Review of Research on the Effectiveness of Early Intensive Handling of Foals

A number of studies have been undertaken to determine the short- and long-term benefit of early, intensive handling of foals, commonly known as "imprint training". There have been no profound differences found in subsequent response to handling between foals that have a regimented early handling procedure and those maintained under routine management. Control subjects in these various systematic studies provide evidence of the ability of foals and young horses to decrease reactivity to handling after a range of routine positive interactions with humans.

1. Introduction

Early intensive handling of domestic animals involves a program of handling of the neonate or young by humans and introduction to domestic procedures and equipment at an early stage. In some cases, the presumption is that efficiency of learning and retention is greater during a purported critical or sensitive developmental period. Early intensive handling in the equine industry is probably best known by the term "imprint training", [1] introduced and popularized worldwide by the veterinarian Dr. Robert Miller. Early intensive handling of foals is a subject that has produced a number of peer-reviewed publications over the last several years, and many seem to be directly addressing Dr. Miller's particular technique of "imprint training".

"Imprint training" in brief requires first taking the foal immediately after birth and maintaining it in lateral recumbency while performing a highly regimented series of handling procedures. Further steps are completed after the foal has stood. Dr. Miller suggested that the handling ideally should be repeated at some intervals within the foal's first 2 wks, along with additional procedures such as tying and trailer loading. The specific handling during recumbency includes the following:rubbing the foal all over its face, body, and limbs; inserting a finger into the mouth, ears, and rectum; tapping on the soles of the feet; running clippers and a plastic bag over the foal; using a spray bottle on or around the foal; and fitting a halter. Once the foal is standing, the handler applies digital pressure until the foal walks forward or backward and moves laterally. The goals are to complete each step until the foal appropriately relaxes or responds to each stimulus and to not allow normal resistance to interrupt the handling.

According to Miller, these steps will provide irreversible learning experiences during a proposed critical period in early development. Furthermore, this specific handling procedure should result in "bonding with humans, desensitization to certain stimuli, sensitization to other stimuli, [and] submission to humans" [1]. Ultimately the goal is to have a foal that is compliant with all subsequent handling and training procedures.

2. What Is Imprinting?

When applied to the usual approach toward early intensive handling, the use of the term imprinting is incorrect, as pointed out by a number of authors in the papers discussed in this review. It is important to clarify the generally accepted meaning of "imprinting" and the use of this terminology as it relates to this specific handling procedure. Imprinting is considered a specific kind of learning during a sensitive period of a young animal's life. Usually imprinting is used to describe the development of social preferences, principally for recognition of kin or future sexual partners [2]. Throughout the behavior literature, there is much argument surrounding imprinting and sensitive periods in all species:is imprinting indeed a special kind of learning? Can behaviors that arise during a sensitive period be modified as the needs of the animal change during development and throughout its life? These questions are important relative to the horse, where a discrete imprinting or sensitive period for the foal has not been clearly identified. In addition, submitting the effect of a sensitive period to a subjective application such as horse training falls outside the bounds of the usual social and sexual benefits derived.

Questions about Miller's "imprint training" procedure intersect broadly over both lay and research interests. Are all of the regimented steps required? Is the procedure standardized and universally applied by researchers and lay people alike? Does early intensive handling improve compliance with later training of the horse? Does early intensive handling have benefits in reducing fear or reactivity, particularly with regard to the foal generalizing this learning toward novel objects, new training situations, or unfamiliar people? Are there intra-species social effects for foals that have been "imprint trained" by a person? Are there any detriments, particularly to the foal's health? Can research in this area establish or guide us to better understanding of an actual critical period for learning in horses and how to benefit from it?

3. What Is the Current Research on Early Intensive Handling?

There Are Many Variations on the Published "Imprint Training" Procedure

Most recent studies designed to specifically evaluate the "imprint training" procedure have performed it, as best as can be discerned, if not to the exact detail at least faithfully to the intent and purpose. Study designs do vary in that they may use single or multiple handling sessions over days or weeks, with later testing for comparison to unhandled foals once or several times. Most of those reviewed here use a handling procedure with individual steps that rather closely follow that outlined by Miller both in substance and objective.

Results Regarding Compliance, Reactivity, and Novelty

There have been no consistent, positive findings connected to early intensive handling with regard to compliance and long-term benefits in training or in reaction to novelty or potentially fearful situations.

