Influence of Maternal DHA Supplementation on Foal Cognition and Memory Tayler Hansen Honors Thesis
Abstract Supplementation of long chain polyunsaturated fatty acids (LCPUFA), specifically docosahexaenoic acid (DHA), has been shown to incre ase cognitive development in humans, mice, goats, and sheep. It was hypothesized that DHA could similarly lead to faster learning and improved trainability in foals. Mares were fed a control (CON, n=10) or a high DHA (LCPUFA, n=10) diet from d 280 of gesta tion through 70 d of lactation. At 2 mo of age, foals from these mares were taught to touch a target object with their nose on command. Memory recall of previous training and rate of learning of new, more complex behaviors were re assessed using the targe t training technique in weaned foals at 185 2 d of age. New behaviors learned included walking from point A to point B, performing target behaviors in an open environment, walking over novel objects, and loading into a livestock hauling trailer on comman d. Weanling response times, performance scores and the number of cues needed to successfully complete each behavior were recorded during 11 training sessions conducted over 6 d. Results indicated weanlings were able to remember previously trained behavior s and learn new ones using the target training method, but DHA supplementation of the mare did not affect training responses. Future research should consider cognitive evaluation of older foals, as well as evaluation after direct supplementation of DHA to the foal.
Background Fatty Acids Fatty acids define the physical and chemical properties of dietary lipids based on the number of carbon atoms and the amount of carbon carbon double bonds. Fatty acids have two main structural elements, a chain of hydro carbons and a carboxylic acid. The chains of hydrocarbons may be short (2 4 carbon atoms), medium (6 10 carbon atoms), or long (12 26 carbon atoms). The length of chain is directly proportional to melting point. Short chain fatty acids have a low melting p oint. A fatty acid is classified as saturated (no carbon carbon double bonds), monounsaturated (one carbon carbon double bond), or polyunsaturated (more than one carbon carbon double bond). A double bond will also decrease melting point and is more reactiv e than a single bond in chemical reactions due to the pi bond a covalent chemical bond that is found in double or triple bonds. The carbon atom close s t to the carboxylic group is defined as the alpha carbon, while the methyl end is defined as the omega o r n end. The number of carbons to the first carbon carbon double bond is also used to describe the fatty acid. For instance, omega 3 (or n 3) fatty acids are classified as such since the first double bond occurs three carbons away from the omega carbon. A nutrient is considered essential when it must be supplied by th e diet to meet cellular demands because t he body is not able to synthesize the nutrient at all or at a sufficient rate. There are two essential fatty acids, alpha linolenic acid ( ALA;18:3 n 3 ) and linoleic acid (LA;18:2 n 6 ), that are used to maintain cell membrane structure, capillary wall integrity, lubrications of the skin and prostaglandin production by serving as a precursor molecule for long chain polyunsaturated fatty acids (PUFA) synthe sis. Synthesis of essential fatty acids to long chain
PUFA is minimal (Umhau and Dauphinals, 2010). This may indicate benefits to health when dietary supply of long chain PUFA supplementation is increased (Horrocks and Yeo, 1999). Omega 3 and omega 6 fat ty acids have complex interactions due to competition for key enzymes in metabolic pathways. In general, increased concentrations of circulating omega 3 fatty acids result in a healthier mix of metabolites as fatty acids enzymes are redirected towards omeg a 3 pathways (King, et al., 2008). Docosahexaenoic Acid Docosahexaenoic acid (DHA ; 22:6 n 3 ) is a PUFA with 22 carbon atoms and 6 carbon carbon double bonds. The modification of ALA with elongase and desaturase enzymes creates the double bonds that are fou nd in DHA (Barboza et al., 2009). The synthesis of DHA from ALA is minimal at best (Umhau and Dauphinals, 2010). Neural tissue has the highest concentration of fatty acids in the body beside adipose tissue (Bourre, 2009). DHA is the primary structure in t he grey matter of the brain ( a major componen t of the central nervous system) and, along with arachidonic acid, it is found in nerve synapses where they function in phospholipid mediated signal transduction (Jones et al., 1997; Horrocks and Yeo, 1999). Sin ce DHA is a key component of neural tissue, it has been examined for cognitive development. Supplementation of DHA has shown many benefits in human health. DHA has anti inflammatory properties through possible alteration of lymphocyte, monocyte, and macrop hage functions, which may be why DHA has positive effects on arthritis cases (Horrocks and Yeo, 1999). Woodward and et al. (2005) also found this trend in horses supplemented with long chain PUFA. Supplementation of DHA and EPA showed a trend towards incre ased stride length at the trot compared to a baseline stride length taken before supplementation started. Additionally,
DHA aids in the treatment or provides benefits to diseases such as hypertension, atherosclerosis, depression, diabetes mellitus, myocard ial inf ar ction, thrombosis, heart disease, and some cancers (Horrocks and Yeo, 1999). DHA and Cognition During fetal development, the mother must insure that the fetus is obtaining sufficient quantities of DHA for proper brain development. Since de novo synthesis of omega 3 and omega 6 fatty acids does not occur in mammals, fatty acids, including DHA, must be supplied via placental transfer in utero (Innis, 2005). The supplementation of DHA in maternal diets has shown increased cognitive development in infants. In addition, infant diets deficient in omega 3 fatty acids may lead to a decrease in learning ability due to the function DHA serves in neural transmission (Horrocks and Yeo, 1999). Studies have shown increased recognition memory, problem solving subtest, Bayley Psychomotor Development Index, and intelligence quotient in diets containing DHA compared to diets without added DHA (Lucas et al., 1992; Jensen et al., 2005; Hendrickson et al., 2008; Jensen et al., 2010). These studies showed DHA dietary supply had an effect on both preterm and term infants. Henriksen et al. (2008) bottle fed very low birth weight infants with human milk supplemented with 32 mg DHA and 31 mg arachidonic acid and found infants receiving the supplemented milk scored higher on problem solving tests and recognition memory tests than a nonintervention group. Testing in older subjects also showed benefits to DHA supplementation early in life. For example, Jensen et al. (2010) supplemented the maternal diet of breast fed infants with DHA during the first four months of lactation and found a significant increase in
sustained attention when the children were examined at five years of age, but the results were significantly different for girls only. The effects of DHA and other long chain PUFA on cognitive development have been studied in several animal species, including mice and goats ( Fedorova and Salem Jr., 2006; Petursdottir et al., 2008; Duvaux Ponter et al., 2008). NaÂ•ve 10 mo old male SAMP8 mice were assigned to a high DHA or low DHA diet and assessed in a T maze foot shock avoidance test, and the high DHA diet mice were significantly better at learning the test and had greater retention when retested 1 week later (Petursdottir et al., 2008). DHA deficient diets lead to slower habituation in open field environments ( Fedorova and Salem Jr., 2006). Based on these studies, dietary DHA supply may be an influential factor in equine cognition. DHA supplementation in maternal diets increases the content of DHA in breast milk along wit h increases in infant plasma phospholipid pattern (Jensen et al., 2005). Similar results were seen with horses, where DHA supplementation increased the DHA content in mare milk (Stelzleni et al., 2006) and in goats, where maternal supplementation of alpha linolenic acid increased the proportion of DHA in the fatty acid profile of 2 week old kid plasma ( Duvaux Ponter et al., 2008). These factors led to the conclusion that DHA supplementation of the maternal diet during late gestation and lactation results in increased DHA availability to the fetus and newborn This increased availability of DHA may be beneficial during brain development and lead to offspring with better cognitive skills than those whose dam did not receive DHA supplementation. Additional supp ly of DHA in the maternal diet during the last months of gestation and early infancy may be the reason for the increased cognitive ability of offspring as supplementation increases DHA concentration in mares milk (Stelzleni et al., 2006) and infant
