Weaving, Headshaking, Cribbing, and Other Stereotypies

1. Introduction

Many stable- and stall-housed horses perform a variety of apparently functionless, repetitive behaviors such as weaving, box-walking (stall-walking), cribbing (also known as crib-biting or windsucking--the latter term not to be confused with pneumovagina), and pawing. In the scientific literature, these behaviors have gone under many names including stereotypic activities, stereotypies, stereotypes, obsessive compulsive disorders, and compulsive disorders; in the popular literature, they are often referred to as vices or habits and might include behaviors like headshaking. Although it may be tempting to ascribe the motivation for these behaviors to general emotional states (such as boredom or frustration) or to poor temperament, it is now clear that specific causal factors exist for specific stereotypic activities in the horse. For convenience, this paper will refer to all these behaviors as stereotypies with the caveat that this does not imply that they necessarily share anything in common other than being repetitive behavior problems that lack an obvious cause. Behavior problems are often an unfamiliar field of practice for many veterinarians, and advice on such matters is often referred to individuals outside the profession. However, many involve medical conditions or at least, the exclusion of medical factors; therefore, it is essential that veterinarians appreciate the fundamental concepts involved, and their clinical skills should be applied to these matters.

Some of these behaviors may be considered an "unsoundness", which can reduce the sale value of the horse. This is probably because it is often thought that these behaviors cause ill-thrift or more specific health problems (e.g., weaving makes a horse more likely to suffer from tendon problems, or horses that crib or windsuck can swallow so much air that they are prone to colic). However, scientific evidence to support this theory is either absent or it points to other conclusions as discussed below. The equine industry should, however, be concerned that these behaviors are typically associated with a poor environment and psychological problems in other captive species and therefore, are a major welfare concern. Given the common lack of understanding of these problems, it is not surprising that many seek to control them by whatever means they can with varying success. These efforts at control can reach quite remarkable levels of cruelty, but this almost always reflects the owner's misunderstanding of the nature of the behavior rather than any ill intent on their part.

In recent years, there has been significant progress in our understanding of the nature of these problems through epidemiological, developmental and, experimental studies. One of the greatest imperatives for improved welfare is the wider dissemination of this knowledge within the horse community and veterinary profession. It is now recognized by welfare scientists that punishment or physical prevention of stereotypies is generally detrimental to the horse's well being without solving the underlying causal factors in the horse's environment.

2. Generalizations about Stereotypic Behavior in Horses

Are Stereotypic Behaviors Inherited?

Behavior is neither purely learned nor purely genetic but a combination of both. Instead, we should question really how significant genetic factors are in the problem.

This is not as simple a question to answer in practice, because it may seem that different breeds are often associated with different uses and management regimens. However, several studies suggest stereotypic behaviors are more common in Thoroughbreds and Warmbloods compared with Standardbreds. Analyses of breed records have also found uneven distributions of the prevalence of some of these behaviors in different families, which strongly suggests a heritable component. However, it must be recognized that it is not the behavior that is inherited but the tendency to perform the behavior. This only becomes apparent in the presence of other immediate and developmental risk factors such as those described below.

The Role of Diet and Feeding Practices

In several studies, diet and feeding regimen have been implicated as a general risk factor for the expression of abnormal behavior in the horse. Specifically, McGreevy et al. [1] found that feeding <6.8 kg of forage per day and feeding hay as opposed to other forage increased the risk of abnormal behavior, especially weaving and woodchewing. Offering forage more than three times a day was also associated with an increased risk of abnormal behavior in this study, but the effect of meal scheduling is quite complex and dependent on the particular behavior of concern. It often interacts in practice with other factors like time spent feeding; this may have the opposite effect to the arousal caused by the delivery of food, and generalizations about the effect of diet on these behaviors as a whole are generally unwise.