Mal et al. [3] looked at minimally handled foals (routine and emergency veterinary care only), intermediate handled foals handled for 10 min twice daily from days 1 to 7, and extensively handled foals handled the same as the intermediate group and additionally once weekly for 10 min until weaning at 120 days. Handling for each 10-min session included rubbing the foal all over its head and body but did not include limb or foot manipulations. All foals were housed together with their mares and had visual contact with humans twice daily for feeding through the duration of the study. They found there were no differences among the groups in manageability (measures related to flight zone with approach of an unfamiliar human) and response to a novel stimulus. This might suggest that the handling procedure was inadequate in timing, duration, and type, whereas another plausible explanation is that the daily visual contact with humans is in fact adequate to establish some type of manageability and decrease reactivity in even unhandled foals. In most studies discussed here, the control foals in fact had some consistent visual or other modest contact with humans related to routine management procedures.

Simpson [4] followed an intensive handling protocol first at 2 - 8 h after birth and once daily thereafter for 5 days. Handled and control foals at 4 mo of age were presented with some novel and other specific stimuli similar to those included in the early handling protocol. Overall, handled foals had more favorable scores for calmness and friendliness. Both handled and control foals equally allowed approach by a person, but the handled foals were more willing to approach. Although two of eight control foals could not complete the testing because of handling concerns, there were no significant differences in compliance with specific stimuli similar to those performed on the handled foals at birth.

Williams et al. [5] evaluated foals handled at various latencies from birth, intervals, and frequencies within the first 72 h compared with unhandled foals. They followed Miller's imprinting procedure closely, except they did not begin with the foals handled" at birth" until the foals had stood and nursed. They found no significant differences in handled compared with control foals when tested at 6 [5] or 1, 2, or 3 mo [6] of age. They tested using stimuli that were similar to the earlier handling procedure and a novel stimulus. Interestingly, they did have an effect of year in one study [5] with both control and handled foals from the second year of this multi-year study being more compliant with nearly all stimuli presented. This suggests the possibility of handler or technician drift (improvement in handling all foals, particularly during the testing procedure or some change in criteria).

Spier et al. [7] used methods very similar to Miller's in timing (handling at 10 min and 24 h after birth) and specific handling procedures. When the foals were tested at 3 mo of age, significant differences were found between handled and control foals only in compliance with handling the legs and feet. The overall compliance score was better for the handled foals compared with the control foals, although not statistically significant (p = 0.097). There were no perceptible or statistical differences between the handled and control foals when presented with novel and potentially noxious procedures (deworming or intranasal and intramuscular vaccinations).

In preliminary studies in our laboratory [a], full sibling Quarter horse foals were submitted to either minimal handling (control) or intensive handling in a single session immediately at birth. At 1 mo of age, all foals were haltered by a single, blinded handler, and at 3 mo of age, they were tested again by the same handler with stimuli mimicking, but of greater intensity than those performed as neonates. At a later date, videos of the sessions were viewed by experienced, lay horse people blinded to the experimental condition of the foals. The foals were ranked with regard to compliance with the procedures. There were no significant differences in the rankings of the foals according to their control or handled status for either test situation. Because this study was directed more toward the question of an absolute critical period for handling the foal, it did not include repetition of the handling over several days as usually recommended by Miller.

In a study using Icelandic horse foals, Sigurjonsdottir and Gunnarsson [8] compared control foals and foals handled at 24 h of age. The handling steps in this study diverged more than others from the standard procedure described by Miller. At the initial handling at 24 h, foals were maintained in recumbency and rubbed all over the body, and the handler inserted a finger into the mouth and ears, in a manner described as "not as often nor as fast as Miller does". Foals were handled an additional three more times at intervals of 1 - 4 days, each time repeating the earlier steps and adding additional steps of sensitization, haltering, and leading. Testing the foals at 5 mo of age included catching, haltering, leading, rubbing the body, and picking up the feet. There were largely no differences, because of high individual variation, in handled and control foals except that the handled foals were less resistant to being caught, haltered, and lead. They also found a correlation within the handled group of foals between resistance to handling and negative temperament scores of their dams. On the strength of this finding, they concluded that the force and restraint required for this type of early handling would be a detriment to foals with a predisposition to resistance to handling (based on dam's temperament). Interestingly, this conclusion somewhat bears out Miller's forewarning that improper or incorrect performance of the procedure on particularly dominant foals will result in a horse that is more difficult to handle later on.

