ORIGINAL ARTICLES AND REVIEWS
Paolo BaragliI; Barbara PadalinoII, III; Angelo TelatinIV
IDipartimento di Scienze Veterinarie, Università degli Studi di Pisa, Pisa, Italy
IIDipartimento di Medicina Veterinaria, Università degli Studi Aldo Moro, Bari, Italy
IIIFaculty of Veterinary Science, Sydney University, NSW, Australia
IVDepartment of Equine Science and Management, Delaware Valley University, Doylestown, PA, USA
Horses were domesticated 6000 years ago and since then different types of approaches have been developed to enhance the horse's wellbeing and the human-horse relationship. Even though horse training is an increasingly important research area and many articles have been published on the subject, equitation is still the sport with the highest rate of human injuries, and a significant percentage of horses are sold or slaughtered due to behavioral problems. One explanation for this data is that the human-horse relationship is complex and the communication between humans and horses has not yet been accurately developed. Thus, this review addresses correct horse training based on scientific knowledge in animal learning and psychology. Specifically, it starts from the basic communication between humans and horses and then focuses on associative and non-associative learning, with many practical outcomes in horse management from the ground and under saddle. Finally, it highlights the common mistakes in the use of negative reinforcement, as well as all the implications that improper training could have on horse welfare. Increased levels of competence in horse training could be useful for equine technicians, owners, breeders, veterinarians, and scientists, in order to safeguard horse welfare, and also to reduce the number of human injuries and economic loss for civil society and the public health system.
Key words:animal learning, training, horse, welfare
In recent decades, the horse has been used in a variety of roles that is unparalleled to other species. Horses are considered workforce and a source of meat in developing countries, whereas in Western society the horse has become a companion animal used for leisure and sports, developing an industry estimated at tens of billions of euros in Europe alone . In addition, the horse has recently played an important role in Animal-Assisted Activities and Therapies [2, 3].
There is a widespread notion regarding horse and human interaction that the horse should be submissive and docile to humans . It is believed that during the introductory training process, the domination by a person and the subjugation of the horse are essential for success . Taming techniques have two common elements: isolation of the horse from its peers, and ensuring that the horse accepts its attempts of avoidance and escape will be unsuccessful . These elements are present even without apparent use of coercion, as in natural horsemanship (codified techniques of horse training, developed in the USA, and known as round pen techniques from the name of the circular enclosure where they are generally implemented) that seem, to the layman's eye, more ethically acceptable [7, 8].
Even though horses have been domesticated for more than 6000 years and different types of taming have been developed, horseback riding is the sport/leisure activity with the highest rate of human injuries [9-11]. Furthermore, many horses are lost due to the onset of behavioral problems that make them unfit to perform the task for which they are bred [12, 13]. Consequently, a high percentage of horses are slaughtered at a young age or before reaching normal retirement age . In the USA, the percentage of horses slaughtered for behavioral problems is about 10% . Consequently, there is a significant loss of economic resources for civil society and the equestrian industry; this is also indicative of compromised conditions of physical and mental well-being for horses . Very recently it has been demonstrated that horses in these types of conditions can show depressive-like behaviors  and learned helplessness could be a consequence of several years of improper management . For instance, older horses used for activities that require calm and non-reactive animals (i.e. riding schools for beginners or equine assisted activities and therapy) may experience negative mood states, although they appear calm and show little behavioral reactivity. All of these issues suggest that the developed communication between human and horse is not yet correctly refined .
Furthermore, while the influence of emotional and cognitive performance is widely studied in the human athlete , demonstrating that mental health and well-being are essential to good performance , this field is still lacking research in horses. Within equitation activities (sports or leisure or animal-assisted activity and therapy), two biologically and evolutionarily different living beings (horse and human) need to collaborate and communicate to achieve positive results . Competitive horseback riding is the only Olympic sport with this unique requirement. It becomes evident that in order to obtain good results within the horse-human relationship, it is essential to consider the psychology of the horse and its emotional involvement. Riders and trainers should carefully take into account all factors affecting the welfare of their horses , especially in the early stages of training when psychological rules of animal learning are critical to develop a balanced relationship with humans for both riding and handling purposes.
This article critically reviews the most recent published papers to provide a comprehensive, detailed, and deeper understanding on how associative and non-associative learning functions when working with horses. Despite the fact that some review articles have been already published on this topic, this review highlights some examples of practical outcomes in training from the ground and under saddle. This was done to underline the application of learning theory in practice; this review may become a useful tool in horse training and management. The non-associative and associative learning rules (see Table 1 for the most relevant definitions) should be deeply understood by all individuals working with horses, not only to improve the physical and psychological welfare of horses, but also to reduce accidents and the number of horses lost to behavioral problems caused by an improper relationship with a person.
The key to resolving these issues and improving the welfare of horses is recognizing the lack of the use of scientific knowledge in the horse-human relationship, regardless of the activity in which they are involved (sports, leisure, assisted activity and therapy, etc.) .
In recent years, studies on equine behavior have made significant progress. Horses recognize each other individually  and recognize humans . They possess sophisticated cognitive strategies to maintain social equilibrium  and have a long  and short term memory capacity . Furthermore, recently patterns of intraspecific communication have been highlighted in horses . Therefore, the complexity of the horse's mind and behavior indicates that learning skills of horses are qualitatively the same as those of other domestic mammals. Once professional and leisure riders are able to properly apply the rules of animal learning to horse training and handling, their scientific knowledge should also be expanded to these more complex cognitive aspects related to horse-human relationships and stable environments, not included in the topics of this review.
This article starts by outlining proper communication with horses through the use of pressure, then presents how rules of learning theory should be applied to the training of horses, explaining how positive and negative reinforcements work. Practical procedures and outcomes are then suggested for handling horses from the ground and under saddle. The review then addresses how classical conditioning may help to reduce stimuli administered to the horse. Finally, the authors discuss common mistakes in equine training, such as the risk of habituation to pressure stimuli and the dangerous boundary between the correct application of negative reinforcement and punishment.