Social Influences on Stereotypies

It is widely believed by horse owners that stereotypies in common with other "stable vices" can be learned from other horses. However, the evidence for social learning is very poor, and it is likely that other processes may underlie this perception. For example, phenomena such as social facilitation, which is the release of an already established behavior in the presence of another engaging in it (this is not the same as imitation) could account for the sudden occurrence of the problem in one horse after the introduction of another expressing the behavior more overtly. This may be relevant, because there is some indirect evidence to suggest that some of these problems may become incorporated into the behavioral repertoire of horses earlier in life and then, may remain latent or unexpressed until later in life [2]. In theory, a cribbing horse might also facilitate the investigation of the top of the stall doors in neighboring horses (through a phenomenon known as stimulus enhancement). As a result, others independently engage in the behavior if it has self-reinforcing effects. This would only be expected to occur in a horse that is already pre-disposed to the behavior. By contrast, it should be recognized that the only demonstrable effect of the presence of a horse with a repetitive behavior problem on others is a reduction of weaving and headnodding [3]. Although social isolation may prevent the spread of undesirable activities through social learning, it may also increase the risk of developing these activities as a response to social deprivation.

Other General Risk Factors

Other factors found to influence the risk of developing abnormal repetitive behaviors include social rank (foals of dominant mares are at greater risk) and weaning (individual stall weaning increases the risk more than paddock- or group-weaning strategies) [4]. Time spent outside has been associated with an increased risk of repetitive behavior in dressage and eventing horses but not in endurance animals [5]. However, endurance horses tend to spend more time outside, confounding the comparison. As a simplification, it would seem that confinement for more than a given number of hours in the day (~4 h) may be the important factor. In addition, weaning and other potentially stressful events, such as the time when horses are started under saddle, seem to be important triggers for the behavior.

Despite some commonly held beliefs to the contrary, there is no evidence of an association between coat color and these behaviors in the horse [2].

In summary, genetic as well as environmental factors, especially around the time of weaning, have major non-specific effects on the risk of an individual developing some form of repetitive behavior. These factors probably alter the risk of stress-induced sensitization of the relevant pathways; however, specific factors are known to pre-dispose individuals to specific problem behaviors as discussed below.

3. Weaving and Box-Walking

Form and Prevalence

Weaving is an obvious lateral swaying movement of the head, neck, forequarters, and sometimes hindquarters. Across a range of surveys, the behavior is reported on average in just >3% of horses [6]. By contrast, box-walking describes the repetitive tracing of a circular route within the stall and is reported in just >2% of horses across all studies [6]. Both of these values are probably an underestimate of the true prevalence of the problems, but there is considerable variation between different populations. The two behaviors are often considered to be related, with some justification, because it has been found that if a horse either weaves or box-walks, it also is much more likely to perform the converse rather than some form of oral repetitive behavior [2]. However, there are clearly other factors that differentiate the expression of one locomotor behavior over another. For example, Luescher et al. [7] in their survey of stables in South-Western Ontario recorded a higher prevalence in Arabs than any other breed group for box-walking (7.32%), but this breed had low prevalences for weaving. No weaving was recorded in Standardbreds or ponies, and no box-walking was recorded in Quarter horses, Warmbloods, or ponies. In a longitudinal study over a 4-yr period, box-walking had a median age of onset of 64 wk, and weaving had a similar median age of onset at 60 wk. These values contrast with the median age of onset of cribbing, which was much earlier at 20 wk [4].

Although it is often said that weaving affects performance, there are no data to support this contention. Some would argue that the behavior puts extra strain on the tendons, but this is somewhat illogical. Horses naturally walk around all day, and the foot falls of the weaving horse are usually in the walk sequence. Thus, given the functional anatomy of the distal limb and its circulatory system, it might be that, compared with standing idle, weaving actually improves circulation to the structures of the distal end of the limbs and therefore, reduces the risk of injury. Because of the implications of these opposing claims, such as the justification of some potentially inhumane measures of control, it would be worth investigating this matter further.