Jezierski et al. [9] compared unhandled and handled stabled and reserve-raised (semi-feral) horses. The handling was not of the same substance and timing as those specifically addressing imprinting, because the handling for stabled foals began at 2 wk of age and for reserve foals began at 8 - 10 mo of age; handling sessions for both stabled and reserve-raised foals thereafter consisted of 10 min/day, 5 days/wk through 24 mo of age. Handling included haltering, rubbing the body, and picking up the feet. Testing at 6 (stabled only), 12, 18, and 24 mo of age included a familiar handler catching, leading, picking up hooves, and holding the horse for approach by an unfamiliar person. Both stabled and reserve-raised handled horses scored better on the tests than both groups of unhandled horses at 18 and 24 mo. However, both stabled and reserve-raised unhandled horses showed dramatic improvement in their test scores from their first to their 18-mo test. Interestingly, all groups declined to some extent in their test scores between 18 and 24 mo.

Lansade et al. [10] also completed a study on unhandled foals handled either immediately at weaning (6 mo) or at 3 wk after weaning. The handling procedure included petting, haltering, picking up the feet, and leading. The foals handled immediately at weaning were in all cases more compliant than the foals handled 3 wk later during the actual handling session. However, during testing at 2 days, 4 mo, 7 mo, 10 mo, and 18 mo after the initial handling session, the two groups of handled foals were similar in compliance, and both earned better scores than the previously unhandled control foals. The control foals and the handled foals did improve in compliance with each testing session. There were largely no differences in reactivity to isolation, a novel person, or a novel object between all handled and control foals on all the testing days after day 2. The authors concluded that handling at weaning was optimal in this study and suggest that the stress associated with weaning presented another sort of critical learning period.

Results Regarding Future Learning Ability

Little work has been done on learning ability of foals handled intensively in the neonatal period. Recently Lansade et al. [11] reported no effect of neonatal handling on learning ability as yearlings, using object discrimination and spatial discrimination tests. Earlier studies, however, have compared handling regimens conducted on older foals and subsequent learning ability using simple maze tests.

Heird [12] studied young horses handled under controlled but routine management practices either at an intermediate or intensive level compared with control horses (haltered during first 3 days of life, only then range-kept). When yearlings they were subjected to a 20-day simple maze learning task. The intermediate-handled group showed the greatest learning in the task. Both handling groups performed better than controls. In a later study, Heird [13] looked at five groups of young horses handled using routine management practices either intermittently or continuously over 18 mo, whereas unhandled horses were kept on range until being haltered just before testing. They used the same maze learning task over 20 days. In the initial 10 days, the continuously handled group performed best (percent correct choices and fewer trials to criterion); however, there were no significant differences among all groups in performance of the task during days 11 - 20.

Are There Effects on Normal Early Mare and Foal Interactions?

Intrusion by extensive human handling during the immediate post-partum period may interrupt normal mare-foal interactions. The early, normal interactions are suspected to play a role in ensuring accurate future recognition between the mare and foal. One interpretation of actual imprinting by foals is the following response, in which neonatal foals navigate toward or attend to any large, looming object, which would ordinarily be the dam [14-17]. This idea is also supported by observations that mares will usually take slow, deliberate steps away from the foal as it first attempts to nurse, which may appear to the untrained eye as resistance [18].

Many veterinarians and horse breeders believe or anecdotally have observed first-hand that intrusion by human handlers during the immediate post-partum period may affect the ability of the mare and foal to bond properly or instigate rejection by the mare. These concerns have not been clearly borne out in research or controlled settings. In a survey of foal rejection rates in a domestic horse population, the presence or number of human observers at the time of delivery did not seem to be associated with rejection [15,19]. Mares facing extreme human intervention in a neonatal intensive care facility, if given the opportunity to observe or touch their foal and provided appropriate mammary gland care, more often than not continue to pay attention to the foal and do not resist later nursing efforts by the foal [ b].

Diehl et al. [18] using Quarter horse foals out of multiparous mares, closely examined the behavior of the dam and her interactions with the foal and the handler during the intensive handling procedure immediately at birth and through 4 h post-foaling. The mares, unrestrained while the foals were handled, did not interfere with the completion of any part of the procedure (Fig. 1). All mares licked and sniffed their foals and vocalized apparently toward their foals. Only one mare showed aggression toward the handler or toward the horse in the adjacent stall. Aggressive behavior was of low frequency (not observed in every mare) and intensity and included mainly kick and bite threats and nips. Likewise, overt mare avoidance of the foal without aggression was low. There were no differences in these behaviors between dams of intensively handled and control foals. The dams of intensively handled foals, however, spent a significantly decreased time eating and drinking and a significantly greater time sniffing their foals compared with the dams of control foals. Spier et al. [7] measured dam disposition during the 3-mo testing of previously early handled and control foals and found no differences between the two groups. None of the studies reviewed here reported any instances of failure of normal mare-foal attachment.

