Horse behavior and facial movements in relation to food rewards

Laize G Carmo; Laís C Werner; Pedro V Michelotto, Jr; Ruan R Daros

doi:10.1371/journal.pone.0286045

. 2023 Jun 12;18(6):e0286045. doi: 10.1371/journal.pone.0286045

Horse behavior and facial movements in relation to food rewards

Laize G Carmo ¹, Laís C Werner ¹, Pedro V Michelotto Jr ¹, Ruan R Daros ^1,^*

Editor: Chris Rogers²

PMCID: PMC10259797 PMID: 37307268

Abstract

Food rewards are believed to have a positive valence in horses. The aim of this study was to assess the effect of food rewards on horse behavior before entering a horse chute, and behavior and facial movements while restrained in it. Thirteen female adult horses were brought once daily to an animal handling facility for three weeks. In week 1, baseline period, no reinforcement was applied. In weeks 2 and 3, experimental phase, half of the horses received positive reinforcement treatment after entering and remaining in the chute; the remaining horses were considered as controls (no positive reinforcement applied). There was a cross-over between the groups during the experimental phase. Horses were individually brought to the restraining chute and videos recorded during 60-sec. The duration and number of entries into the area close to the gate leading to the chute were measured before restraining and body posture, neck position, and tail swinging were recorded in the chute. Facial movements were also recorded and scored using EquiFACS methodology. Multilevel linear and logistic models were built to assess behavioral changes from baseline to the treatment phase and between phases (control and positively reinforced). Horses did not change their body posture or tail swings across the different phases (P > 0.1) and were less likely to show lowered neck during the positively reinforced phase (OR: 0.05; CI95%: 0.00–0.56; P = 0.05) compared to baseline. The likelihood of a lowered neck did not differ between the positive reinforcement and control phases (P = 0.11). In the positively reinforced phase, horses seemed to be more attentive (ears forward) and active (less eye closures, more nose movements) than in the control phase. A three-day positive reinforcement phase did not elicit major changes in body behavior in the chute but affected the facial movements of group-housed mares.

1. Introduction

Research on affective states of animals has historically focused more on negative than on positive affects, but the explicit incorporation of positive affects in animals into the animal welfare concept [1] has led to significant development of this topic in recent decades [2, 3]. Animal behavioral responses studies can provide insights into how an animal feels under a given circumstance [4] and have been successfully applied to study positive affective states in animals [5, 6] including in horses [7].

Positive reinforcement can promote motivated behavior to access rewards and can be used to train animals to voluntarily participate in activities that are necessary (like veterinary care) but not pleasurable to the animal [8, 9]. Equids trained using positive reinforcement show fewer behaviors associated with discomfort, escape, and frustration and respond more rapidly to the activity imposed, expressing more behaviors associated with positive affective states than those trained with negative reinforcement [10, 11].

The analysis of facial behaviors has contributed extensively to the field of affective states of animals study. In equids, facial movements of pain have been described, resulting in the development of the Horse Grimace Scale (HGS) [12]. More broadly, the Equine Facial Action Coding System (EquiFACS) [13] allows the evaluation of all possible facial movements of horses, which is useful for a range of affective situations. However, little research has been done to date and most studies have only applied EquiFACS in the context of negative affective states like workday load [14], transportation and social isolation [15].

This study aimed to assess the behavioral and facial movements of horses being provided positive reinforcement through food rewards to execute a routine procedure. Horses are expected to change their behavior before entering the chute (e.g., spending more time close to the gate that leads to the horse chute) while body postures in the chute are expected to be associated with behaviors of expectancy (less resting, neck above withers) during the positively reinforced phase compared with the non-reinforced phase. Only few studies have assessed facial movements of horses in positive emotional contexts. Therefore, we only predict that horses will show differences in facial movements between the positive and non-reinforced phases.

2. Materials and Methods

2.1 Animal handling

This study was approved by the Animal Use Ethics Committee of the Pontifícia Universidade Católica do Paraná (CEUA-PUCPR #01749) and was conducted at the Fazenda Experimental Gralha Azul of PUCPR, located in the municipality of Fazenda Rio Grande, Paraná, Brazil. Thirteen healthy mixed-breed non-pregnant mares, ranging from 4 to 22 years of age, were used. The mares were housed together in large grazing paddocks with free access to shade and water. Hay and concentrates were fed twice daily.