Summary of classic presentation and clinical findings in a weaver:

• Weaving often precedes an arousing event.
• Close social contact is restricted but may have visual access to other horses across the walkway.
• Exercise requirement and daily food allowance are not properly matched.
• The management routine is highly predictable.
• The bedding is something other than straw.
• There is no lameness apparent.

Risk Factors and Treatment Opportunities

Housing

The housing environment is usually the focus of attention when considering these behaviors, and standing stalls are associated with a lower prevalence of weaving [7]. Weaving is also less common on larger (>75 horses) facilities where horses are fed a forage other than hay, such as silage etc. , where more forage is fed, where horses have visual and tactile contact with con-specifics, and where horses are bedded on straw [1]. Not only can increasing social contact reduce weaving [3] but this effect can be replicated with mirrors [8]. A horse's response to a mirror can be variable, but the effect on weaving and box-walking can be very durable; in a small proportion of cases, the effect seems to wear off with time, but in other cases, mirrors are not effective for some time. However, effect on weaving is usually fairly instantaneous; we recorded cessation within 24 h in subjects that have been known to weave for 15 - 20 yr. From our long-term follow-up of field cases, it seems that mirrors can be used just as effectively to help control box-walking. The use of mirrors is not a panacea and should not be used in anti-social horses who do not get along with other animals. It is also only indicated for animals spending long or frequent periods in isolation. We do not know the importance of the location of the mirror, but we always recommend that it be placed toward the front of the stall so that a reflection can be seen as the horse comes forward at a time when it might be expected to weave. Some cases of stall-walking may also be redirected into something more closely resembling ambulatory grazing behavior through the strategic distribution of forage around the stall [9]. This might also explain the reduction in box-walking reported in horses with foraging devices.

Because bedding on a substrate other than straw seems to be associated with an increase in the risk of weaving [1] straw bedding is recommended, although some feel this might increase the risk of gut impaction. It is unclear whether the provision of straw encourages other behaviors in preference to the problem behavior or whether it provides nutrients that have a suppressing effect on the stereotypy.

Exercise

Repetitive locomotor behaviors are reportedly more common among horses kept in small- to medium-sized (<1. 5 ha) paddocks [10] and paddocks of this size are associated with restricted cantering [11]. It is worth speculating that the opportunity to express certain gaits might be important to horses. It has also been found that exercise deprivation is associated with a rebound increase in activity when the opportunity arises again [12]. Box-walking is more common among horses turned out more frequently in the summer and for longer periods of time in spring, summer, and autumn [7]. Further evidence for a relationship between exercise and box-walking in particular also comes from the finding that endurance horses are more likely to box-walk but less likely to weave than eventing or dressage horses [5].

Thus, the need for exercise may not be entirely based on external cues but partly on previous experience or internal motivational factors. This obviously has significant implications for horse welfare and suggests that horses should be given sustained periods of turn-out throughout the year if at all possible. As long as the animals have water and shelter from wind and rain, they can usually adapt to quite adverse weather conditions.

Evaluation of Other Treatments

Operant feeding devices [13,14] that encourage foraging behavior by providing concentrate rations as a result of the manipulation of a horse-feeding "toy" can be helpful in some cases, but in many instances, they are emptied so quickly that they have little practical use in the longer term. Nonetheless, if used carefully, they can help to remove pre-feeding cues that can increase arousal and trigger the behavior [15]. A simpler alternative is simply to scatter the concentrate ration in the stall throughout its bedding. Logically, such methods must be more welfare friendly than those like anti-weaving bars that seek to physically prevent the behavior and from our preliminary studies, seem to produce further frustration if they are effective at all.

Summary of treatment regimens:

• Increase turn-out and exercise.
• Match exercise and turn-out with sporting requirements.
• Use feeding stations around the stall or scatter feed around the stall.
• Try to reduce predictors of the arrival of food. Consider using a range of foraging toys with variable amounts of concentrate in them if concentrate feeding is necessary.
• Introduce a stall mirror if the horse is sociable.
• Try a foraging devices, but it may increase other undesirable behaviors such as door kicking.
• Avoid using anti-weaving grills and drugs.