COMMUNICATION WITH THE HORSE THROUGH PRESSURE
Practical experience tells us that horses often exhibit unexpected and defensive behavior without a plausible explanation. There is the possibility that these behaviors are simply due to communicative signals performed by the human, signals that humans are often unaware of giving to the horse.
The actions performed by humans with the legs on the horse's side, hands on the reins, and body position in the saddle affect the ridden behavior of the horse . Human actions also affect the horse when led from the ground with a halter or during handling for medication or clinical investigations. During these activities, most gestures of the person are physical actions transmitted to the horse, thus representing a form of physical communication. Every time a rein is pulled, the whip is used, the horse is touched with the rider's leg or led by hand with the halter, the mechanical action performed with the hands or other parts of the human body is transmitted to the horse's body [30-32]. Essentially, physical forces are applied to induce behavioral responses in horses and become pressures, stimuli on the body of the horse, causing change in its behavior . For example, simply inserting a finger between the horse's skin and the various instruments (i.e. bit, noseband, bridle, saddle, twitch, curb chain, bit-less...) is a useful and easy exercise to understand the pressure exerted on the horse's body by every action performed by humans . These pressures can be measured in some cases [30, 35] and are often amplified by the use of martingales, draw reins and other training devices . The pressure applied on the body of the horse (Figure 1) is an aversive stimulus, which creates discomfort . This discomfort can be of varying intensity, from very light and almost imperceptible at the sensory level, to very strong and eventually painful if legs and hands are used improperly [31, 38]. The pressure (hence the discomfort) is then removed when the horse performs the desired behavior. The horse stops when the rider pulls the rein because this behavior helps it to reduce the pressure on the muzzle, nose, or mouth . Therefore, during training, horses learn that pressure applied by the rider through the reins disappears when they stop or slow down the pace, the pressure of the rider's legs disappears when they move forward, and so on.
THE TRAINING OF HORSES
To be effective and ethologically correct, training should include the application of pressure and its immediate removal when the horse behaves as requested . When this does not happen and the release of pressure (the removal of discomfort) is performed incorrectly, horses may develop conflict and defensive behaviors, making it difficult to manage them. This mainly occurs when pressure becomes pain [39, 40]. These distinctions are particularly important in the early stages of training. Despite the large number of sports and recreational activities in which horses are involved, there are four basic responses on which the training of horses is built: forward, stop, turn on the forehand, and turn on the haunches . At the beginning of the training, which is the critical phase of the proper relationship between horse and human, horses learn these basic responses through negative reinforcement . Most actions that take place in handling a horse from the ground and from the saddle are related to the use of negative reinforcement . Essentially, the training is based on the capabilities of horses to avoid discomfort  and negative reinforcement can also be used without touching the horse. An example of this is the use of fear as an aversive stimulus, which results in the horse galloping around the perimeter of the round pen .
The main purpose of training is to provide signals that the horse learns to associate with a specific behavior. When this does not happen and the horse performs unexpected or unwanted behaviors, this result means that the training is not adequate in establishing the association between the desired signal (the pressure) and the desired behavior, and the reinforcement used to create this association is insufficient in overcoming the horse's motivation to perform other behaviors . Therefore, unexpected and defensive behavior may be the result of communication signals that are unclear to the horse . Even so, in the face of many behavioral responses required by training, the points of the horse's body where riders can apply pressure are relatively few . That is, during the training process the rider has at his disposal a limited vocabulary for asking the horse a considerable number of behavioral responses. As discussed, the communication between human and horse from the ground and from the saddle takes place through the application of pressure, administered with the aid (either natural: such as hands, legs or seat, or artificial: spurs, whip and other training devices). This communication is actually a binary language of Yes, this is what I want (release of pressure) and No, this is not what I want (application of pressure). The question is how can we effectively communicate with the horse using only Yes and No? Take, for example, the game of finding a hidden object following the hot (Yes) or cold (No) clues, depending on whether the person playing moves closer to the goal. The secret to finding the object lies in the skill of the person who is driving the game by saying hot (Yes) as soon as the subject makes the slightest movement in the correct direction and then continues to be more and more accurate. If the person who is driving the game is not ready to say hot (Yes) at the slightest hint of movement in the correct direction, the subject will move based on chance. This lack of clarity with orders will lead to confusion and frustration, as most likely happens for horses . Therefore, the key to horses' welfare is to understand how the reinforcement works, especially the timing when the pressure applied by the rider or caretaker must be released (or the food provided) during training and handling management.
Positive and negative reinforcement in the training of horses
There are two main types of reinforcement: positive and negative (Table 1). Reinforcement, by definition, is always advantageous, because it has a value for the animal. Depending on the context of the training and its goals, positive and negative reinforcement can be easily implemented from the ground, whereas when working from the saddle, riders cannot adequately apply positive reinforcement using food [6, 45]. Despite these issues, positive reinforcement should be considered while handling the horse from the ground and a mixed strategy (using both positive and negative reinforcements) could present a balanced solution for considering the horse's welfare .
The use of aversive stimuli has been common practice when teaching horses to load on a trailer or correcting loading issues. Some halters and head-collars, for example, are designed to exert pressure at various points on the head of the horse when it pulls away from the caretaker . Further practical evidence suggests wrapping a long rope around the hindquarters of a non-collaborative horse when attempting to load it, or prodding the horse with the bristles of broom, in order to pressure it from behind. Subsequent correct responses of approach to the loading ramp are therefore negatively reinforced as they result in the termination of these pressures.