Data was collected during the last three weeks of February 2021, on the afternoons of Wednesday, Thursday and Friday of each week. Week 1 was considered the baseline and habituation phase, and weeks 2 and 3 were considered the experimental phases. During the second week half horses received a food reward after entering the restraining chute and no food reward was supplied to the rest; a cross-over between these groups was applied in the third week (Fig 1). All horses were used as controls. Details of each phase are provided in the following sections.

During data collection days, at approximately 2 pm, all horses were walked from the grazing paddock to the horse management facility (Fig 2; approximately 150 m away from the paddock), consisting of a holding area (approximately 145 m²) and a restraining area fitted with two similar horse chutes. Both areas were concrete floored. The horses had ad libitum access to water and hay in their holding area. The restraining area was fully covered, whereas the holding area covered 55% of its area.

2.2 Data collection setup

Three cameras were used for recording the study. A wide-lens camera (GoPro, model Hero3, GoPro Inc., California, USA) was positioned 4 m high adjacent to the management facility, to capture horse behavior in the holding area. The camera remained on during all data collection. To video record the behaviors in the horse chute, a hand camera (Sony, model DCR-PJ5, Sony Group Corporation, Tokyo, Japan) was placed at 1.1 m high, 1 m away from the back of the horse chute (Fig 3). Another camera (Canon, model T3i, Canon Inc., Tokyo, Japan) was positioned in front of the horse chute at a 45° angle, 1.5 m away from the chute and 1.5 m high, to capture the horse face (Fig 3). The video recording of these cameras (Sony and Canon) was synchronized and started when the horse entered the chute and stopped when the horse exited the chute.

The horses were free in the paddock before the start of data collection and were led in a group to the horse management facility. During the route, horses could assume different leadership positions and entry orders in the management facility according to their social configuration. The holding area of the management facility was divided into two sections. The section of the holding area that was closer to the gate accessing the restraining area was considered the central area, whereas the remaining sections of the holding area were considered peripheral (Fig 2). The horses were free to move between all sections and had access to hay and water from both peripheral and central areas. The GoPro captured the horses’ movements in both sections.

After all the horses were in the management facility, they were taken individually by a lead rope to the chute. The chute entering order was pseudo-randomized (alternating between control and reinforced horses) and kept constant during the three weeks, i.e., horses were brought to the chute in the same order every day.

Before the study, horses were familiar to management facility but not to the feed rewards used in this study. Thus, one day before the beginning of the study, all horses were offered carrots, apples, and molasses treats. The procedure was performed twice. All horses promptly ate their carrots. Nine horses ate apples, and the molasses treat was eaten by three horses. None of the horses showed any sign of aversion (i.e., backing up after sniffing the food).

2.3 Habituation phase

This phase was completed during the first week of the study. The herd was brought to the management facility and allowed a period of 10 min without disturbance before the first horse was led to the restraining chute. All horses were familiar with the testing environment prior to the experiment and were accustomed to entering the chute for veterinary treatment and daily reproductive management, but the individual experiences of each horse and their perceptions based on personality are unknown by the authors. A member of the research team familiar with the horses then entered the holding area, placed a halter on a horse, and led her to the chute. The horse was left in the chute without physical contact for 1 min. During this period, both cameras (Sony and Canon) were on, and the researcher observed the horse for selected behaviors live.

The observer stayed approximately 1.5 m away in front of the horse (Fig 3) without looking directly into her eyes and behaving in the most disinterested way possible. The videos were always recorded at the same place lasting 1 minute, and during video recordings, there was no movement of people or other animals close the horses. After the time in the chute, the horse was returned to the holding area and her halter removed. This procedure was repeated for each horse in this study.

2.4 Experimental phase

This phase was conducted in the second and third weeks of the study. The same handling procedures as described previously were used in this phase; however, seven horses received a food reward after entering the restraining chute and no food reward was supplied to the rest, i.e., they were in the control group in week 2. In week 3, a crossover was applied (i.e., the seven horses that received food rewards in week 2 were now in the control group and vice-versa).

The mares were offered a total of six food rewards (two small pieces of carrots, apples, and molasses treats). All food items were hand-fed one at a time starting with the carrot, apple, and then the molasses treat. If the horse refused any item, it was put aside and the next piece was offered. The rationale was to offer the horses a chance to eat or pass as they wish. The food items were offered immediately after the front gate of the restraining chute was closed and the horse was not pushing against the gates; thus, reinforcing the short sequence of behaviors that happens after the chute has been closed and the food offering. We highlight that the person closing the chute was not the same as the one offering the food item. This procedure and timing were applied consistently throughout the study.