Prevention of weaving and box-walking:

• Manage from pasture as much as possible.
• Consider the potential for social housing.
• Wean gradually rather than abruptly, and wean in groups or with a mirror companion.

Conclusion

It seems that many repetitive locomotor behaviors arise from internal and external factors affecting motivation for locomotor activity. This results in the release of goal-directed behavior programs involving movements that are prevented from completion by the confinement of the stable. Factors such as the presence of a potential social partner or feeding cues seem particularly important to the regulation of weaving, whereas factors such as a requirement for exercise based on experience or the motivation to forage may be more relevant to box-walking. However, there is undoubtedly some cross-over between these factors. Sensitive management should consider reducing cues and routines that could cause an expectation of feeding or exercise while allowing adequate exercise, feed, and social interaction in a controlled setting.

4. Cribbing and Windsucking

Form and Prevalence

The terms cribbing, crib-biting, and windsucking are used to describe basically similar behaviors. If a distinction is to be made, cribbing or crib-biting refers to when the horse grasps a fixed object with its incisors, pulls back, and appears to draw air into the cranial esophagus while emitting a characteristic pharyngeal grunt; in windsucking, the behavior is performed without grasping an object. In keeping with other authors, we will not distinguish between them. Surveys on different populations suggest the prevalence may range from 0% in Standardbreds in Canada [7] to 10.5% in young U.K. Thoroughbreds [4].

These behaviors are also described as forms of aerophagia, but this seems to be a misnomer; fluoroscopic studies suggest that very little, if any, air travels down the esophagus during cribbing [16]. The characteristic grunt associated with the behavior seems to be more like a burp than a gulp. This questions the belief that it is an engulfing of air that leads to colic as a result of cribbing; this idea has been used to justify surgical interference in many cases. It seems more likely that if these horses are at increased risk of colic (and this is equivocal in most cases with perhaps the specific exception of epiploic foramen entrapment), it is because there is a common gastro-intestinal factor that is causing both problems, rather than one causing the other [17].

It is often reported that cribbers fail to thrive, and it has been shown that these animals do tend to rest less than normal controls and may have more difficulty in maintaining body weight on comparable diets [17]. Cribbing can cause serious tooth wear, but this is rarely of clinical significance. Thus, it would seem that there is little justification for the range of surgical procedures that have been developed to prevent the physical expression of the behavior. Not only are there risks of complications from surgery, but prevention of the behavior is associated with a rebound response when the behavior is allowed and an increase in a variety of stress parameters during prevention of the behavior [18]. Thus, there can rarely be an ethical justification for physical prevention with collars or surgery or for aversion therapy (including the use of electric collars). Given that the behaviors may have some adaptive value, as discussed below, it is preferable to allow the animal to continue to perform the behavior on a suitable substrate than to simply prevent the behavior.

Summary of classic presentation and clinical findings in a cribber:

• Characteristic behavior with oropharyngeal grunt.
• History of intermittent colic.
• Evident wear of the labial surface of the upper incisors.
• Difficulty in maintaining weight.
• Restlessness in the stable.
• Hypertrophy of the ventral neck muscles.
• Gastric ulceration.
• Lower than normal fecal pH.
• Lower parasympathetic tone at rest, indicating a shift in sympathetic dominance in the control of resting heart rate.
• In cases of field cribbers, the fence posts may be worn but will be flat rather than pointed on top.