plasma (Jensen et al., 2008) In humans, brain development begins in the last trimester of pregnancy and throughout the first year before slowing pace (Horrocks and Yeo, 1999). Since DHA makes up a large portion of neural tissue, inadequate quantities will hinder development. As research continues, it will become apparent if quantities over the minimal DHA nutritional requirement enhance brain development more so than the dietary requirements. Horse Learning and Memory Ability Horse trainers are able to obtain ce rtain responses from horses through understanding horses behavioral characteristics (McGreevy, 2007). Horses are prey animals, therefore their typical response to increased pressure or sudden movement is to flee. These situations occur often with domestic horses, such as a rider applying pressure with their legs to ask the horse to move forward. Trainers, riders, and handlers use rewards, in this example the removal of pressure once the horse moves forward, to override the horses basic drive to flee from pr essure (McGreevy, 2007). The use of stimulus response reinforcement chains has allowed horse trainers, riders, and handlers to establish many different behaviors that allow horses to be used in a variety of disciplines. In addition, the horse's ability to distinguish between stimuli allows for very complex riding movements to be formed along with scientific testing of horses cognition and memory (Hangii, 2005). Horse memory is considered good and in some instances may be better than humans (McGreevy and M cLean, 2010). Marinier and Alexander (1994) successfully showed that horses could remember a maze at 1 week and 2 months after initially completing the maze. Horses have been able to remember a task four years after the initial training and le arning of the task (Wolff and Ha usberger, 1996). These findings support what is found in the industry. Many horses
experience extended time without being ridden, and besides the initial display of extra energy, their riders can work them almost immediately. Assessing learning ability and memory in horses has required the use of specialized equipment such as mazes and food dispensing apparatuses and in some studies, specially trained horses (Mal et al., 1993; Marinier and Alexander, 1994; Wolff and Hausberger, 1995; Ha ngii, 2005). Food dispensing apparatuses have been effective tools in horse learning trials, but in one trial appetitive conditioning tasks were not effective with weanling horses (Mal et al., 1993; Hausberger et al., 2004). Novel object tests and open fie ld observations are often used to assess emotionality and behavior in horses and other animals, but fail to evaluate trainability of horses (Hausberger et al., 2004; Fedorova and Salem Jr., 2006). Operant Conditioning Operant conditioning is used to train animals to respond to stimuli in a specific manner. The animal must voluntarily work within the environment such as moving towards an object, to receive a reward (McGreevy, 2007 ). The reward may be positive reinforcement such as food, or the termination o f negative reinforcement such as the end of applying pressure when an animal moves. Advanced behaviors are most often established with successive steps that are less complex, known as approximations. Unlike classical conditioning, where the reward is assoc iated with the stimulus, the reward is associated with the response in operant conditioning (McGreevy, 2007). An example of operant conditio ning is when a dog owner teache s their dog to sit down. The owner says the word sit, and when the dog sits, the volu ntary behavior, the dog is rewarded. Horses excel at both classical and operant conditioning (Hanggi, 2005) with either positive of negative reinforcement; however, most training in research is limited to positive
reinforcement (Hausberger, et al., 2004). This limitation should be addressed in future research as training in the industry relies primarily on negative reinforcement. Target Training Target training uses operant condition ing to teach an animal to touch an object which then can be moved to dif ferent locations and used to develop more complex behaviors Typically, target training requires movement from the animal. This method of training is ideal in zoo settings with protective contact. A zookeeper can move an animal into stocks, have the animal change direction, and stand for examination with th e use of a target without needing to be in the same pen as the animal. Target training may be an effective method for assessing both learning ability and trainability in horses. While target training an d positive reinforcement has been used in the equine industry, research on target training is limited. Target training has been used to load horses into trailers, for groundwork, and to teach horses to stand quietly for grooming and veterinary procedures ( Hangi i, 2005). While there is limited research on target training in horses the effectiveness of target training in othe r animals has resulted in more reliable animals and decrease d data collection time in laboratory settings (McKinl e y et al., 2010). Trai ler Loading Loading into a trailer can be a source of stress to horses and also cause injury to both horse and handler (Shanahan, 2003). Typically, unwillingness to load is met with incre ased pressure on the lead line, using a lead with a chain over or und er the nose or applying pressure to the animal from behind. Occasionally, a handler will use whips, ropes, and tranquilizers to load a horse. The horse may respond to negative stimuli undesirably, from rushing onto the
trailer to becoming more resistant t o loading whereas horse taught to load using positive reinforcement load readily and eagerly (Hangii, 2005) The use of target training has been applied to horses to load into trailers Ferguson and Rosales Ruiz (2001) used targeting to load horses that had previously been described as problem loaders, taking longer than 3 hours to load into trailers. After an initial period establishing the target behavior (began with placing the target within 0.3 and 0.6 m of the horse's nose and then moving it to anoth er area near the horse), the researchers used a series of approximations (from the horse approaching within 3 m of the trailer to the horse being complete ly loaded with all four feet in the trailer and door closed) that resulted in successful trailer loadi ng (Ferguson and Rosales Ruiz, 2001 ). While Shanahan (2003) did not specifically use a target, their training method imitates successive approximations used in target training by leading the horses through obstacles simulating trailer elements ( i.e., stepp ing onto a raised platform, over a tarp, and through tight spaces). Compared to pre training levels, loading time, cortisol levels, and heart rate decreased during trailer loading after training indicating successive approximations were useful in decreasin g stress in trailer loading (Shanahan, 2003). In conclusion, DHA supplementation directly to the infant or indirectly via the maternal diet appears to benefit cognitive development in humans Similar benefits may be seen in foals, as DHA supplementation ha s improved cognition in other animals. Increased cognition would benefit the equine industry by possibly increasing trainability of horses. Increased trainability would enhance athletic performance, decrease behavior problems in daily care, and improve the interactions between horses and humans. Target training and positive reinforcement are effective training tools and shown to increase compliance in training situations compared to traditional
methods. Progress through approximations of a complex behavior may indicate learning ability in foals, an area of limited research in horse cognition.