Since chewing influences facial movements and other behaviors, the horses were recorded during the whole period that they were in the chute, but only a 60-sec video clip starting immediately after horses had finished chewing was used. The time spent in the chute from entering to release ranged from 4 to 6 min for horses in the experimental group. In the control group, no waiting period was applied, so the 60-sec video clip started immediately after the horse was closed in the chute. This approach was used to not create a situation that was novel and, possibly negative, to the horse (i.e., staying in the chute for 4 to 6 min without anything happening), potentially invalidating comparisons between phases. The observation time was standardized for each horse—i.e., a synchronized 60-sec video clip in both cameras (Sony and Canon).

2.5 Behavioral analysis

The holding area behaviors investigated in this study (evaluating data from the GoPro camera) were arrival order from the paddock to the holding area, time in each section of the holding area and frequency of entering the central section of the holding area before and after passing through the chute. Time variables were further categorized into a proportion, as horses who were first to enter the horse chute would, by default, have less time in the holding area before passing through the chute. Holding area behavioral data were assessed through video analysis by a single observer. Each video was analyzed for each horse and was manually recorded in a spreadsheet.

During the restraining period, several behaviors were recorded live and by video (Sony and Canon cameras). The selection of behaviors in the chute was based on and adapted from Young et al. [16]. Descriptions of each evaluated behavior are presented in Table 1. Behaviors described were collected continuously for 60-sec and annotated in a spreadsheet by an observer who was not blinded to the treatments and was always present being possible to observe the entire horse while in the chute. Pawing, kicking, rear and defecation/urination were recorded according to the frequency they occurred, while standing, resting and neck position and tail movement were recorded according to the duration.

Table 1. Behaviors assessed during the restraining period in a horse chute.

Behavior	Behavioral description
Standing	Standing with the weight resting on 4 legs.
Resting	Standing with the weight resting on 3 legs.
Pawing	Striking a vertical or horizontal surface or the air with forelimbs.
Kicking	One or both hind legs thrust backwards or to the side, contact with part of the stable may be achieved.
Rear	Front legs raised off the ground with forehand higher than hindquarters.
Neck above withers	Eye-level elevated above the height of withers, but below 45 degrees
Neck below withers	Eye-level parallel or below the height of withers.
Neck over 45 degrees	Neck raised over 45 degrees.
Tail neutral	Fleshy part of tail relaxed against the body.
Tail swish	Tail is flicked to one side and/or the other of the quarters.
Defecation/Urination	Elimination of feces and urine.

Open in a new tab

Video clips from the last two days of the experimental phase (week 2 and 3, totaling 52 video clips of 60-sec) were analyzed to extract data regarding the facial action units (FAUs) of each horse in the chute.

A total of 17 Action Units (AUs) and 14 Action Descriptors (ADs) were assessed, following the EquiFACS methodology described by Wathan et al. [13]. The start and end of each assessed FAU was manually annotated by the same FACS coder watching on a computer. Each video was watched multiple times as only one FAU was observed and was rewound or slowed down when in doubt, being scored only when visible. Our objective was to evaluate the occurrence of all AUs during the 60-sec, however, in the moments when it was not possible to score AUs, no data was generated. We emphasize that most of the videos provided good visibility and good angle as the horses remained still. All the data were manually annotated using a spreadsheet. Facial recordings were only possible in the chute as the camera recording the holding area behavior was set to capture images from the whole herd.

The categorization of each FAU was done following the EquiFACS methodology [12], though we adapted the measurement of FAUs that are short lasting to measure frequency only (eye closure, blink, half blink, lower lip depressor, chin raiser and lip presser). FAUs that lasted longer were assessed as duration and frequency. The time required to correctly classify each FAU was always accounted for. In Table 2, we inform which FAUs were observed, using the codes AU, AD, AUH and EAD, as described in Wathan et al., 2015. The duration and frequency of each FAU were analyzed separately. Thus, the total duration (in seconds) and the frequency of each FAU are summarized per horse per day. The sequences of activation of each FAU and the co-occurrence during the 60-sec video clip were not considered.

Table 2. Facial action units (FAUs) assessed in the study using EquiFACS: The equine facial action coding system.