Negatively reinforcing forward motion by removing an aversive stimulus (pressure) is effective only if it can be performed consistently and with correct timing. However, a number of other potentially dangerous avoidance behaviors (such as rearing or bolting) that do not result in loading could also lead to the termination of pressure for the horse . The risk is that an inexperienced trainer could negatively reinforce incorrect or dangerous behaviors, thus cultivating subsequent loading problems. Furthermore, increasing aversive stimuli is likely to induce habituation, which leads to a reduction in response . This would render strategies employed ineffective and potentially lead to abuse of the animal . Overall, the aversive treatment of problematic behavior (through negative reinforcement), such as trailer loading, could lead to further undesirable behaviors and could be inadequate in overcoming common problems . Animals trained through positive reinforcement seem to be easier to load and appear less stressed . Thus, it is essential that aversive (i.e. negative reinforcement) and non-aversive (i.e. positive reinforcement) training methods are balanced in order to optimize learning abilities of horses and improve their welfare  and to evaluate learning performance people should take into account visual attention of horses .
In a comparative study between positive and negative reinforcement training, horses trained with positive reinforcement were more likely to display the target behavior and show increased motivation to participate in training activities . Therefore, applying positive reinforcement during common handling procedures could lead horses to better cope with training-induced stress , mainly from the ground. Horses can be trained to perform various tests to obtain a food-based reinforcement, such as pressing a lever , completing a maze , and choosing between containers of different colors [55, 56] or forms . The same principles can be applied to training a horse to load on a trailer by means of positive reinforcement .
Handling the horse from the ground (with practical outcomes)
When a horse performs an undesirable behavior, instead of attributing anthropomorphic adjectives (evil, vindictive, submissive, dominant, etc.) to it, we should remember that horses tend to repeat any behavior that precedes a reward such as food, or relief from discomfort with greater intensity and frequency . Therefore, firstly the reinforcements (often unobvious) that led the horse to behave that way should be determined. With the application of a sudden pressure, the horse can be induced to display a defensive or escape behavior to stop it. The more unexpected and severe the pressure is, the greater the possibility that the horse will perform one of these behaviors [6, 58].
Most of the actions by the veterinarian put pressure on the horse, often of high intensity. For example, an injection is a sudden and high intensity pressure, exceeding the threshold of pain. The insertion of the needle (stimulus), especially initially, can induce the horse to perform defensive behaviors, which may be mild (moving the muscles of the neck, shaking the head, etc.) or strong (rearing, etc.) (response). If one of these behaviors during the injection causes the needle to slip off the skin, the horse learns that this behavior functions to remove the discomfort created by the needle (reinforcement), so the possibility that this behavior will be repeated by the horse at the next injection becomes high (Figure 2). By applying the rules of associative learning, we can train the horse to perform a more functional behavior, rather than a defensive behavior (dangerous for both horse and human) . In the case of injections, this would be the behavior of standing still. If the horse reacts to the injection, instead of immediately inserting the needle, the handler should tap the horse's neck continuously with a finger, simulating the syringe and continuing this pressure as soon as the horse's first movement occurs. As a result the horse will not associate his movement with the prevention of the injection. As soon as the horse stands still, at least for a moment, the handler should stop tapping the finger on its neck. With a sufficient number of repetitions of this sequence, the horse will learn that the behavior which stops the simulated pressure of the syringe is standing still (Figure 2). Another example involves horses that shake their heads, moving and trying to rear during clipping. In this case it is non-associative learning (habituation) and negative reinforcement that regulate the behavioral response . Through habituation, in most cases horses learn that the passing of the clipper on the body is not an unpleasant stimulus. For this to happen, it is important that the horse stands still while clipping; but if it shakes, moves to the side, or performs other aversive behaviors, and the clippers are stopped in that moment, the horse will learn that there are behaviors that it can perform to stop the clipping. In other words, when the clipping begins and the horse moves, the handler should continue the clipping without interruption. When none of the horse's aversive behaviors cause the clipping to stop, the horse becomes habituated to the clippers (which, contrary to the syringe, usually does not induce pain) and thus, stops reacting. This sequence refers to the application of non-associative learning and works well with horses clipped for the first time, especially when clipping the thorax and abdomen, because even if the horse moves, generally the contact with its body can be maintained. The process becomes more complicated when the clipping gets closer to the head or legs. While clipping these areas (stimulus), the horse is able to get away with a sudden movement (response) and stop the action of the clipper (reinforcement). If this occurs, even with the continuation of the work of desensitization, it will be difficult to solve the problem because, through operant conditioning, the horse has just learned a behavior to avoid the action of the clipper. To prevent the occurrence of this problem, it is necessary to employ negative reinforcement, releasing the pressure (i.e., by interrupting the action of the clipper) at the proper time. The handler should turn on the clippers and get closer to the horse. As soon as the horse begins to show defensive behaviors, even minor ones, the handler should stop and maintain the position with the clippers switched on. When the horse stands still, the handler should immediately turn off the clippers. This sequence needs to be repeated several times until the horse associates the position (standing still) with the action of turning the clippers off. It is important to turn off the clippers often when the horse is still to reinforce its behavior and clip intermittently, rather than continuously as when the horse is already accustomed to standing still. However, once the association has taken place, the handler should simply turn off the clippers at intermittent periods during clipping so the horse will stand still for a longer time while the clippers are running, because it has learned that eventually the clippers will stop (according to the rules of variable schedule of reinforcement). It is important to remember that correct application of the rules of animal learning requires concentration on what the horse is doing and what the handler is doing. To correctly reinforce a behavior, the handler must intervene in a span of a very few seconds (the closer the reinforcement, i.e. the removal of the pressure, is to the behavior, the greater the possibility of creating the association) immediately after the horse performed the correct behavior .