Risk Factors and Treatment Opportunities

Dietary Factors

As with weaving and box-walking, the environment plays a critical role in the development and expression of these behaviors. Generally, cribbing has an earlier onset than the repetitive locomotor behaviors, and the feeding of concentrates after weaning has been found to increase the risk of cribbing four-fold [4]. The stomachs of crib-biting foals are more likely to be inflamed and ulcerated than those of normal foals [19]. This too may relate to the feeding of concentrates that rapidly become acidic in the stomach and reduce the amount of saliva produced each day as a result of the reduced chewing. It has been hypothesized that cribbing is a functional behavior, acting as a substitute for the normal level of chewing in an attempt to increase total saliva production and buffer against the increased gastro-intestinal acidity [6]. This does not seem to be the whole story. Nonetheless, feeding a diet supplemented with an antacid does seem to reduce cribbing behavior in some horses. The effect is seen most markedly on post-feeding cribbing [19,20]. The long-term effects of mineral-based supplementation remain unknown and deserve careful attention. It is also worth noting that despite the widespread publicity surrounding the efficacy of proton pump inhibitors for the control of gastric ulcers in horses, there have been no reports of an associated reduction in cribbing. Even without studies targeted at this indication, it would be expected that reports linking the two would appear if an association had been noted. Other dietary alterations aimed at reducing gastric acidity, such as the use of more fiber and the use of fat rather than carbohydrate as the primary energy source of concentrates, may also help and are worth considering if concentrates cannot be eliminated from the diet altogether.

Foraging devices may also be used to reduce the rate of concentrate consumption and its subsequent effect on gastro-intestinal pH, but again, scatter feeding is preferable, because cribbing horses are more likely to show signs of being frustrated with such devices when they are empty. In some cases, they may also increase the problem. However, horses kept at grass may also crib, and there are several possible reasons for this. It might be that once established, cribbing is persistent even when the primary cause has been removed. However, a physical trigger for the behavior present may remain in the pasture environment that can lead to problems relating to gastro-intestinal acidity or gastric ulceration.

Other Factors

Indirect evidence for the role of other factors in cribbing comes from a study of the use of acupuncture to treat the condition [21]. Seven acupuncture points were chosen, including three specifically used to treat gastro-intestinal disorders in the horse. After treatment, 64% (14 of 22) of the horses with gastro-intestinal symptoms improved or stopped the behavior, whereas none without these signs (0 of 11) showed a response. The role of acupuncture in veterinary medicine is gaining wider acceptance, and this apparent association is certainly worthy of note when trying to hypothesize on the nature of the behavior. It is the author's opinion that cribbing is a response to some form of visceral pain; this may be caused by gastric ulceration or other factors. These cases, therefore, require careful medical work-up. Thus, the often-quoted association with colic is caused by the fact that the two conditions have a common primary cause.

Summary of treatment regimens:

• Carefully examine and treat any underlying visceral pain, especially gastric ulcers.
• Reduce concentrate and sweet feed to a minimum; if concentrates are essential, fat- or oil-based rations may be preferable to carbohydrate-based ones.
• Increase forage and turn-out on pasture.
• Provide cribbing boards rather than anti-cribbing collars.
• Consider surgery as a last resort when there is a clear and serious risk to the animal's health if the behavior is allowed to continue.

Prevention of cribbing:

• Avoid creep pre-weaning.
• Minimize concentrates or supplement with an antacid.
• Manage from pasture as much as possible.

5. Woodchewing

Form and Prevalence

Woodchewing is a common repetitive behavior that is exhibited by ~one-third of young Thoroughbreds [4] and is frequently recorded in all other captive populations. There has been debate over whether this should be considered a stereotypy; unlike the other behaviors so far described, it is frequently seen in the field and may be considered a normal, functional foraging behavior. The behavior has a diurnal pattern, being more common at night; [22] this and other studies suggest that gut-fill or lack of it may trigger the behavior [23,24]. Roughage in the diet seems to be the best predictor of tendency to woodchew in general [25] with a lower incidence in horses on high-fiber rations.