FAU¹	Description
AU101	Inner Brow Raiser
AU143	Eye Closure
AU145	Blink
AU47	Half Blink
AU5	Upper Lid Raiser
AD1	Eye White Increase
AU10	Upper Lip Raiser
AU12	Lip Corner Puller
AU113	Sharp Lip Puller
AUH13	Nostril Lift
AU16	Lower Lip Depressor
AD160	Lower Lip Relax
AU17	Chin Raiser
AU18	Lip Pucker
AU122	Upper Lip Curl
AU24	Lip Presser
AU25	Lips Part
AU26	Jaw Drop
AU27	Mouth Stretch
EAD101	Ears Forward
EAD102	Ear Adductor
EAD103	Ear Flattener
EAD104	Ear Rotator
AD29	Jaw Thrust
AD30	Jaw Sideways
AD133	Blow
AD38	Nostril Dilator
AD55	Head Tilt Left
AD56	Head Tilt Right
AD57	Nose Forward
AD58	Nose Back

Open in a new tab

¹ Detailed description of each FAU can be found in Wathan et al. [13]. AU = Action Unit; AD = Action Descriptor; EAD = Ear Action Descriptor.

Interobserver reliability was calculated (irr, R package) using data collected independently on five 60-sec video clips selected randomly from horses participating in the study. None of the video clips were from the experimental phase of the study. In these videos, the frequency of each FAU was calculated and the intraclass coefficient (ICC) was used to determine the level of agreement between observers [17]. To account for the non-random selection of observers for this study and consistency across observations, the experimental design was such that the ICC was set for “two-way” and “consistency” as suggested by Hallgren [17]. ICC values ranged from -1 to 1 (complete disagreement to excellent agreement; zero being agreement no different than chance). The ICC was 0.73 (95% CI: 0.63–0.80); ranging from good to excellent agreement [17].

2.6 Statistical analysis

Statistical analysis was performed using R (version 4.0.4; R Core Team, 2021). Original data (.csv format) were imported into R and organized using the tidyverse package [18]. The original data, R scripts, and output can be found at https://doi.org/10.6084/m9.figshare.21714233.v1.

For numeric outcome variables and binary outcome variables (i.e., all behaviors except FAUs), mixed linear regression and mixed logistic regressions were applied, respectively [19]. As the study was divided into a baseline and an experimental phase, two sets of models were developed: 1) comparisons between the control and positive reinforcement phases, and 2) comparison between the baseline and the positively reinforced phase. Only the third day of the control and positively reinforced phases was used for comparison between data collected in the holding area, as that is the day we were most certain that the mare was habituated the experimental phase she was in. All three days of the baseline phase were used to compare the behaviors collected live and continuously. Horses were included as a random effect to allow for repeated measures. The order of treatments (control/reinforced) in the experimental phase was included in all models at first, but it did not affect the results (P > 0.1); thus, it was dropped from all final models. Age was also included but no effect was detected and thus dropped from the final models.

A similar statistical approach was used for each FAU; however, comparisons were only made between the control and positively reinforced phases using data from the last two days of each phase. The analysis of FAUs that presented high skewness (graphically assessed) was complemented with an additional generalized mixed model using Poisson distribution. The model assumptions of homogeneity and linearity were assessed graphically.

3. Results

3.1 Holding area behavior

It was expected that during the positively reinforced days, horses would arrive at the management facility before those that were under the control phase. However, this hypothesis was not confirmed because horses did not significantly change their arrival order (t = 0.44; P = 0.66) from the baseline to the positively reinforced days. When comparing the positive reinforcement with the control phase, no differences were detected in the order of arrival at the holding facility (P = 0.80).

There were no differences in the time spent in the central section of the holding area before passing through the chute between the control and positively reinforced phases (t = 0.87, P = 0.39). However, when comparing the last day of the positively reinforced phase against the baseline, there was a tendency for difference between these phases; during the positively reinforced days, horses spent 40% of the time in the central area before passing through the chute while they only spent 28% in the baseline phase (t = 1.81; P = 0.08).

On average, horses entered the central section of the holding area five times. This variable did not differ between the control and positively reinforced phases (t = 0.07; P = 0.94) or between the baseline and positively reinforced phases (t = 0.46; P = 0.65).

3.2 Behavior in the chute

When provided with a food reward, all horses ate at least part of the treats. They did not differ in most of the behaviors assessed while in the chute. However, defecation/urination, rearing, elevated neck, pawing, and kicking had very low frequencies or did not occur during any day of the study; therefore, they were not analyzed further (Table 3).

Table 3. Total events of infrequently occurring body behaviors of all 13 mares observed during all sessions of 60-sec interval in a restraining chute.