Introduction Essential fatty acids are the polyunsaturated fatty acids (PUFA) alpha linolenic acid (ALA), an 18 carbon omega 3 fatty acid, and linoleic acid (LA), an 18 carbon omega 6 fatty acid; however it is the longer chain omega 3 and omega 6 fatty that possess biological activity. For example, arachidonic acid (n 6, 20:4) is a precursor molecule for prostaglandin production. The body can synthesize long chain PUF A omega 3 or omega 6 fatty acids from the essential 18 carbon precursors with the use of elongase and desaturase enzymes; however, the conversion to longer forms of fatty acids that are most abundant in cell membranes occurs only in limited quantities. Sup plementation of long chain PUFA shows benefits in many areas of health and nutrition. Increased intake in the diet provides benefits to the cardiovascular system, control of diabetes mellitus and depression (Horrocks and Yeo, 1999). The addition of long ch ain polyunsaturated omega 3 fatty acids has shown increases in cognition during human infancy and early childhood (Horrocks and Yeo, 1999; Jensen et al., 2005; Henriksen et al., 2008). Specifically, supplementation with the long chain omega 3 fatty acid do cosahexaenoic acid (DHA) in the maternal diet has shown increased attention span in children at five years of age (Jensen et al., 2010). DHA is a key component of neural phospholipids, and dietary supply of DHA results in greater DHA incorporation into bra in phospholipids including the hippocampus, an area important for recall memory (Petursdottir et al., 2008). Supplementation of long chain omega 3 fatty acids has been studied in other species for its effects on behavior and learning ability, including mi ce and goats ( Petursdottir et al., 2008; Duvaux Ponter et al., 2008). If supplementation increases cognitive development in foals, this may lead to increased trainability. Greater trainability of horses allows for improved utilization
of horses in the indu stry, including better results in athletic performance to more suitable fit into therapeutic programs. Additionally, increased trainability may reduce the amount of unwanted horses, since they may find use in another setting. Limited study on training abil ity and cognition has been conducted in horses as compared to other species. Maze testing has been used in several studies to quantify cognition (Marinier a nd Alexander, 1994; Wolff and Ha usberger, 1996), but it fails to address the trainability of horses. Target training has been used successfully with horses and other animals to train them to complete a variety of tasks including trailer loading (Ferguson and Rosales Ruiz, 2001) The current study aimed to identify if maternal supplementation with DHA du ring late gestation and early lactation influenced the cognitive ability and memory of foals after they had been weaned. In order to assess cognitive development and memory, we used target training behaviors established when the foals were 2 mo of age as p art of a previous study. During the period between the two projects, the foals did not experience any additional training or see the target stick. The first two training sessions were used to evaluate the weanlings' memory recall of previous target trainin g behaviors. This was followed by 9 sessions of target training in which the weanlings were asked to perform a series of new tasks that culminated in loading the weanlings into a stock type trailer used for transporting horses. We hypothesize d that w eanlin gs from mares that received DHA supplementation would remember previous training faster and proceed through new training at a quicker rate with greater compliance, indicating DHA improved cognition and memory in weanlings.
Material s and Methods Animals Th is study utilized 20 stock type weanlings ( Quarter Horse, Paint, or Quarter Horse cross ; 13 fillies, 7 colts) aged 185 2 d (mean SE) weaned at 161 2 d (mean SE) During this study, w eanlings were maintained on a diet consisting of 1 1.5% body weigh t (BW) of a custom grain mix concentrate formulated for growing horses ( OBS Feed and Supply Ocala, FL) and mixed bahiagrass pasture available ad libitum Daily amounts of concentrate were split into two feedings offered at 0730 and 1530 h. Training sessio ns coincided with feeding times. On days that weanlings participated in training, a partial ration of concentrate was fed prior to a session beginning. Once the training session was completed the remainder of the ration was given to the weanling to reinfo rce a successful training session. Study Treatments Assignment of weanlings to treatment was based on maternal diet prior to weaning. Weanlings were born to mares fed 120 mg/kg BW of either a placebo supplement ( CON; n=10) or mares supplemented with long chain omega 3 fatty acids ( LCPUFA; n=10) as part of a previous study Supplementation of mares began 60 d prior to expected foaling and continued through the first 70 d of lactation. The basal diet of m ares consisted of the same custom grain mix concentrat e fed to weanlings in the current study, along with Coastal b ermuda grass hay or mixed bahiagrass pasture. Beginning at 2 mo of age, foals were offered creep feed consisting of the same grain mix concentrate fed to mares. Although foals were not directly fe d the treatment supplement, they were not restricted from accessing their dams' feed. Nutrient composition of the CON and LCPUFA supplements is presented in Table 1. Investigators in the current study were blinded to weanling treatment.