Rules of animal learning can also be useful in the stable for horses that kick stall doors, bite bars, and/or attack people and neighbors when approaching feeding time. Keeping a horse in the stable alters its natural behavior [61, 62]. In the stall, the horse stands still for most of the day, often without social contact, and eats for a shorter period than when out on pasture. This type of management can lead to the onset of destructive and abnormal behaviors , which offset natural behaviors such as walking and chewing and, with the passage of time, can develop into fixed behaviors, becoming compulsive and then stereotyped . In natural life there are different patterns of behavior related to foraging (scratching with the foreleg, threatening herd members, etc.) that the horse tends to repeat in the stall, where the situation is exacerbated by the fact that food is available only for short time periods during the day . This may lead to the occurrence of these behaviors (scratching, kicking, biting, etc.) at every mealtime. It may happen that the caretaker administering the food does not pay attention to the horse's behavior. In these cases, operant conditioning teaches us that if we put the food into the feed bucket when the horse is performing a destructive behavior (with perfect, but unintended, timing of reinforcement), that behavior is then reinforced. For example, if we put the food in when the horse is kicking, we are teaching it that a kick is rewarded with food. With time, this association will become fixed and the horse will use more and more of this behavior as the meal approaches. The unaware action of the caretaker (who neglects to notice behavior of the horse while feeding it) means that, once learned, these behaviors are reinforced with an intermittent reinforcement variable ratio, which is one of the strongest in maintaining fixed behavior because it introduces a certain degree of unpredictability in the arrival of the reinforcement . This is similar to the case of the human being who plays slot machines, where an award intermittently and by chance from the behavior of pulling the lever is sufficient to induce people to repeat this movement regularly .
The best strategy to preventing the horse from learning these types of behaviors is a knowledge base of associative learning rules and avoiding unconsciously reinforcing them. However, if these undesired behaviors are unconsciously reinforced, the intermittent positive reinforcement of other behavior through operant conditioning techniques allows us to teach new behaviors that can replace the unwanted ones .
Working under saddle (with practical outcomes)
The rules of learning are the same with mounted horses. For example, we can analyze how to teach the horse to leg yield. The rider applies the pressure of the inner leg while resisting with reins. At this point the horse will probably begin to move forward, as this is the previously learned behavior associated with the pressure of the leg. This time, instead of releasing the pressure, the rider will maintain it (No, this is not what I want) until the horse will show even minimal lateral movement. At that time the rider will immediately release the pressure (Yes, this is what I want). The rider will then have properly reinforced the slightest lateral movement of the horse through the chain: stimulus (pressure) response (lateral movement) reinforcement (immediate release of pressure). What errors may be committed? It is possible that the first time the pressure is applied, the horse will perform other behaviors (shakings its head, trying to bite the rider's leg, or, in extreme cases bucking or rearing). The horse may perform these behaviors to stop the pressure from the reins and legs before the desired behavior (to perform a lateral step) . If one of these incorrect behaviors results in removal of the pressure, the horse will learn that bucking or rearing relieves it from the pressure (Figure 3). There are other errors the rider can commit. If the horse is believed to be performing the leg yield because the pressure of the inner leg moves it to the side, the rider will continue to press while it's moving. In this way the rider is maintaining the pressure while the horse performs the desired behavior. Therefore, despite the horse giving the correct answer, the rider is saying to it, No, this is not what I want. This will probably create confusion to the horse, pushing it to try other behaviors to gain relief from the pressure. These defensive behaviors will be shown with greater frequency and intensity when the same pressures arise again in the future (for an overview about behavior of mounted horses see [32, 35, 66, 67]). It is essential to underline that horses are unlikely to be able to assimilate all the components of a complex behavior at once. Therefore, the training process requires putting together a step-by-step sequence comprised of single elements of the final complex behavior required, with successive approximations until the final result is achieved, following the shaping procedure .
THE POWER OF CLASSICAL CONDITIONING AND ITS IMPLICATION FOR THE WELFARE OF HORSES
Through classical conditioning, a stimulus that originally has no meaning for the animal is associated with an unconditioned stimulus that can be used as a reinforcement (or punishment). This means that once a behavior has been learned by negative reinforcement, the horse will use other stimuli to predict the arrival of the reinforcement, i.e. the removal of discomfort. Riders use classical conditioning when replacing pressure stimuli used in the initial stages of training with new stimuli that were initially neutral stimuli (they evoked no response) . This is called secondary reinforcement-stimuli which can be words, gestures, or other pressure, such as the use of seat position . This could have welfare implications for horses because the classical conditioning permits the rider to substitute the adverse stimuli of the negative reinforcement sequence with the use of light pressure or other non-invasive stimuli .
Due to classical conditioning associations, a horse can also show defensive or escape behaviors in an unexpected way without seemingly plausible reason . When defensive behaviors have been learned, the horse can, through classical conditioning, predict the injection event because it will have associated a series of stimuli (a certain person who enters the stall, the smell, the sight of the syringe, etc.) that predicts the injection and it will show defensive behavior in advance. The same can happen when working under saddle. If a horse has learned to rear because of errors in the removal of the pressure by the rider during the application of negative reinforcement, it will associate other stimuli as prediction of the pressure's appearance through classical conditioning (even the simple entry into a competition or training arena) , and it will display the behavior of rearing before the onset of pressure that caused it has been reached. The lack of knowledge of these rules could be a source of behavior problems, because the stimulus that the horse chooses to predict the arrival of a pressure that has been already associated with defensive behaviors may be an unrecognizable stimulus, and therefore beyond our control. These defensive behaviors without notice, in advance and without apparent reason, are also called anticipatory behavior .
Some variants of classical conditioning are blocking and overshadowing. They indicate how the simple coupling of a stimulus with reinforcement is not sufficient to produce the necessary associative learning, but rather that the animal makes a choice regarding which stimulus best predicts the arrival of the reinforcement . These phenomena are very important for horses' training, both while working under saddle and from the ground [49, 71]. The more stimuli (pressures) that are concurrently given by the rider, the greater the chance is that the horse performs its selection, regardless of the will of the rider, about the stimulus which best predicts the removal of the pressure and implements behaviors that are unexpected to the rider . An example can be seen in the canter departure that is requested with light pressure from the leg aids . While applying pressure with legs, the rider usually perform a thrusting movement with the pelvis on the saddle that also produces light pressure. By the blocking/overshadowing (in this case it is difficult to say which of the two singularities is actually involved), the horse could choose to respond to the pelvic movement and fail to pick up the canter from the leg aids. As a result the rider, trying to reestablish the correct negative reinforcement sequence to obtain the canter departure, may unknowingly think that the horse is unresponsive to the leg aid and consequently increases the leg pressure with the use of spurs or a whip, often exceeding the horse's threshold of pain. Similar situations to this one may lead to training the horse in a frequent state of pain, causing welfare concerns.