Experimental phase / Behavior	Elevated neck	Pawing	Kicking	Rearing	Elimination
Positive reinforcement	1	1	0	0	1
Control	4	2	1	0	0

Open in a new tab

Comparisons between the control and positive reinforcement phases and the baseline and positive reinforcement phase were performed to assess the frequency of lowered neck, resting all time, never resting, and tail swish. Only one behavior showed difference between the phases. Horses had 95% reduced odds of showing a lowered neck during the positively reinforced phase (OR: 0.05; CI95%: 0.00–0.56; P = 0.05) compared to the baseline phase. However, the likelihood of a lowered neck did not differ between the positive reinforcement and control phases (OR: 0.0; CI95%:0.00–38.0; t = -1.59; P = 0.11). We found no evidence of order effect of groups in any of our models.

3.3 Facial action unit—EquiFACS

The comparisons between the control and positively reinforced phases for each FAU are presented in Table 4. In brief, out of the 31 FAUs assessed (AUs and ADs), 20 presented differences between phases, seven did not differ between phases, and four were not observed in any phase. Among those that presented differences between the phases, there was a higher frequency or duration of eye closure (AU143), inner brow raiser (AU101), lower lip relax (AD160), lips part (AU25), ear rotator (EAD104), and nostril dilator (AD38) in the control phase. During the positive reinforcement phase, eye white increase (AD1), lip corner puller (AU12), chin raiser (AU17), lip pucker (AU18), upper lip curl (AU122), jaw drop (AU26), ears forward (EAD 101), head tilt right (AD55), and nose forward and back (AD57, AD58) were seen more often or for longer. Mouth stretch (AU27) was only observed in the positively reinforced phase. Upper lip raiser (AU10), sharp lip puller (AU113), and jaw thrust and sideways (AD29, AD30) were not observed during any phase.

Table 4. Estimated means for the positively reinforced (PR) and control phases from mixed linear regression models of each facial action unit (FAU) evaluated in 13 mares.

		Duration (seconds)			Frequency
FAUs	Description	PR	Control	P-value	PR	Control	P-value
AU101	Inner Brow Raiser	6.5 ± 1.6	11.1 ± 1.6	0.05 ¹	3.2 ± 1.5	3.8 ± 1.5	0.75 ¹
AU143	Eye Closure	-	-	-	0.2 ± 0.3	1.0 ± 0.3	0.03 ¹
AU145	Blink	-	-	-	18.7 ± 1.7	17.0 ±1.7	0.22
AU47	Half Blink	-	-	-	6.4 ± 0.8	6.0 ±0.8	0.68
AU5	Upper Lid Raiser	5.3 ± 1.9	3.5 ± 1.9	0.33	4.6 ± 1.1	1.6 ± 1.1	0.01 ¹
AD1	Eye White Increase	3.4 ± 1.1	2.0 ± 1.1	0.23 ¹	1.9 ± 0.6	0.76 ± 0.6	0.06 ¹
AU10	Upper Lip Raiser	NO²	NO²	-	NO²	NO²	-
AU12	Lip Corner Puller	2.1 ± 0.9	1.3 ± 0.9	0.35 ¹	0.33 ± 0.2	0.16 ± 0.2	0.44
AU113	Sharp Lip Puller	NO²	NO²	-	NO²	NO²	-
AUH13	Nostril Lift	3.9 ± 2.9	7.8 ± 2.9	0.18	0.1 ± 0.1	0.2 ± 0.1	0.33
AU16	Lower Lip Depressor	-	-	-	5.0 ±2.0	4.8 ±1.9	0.95
AD160	Lower Lip Relax	24.9 ± 4.4	37.3 ± 4.3	0.03	4.0 ± 1.6	4.4 ± 1.6	0.81
AU17	Chin Raiser	-	-	-	4.8 ±1.8	3.4 ±1.8	0.59 ¹
AU18	Lip Pucker	8.6 ± 2.5	2.8 ± 2.5	0.02 ¹	4.2 ± 1.5	1.9 ± 1.5	0.17 ¹
AU122	Upper Lip Curl	3.3 ± 1.7	1.2 ± 1.7	0.40 ¹	3.0 ± 1.6	1.0 ± 1.6	0.40 ¹
AU24	Lip Presser	-	-	-	8.6 ± 2.7	7.4 ± 2.7	0.75
AU25	Lips Part	2.1 ± 2.8	9.2 ± 2.8	0.02 ¹	1.2 ± 1.6	5.2 ± 1.6	0.05 ¹
AU26	Jaw Drop	5.7 ± 1.3	2.3 ± 1.3	0.06 ¹	0.7 ± 0.9	1.34 ± 0.9	0.59
AU27	Mouth Stretch	2.5 ± 1.3	NO ²	-	2.7 ± 1.7	NO ²	-
EAD101	Ears Forward	30.2 ± 4.1	13.8 ± 4.0	< 0.01 ¹	3.2 ± 1.5	3.8 ± 1.5	0.79
EAD102	Ear Adductor	12.0 ± 3.4	13.5 ± 3.3	0.65 ¹	2.8 ± 0.7	2.0 ± 0.7	0.43
EAD103	Ear Flattener	NO ²	0.8 ± 0.6	-	NO ²	0.8 ± 0.6	-
EAD104	Ear Rotator	14.7 ± 4.0	31.2 ± 3.9	< 0.01 ¹	4.4 ± 1.4	4.6 ± 1.4	0.93
AD29	Jaw Thrust	NO ²	NO ²	-	NO ²	NO ²	-
AD30	Jaw Sideways	NO ²	NO ²	-	NO ²	NO ²	-
AD133	Blow	0.38 ± 0.1	0.12 ± 0.1	0.14	0.2 ± 0.2	0.5 ± 0.2	0.32
AD38	Nostril Dilator	9.4 ± 2.2	12.6 ± 2.2	0.29 ¹	16.6 ± 2.2	16.4 ± 2.1	0.96
AD55	Head Tilt Left	0.3 ± 0.1	0.1 ± 0.1	0.16	0.3 ± 0.1	0.1 ± 0.1	0.12
AD56	Head Tilt Right	0.4 ± 0.1	0.1 ± 0.1	0.05 ¹	0.6 ± 0.2	0.1 ± 0.2	0.04 ¹
AD57	Nose Forward	2.9 ± 0.8	0.5 ± 0.8	0.02	2.8 ± 0.7	0.6 ± 0.7	< 0.01¹
AD58	Nose Back	2.2 ± 0.5	1.2 ± 0.5	0.20 ¹	1.3 ± 0.3	0.4 ± 0.3	0.04 ¹