Relationship to Cribbing

It has been suggested that woodchewing and crib-biting are related, because horses that crib-bite tend to woodchew first [4]. Furthermore, virginiamycin seems to suppress both behaviors [26]. However, this relationship is not simple, because the provision of hard feed (grain) post-weaning is associated with a reduced risk of woodchewing but an increased risk of crib-biting [4]. Housing foals post-weaning (as opposed to keeping them at grass), group housing at this time, and a socially dominant mother are also clearly associated with an increased risk of woodchewing [4]. Woodchewing has a variety of causal factors, which in some cases may not be related to the stable environment. For example, it has been observed that after a spring flush, feral horses may preferentially strip the bark of willow trees as if seeking analgesia. Careful assessment of the behavior rather than simple generalizations are warranted.

Conclusion

Thus, repetitive oral-based behavior seems to be most accurately predicted by some form of discomfort (possibly visceral rather than musculo-skeletal) that may range from hunger to acute pain. It would be useful to evaluate other potential sources of such pain in these animals and those that do not respond to management of gastro-intestinal pain. This is in stark contrast to the pharmacological route of control being explored by some that seems to be focused on blocking analgesic opiate receptors [27].

6. Headshaking and Nodding

Form and Prevalence

Although these behaviors may seem superficially similar, they are quite different in both structure and cause and therefore, will be discussed separately. First, we will briefly consider nodding.

Headnodding describes a repetitive up and down movement of the head, and it commonly occurs in two contexts. In the housed horse, it often occurs at similar times to weaving and is similarly reduced by increased social contact [3]. In the exercised horse, it also occurs in response to lameness in either the forelimb or hindlimb and is usually most evident at the trot.

The term headshaking is used to describe both a syndrome including many other behaviors and a specific behavior within this collection. Horses with a headshaking problem typically present with recurrent, intermittent, sudden, and apparently involuntary bouts of head tossing that may be so extreme as to throw the horse and rider off balance. For this reason, headshakers can be dangerous to ride, and owners commonly report an inability to train or compete effectively on them [28]. The headshaking is frequently accompanied by sneezing or snorting and attempts by the horse to rub its nose on the ground, on a fore leg, or on nearby objects [29,30]. Although the condition can occur at rest, more cases are obvious during exercise, especially at the trot [29,31,32]. The prevalence of headshaking remains unknown, but it seems to be a problem of increasing commonality. Median age of onset (~5 yr of age) is generally later than that of the other behaviors described here, but it may first appear at any age [30,33].

Of the repetitive behavior problems seen in the horse, the etiology of headshaking has historically received most attention from the veterinary profession. This is because it is primarily seen to be a response to pain or irritation, although it has been suggested that, in some cases, it may be a purely behavioral problem, because a definitive medical diagnosis is often difficult. However, the common occurrence of other behavioral features may suggest irritation of a specific area. For example, "flipping" the nose, sneezing, snorting, "clamping" the nostrils as if to close them, and attempting to rub the face on a fore leg may be responses to naso-facial irritation or pain [30]. A statistical analysis of the collection of signs that frequently accompany headshaking has suggested five dimensions; these dimensions may be used to describe animals with this syndrome that are [33] potentially reflecting pain in different regions of the head. Careful attention to the exact behaviors expressed using video support may be a useful diagnostic guide.

The condition seems to be seasonal in ~60% of cases [30,33]. In about one-quarter of seasonal cases, the length of time the animal displays the behavior tends to increase year after year [33]. The seasonal pattern of symptoms has been used to suggest the importance of medical conditions linked to photoperiod and allergic rhinitis in the cause [29,32]. However, many other factors also correlate with this time of year, including an increased tendency to ride out and a higher aerial dust burden, both of which may be implicated in irritation leading to headshaking; however, the former has been recently shown to not necessarily vary with headshaking occurrence [28].