THE RISK OF THE HABITUATION
Non-associative learning is widely used in training to ensure a decrease in the frequency and intensity of undesirable behaviors performed by the horse, such as escape and defensive behaviors, and to habituate to new objects and environments [22, 72]. In horses, a common response to unfamiliar situations is to put distance between themselves and potential dangers by avoidance behavior . With repeated exposure, horses habituate to their surroundings and cease to avoid nonthreatening stimuli . For example, a naive horse may shy away when touched by a novel stimulus (such as a saddle pad or splint boots), but the response will diminish with repeated exposure . In order to effectively train a horse with habituation, the stimulus offered must be equal and constant throughout the process . If the stimulus changes shape, intensity, or presentation modality, the horse will react as in the presence of a new stimulus and the process of habituation will be invalidated. It appears that horses either adopt habituation (reduced behavioral response) or sensitization (increased behavioral response) to repeated exposure of a novel stimulus .
Through non-associative learning, the horse can also become habituated to pressure, provided that it is light (because more pressure results in greater discomfort, with a greater intensity of discomfort, it is less likely that the horse will become habituated) . This introduces the risk of riding a horse on contact, which is pressure. (In order to communicate with the horse, we need to establish a low baseline of contact [light pressure]. The horse will accept a consistent, light pressure and consider this a new baseline.) However, it is essential to know the result that constant pressure on the mouth (or on other parts of its muzzle based on the bridle used) has on the horse. Without knowledge of non-associative learning and without the necessary attention to timing, a rider may not be able to adequately assess the extent of contact with the horse's mouth and without intent, can apply pressure of some intensity . A rider could consistently apply a level of pressure similar to one the horse has associated with the stimulus to certain behaviors, but without consciously giving any command. As the work proceeds, the horse would then become habituated to this constant pressure on its mouth, creating a new, higher baseline of pressure, which would result in the pressure losing its effect . The similar intensity of two different pressure stimuli (one of which is used for communication) is a risk because the horse may lose the ability to distinguish between them . Thus, it may not be able to perceive the difference between the pressure signal associated with a specific behavior and the baseline contact with the mouth, which instead has no meaning .
THE REINFORCEMENT AND THE PUNISHMENT: A NOT WELL DEFINED BOUNDARY
As previously mentioned, negative reinforcement is the removal of an aversive stimulus (that creates discomfort) from the horse. Any aversive stimulus (including pressure implemented by humans) is nothing more than a punishment, even when light . In the human mind, the word punishment conjures up negative scenarios, relating the use of whips or other means of coercion, but it should not be understood as simply inflicting physical pain or mental coercion. Light punishment is employed every time pressure is applied to the horse (No, this is not what I want), the removal of which (Yes, this is what I want) allows implementation of negative reinforcement. It is essential to highlight this concept because only awareness of our behavior allows us to apply negative reinforcement correctly. Through a total understanding of the difference between negative reinforcement and positive punishment, we see that if the pressure is not removed immediately after the horse performs the correct behavior, it would actually result in a punishment for that behavior by saying to the horse No, this is not what I want. Therefore, the non-application of the reinforcement (failure to remove pressure) can be interpreted by the horse as a punishment  thus indicating how thin the line is between proper communication and sending contradictory signals. To eliminate an undesired behavior of a horse, it is important to know the psychological rules of the punishments and consequences to avoid errors in their application. The effectiveness of the punishment depends partially on its intensity. This does not mean that full strength should be used, but rather that the intensity should be suitable for that specific horse at that specific time (any horse, like any person, has a threshold at which the punishment is effective). The other critical element is the time interval between the behavior that needs to be punished and the administration of the punishment. The more time that passes between the two, the greater the chance that the punishment is ineffective and has only undesirable consequences. As for correct application of reinforcement, learning is effective only within a timeframe of a few seconds, outside of which it is likely that the punishment takes on a different meaning . For example, if the horse is whipped following a refusal of a fence, it may not know whether to associate the stimulus with the refusal itself or with some other part of the ongoing situation. The horse could associate the punishment with the fence, the jumping, or even the riding. A very high percentage of punishments administered as such are incorrectly applied regarding timing and intensity that is, without the consideration of all conditions in which punishments are unconsciously administered . The punishment can quickly be associated with the person who performed it, and this will bring the horse to implement defensive or escape behaviors as soon as this person is present in the training process. In the same way, that association can be extended to places or objects . Repeated punishment with the wrong timing of application  induces reduction of spontaneous activity, which is often associated with decreased attention and learning .
The reduction of spontaneous activity can lead to a negative perception of the environment even if the horse is stabled appropriately for its needs, consequently reducing the welfare state. These two conditions may result in pathological forms of apathy and an inability to adapt to the environment, that could lead horses toward depressive-like disorders [17, 18, 48]. Finally, the most dangerous effect induced by punishment is aggressiveness toward humans . Therefore, for all the aforementioned reasons, positive punishment should be avoided as much as possible.
In conclusion, to ensure an adequate level of equine welfare, it is necessary that the psychological rules of animal learning are applied during training and management. This will lead to positive effects for all sports and activities in which the horse is employed. Furthermore, this will also reduce the possibility of injuries for humans involved in the equestrian industry, with a significant reduction of costs for civic society and the national health system. This review summarized the actual scientific and practical knowledge of learning theory to practically apply them in the field. These concepts could be easily taught; therefore, the authors suggest that all individuals who work or manage horses by profession or passion, should attend specific training programs on psychological rules of animal learning in order to apply this knowledge during their daily work with horses.