Open in a new tab

“—” Not assessed.

¹ This facial action unit presented a difference (P < 0.05) between phases when analyzed via mixed Poisson regression. All statistical details can be found in the Supplementary statistical files.

² Not observed.

Ear movements were the behaviors that lasted longest, with calculated mean during the positively reinforced phase showing that mares spent 50% of the observed time pointing their ears forward (EAD101), while during the control phase they spent only 23% of the observed time. Conversely, the ear rotator (EAD104) was activated for a longer period during the control phase (52%) of the time compared with the positively reinforced phase (25%). Ear adductor (EAD102) was observed on average for 20% of the time in both phases, and ear flattener (EAD103) was observed only once during the control phase and remained active for 20 s.

4. Discussion

Contrary to our hypothesis, horses did not spend more time in the central area of the handling facility during the reinforcement phase and there were no major differences in the behaviors displayed in the chute during the study, except for a reduction in the likelihood of horses showing a lowered neck in the positively reinforced phase. However, several specific changes in facial movements were detected in the positively reinforced phase.

4.1 Behavior in the holding area

It was expected that horses receiving food rewards would be more motivated to arrive faster at the management facility. The speed of relocation from the paddock to the facility could have been a better indicator of motivation, which could not be measured. However, there is no evidence that the horses changed their behavior as a function of food rewards. This may have been influenced by the low number of trials (three days) receiving food rewards or that the quantity or quality of the rewards was not sufficient to show marked behavioral changes. However, horses expecting a reward may approach a test area more rapidly than horses under standard management without being given rewards, which was shown within the first few days of the start of reward provisioning [20] and a study demonstrated that horses may also learn to discriminate between compartments with and without a reward within 1 d after being positively reinforced [21].

A variety of food types were used as rewards and allowed the horses to refuse any item if desired. Since all mares ate most of the food items offered in all trials, the food reward was adjudged as non-aversive. The social hierarchy may have influenced the location of the horses in the holding area; thus, horses motivated to approach the central section may have been discouraged from doing so by a more dominant mare [22, 23]. This effect may have been exacerbated by the small holding area. Social hierarchy or interactions were not measured in the holding area.

4.2 Behavior in the chute

Horses had a low probability of showing a lowered neck during the positive reinforcement phase. Neck position has been linked with excitation levels in horses; thus, elevated necks may reflect a state of high arousal, hypervigilance, or attention towards specific stimuli [24]. Considering that no sonorous stimulus was applied and there was no movement of people or any local noise, we ruled out that horses showed less lowered neck more often during the positive reinforced phase because of differences in attention-related behaviors between phases. Lowered necks have been associated with more relaxed and calm states [25, 26]. Thus, it is likely that horses being offered food rewards were more attentive and remained with their neck more elevated, but not hypervigilant.