Risk Factors and Treatment Opportunities

Disease

Physical causes are frequently recorded in the literature during retrospective case studies and feature prominently in the classic list of Cook [34] that suggested 58 conditions that may result in the headshaking syndrome. These are summarized in Table 1 together with a review of the literature. However, Lane and Mair [29] working in a referral center, identified a potential cause in only 11% of their cases and found correction of these problems alleviated the problem in only 2% of cases; this highlights the fundamental difficulty with these cases. In many instances where no gross physical lesion is apparent, the role of the trigeminal nerve, which provides the sensory input from a large part of the muzzle and face, has attracted the most interest, because acute stimulation of this nerve is likely to result in reflex motor responses such as those seen in the headshaking syndrome.

In the absence of any overt pathology, other risk factors such as respiratory allergies have been suggested [29,31] but a case-controlled study [35] failed to find any evidence for an increased risk of respiratory allergies in headshakers. However, there was a suggestion that they were at a greater risk of non-respiratory allergies.

Environmental Factors

Cook [36,37] has suggested that the bit may play a pivotal role in the development of headshaking as a problem, because bit pressure may damage the mandibular branch of the trigeminal nerve and cause pain that could possibly be referred to other branches innervating the muzzle both externally and internally. Evidence of the reaction that the bit can cause in underlying tissues comes from the bony reaction noted in this region of the mandible in many ridden horses. This does not discount the possible role of light, which has been emphasized by some workers in the field [32]. Photic sensitization of sensory aspects of the trigeminal nerve is a well-established phenomenon [38]. Based on a hypothesis of photic mechanism, Madigan et al. [32] found cyproheptadine to be quite effective in five of seven cases; a U.K. study found no evidence of significant improvement in >20 horses using an ultraviolet (UV)-filtering mask identical to one used effectively by Madigan et al. [28,32] in the United States. It seems likely that there may be a difference in the importance of different potential trigger factors in the United Kingdom and the United States, even if the ultimate response is the same. The involvement of these factors does not exclude potential involvement of the posterior ethmoidal branch of the trigeminal nerve that has been emphasized by others [39]. Indeed, the hypothesis of referred pain exacerbated by many potential sources helps to explain the disappointing and frequently temporary effect of surgical intervention on different parts of the trigeminal nerve [39-41]. If the problem relates to the referral of pain to various branches of the trigeminal nerve as a result of stimulation by the bit or other riding gear, it is likely that the problem will persist until the sources of stimulation are addressed, because removal of one branch of the nerve simply results in the problem arising in another branch. Interestingly, small but significant improvements in headshaking severity have recently been found in headshakers exercised without a bit for a short period of time (i.e., riding in a bitless bridle [28]). Long-term studies are justified. Interestingly, Newton et al. [39] in the United Kingdom also reported a good response with carbamazepine, which is used diagnostically in man to assess trigeminal neuralgia, and no response with cyproheptadine.

It is frequently reported that headshaking is twice as common in geldings as mares [29,30,33] and Thoroughbreds seem to be at greater risk of developing the behavior [29,30] even though it is rarely reported in active racehorses. Although Cook [36] suggested that this might also be explained by the role of the bit in this condition, a recent survey [28] suggests that the apparent bias toward males in headshaking samples may, in the United Kingdom at least, simply be a reflection of the normal sex ratio in the general horse population. Other risk factors such as use have been suggested [31] but a case-controlled study of headshakers [35] has failed to find any evidence for management factors, including bedding substrate, diet, or type of use of the horse, associated with an increased risk of the behavior.

Alternative Treatments

Mills et al. [42] in a survey of 245 owners of headshakers, found about three-quarters of owners had consulted a veterinary surgeon about their horse's condition, and traditional veterinary therapy was used in ~52% of cases with ~30% reporting improvement. Herbal and homeopathic feed supplements have been reported to be popular alternatives to veterinary intervention with similar levels of improvement reported for all. These findings were mirrored in a later survey of 200 horses [28] which is presented in Table 2.