The authors are grateful to Cory Kieschnick (Equine Science Department, Delaware Valley College) and Ronald DePeter (Writing Center, Delaware Valley College, USA) for proofreading and English revision.
Conflict of interest statement
There are no potential conflicts of interest or any financial or personal relationships with other people or organizations that could inappropriately bias conduct and findings of this study.
1. Liljenstolpe C. Horses in Europe. Uppsala, Swedish University of Agricultural Sciences: 2009. Available from: www.wbfsh.org/files/EU%20Equus%202009.pdf.
2. Cerino S, Cirulli F, Chiarotti F, Seripa S. Non-conventional psychiatric rehabilitation in schizophrenia using therapeutic riding: the FISE multicentre Pindar project. Ann Ist Super Sanità 2011;47(4):409-14. DOI:10.4415/Ann_11_04_13
3. Cirulli F, Borgi M, Berry A, Francia N, Alleva E. Animal-assisted interventions as innovative tools for mental health. Ann Ist Super Sanità 2011;47(4):341-8. DOI: 10.4415/ANN_11_04_04.
4. McLean AN, McGreevy PD. Ethical equitation: Capping the price horses pay for human glory. J Vet Behav 2010;5:203-9. DOI:10.1016/j.jveb.2010.04.003
5. Goodwin D, McGreevy P, Waran N, McLean A. How equitation science can elucidate and refine horsemanship techniques. Vet J 2009;181:5-11. DOI:10.1016/j.tvjl.2009.03.023
6. Waran NK, Casey R. Horse training. In: Mills DS, McDonnell SM (Eds). The domestic horse: the origins, development and management of its behaviour. Cambridge: Cambridge University Press; 2005.
7. Warren-Smith AK, McGreevy PD. Preliminary investigations into the ethological relevance of round-pen (round-yard) training of horses. J Appl Anim Welfare Sci 2008;11:285-98. DOI:10.1080/10888700802101304
8. Henshall C, McGreevy PD. The role of ethology in round pen horse training- a review. Appl Anim Behav Sci 2014;155:1-11. DOI:10.1016/j.applanim.2014.03.004
9. Hausberger M, Roche E, Henry S, Visser K. A review of the humanhorse relationship. Appl Anim Behav Sci 2008;109:1-24. DOI:10.1016/j.applanim.2007.04.015
10. Hawson LA, McLean AN, McGreevy PD. The roles of equine ethology and applied learning theory in horse-related human injuries. J Vet Behav 2010;5:324-38. DOI:10.1016/j.jveb.2010.06.001
11. Lyn CY, Wright J, Bushnik T, Shem K. Traumatic spinal cord injuries in horseback riding. A 35-Year Review. Am J Sports Med 2011;39:2441-6. DOI:10.1177/0363546511419280
12. Ödberg FO, Bouissou MF. The development of equestrianism from the baroque period to the present day and its consequences for the welfare of horses. Equine Vet J 1999;28:26-30.
13. Hayek AR, Jones B, Evans DL, Thomson PC, McGreevy PD. Epidemiology of horses leaving the thoroughbred and standardbred racing industries. In: 1st International Equitation Science Symposium, Broadford, Victoria, Post- Graduate Foundation in Veterinary Science, Proceedings. Sydney: 2005. p. 84-9.
14. Wallin L, Strandberg E, Philipsson J and Dalin G. Estimates of longevity and causes of culling and death in Swedish warmblood and coldblood horses. Livest Prod Sci 2000;63:275-89.
15. Grandin T, McGee K, Lanier J. Prevalence of severe welfare problems in horses that arrive at slaughter plants. J Am Vet Med Ass 1999;214:1531-3.
16. Hausberger M, Gautier E, Biquand V, Lunel C, Jégo P (2009). Could work be a source of behavioural disorders? A study in horses. PloS One 4:e7625.
17. Fureix C, Jego P, Henry S, Lansade L, Hausberger M. Towards an ethological animal model of depression? A study on horses. PLoS ONE 2012;7(6):e39280. DOI:10.1371/journal.pone.0039280
18. Baragli P, Banti L, Vitale V, Sighieri C. Effect of aging on behavioural and physiological responses to a stressful stimulus in horses (Equus caballus). Behav 2014;151:1513-33. DOI:10.1163/1568539X-00003197
19. McCarthy PJ. Positive emotion in sport performance: current status and future directions. Int Rev Sport Exerc Psychol 2011;4(1):50-69. DOI:10.1080/1750984X.2011.560955
20. Adie JW, Duda JL, Ntoumanis, N. Achievement goals, competition appraisals, and the psychological and emotional welfare of sport participants. J Sport Exerc Psychol 2008;30:302-22.
21. Keaveney SM. Equines and their human companions. J Bus Res 2008;61:444-54. DOI:10.1016/j.jbusres.2007.07.017
22. McBride SD, Mills DS. Psychological factors affecting equine performance. BMC Vet Res 2012;8:180-91. DOI:10.1186/1746-6148-8-180. DOI:10.1186/1746-6148-8-180
23. Proops L, McComb K, Reby D. Cross-modal individual recognition in domestic horses (Equus caballus). Proc Nat Acad Sci of the United States of America 2009a;106:947-51. DOI: 10.1073/pnas.0809127105
24. Proops L, McComb K. Cross-modal individual recognition in domestic horses (Equus caballus) extends to familiar humans. Proc Roy Soc B 2012;279:3131-8. DOI:10.1098/rspb.2012.0626
25. Cozzi A, Sighieri C, Gazzano A, Nicol CJ, Baragli P. Post-conflict friendly reunion in a permanent group of horses (Equus caballus). Behav Process 2010;85:185-90. DOI:10.1016/j.beproc.2010.07.007
26. Hanggi EB, Ingersoll JF. Long-term memory for categories and concepts in horses (Equus caballus). Anim Cogn 2009;12:451-62. DOI:10.1007/s10071-008-0205-9
27. Baragli P, Vitale V, Paoletti E, Mengoli M, Sighieri C. Encoding the object position for assessment of short term spatial memory in horses (Equus caballus). Int J Comp Psych 2011;24:284-91.