4.3 Facial action units

To date facial expressions in horses have mostly been described in the context of painful or stressful experiences. However, certain features of the face have already been identified in a positive context [7, 27], although these studies have not did not use EquiFACS methodology. Lansade et al. [7] found that analyses of facial movements during grooming revealed significant differences between groups. However, a negative stimulus (spraying a mild perfume close to the horse) induced a more marked change in facial behaviors compared to positive (grooming) and neutral (observer close to the horse, without physical contact) contexts [27]. We emphasize that our study did not aim to evaluate the facial expression of horses, however, we presented the activation response (duration and frequency) of each FAU separately. Isolated facial movements can also be promising to identify a prominent action (either in duration or frequency) in situations of positive affective valence. We understand that only together could they reflect a “facial expression”, therefore, in future studies, the FAUs should be analyzed in co-occurrence.

Ear position has been associated with emotional valence in horses [28]. During the reinforcement phase, horses spent more time with their ears forward (EAD101), suggesting that they were more attentive. This behavior has been linked to both positive [29] and negative [14] situations in horses. Thus, it is likely that forward ears alone are not sufficient to indicate the emotional valence of horses. Conversely, backward ears (EAD104) may reflect an aversive or painful state [11, 30, 31]. During the control phase, horses spent longer with their ears backward. To avoid any confounding with mares paying attention to other sources of noise or people movement, we carefully chose to conduct our data collection during the period of low area usage and guided all staff and students to stay clear during our data collection.

Horses have presented nostril lifts (AUH13) in slightly negative contexts, such as transportation and social isolation [14]. There were no differences in nostril lift between the study phases, but nostrils were more likely to be dilated for longer during the control phase, perhaps indicating that standing in a chute without stimuli away from the herd (although visual contact was possible) was negative for the horses. Conversely, the nose forward (AD57) and nose back (AD58) were observed more often in the positively reinforced phase. We are unaware of studies that have linked this behavior with an equine affective state.

Most of the studies that assessed eyelid movements did not provide a thorough description of the eyelid movement [32, 33], making it difficult to compare our results. We found that during the control phase, horses had higher frequencies of eye closure (AU143) compared to the reinforced phase (a movement different from blinking or half blinking). It is possible that the control horses closed their eyes more as a result of being in a more relaxed state or engaging in resting behavior. However, blink (AU145) and half blink (AU47) showed no difference between the phases. It seems that blinking frequency may not be a good indicator of emotional valence as there are conflicting results regarding it. Increased blink frequency has been associated with slightly negative situations, such as transportation and social isolation [14, 32] but one study found that the total number of blinks was reduced in horses when social isolation or feeding was withheld for a short period [33]. Future studies should assess the role of arousal on blinking frequency.

Lower lip relax (AD160) was more active in the control than in the reinforced phase. In our study, videos were taken immediately after horses had swallowed the food rewards; therefore, the tension in the lips likely lasted for several seconds during the video recording phase. Indeed, when looking at the frequencies of lower lip relaxation between phases, horses showed no differences between phases.

Head tilt right (AD56) occurred more often and was longer in the positively reinforced phase. Although this may have been associated with laterality [34, 35], we are cautious of this interpretation, as in our study, the structures on both sides of the chute were different and may have attracted more horses immediately after receiving their food reward.

Our control phase consisted of a low-interaction phase, possibly putting horses in a neutral or low-arousal emotional state. It could be argued that horses had generally negative experiences in the chute before the study and experienced social isolation during the study, which may have masked the potential positive effect of the food reward. However, horses were not reluctant to enter the chute even in the baseline phase. Furthermore, the use of food rewards may not have contributed to horses’ affective states in the way we expected. Future studies should assess horse preferences and motivations for accessing food rewards in comparison to other positive stimuli, such as grooming and verbal praise [36]. Blinded observers should be considered to reduce potential bias and the FAUs should be analyzed in co-occurrence. Finally, our positive reinforcement may only have affected the arousal level of the mares and did not specifically affect emotional valence.

5. Conclusions

The use of food rewards did not generate marked behavioral differences in horses and did not elicit major changes in behavior while restrained in a chute. However, several changes in horse facial movements were detected. After providing food rewards, horses entered a more attentive state with ears forward, less eye closure, and less nostril dilation but increased nose movement and more right head tilt, possibly indicating a feed-driven emotional state. However, we were unable to verify whether food reward affected the emotional valence of mares.