Previous Target Tra ining Prior to the start of the current study foals had been introduced to operant conditioning at 2 mo of age After an initial habituat ion phase, the trainer pair ed a primary re inforcement ( food or scratching ) with a bridge, in this case, the noise from a training clicker. The foals were taught a series of behaviors using a target which was a 0.4 m stick with a small buoy attached to one end. The trainer would present the target and give the verbal cue "target." W hen the foal touched the target with its nose, the trainer would use the bridge (clicker) and then offer the primary re inforcement The trainer worked with each foal by positioning the target in different locations near the foal to elicit different behaviors Behaviors included: t arget straight (directly in front of foal, approximately 1 .0 m above the ground), target left (on the left side of the foal), target right (on the right side of the foal), target up (above the foal's head), target down (target on the ground or close to the ground), targe t follow (the target would be placed at a distance so that the foal would need to move 2 or more steps to reach it) and t arget r andom (any of the behaviors described above were requested in random order ) Setting and Materials Weanlings were housed in a 1 6.2 ha pasture equipped with 3.5 m x 3.5 m outdoor feeding pens positioned along one fence line for individual feeding of concentrate. Training sessions took place in either the outdoor feed ing pens or in the pasture itself M aterials were added to this en vironment for Phase II and Phase IV training, including: a two horse stock trailer (interior dimensions 4.2 x 1.4 x 1.8 m with a 0.3 m step up at the entrance) that was placed in the pasture 7 d before the start of training; a black rubber mat (1 .0 x 0.8 m) that was placed on the floor, off to one side of the feed ing pen just prior to Phase II; a separate black rubber mat ( 1.2 x 0 .9 m ) and a wooden trail course bridge that were placed in the pasture for Phase IV training The bridge
had an overall dimensio n of 1.2 x 2.4 m and was structured with a ramp at each of the two shorter sides that lead up to a 1.2 x 1.0 m level platform with a peak height of 0.2 m. The same training target used for previous training, as described above, was used in the current stud y. Weanling Target Training Overview A total of 11 training sessions which asked for 10 cues were used in the current study Training sessions took place over 6 consecutive days. Each day included two consecutive sessions except day 1, which featured a s ingle session. Sessions began by offering two "clicks" from the bridge (clicker) followed by the primary re inforcement Training was separated into four phases with the ultimate goal of teaching the weanling to load into a trailer designed to transport ho rses A weanling would advance to the next phase after successful completion of the previous phase. Successful completion of a phase occurred when the weanling received five scores of 5' on a predetermined scale (Table 1 and Table 2), except for Phase I, where the weanling had to complete a fixed number of behaviors in two sessions before progressing to Phase II. Phase I focused on the weanling's me mory of target behaviors learned from previous training at 2 mo of age Phase II introduced a new behavior (" A to B") where the weanling learned to leave the primary trainer and move to another trainer. Phase II also introduced a black rubber mat that the weanling was required to stand on. When the weanling successfully completed Phase II, training moved into th e larger pasture environment for Phases III and IV. New sets of behaviors were then shaped (Phase III and IV) that prepared the weanlings for entering the trailer (Phase IV). Weanlings were trained individually and progressed at a rate that matched their level of success. This resulted in some weanlings spending more time in an
individual phase and as a result, some weanlings did not finish Phase IV. All training, with the exception of Phase II, which required a second trainer, was performed by the same p erson. Phase I Memory Testing The purpose of Phase I was to assess the speed and ease with which weanlings remembered previous target training behaviors learned at 2 mo of age The primary reward (either scratching or food) used during the initial traini ng at 2 mo of age was used as the primary reward for weanlings during Phase I of the current study Phase 1 consisted of two sessions of 10 behaviors each. In the first session, t he target w as presented in the following order: left, right, up, down, and fo llow, and the weanling had 30 sec to respond to each cue. Response time ( timed from the verbal cue to the sound of the clicker) was recorded by an observer and the weanling's response to each request was scored on a scale of 1 to 5 (T able 2) The first 5 behaviors in each of the two sessions were timed and the following 5 behaviors were only scored Successful completion of Phase I testing was achieved when the weanling received five scores of 5 beginning with the second training session. Weanlings who w ere successful in Phase I were allowed to progress to Phase II. However, before progressing to Phase II, food and scratching was offered at the end of the second session to determine if the weanling's preference for reward had changed since the initial tr aining at 2 mo of age Phase II A to B Training Phase II had two objectives: 1) teaching the foals to go from point A to point B; and 2) habituating weanlings to the flooring of the trailer by having them willingly stand on a black rubber stall mat. The first task weanlings had to complete in Phase II addressed moving from point A to point B. Two t rainers ( A and B ) were positioned on opposite sides of the feed ing pen. While the weanling was in the pen, Trainer A gave the target command and waited for the
weanling to approach and successfully touch the target. Once completed, Trainer A backed approximately 1 m away and removed the target from the weanling's line of vision. Trainer A pointed towards Trainer B and gave a new verbal cue "Go." This was followe d immediately by t rainer B who gave the normal verbal ("target") and visual cues (presentation of the target buoy). The weanling's response was scored on a scale of 1 to 5 (Table 3) W eanling s w ere required to receive five scores of 5' in a single sessio n before progressing to the second task in Phase II The second task in Phase II was designed to habituate the weanling to standing on a rubber stall mat located on the floor of the feeding pen. A to B and target follow cues were used to encourage the weanling to place all four feet on the mat. The weanling's response to each request was scored 1 to 5 (Table 3) A weanling had to receive five scores of 5' in a single session before they progressed to Phase III. Phase III Training in a New Location Th e goal of Phase III was to accustom the weanling to working in an open environment. When weanlings reached Phase III the training session started by letting the weanling out of its feed ing pen and into the adjoining pasture. The weanling had a maximum of 3 min to come to the trainer willing to work and all weanlings began training within the allowed time after being released into the pasture. When the weanling exhibited more interest in the trainer than the new surroundings, the trainer asked for previousl y established target random behaviors (i.e., randomly holding the target buoy straight, left, right, up, or down). The weanling's response to each request was scored 1 to 5 (Table 3 ). A weanling had to receive five scores of 5' in a single session before they could progress to Phase IV.
Phase IV Trailer Loading Phase IV consisted of three tasks designed to prepare weanlings for entering the trailer, followed by the last task of trailer loading. All tasks were performed with weanlings in the pasture. I n order of sequence, tasks consisted of the following: 1. Target Follow in Pasture While in the pasture, the target was placed a distance away from the weanling that required them to walk more than two steps to reach it. The weanling's response to eac h request was scored 1 to 5 (Table 3 ). A weanling had to receive five scores of 5' in a single session before they could progress to the next task. 2. Black Rubber Mat A black rubber mat was placed in the pasture prior to releasing the weanling from it s feeding pen The trainer asked the weanling to come to the area near the mat using the target follow procedure described above ( Phase IV Task 1) Next t he trainer cross ed over the mat and gave the target command while presenting the buoy to the weanli ng to encourage the weanling cross the mat. The weanling's response to each request to step onto or over the mat was scored 1 to 5 (Table 3 ). A weanling had to receive five scores of 5' in a single session before they could progress to the next task. 3. Bridge A wooden bridge was placed in the pasture prior to releasing the weanling from the feeding pen The trainer ask ed the weanling to come to the bridge area by using the target follow procedure described above ( Phase IV Task 1 ) Once the weanling was near the bridge, the trainer ask ed for approximations to encourage the weanling to cross the bridge. The first approximation was to approach within 1 m of the bridge Next the trainer asked for one foot on the bridge, then two feet, then four feet, an d finally over the bridge The weanling's response to each request was scored 1 to 5 (Table 3 ). A full crossing was characterized as when the weanling stepped from the ground onto the bridge with its front feet on one side of the bridge
and end ed with the weanling stepping off the bridge to the ground on the opposite side of the bridge A weanling had to receive five scores of 5' on full crossing in a single session before they could progress to the next task 4. Trailer Loading A two horse stock traile r was placed in the pasture 7 d before the start of Phase I training The weanling was lead to the area near the trailer using the target follow behavior as described above ( Phase IV Task 1) Once the weanling was near the trailer, the trainer asked for approximations that would ultimately result in the weanling entering the trailer (i.e., loading) The first approximation was to approach within 1 m of the trailer. Next, the trainer entered the trailer and turned to face the weanling. The buoy was presen ted and target cues given so as to encourage the weanling to put one foot, two feet, or all four feet on the trailer. The target was also used to get the weanling to exit the trailer (i.e., unload) The weanling's response to each request was scored 1 to 5 (Table 3 ). The weanling was considered to fully load in the trailer when they started with all four feet off the trailer and finished standing in the trailer. A weanling had to receive five scores of 5' on fully loading in a single session for successful completion of this final task If all phases of training were completed before all 11 sessions ended, the remaining behaviors were filled with the trainer's random choice of behaviors taught during training though an emphasis was placed on trailer loadin g and unloading for these remaining behaviors. Statistical Analyses Response time and behavior score s were collected for 20/20 weanlings during Phase I. B ehavior score s for individual tasks and the number of cues required to pass each phase or task was re corded for 20/20 weanlings during Phases II, III, IV. A Shapiro Wilk test was used to determine normality of data, and d ifferences in response time behavior score s and number of
cues were determined by a two tail t test or Wilcoxon Ranked Sum in the prog ram "R" (Version 2.31.1; The R Foundation for Statistical Computing). Differences in response were evaluated based on dietary treatment and sex. For all analyses, p 0.1 is considered a trend and p 0.05 is considered significant.