Figure 1. Unrestrained mare sniffing her foal during the intensive handling procedure.

Are There Effects on the Neonatal Foal's Health?

Reasonable concern can be raised regarding the effects of intensive handling on the health of the neonatal foal. While restraint of the foal normally requires more advantageous positioning than brute strength, there are concerns and anecdotal reports of physical injury to the foal and prolonged course of the procedure, resulting in weak foals that fail to nurse. In Diehl et al. [18] the handler subjectively found it extremely difficult to maintain appropriate restraint against the foal's effort to stand, particularly with strong foals (Fig. 2 and Fig. 3).

Figure 2. Foal remaining quiet during intensive handling session.

Figure 3. Foal showing typical, moderate resistance to part of the intensive handling session.

Of perhaps greater concern is the potential delay in nursing and the timely ingestion of colostrum. Delayed ingestion of colostrum in the presence of environmental microbial challenge could result in grave susceptibility to infections [20]. Few studies have looked at specific physiological or behavioral measures related to neonatal foal health.

Diehl et al. [18] measured foal nursing behavior and immunoglobulin absorption in full sibling foals submitted to either intensive handling or control (limited) handling at birth. The handling procedure began between 3 and 18 min after expulsion of the foal, and the mean time for completion of the entire procedure was 45 min. As would be expected, the intensively handled foals had a significantly longer latency to first stand. Interestingly, there was no significant difference between the handled and control foals for the latency to first nurse. All mares had a colostrometer specific gravity reading of 1.040 or greater (results during the second year only). Blood immunoglobulin G measures using a radial immunodiffusion test at 4 h of age (regardless of time nursing) ranged from <200 to 400 mg/dl. The blood immunoglobulin G concentrations at 18 h of age ranged from 400 (n = 1) to 3600 mg/dl (most >1000 mg/dl). No foals showed clinical signs that would be associated with sepsis.

Spier et al. [7] measured serum cortisol and glucose before and after handling foals at their 3-mo testing sessions, presumably as a physiological parameter for stress. There were no differences between handled and control foals in these measures, although in both groups, serum cortisol was increased after the testing session.

4. Research Conclusions

There seem to be a number of assumptions made, somewhat based on the strong assertions surrounding the value and success of the "imprint training" procedure, in the design and methods of these studies. One is that there does in fact exist a critical period for neonatal foal learning that will somehow influence behavior later in life, relative to expectations of owners, trainers, and riders. Another is that a strict protocol must be followed with regards to substance, timing, and duration to effect the later benefits. What is learned in this critical period is presumed to be measurable and not subject to change or reversion later in life. Finally, the absence of the correct protocol at the correct time should result in a foal unable to accomplish the same learning in a manner that can be measured. Each of these assumptions in fact deserves intensive study on their own merit, and none have been supported by research up to this point.

Most of the behavior measures for compliance and reactivity in these studies relied on scoring systems that may be subject to significant interpretation and bias. Future studies measuring specific behaviors may shed greater light on the precise responses elicited by the test situation. Furthermore, there is unlikely to be complete agreement as to the behaviors, such as those suggesting "reactivity", that are most beneficial or in fact desirable in all training situations. For example, in the study of Heird [12] the intermediate handling group achieved greater success in a maze learning task compared with the intensively handled horses that were noted to be more docile. If one were to make very detailed observations, it is likely we would see many specific behaviors that vary among horses to produce a similar training endpoint.