28. Wathan J, McComb K. The eyes and ears are visual indicators of attention in domestic horses. Curr Biol 2014;24(15): R677-9. DOI:10.1016/j.cub.2014.06.023
29. Mills DS. Applying learning theory to the management of the horse: the difference between getting it right and getting it wrong. Equine Vet J Suppl 1998;27:44-8.
30. Clayton HM, Singleton WH, Lanovaz JL, Cloud GL. Measurement of rein tension during horseback riding using strain gauge transducers. Exp Techniques 2003;27:34-6. DOI:10.1111/j.1747-1567.2003.tb00112.x
31. Newton A, Nielsen AM. A review of horse-related injuries in a rural Colorado hospital: implications for outreach education. J Emerg Nurs 2005;31:442-6. DOI:10.1016/j.jen.2005.08.009
32. Hall C, Huws N, White C, Taylor E, Owen H, McGreevy P. Assessment of ridden horse behavior. J Vet Behav 2013;8:62-73. DOI:10.1016/j.jveb.2012.05.005
33. McGreevy PD. The advent of equitation science. Vet J 2007;174:492-500. DOI:10.1016/j.tvjl.2006.09.008
34. Baragli P. I cavalli di Federico. Guida pratica di etologia applicata al cavallo. Pisa: University Press; 2010.
35. Egenvall A, Eisersiö M, Roepstorff L. Pilot study of behavior responses in young riding horses using 2 methods of making transitions from trot to walk. J Vet Behav 2012;7:157-68. DOI:10.1016/j.jveb.2011.08.006
36. Heleski C, McGreevy PD, Kaiser LJ, Lavagnino M, Tans E, Bello N, Clayton HM. Effects on behaviour and rein tension on horses ridden with or without martingales and rein inserts. Vet J 2009;181:56-62. DOI:10.1016/j.tvjl.2009.03.011
37. McCall CA. A review of learning behaviour in horses and its application in horse training. J Anim Sci 1990;68:75-81.
38. Lesimple C, Fureix C, Menguy H, Hausberger M. Human direct actions May Alter Animal Welfare, a Study on Horses (Equus caballus). PLoS ONE 2010;5(4):e10257. DOI:10.1371/journal.pone.0010257.
39. Lesimple C, Fureix C, De Margerie E, Sénèque E, Menguy H, Hausberger M. Towards a postural indicator of back pain in horses (Equus caballus). PLoS ONE 2012;7(9):e44604. DOI:10.1371/journal.pone.0044604
40. Lesimple C, Fureix C, Biquand V, Hausberger M. Comparison of clinical examinations of back disorders and humans' evaluation of back pain in riding school horses. BMC Vet Res 2013;9:209-18. DOI:10.1186/1746-6148-9-209
41. McLean AN, Mclean MM. Academic horse training: equitation science in practice. Australian Equine Behaviour Centre, Victoria, Australia; 2008.
42. McLean AN. The positive aspects of correct negative reinforcement. Anthrozoös 2005;18:245-54. DOI:10.2752/089279305785594072
43. McCall CA, Salters MA, Simpson SM. Relationship between number of conditioning trials per training session and avoidance learning in horses. Appl Anim Behav Sci 1993;36:291-9.
44. Richard-Yris MA, Hausberger M, Henry S. Bases éthologiques de l'apprentissage. In: Compte-rendu de la 30ème Journe'e de la Recherche Equine. Proceedings. Paris: 2004. p. 179-87.
45. Heleski C, Bauson L, Bello N. Evaluating the addition of positive reinforcement for learning a frightening task: a pilot study with horses. J Appl Anim Welf Sci 2008;11(3):213-22. DOI:10.1080/10888700802100942
46. Innes L, Mcbride S. Negative versus positive reinforcement: An evaluation of training strategies for rehabilitated horses. Appl Anim Behav Sci 2008;112:357-68. DOI:10.1016/j.applanim.2007.08.011
47. Slater C, Dymond S. Using differential reinforcement to improve equine welfare: shaping appropriate truck loading and feet handling. Behav Process 2011;86:329-39. DOI:10.1016/j.beproc.2011.02.001
48. Hall C, Goodwin D, Heleski C, Randle H, Waran N. Is there evidence of learned helplessness in horses? J Appl Anim Welfare Sci 2008;11:249-66. DOI:10.1080/10888700802101130
49. McGreevy PD, McLean AN. Roles of learning theory and ethology in equitation. J Vet Behav 2007;2:108-18.
50. Hendriksen P, Elmgreen K, Jan Ladewig J. Trailer-loading of horses: is there a difference between positive and negative reinforcement concerning effectiveness and stress-related signs? J Vet Behav 2011;6:261-6. DOI:10.1016/j.jveb.2011.02.007
51. Sankey C, Richard-Yris M, Henry S, Fureix C, Nassur F, Hausberger M. Reinforcement as a mediator of the perception of humans by horses (Equus caballus). Anim Cogn 2010;13:753-64. DOI:10.1007/s10071-010-0326-9
52. Rochais C, Henry S, Sankey C, Nassaur F, Gorecka-Bruzda A, Hausberger M. Visual attention, an indicator of human-animal relationship? A study of domestic horses (Equus caballus). Front Psychol 2014;5:1-10. DOI:10.3389/FPSYG. 2014.00108.
53. Dougherty DM, Lewis P. Stimulus-generalisation, discrimination-learning, and peak shift in horses. J Exp Anal Behav 1991;56:97-104.