Table 1: Nutrient composition of the placebo (CON) and long chain omega 3 fatty acid (LCPUFA) supplements fed to mares 60 d before foaling through 70 d postpartum. Treatment assignment of weanlings in the current study was based on these dietary treatments assigned to their dams prior to weaning. Supplement 1 NUTRIENT CON LCPUFA Dry matter (%) 100 100 Digestible energy (Mcal/kg) 4.2 4.2 Crude pr otein (%) 14.0 13.0 Crude fat (%) 37.0 37.0 Vitamin E (IU/kg) 245 16,600 Total omega 6 fatty acids (%) 11. 2 5.4 Total omega 3 fatty acids(%) 1.2 20.9 Total DHA (%) 0 3. 9 omega 6 : omega 3 1 0 :1 0.3 : 1 1 Except for DM, all nutrients on 100% DM basis Table 2: Description of the s coring system used to evaluate weanling responses to training cues given in Phase I training Score Description 1 Weanling r eact ed negatively by taking more than 2 steps away (backwards) from the trainer and/or showed sign s of being extremely nervous/upset. 2 Weanling r eact ed negatively by taking 1 2 steps away from trainer and/or stood stationary and showed signs of being slight nervous ( e.g., pawing, subtle trembling looking away ) 3 Weanling r eact ed neutrally by standin g still and not paying attention (ignoring) and/or not making any effort to move toward target. No steps taken. 4 Weanling r eact ed positively by moving (lowering/stretching out) its head towards the target (no steps) and/or took 1 or more steps towards t he target. Include d weanling's nose coming close to the target or touching the target after 6 sec 5 Weanling r eact ed positively by responding readily once cued and touche d the target buoy with its nose within 5 sec
Table 3: Description of the s cori ng system used to evaluate weanling responses to training cues given during Phase II III, and IV training Score Description 1 Weanling reacted negatively by taking more than 2 steps away (backwards) from the trainer and/or showed signs of being extremel y nervous/upset 2 Weanling reacted negatively by taking 1 2 steps away from trainer and/or stood stationary and showed signs of being slight nervous ( e.g., pawing, subtle trembling looking away from trainer ) 3 Weanling reacted neutrally by standing sti ll and not paying attention (ignoring) and/or not making any effort to move toward the target. No steps taken. 4 Weanling reacted positively by moving (lowering/stretching out) its head towards the target (no steps) and/or took 1 or more steps towards the target. Includes weanling 's nose coming close to the target or touching the target after 11 sec. 5 Weanling reacted positively by responding readily once cued (within 10 sec) and touche d the target buoy with its nose within 10 sec Table 4: Behavior sh aping plan with descriptions of approximations (Approx.) used in during weanling target training. Note that not all approximations were used for each weanling. For example, a weanling could go directly from Approx. 24 (1 foot in trailer) to Approx. 26 (all 4 feet in trailer). Approx. Description Phase II 1 to 3 "A to B" 1 Weanling moves in the direction of Trainer B when Trainer A gives the "Go" command. Trainer B will give the target command 2 Weanling moves closer to Trainer B than Trainer A after "G o" command and target command are given 3 Weanling leaves Trainer A completely and touches target where Trainer B is located. Considered ful l behavior 4 to 8 Black Rubber Mat In Feeding Pen 4 Weanling moves within 1 m of black mat in feeding slip 5 W eanling places 1 foot on black mat in feeding slip 6 Weanling places 2 feet on black mat in feeding slip 7 Weanling places 3 feet on black mat in feeding slip 8 Weanling places 4 feet on black mat in feeding slip. Considered full behavior Phase III 9 Target Random In Pasture 9 Weanling touches target. No approximations used Phase IV 10 Target Follow In Pasture 10 Weanling touches target after moving at least two steps in any direction except backwards
11 to 16 Black Rubber Mat 11 Weanling m oves within 1 m of black mat in pasture 12 Weanling places 1 foot on black mat in pasture 13 Weanling places 2 feet on black mat in pasture 14 Weanling places 3 feet on black mat in pasture 15 Weanling places 4 feet on black mat in pastu r e 16 Weanling begins with front feet off the mat, and places all four feet on the mat while crossing over. Considered full behavior 17 to 22 Bridge 17 Weanling moves within 1 m of bridge in pasture 18 Weanling places 1 foot on bridge in pasture 19 Weanling places 2 feet on bridge in pasture 20 Weanling places 3 feet on bridge in pasture 21 Weanling places 4 feet on bridge in pastu r e 22 Weanling begins with front feet off the bridge, and places all four feet on the bridge while crossing over. Considered full behav ior 23 to 27 Trailer Loading 23 Weanling moves within 1 m of trailer in pasture 24 Weanling places 1 foot on trailer in pasture 25 Weanling places 2 feet on trailer in pasture 26 Weanling places 4 feet on trailer in pasture 27 Weanling begins with fr ont feet off the trailer, and places all four feet on the trailer in loading. Considered full behavior
Results Two LCPUFA weanlings did not finish all training for reasons unrelated to the study. One of these LCPUFA weanlings completed Phases I and II, but was removed due to injury and did not complete subsequent training (Phase III and IV). The second LCPUFA weanling had an adverse reaction to a routine vaccination and was unable to complete the last training session of the study and, as a result, was unable to complete Phase IV. Data collected from these two weanlings up to the time they were removed from the study was included in the statistical analysis. Forty percent of all weanlings were able to successfully complete all tasks in the study ( 50% CO N 30% LCPUFA ) A total of 1 5% of all weanlings were able to progress to the trailer loading task of Phase IV in 6 days but c ould not pass the final trailer loading task (Table 5). Phase I Memory Testing All weanling s were able to complete the Phase I Memory Testing (earning 5 scores of 5' during a single session) by the end of the second session. There was no effect of dietary treatment or sex in the number of cues needed to pass the memory phase (p > 0.05). During session 1, weanling treatment and s ex had no effect (p > 0.05) on first cue response time, individual cue response time, mean response time for all cues or mean score for all cues (Figure 1) For both treatments, target up had a trend for shortest mean response time during session 1 (p = 0.055) and session 2 (p = 0.092) There was no treatment or sex effect on total session score or mean session score. In the second session, mean response time for target follow was affected by treatment ( p = 0.047 ) where CON weanlings responded faster than LCPUFA weanlings (F igure 2). R esponse times for other cues were not affected by treatment (p > 0.05). There was a trend for