Interestingly, several studies have shown specific evidence that foals become compliant with repeated handling without regard to having any significant, regimented neonatal handling. As noted earlier, the study of Lansade et al. of foals handled around weaning at 6 m showed control foals improved in compliance over time [10]. In another study, Lansade et al. [11] compared control foals and foals receiving handling at birth and each day for 14 days. They were given repeated handling tests at 16 days, 3 mo, 6 mo, and 1 yr of age. By 3 mo of age, there were no differences between groups in time required to pick up feet, number of defensive reactions, and flight response; by 6 mo of age, time required to fit a halter was the same; and by 1 yr, compliance with walking on a lead was the same. While there was no analysis to assess the changes over time within each group, it is clear from the data shown that the compliance and behavior of the control foals in most tests improved over time, and thus their improvement was likely a strong component in the progressive similarities between the two groups. The improvement in compliance over time by the control foals suggest equivalent learning compared with the handled foals and short-lived benefit to early intensive handling. In the study of Simpson [4] there were no overall differences in behavior responses between the handled and control foals. The test stimuli were consistently presented at the left side of the foal first. Although the handled foals were more compliant and had lower heart rates while tested on the left side, the control foals showed more favorable responses to stimuli on the right side. This suggests a relatively rapid learning, desensitization, or habituation effect of the immediate handling going on during the test. This improvement from one side to the next was not observed in the intensively handled foals.

In all studies, there is a statement or inference that can be made regarding the existence of some kind of handling of the control foals (e.g., "minimal" handling for routine management such as dipping the umbilical stump, routine physical examination one time) or at least routine contact with humans (at feeding time, moving to new pastures, etc.). Perhaps the visual contact with people is adequate to condition any foal to various in-contact procedures. If this routine contact is coupled with unintentional positive reinforcement (feeding), this is even more likely so. In many of the studies reviewed the control, unhandled foals had apparent and/or statistically significant improvement in behavior measures over time. Further studies using completely unhandled horses, without benefit of positive and possibly rewarding human contact, would be beneficial in identifying critical periods and various positive effects following presentation of specific stimuli.

5. Practical Considerations and Recommendations

There is anecdotal evidence that, in reality, a large number of people performing the "imprint training" procedure do not have a good understanding of its principles nor do they have the ability or desire to carry it out as defined by Miller. This author and a number of colleagues have routinely questioned professional and lay horsemen who purport to carry out "imprint training" on their foals. It is the rare individual that even approaches the intensity and specificity of the procedure as described by Miller or in the research publications described above. One person, for example, who firmly supported the use of "imprint training" stated, "I just go in and love on them". Williams et al. [5] also acknowledged this disjoint between reality and the described procedure. On one hand, Miller has stated that the procedure is fairly robust and, even if performed incorrectly (and to what extent is unclear), most foals still respond well, with the exception of the particularly "dominant" or "willful" foal. On the other hand, he has also raised concerns that, if performed incorrectly, research results will be tainted [21]. It is likely that most uses of the "imprint training" procedure, in practice and in research, can technically be called incorrectly performed, and, therefore, unsatisfactory results might be too easily refuted.

Although there has been concern that improper performance of "imprint training" may result in injuries to foals and significant delay in colostrum intake, this has not been shown in the research reviewed here. Subsequent to his original publication of the procedure, Miller has recommended feeding the foal colostrum while it is still recumbent and before the start of the "imprint training" procedure [21]. This would, not surprisingly, function to relieve the time pressure on the handler but also might decrease the foal's urge to stand and nurse, thus making restraint in recumbency easier. It is incumbent on the handler, however, to carefully administer the colostrum to decrease the risk of aspiration by the foal.

Another unsubstantiated concern is the encouragement some handlers may feel, based on their interpretation of the "imprint training" procedure, to interact with the foal too much. Commonly, orphaned foals hand-raised by people become less respectful of handlers and are often described as "spoiled". While the "imprint training" procedure does not specifically encourage such over-handling, the use of the term bonding and the recommended intensity of handling may be misleading to the inexperienced lay person. These health concerns and the prospect of over-handling bear emphasis when electing to perform or recommending early intensive handling of foals.

Certainly credit should be given to the "imprint training" procedure if it has led to horse breeders and owners having a greater awareness of simple learning mechanisms used by horses of any age. In addition, perhaps it has identified "breaking" horses with unnecessarily rough handling that may still occur in the industry. With early intensive handling, there may be certain specific, although perhaps temporary, benefits conferred with respect to a young horse's later compliance and response to novel stimuli. Thus far, however, no studies have identified clear and convincing, overwhelming, broad long-term benefits that could not possibly be achieved by handling at other times and using different protocols.

This work was supported by Project 3817 of the PA Agricultural Experiment Station and the Department of Dairy and Animal Science at the Pennsylvania State University. All research conducted under Project 3817 was approved by the Institutional Animal Care and Use Committee at the Pennsylvania State University.

Footnotes

1. Diehl NK. Unpublished data. December 2004.
2. Wilkins P. Personal communication. 2001.