54. Kratzer DD, Netherland WM, Pulse RE, Baker JP. Maze learning in Quarter horses. J Anim Sci 1977;45:896-902.
55. Clarke JV, Nicol CJ, Jones R, McGreevy PD. Effects of observational learning on food selection in horses. Appl Anim Behav Sci 1996;50:177-84.
56. Sappington BKF, McCall CA, Coleman DA, Kuhlers DL, Lishak RS. A preliminary study of the relationship between discrimination reversal learning and performance tasks in yearling and 2-year-old horses. Appl Anim Behav Sci 1997;53:157-66.
57. Proops L, Burden F, Osthaus B. Mule cognition: a case of hybrid vigour? Anim Cogn 2009b;12(1):75-84. DOI:10.1007/s10071-008-0172-1
58. Keeling LJ, Jonare L, Lanneborn L. Investigating horse-human interactions: The effect of a nervous human. Vet J 2009;181:70-1. DOI:10.1016/j.tvjl.2009.03.013
59. Telatin A. Confronto fra le tecniche di addestramento del cavallo e la psicologia dell'apprendimento. Padova: Tesi di Laurea, Facoltà di Scienze Fisiche, Matematiche e Naturali; 1995.
60. Hockenhull J, Creighton E. Training horses: positive reinforcement, positive punishment, and ridden behavior problems. J Vet Behav 2013;8(4):245-52. DOI:10.1016/j.jveb.2012.06.002
61. Baragli P, Paoletti E, Pacchini S, Martelli F, Sighieri C. Influenza di un protocollo di arricchimento ambientale sul comportamento dei cavalli in scuderia. Ippologia 2009;20(2):23-33.
62. Fureix C, Benhajali H, Henry S, Bruchet A, Prunier A, Ezzaouia M, Coste C, Hausberger M, Palme R, Jego P. Plasma cortisol and faecal cortisol metabolites concentrations in stereotypic and non-stereotypic horses: do stereotypic horses cope better with poor environmental conditions? BMC Vet Res 2013;9:3-13. DOI:10.1186/1746-6148-9-3
63. Cooper JJ, Albentosa M. Behavioural adaptation in the domestic horse: potential role of apparently abnormal responses including stereotypic behaviour. Livest Prod Sci 2005;92:177-82. DOI:10.1016/j.livprodsci.2004.11.017
64. Clark L, Limbrick-Oldfield EH. Disordered gambling: a behavioral addiction. Curr Opin Neurobiol 2013;23(4):655-9. DOI: 10.1016/j.conb.2013.01.004.
65. Fox AE, Bailey SR, Hall EG, St. Peter CC. Reduction of biting and chewing of horses using differential reinforcement of other behavior. Behav Process 2012;91:125-8. DOI:10.1016/j.beproc.2012.05.001
66. Christensen JW, Beekmans M, van Dalum M, Van Dierendonck M. Effects of hyperflexion on acute stress responses in ridden dressage horses. Physiol & Behav 2014;128:39-45. DOI:10.1016/j.physbeh.2014.01.024
67. Hall C, Kay R, Yarnell K. Assessing ridden horse behavior: professional judgment and physiological measures. J Vet Behav 2014;9:22-9. DOI:10.1016/j.jveb.2013.09.005
68. Cooper JJ. Comparative learning theory and its application in the training of horses. Equine Vet J Suppl 1998;27:39-43.
69. Peters SM, Bleijenberg EH, van Dierendonck MC, van der Harst JE, Spruijt BM. Characterization of anticipatory behaviour in domesticated horses (Equus caballus). Appl Anim Behav Sci 2012;138:60-9. DOI:10.1016/j.applanim.2012.01.018
70. Vallortigara G. Other Minds. Comparative Study of Animal Cognition. Bologna: Il Mulino; 2000.
71. McLean. Overshadowing: a silver lining to a dark cloud in horse training. J Appl Anim Welf Sci 2008;11(3):236-48. DOI:10.1080/10888700802101064
72. Christensen JW, Rundgren M, Olsson K. Training methods for horses: habituation to a frightening stimulus. Equine Vet J 2006;38(5):439-43. DOI:10.2746/042516406778400574
73. Nicol C. Learning abilities of the horse. In: Mills DS, McDonnell, SM (Eds). The Domestic Horse: The Origins, Development and Management of Behaviour. Cambridge: Cambridge University Press; 2005.
74. Leiner L, Fendt M. Behavioural fear and heart rate responses of horses after exposure to novel objects: Effects of habituation. Appl Anim Behav Sci 2011;131:104-9. DOI:10.1016/j.applanim.2011.02.004
75. Terada K, Clayton HM, Kato K. Stabilization of wrist position during horseback riding at trot. Equine Comp Exerc Physiol 2006;3:179-84. DOI:10.1017/S1478061506337255
76. McGreevy PD, McLean AN. Punishment in horse-training and the concept of ethical equitation. J Vet Behav 2009;4:193-7. DOI:10.1016/j.jveb.2008.08.001
77. Seligman MEP, Maier SF. Failure to escape traumatic shock. J Exp Psychol 1967;74:1-9.
78. Hinde R. Towards Understanding Relationships. Londres: Academic Press; 1979.
79. Barnlund DCA. Transactional model of communication. In: Mortensen CD (Eds). Communication theory; 2nd ed. New Brunswick, New Jersey: Transaction; 2008.
80. McGreevy PD, Boakes RA. Carrots and Sticks: Principles of Animal Training. Cambridge: Cambridge University Press; 2006.
81. Skinner BF. The Behaviour of Organisms. New York: Appleton-Century-Crofts; 1938.
82. Pavlov IP. Conditioned reflexes. Oxford: Oxford University Press; 1927.
83. Thorndike EL. An experimental study of the associate processes in animals. Psychol Rev 1898;2(4)1-8.
Address for correspondence:
Dipartimento di Scienze Veterinarie, Università di Pisa,
Viale delle Piagge 2, 56124 Pisa.
Received on 12 May 2014.
Accepted on 24 February 2015.