colts to be quicker to respond for the behavior target follow than fillies (p = 0.096 ). There was no effect of treatment or sex on total score or mean score for the second session of memory testing When comparing the two memory sessions, mean response time during session 1 ( 8.58 1.00 sec ) was greater (p = 0.031 ) than th e mean response time on during session 2 ( 4.09 0.50 s ec ) (F igure 3). Each individual cue response time was also greater during session 1 than session 2; target left (Session 1 11.73 2.29 sec Session 2 4.79 0.86 sec; p = 0.030), target right (9.44 1.82 sec 4.83 0.71 sec; p = 0.068), target up ( 3 .36 0.43 sec 2.12 0.19 sec; p = 0.009), target down ( 8.19 1.90 sec 4.16 1.40 sec; p = 0.069), and target follow ( 10.17 1.75 sec 4.58 0.62 sec; p = 0.006). Phases II IV New Behaviors Treatment and sex did not have a significant effect on mean score of new behaviors Phases II through IV (p > 0.05). All 20 weanlings completed the behavior A to B in Phase II There was no treatment effect on the number of cues it to ok to successfully complete this phase There was a trend for colts (7.6 cues) to complete the behavior with fewer cues (p = 0.068) than fillies (11.23 cues). In the same phase, all weanlings successfully complete d behaviors involving stepping onto the black mat in the feed ing pen but no treatment or sex effect was observed (p > 0.05). There was no effect by treatment in the mean score of behaviors during Phase II, but there was a trend for colts to score higher (4.80) than fillies (4.67) (p = 0.055). All weanlings successfully completed Phase III and parts of Phase IV includ ing pasture random and pasture follow behaviors, as well as crossing the black rubber mat in the pasture.
However, t here was no significant difference by treatment or sex in the number of cues it took to pass these behaviors or the mean score for these behaviors Twelve of the 19 weanlings were able to pass the bridge behavior ( 70% CON and 55.6% LCPUFA ) Of the weanlings that completed the bridge behavior, the CON weanlings were able to pass in fewer cues than LCPUFA (p = 0.037 ). Dietary treatment and sex had no effect on the number of cues it took to pass the traile r phase (p > 0.05). There was no difference in mean score for Phase IV (p > 0.05).
Table 5: Completion percentages by behavior for all weanlings, control (CON) and long chain polyunsatura ted fatty acid (LPUFA). = Percentage of 19 weanlings available for this behavior; ** = Percentage of 18 weanlings available for this behavior; Â = Percentage of 9 LCPUFA weanlings available for this behavior; = Percentage of 8 LCPUFA weanlings availabl e for this behavior. Behavior Memory A to B Black Mat in Feed Slip Pasture Random Pasture Follow Black Mat in Pasture Bridge Trailer All Weanlings 100% 100% 100% 100%* 100%* 100%* 63.16%* 44.44%** CON 100% 100% 100% 100% 100% 100% 70% 50% LPUFA 100% 100 % 100% 100%Â 100%Â 100%Â 55.56%Â 37.50% Figure 1: Mean cue response time for each behavior during session 1 in control (CON) and long chain polyunsaturated fatty acid (LCPUFA) weanling horses. !"!!# $"!!# %"!!# &"!!# '"!!# (!"!!# ($"!!# (%"!!# (&"!!# )*+,-.# /-0.# )*+,-.# 12,3.# )*+,-.#45# )*+,-.# 6789# )*+,-.# :7;;78# !"#$%#&'()$*+,#$-./$ /<=4:># @#
Figure 2: Mean cue response time for each beha vior during session 2 in control (CON) and long chain polyunsaturated fatty acid (LCPUFA) weanling horses. *Treatments differ (p < 0.05) Figure 3: Comparison of mean cue response time for each behavior between session of Phase I memory testing. The m ean response time was greater in session 1 than session 2. *Session response time differs (p < 0.05); Trend for sessions to differ (p < 0.1) !# (# $# A# %# B# C# )*+,-.#/-0.# )*+,-.# 12,3.# )*+,-.#45# )*+,-.# 6789# )*+,-.# :7;;78# !"#$%#&'()$*+,#$-./$ /<=4:># @# D# !"!!# $"!!# %"!!# &"!!# '"!!# (!"!!# ($"!!# (%"!!# )*+,-.# /-0.# )*+,-.# 12,3.# )*+,-.#45# )*+,-.# 6789# )*+,-.# :7;;78# !"#$%#&'()$*+,#$-./$ E-FF279#(# E-FF279#$# D G D G# D
Figure 4: Comparison of the number of cues needed to successfully complete specific behaviors between co lts and fillies. AB = going from trainer A to B; FM = standing on black rubber mat in the feeding pen; PN = pasture random; PF = pasture follow; BM = crossing a black rubber mat in the pasture; BR = crossing wooden bridge; TR = loading in the trailer. Tre nd for sex difference (p < 0.1) Figure 5: Comparison of the number of cues needed to successfully complete specific behaviors between control (CON) and long chain polyunsaturated fatty acid (LCPUFA) weanlings. AB = going from trainer A to B; FM = stan ding on black rubber mat in the feeding pen; PN = pasture random; PF = pasture follow; BM = crossing a black rubber mat in the pasture; BR = crossing wooden bridge; TR = loading in the trailer. *Treatments differ (p < 0.05) !# B# (!# (B# $!# $B# >H# :I# =@# =:# HI# H1# )1# 0",1#2$(3$!"#&$ :-J*;-F# I*;-F# G# !# B# (!# (B# $!# $B# A!# AB# >H# :I# =@# =:# HI# H1# )1# 0",1#2$(3$!"#&$ /<=4:># @# D
Discussion The results of th is study indicate that training introduced at a young age (2 mo) is retained for at least four months in horses. These finding are useful to the equine industry, as most previous research has not described memory ability of horses younger than one year of age (Marinier and Alexander, 1994; Wolff and Hausberger 1996). Mal et al. (1993) examine d learning ability in weanlings, but reevaluated the behavior only one day later. In the current study, memory of previously learned behavior s may have been strengthe ned due to use of positive reinforcement (Sankey et al., 2010). Additionally, weanling response time was improved during the second training session, indicating top performance may not be obtained the first day back in training but can be quickly recovere d The behavior target up had the shortest mean response time during both session 1 and session 2 of memory testing (Figure 3) Holding the target buoy above the weanling's head, as was done for the "target up" behavior places the object in the horse' s binocular visual field, which may have lead to a faster mean response time (Murphy et al., 2009). During session 2 of memory testing there was both a treatment and sex effect on the behavior target follow. The difference due to treatment may have bee n confounded because the CON group had 5 colts, whereas the LCPUFA only had 1 colt. Although study of memory in horses is limited research has generally not found gender to affect memory Marinier and Alexander (1994) did not have sufficient numbers to co nclude a sex affect in mature horses (only 3 males in a study of 9 horses), but Hausberger et al. (2004) studied multiple possible predictors of learning ability and emotionality and found sex did not affect learning ability or emotionality in mature horse s Sire was also shown to have influential effect on learning ability, but this study minimized the affect by using a limited number of sires (n = 7 ) (Hausberger et al., 2004).