Although some of these interventions may help in individual cases, the range of treatments that seem to have a significant effect is notable. This could also be interpreted to provide further evidence of the importance of trigeminal neuralgia in the condition, because, in humans, unpredictable and intermittent remission is a common finding. Coincidental remission with the use of a therapy may, therefore, lead to an unfounded causal association between the two in the owner's mind and the popularization of a wide range of suggested "treatments" as a consequence. Two recent placebo-controlled field trials of alternative therapies popularly used for headshaking found evidence that would also support this contention [28]. Neither treatment was reported to have exceeded the improvement reported with the placebo, but the improvement with any intervention was significant with similar levels of improvement to those reported in these owner surveys [28,42]. Perhaps of greatest concern is that trials with a similar level of control have not been reported in the scientific literature for any traditional veterinary therapies, and therefore, their efficacy may also be more questionable than generally assumed. In the absence of a placebo control, it is necessary to not only achieve a consistent effect but also to maintain it; current reports generally fail to achieve this standard.

The most consistently effective management aid investigated to date seems to be the nose net, with more than one-third of all users reportedly achieving substantial, if not complete, success; this suggests that response is not coincidental [28,33] (see Table 2). Nose nets have traditionally been thought to be effective by acting as a filter, but the success of an alternative "half-net", with a large mesh [43] suggests an alternative mechanism may be involved in their success. One possible alternative explanation relates to a reduction in turbulent airflow through the nose as a result of the mesh. The theory is that the mesh increases laminar air flow. Turbulent airflow might directly result in stimulation of sensitized branches of the trigeminal nerve within the nose, but it will also encourage the seeding out of potentially irritant particle deposits. This may also explain why the problem is more common at the trot and possibly, in dressage horses, because the angle of the head during these exercises may encourage a more turbulent airflow pattern. Turbulence in the nares would also explain why the radical procedure of tracheostomy seems to have a high rate of success in relieving the signs in the small number of cases in which it has been tried [39]. The value of other surgical procedures has already been discussed, but whenever surgery is being considered in these cases, it must focus on relieving the pain of the condition and also recognize the potential impact on the quality of life in the event of any complications.

Conclusion

Headshaking has many causes and careful evaluation of any given case is warranted to determine obvious physical causes. At present, most cases without an overt physical lesion seem to be caused by hypersensitivity of one or more branches of the trigeminal nerve, which is exacerbated by environmental triggers including ill-fitting or improperly used tack. At present, there is a lack of good controlled studies that show consistent efficacy of any specific intervention, although management changes may be just as important as medical or surgical interventions.

7. Conclusion

Generalizations about the causes and treatments of these problems are common but of limited use and potentially misleading. Repetitive behaviors in the horse cannot be explained simply in terms of either learned or inherited traits alone; additionally, their motivation cannot usefully be ascribed to general terms such as boredom, frustration, or coping without careful consideration of the causes and effects of each pattern of behavior. They are multi-factorial, arising from a combination of environmental and genetic risk factors in a given individual at a given time. Prevention should aim to reduce these risk factors. There is undoubtedly a need for further research and for the veterinary profession to look beyond the traditional treatment paradigms of medicine and surgery if we are to fully understand these problems and offer the most effective and humane treatment options. It is perhaps surprising how little has been invested into researching the nature of these problems given their widespread prevalence and their clear relevance to horse welfare. Clearly there is also a need for a shift in the attitude of funding bodies if we are to realize the potential for new and humane treatments, especially given the promising results of some recent studies.

The Animal Behaviour, Cognition and Welfare Group of the University of Lincoln acknowledges the support of the Universities Federation for Animal Welfare for their work on equine stereotypic behavior over the years, which led to the development of the Lincoln Stable mirror. This is now available commercially by Jacksons Arenas in return for a royalty that supports our equine welfare work. We also acknowledge the financial support of Equilibrium Products in the assessment of the half net discussed in this article. In all studies reported here, the University maintained complete ownership of all intellectual property and publication rights.