in the presence of one stimulus, but not the other. In one study a specific horse learnt 20 pairs of discriminations. This horse also "learnt to learn" by using a general pattern to more easily solve subsequent tests. After 6 months he had also retained 77.5% of the discriminations. The "learning to learn" phenomenon has been noted in many other studies. It is a worthwhile tool in training. Often horses are restricted to performing only within a particular discipline. Racehorses don't jump; and dressage horses rarely set foot on a trail. These limitations prevent horses from learning about a variety of stimuli and such animals are less able to deal with novel situations. The more positive stimulation a horse experiences, the more readily it learns in new situations. It is also better adjusted in a variety of conditions and environments.
 
Spatial cues seem to be more readily discriminated by horses than any other stimuli. This is supported by observations made by vigilant horse owners of how well horses find their way around in areas they have only visited infrequently. Horses also react noticeably when objects in their environment are moved. This demonstrates an awareness of spatial surroundings.
 
The ease with which horses discriminate visual stimuli makes it easier for researchers to study cognition; and also to measure perceptual abilities and therefore equine vision. It has been experimentally demonstrated that a horse's visual acuity is 20/30 on the Snellen scale. (Human 20/20, dogs 20/50).
 
Experiments regarding colour vision in horses showed that they could discriminate red and blue from gray. Others showed that they could also discriminate between green and yellow. Recent research by Hanggi and Waggoner showed that horses responded to colour in the same manner as some red/green colour blind humans.
 
A myth that surfaces repeatedly is that horses cannot recognize what they have seen with one eye with the other eye. Anatomical dissections have proven that the horse's two cerebral hemispheres do indeed have a functional pathway to convey information, and a behavioural study proved that they do have inter ocular transfer. This study used multiple two choice discrimination tests with one eye blindfolded. Once the discrimination was learned, the blindfold was placed over the other eye. Horses immediately responded to the same stimulus, clearly demonstrating interocular transfer.
 
Horses often startle at an object they have seen before when approaching it from a different angle. This can be explained by the fact that the shape of the object appears differently when seen from another perspective. Experiments with different objects demonstrated that recognition of rotated objects was good under some but not all conditions.  
 
With generalization a behaviour previously conditioned to one stimulus, transfers to other similar stimuli. Riding school horses regularly utilize this form of learning when making sense of a large assortment of inexact hand, leg and seat cues from riders of varying skill levels.  Dressage horses, in contrast, learn to discriminate highly precise cues from their riders. As a result they go through mechanical motions that rarely enhance cognitive skills. A horse's generalization abilities can be enhanced by incorporating variety into their programs.  As social animals, horses are most comfortable in the company of their own kind. Social interactions can facilitate the learning of new behaviour by observing other horses
 
The cognitive abilities of horses need to be understood to ensure good husbandry and treatment.
It is as harmful to confine a thinking animal to a dark, dusty stable with little or no social interaction as it is to provide inadequate feed or use abusive training methods. It is certainly in the best interests of both horses and handlers to better understand the scope of equine thinking, and enrich horses' environment. In the not too distant past little consideration was given to why horses behaved like they did. This limited the provision of adequate care and welfare. More recently research into equine cognition and perception has made noteworthy advances. This has led to a greater interest in training methodology and management. Even so, we are still a long way from fully understanding what it is to be a horse. More studies are needed to fully understand this remarkable animal and provide it with the best environment and training for its and our needs.