The influence of DHA supplementation on memory and cognitive ability is thought to be due to an increase in DHA supply to the brain; however, no equine studies have been conducted to assess if dietary supply affects DHA concentration in the brain. Factors supporting DHA supplementation increasing cognitive ability are that DHA is the primary component of phospholipids in brain grey matter and i ncreased grey matter increases intelligence q uotient in children and adolescents (Reiss et al., 1996). In the equine brain, DHA is the highest concentrated ethanolamine phosphoglycerides (EPG) p olyenoic fatty acid (Crawfor d et al., 1976 ). This may mean DHA also influences cognition in horses, but further research should be conducted to establish this claim in non human species. The results of the current study show ed that DHA supplementation of t he mare did not enhance memory or learning ability in their foals after weaning. Th e lack of an effect of DHA on foal cognition could be due to a delayed effect of DHA supplementation on cognition, which was shown in Jensen et al. where increases to attent ion span were not observed until five years of age in the DHA supplemented group (2005, 2010). Length of gestation may also be key to the role DHA supplementation has on cognitive development. Preterm infants that received supplementation of DHA and arachi donic acid in milk after birth performed better on recognition memory and problem solving test at 6 months of age than those that didn't receive DHA supplementation (Henricksen et al., 2008). In the last semester of pregnancy, brain development occurs rapi dly and supplementation of DHA in preterm infants may be more critical than term infants (Horrocks and Yeo, 1999). While an exact timeline for brain development in horses does not exist, the brain vesicles transition to the brain between 40 d and 107 d aft er ovulation (Franciolli et al., 2011). Rattr a y et al. (1975) used an ovine model and determined there was active hyperplastic and hypertrophic brain growth during late gestation; therefor e, brain growth
should be further investigated in horses to determin e the optimum time frame for DHA supplementation. Target training was a useful training method to get weanlings to load onto a trailer; however, as a tool for assessing cognition, the procedure may need modification. The mean score for all behaviors range d from 4.51 to 4.93 on a 5 point scale The scores in this project were adapted from the scoring system used at 2 months of age when foals were more likely to have a negative reaction towards people and training. In future studies, the scoring system could combine the 1 3 scores into one score to have a greater range for positive responses to a cue. For example, a score of 5 could describe touching the target under 5 seconds, 4 between 5 10 seconds, 3 10 15 seconds, etc. Response time could also be take n for all cues. Additionally, while the use of successive approximations is a valid training tool, when attempting to measure cognition and trainability, it may have hindered results. The successive approximations used in this study were established so th at the weanlings would remain confident as the black rubber mat, wooden bridge, and trailer were introduced. Indications of treatment affects may have been masked with this approach, as each weanling was able to do well. In future studies, the number of ap proximations (Table 4) could be reduced and only rewarded in a set order. For example, the bridge behavior approximations could be modified to use only three approximations before completing the full behavior, the weanling coming within 1 m of the bridge, placing 2 feet on the bridge, and finally walking completely over the bridge. In this example, the number of cues needed to pass a behavior can still be recorded, but it also eliminates completed behaviors of weanlings that needed many appr oximations to pa ss the behavior.
The bridge behavior appeared to by the most difficult behavior, as most weanlings that passed the bridge behavior were able to pass the trailer behavior. This may be due to the novelty of the object, as the weanlings had no access to it e xcept during training sessions. It may have taken more time for the weanlings to become familiar with the bridge than other materials brought in since the black mat was placed in their feed slip before training on the second day and the trailer was placed in the pasture 7 d before training. Also, the bridge had no side enclosures, which allowed weanlings to walk around the bridge to reach the target. Putting the bridge against the pasture fence line, creating a r ailing on each end, or increasing the length of the bridge could prevent the weanlings from walking around the bridge in futur e studies. In addition, the LCPUFA colt that completed the bridge behavior required 53 cues to pass the behavior where as the mean number of cues for LC PUFA to complete the be havior was 29.6 cues and may have influenced the statistical analysis for the bridge behavior. In conclusion, t he current study provided evidence that weanlings were able to remember previously trained behaviors, but DHA supplementation of the mare did no t appear to further augment cognitive development in weaned foals There were a couple of behaviors that appeared to be affected by sex, but no overt differences in response time or number of cues needed to learn the behavior successfully between colts and fillies. T his finding agrees with other research indicating sex does not affect learning ability and memory (Wolff and Hausberger, 1996). The results of the current study also highlight the individual variability between weanling s; thus, use of a greater number of horse s would benefit future studies.
References Barboza P S K.L. Parker and I.D. Hume 2009. Integrative Wildlife Nutrition, pp. 119 131. Springer Verlag, Berlin. Bourre J M. 2009. Diet, Brain Lipids, and Brain Functions: Polyunsaturated Fat ty Aci ds, Mainly Omega 3 Fatty Acids. In:[G. Testament and G Goraccia (eds.)] Neural Lipids pp 411 420 Springer Science+Business Media, New York Crawford M.A., N.M. Casperd, A.J. Sinclair. 1976. The long chain metabolites of linoleic and linolenic acid s in liver and brain in herbivores and carnivores. Comp. Biochem. Physiol. 54B:395 401 Duvaux Ponter C K. Rigalma S. Roussel Huchette, Y. Schawlb and A.A. Ponter 2008. Effect of a supplement rich in linolenic acid, added to the diet of gestating and l actating goats, on the sensitivity to stress and learning ability of their offspring. Applied Animal Behavior Science 114:373 394 Fedorova I. and N. Salem Jr. 2006. Omega 3 fatty acids and rodent behavior. Prostaglandins, Leukotrienes, and Essential Fatty Acids 75:271 289. Ferguson D L and J. Rosales Ruiz 2001. Loading the problem loader: the effects of target training and shaping on trailer loading behavior of horses. Journal of Applied Behavior Analysis 34:409 424 Franciolli A., B. Cordeiro, E.T. da F onseca, M.N. Rodrigues, C.A. Sarmento, C.E. Ambrosio, A.F. de Carvalho, M.A. Miglino, L.A. Silva. 2011. Characteristics of the equine embryo and fetus from days 15 to 107 of pregnancy. Theriogenology 76(5):819 832. Hangii E B. 2005. The Thinking Horse: C ognition and Perception Reviewed. AAEP Proceedings 51:246 255 Hausberger M C. Bruderer N. Le Scolan and J.S. Pierre 2004. Interplay between environmental and genetic factors in temperament/personality traits in horses ( Equus caballus ). J ournal of Co mparative Psychology 118:434 446 Henricksen C K. Haugholt M. Lindgren A.K. AurvÂŒg A. Rnnestad, M. Grnn, R. Solberg,
A. Moen B. Nakstad, R.K. Berge, L. Smith, P.O. Iversen and C.A. Drevon 2008. Improved cognitive development among preterm infants attributable to early supplementation of human milk with docosahexaenoic acid and arachidonic acid. Pediatrics 121:1137 1145 Horrocks L A and Y.K. Yeo. 1999. Health Benefits of Docosahexaenoic Acid (DHA). Pharmacological Research 3:212 225 Innis S.M. 2 005. Essential Fatty Acid Transfer and Fetal Development. Placenta 26:S70 75. Jensen C L R.G. Voigt, T.C. Prager, Y.L. Zou, J.K. Fraley, J.C. Rozelle, M.R. Turcich, A.M. Llorente, R.E. Anderson, and W.C. Heird. 2005. Effects of early maternal docosahexae noic acid intake on visual function and neurodevelopment in breastfed term infants. Am J Clin Nutr 82:125 132 Jensen C L R.G. Voigt, A.M. Llorente, S.U. Peters, T.C. Prager, Y.L. Zou, J.C. Rozelle, M.R. Turcich, J.K. Fraley, R.E. Anderson, and W.C. Heird. 2010. Effects of early maternal docosahexaenoic acid intake on neuropsychological status and visual acuity at five years of age of breast fed term infants. J Pediatr 157:900 905 Jones C R T. Arai, and S.I. Rapoport. 1997. Evidence for the Invol vement of Docosahexaenoic Acid in Cholinergic Stimulated Signal Transduction at the Synapse. Neurochemical Research 22 (6): 663 670 King S S A.A. AbuGhazaleh, S.K. Webel, and K.L. Jones. 2008. Circulating fatty acid profiles in response to three levels of dietary omega 3 fatty acid supplementation in horses. J Anim Sci 86:1114 1123. Lucas A R. Morley, T.J. Cole, G. Lister, and C. Leeson Payne. 1992. Breast milk and subsequent intelligence quotient in children born preterm. Lancet 339:261 264 Mal M.E. C.A. McCall, C. Newland, and K.A. Cummins. 1993, Evaluation of a one trial learning apparatus to test learning ability in weanling horses. Appl. Anim. Behav. Sci. 35:305 311 Marinier S L and A.J. Alexander. 1994. The use of a maze in testing learning an d memory in horses. App Anim Beh Sci 39:177 182
McGreevy P D. 2007. The advent of equitation science. The Veterinary Journal 174:492 500 McGreevy P D and A.N. McLean. 2010. Equitation Science, pp. 6 35. Wiley Blackwell West Sussex. McKinley J., H.M Buchanan Smith, L. Bassett, K. Morris. 2003. Training common marmostes ( Callithrix jacchus ) to cooperate during routine laboratory procedures: Ease of training and time investment. J. Appl. Anim. Welfare Sci. 6(3);2009 220 Murphy J., C. Hall, S. Arkins. 2 009. What Horses and Humans See: A Comparative Review. International Journal of Zoology 14 pages. Petursdottie r A L S.A. Farr, J.E. Morley, W.A. Banks, and G.V. 2008. Skuladottir. Effect of dietary n 3 polyunsaturated fatty acids on brain lipid fatty a cid composition, learning ability, and memory of senescence accelerated mouse. J Gerontol A Biol Sci Med Sci 63A(11):1153 1160 Rattray P.V., D.W. Robinson, W.N. Garrett, R.C. AshmorÂŽ. 1975. Cellular changes in the tissues of lambs during fetal grow th. J. Anim. Sci. 40(4):783 788 Reiss A.L., M.T. Abrams, H.S. Singer, J.L. Ross, M.B. Denckla. 1996. Brain development, gender and IQ in children. Brain 119:1763 1774 Sankey C., M.A. Richard Yris, H. Leroy, S. Henry, and M. Hausberger. 2010. Positive inter actions lead to positive memories in horses, Equus caballus Animal Behaviour 79:869 875 The Program "R." The R Foundation for Statistical Computing. Version 2.31.1. Stelzleni, E.L., L.K. Warren and J. Kivipelto. 2006. Effect of dietary n 3 fatty acid sup plementation on plasma and milk composition and immune status of mares and foals. Joint Annual ASAS ADSA PSA Meeting, Minneapolis, MN. J. Anim. Sci. 84(Suppl. 1):392 393.
Shanahan S. 2003. Trailer loading stress in horses: behavioral and physiological effe cts of nonaver sive training (TTEAM). Journal of Applied Animal Welfare Science 6(4):263 274 Umhau J C and K.M. Dauphinais. 2007. Omega 3 Polyunsa turated Fatty Acids and Health. In:[ L. L'abate (ed)] Low cost Approaches to Promote Physical and Mental Heal th pp 87 101. Springer Science+Business Media, New York. Wolff A and M. Hausberger. 1996. Learning and memorization of two different tasks in horses: the effects of age, sex and sire. Applied Animal Behaviour Science 46:137 143 Woodward A D B.D. Niel son, C.I. O'Connor, S.K. Webel, M.W. 2005. Orth. Dietary long chain polyunsaturated acids increase plasma eicosapentaenoic acid and docosahexaenoic acid concentrations and trot stride length in horses. Proc 19 th Equine Sci Soc 101 106
Appendix 1 D efinition of Terms Behavior A particular response required of the weanling when presented with the target. Behaviors in this study were target left, target right, target up, target down, target follow, target random, A to B, black mat in feed slip, pastu re random, pasture follow, black mat in pasture, bridge, and trailer. Cue The presentation of the target for one behavior. A cue required the target to be presented, the weanling to touch the target, the secondary reward offered, and then the primary rew ard offered, or the presentation of the target and no response/negative response of the weanling. Primary Reinforcement A reward that does not require pairing/association with another stimulus to hold value as a reward. In this study, the primary reward was eit her scratching or food. Bridge (Secondary Reinforcement) A reward that has gained function by pairing it with another stimulus. In this study, the secondary reward was the noise a training clicker made. Session A series of behaviors that took place over a short period of time. Sessions had 10 cues. Target Any object that and animal can easily see and touch. In this study, the target was a 0.4 m stick with a small buoy attached to one end.