Acknowledgments

We thank the Fazenda Experimental Gralha Azul (FEGA-PUCPR) staff for their help in setting up the experiment, and the students who helped with data collection: Alécia Mendes Araújo da Cunha, Alice Marega Reis, Ana Laura Gervinski, Bianca Barbosa, Camila Trojanovski das Neves, Jessica dos Santos Moreira, Lorena Tonetti Assis, Matheus Borges de Carvalho, and Renata Aline Gagliano. LGC received scientific scholarships (PUCPR-PAIC200590 and PUCPR-PAIC210069) from the Brazilian National Council for Scientific and Technological Development (CNPq) during the study period and manuscript writing.

Data Availability

The original data, R scripts, and output can be found at URL: (https://doi.org/10.6084/m9.figshare.21714233.v1) DOI: (10.6084/m9.figshare.21714233.v1).

Funding Statement

The authors received no specific funding for this work.

References

1.Mellor DJ. Updating animal welfare thinking: Moving beyond the “Five Freedoms” towards “a Life Worth Living”. Animals. 2016; 6(3): 21. doi: 10.3390/ani6030021 [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Boissy A, Manteuffel G, Jensen MB, Moe RO, Spruijt B, Keeling LJ, et al. Assessment of positive emotions in animals to improve their welfare. Physiol Behav. 2007; 92(3): 375–397. doi: 10.1016/j.physbeh.2007.02.003 [DOI] [PubMed] [Google Scholar]
3.Rault JL, Hintze S, Camerlink I, Yee JR. Positive welfare and the like: Distinct views and a proposed framework. Front Vet Sci. 2020; 7: 370. doi: 10.3389/fvets.2020.00370 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Weary DM, Droege P, Braithwaite VA. Behavioral Evidence of Felt Emotions: Approaches, Inferences, and Refinements. Adv Study Behav. 2017; 49: 27–48. doi: 10.1016/bs.asb.2017.02.002 [DOI] [Google Scholar]
5.van der Harst JE, Spruijt BM. Tools to measure and improve animal welfare: Reward-related behaviour. Anim Welf. 2007; 16(Suppl.): 67–73. [Google Scholar]
6.Yeates JW, Main DC. Assessment of positive welfare: a review. Vet J. 2008; 175(3): 293–300. doi: 10.1016/j.tvjl.2007.05.009 [DOI] [PubMed] [Google Scholar]
7.Lansade L, Nowak R, Lainé AL, Leterrier C, Bonneau C, Parias C, et al. Facial expression and oxytocin as possible markers of positive emotions in horses. Scientific reports. 2018; 8(1), 14680. doi: 10.1038/s41598-018-32993-z [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Laule GE, Bloomsmith MA, Schapiro SJ. The use of positive reinforcement training techniques to enhance the care, management, and welfare of primates in the laboratory. J Appl Anim Welf Sci. 2003; 6(3): 163–173. doi: 10.1207/S15327604JAWS0603_02 [DOI] [PubMed] [Google Scholar]
9.Lambeth SP, Hau J, Perlman JE, Martino M, Schapiro SJ. Positive reinforcement training affects hematologic and serum chemistry values in captive chimpanzees (Pan troglodytes). Am J Primatol. 2006; 68(3): 245–256. doi: 10.1002/ajp.20148 [DOI] [PubMed] [Google Scholar]
10.Hendriksen P, Elmgreen K, Ladewig J. Trailer-loading of horses: Is there a difference between positive and negative reinforcement concerning effectiveness and stress-related signs?. J Vet Behav. 2011; 6(5): 261–266. doi: 10.1016/j.jveb.2011.02.007 [DOI] [Google Scholar]
11.Freymond SB, Briefer EF, Zollinger A, Gindrat-von Allmen Y, Wyss C, Bachmann I. Behaviour of horses in a judgment bias test associated with positive or negative reinforcement. Appl Anim Behav Sci. 2014; 158: 34–45. doi: 10.1016/j.applanim.2014.06.006 [DOI] [Google Scholar]
12.Dalla Costa E, Minero M, Lebelt D, Stucke D, Canali E, Leach MC. Development of the Horse Grimace Scale (HGS) as a pain assessment tool in horses undergoing routine castration. PLoS One. 2014; 9(3): 1–10. doi: 10.1371/journal.pone.0092281 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Wathan J, Burrows AM, Waller BM, McComb K. EquiFACS: The equine facial action coding system. PLoS One. 2015; 10(8): 1–35. doi: 10.1371/journal.pone.0131738 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Trindade PHE, Hartmann E, Keeling LJ, Andersen PH, Ferraz GDC, Paranhos da Costa MJR. Effect of work on body language of ranch horses in Brazil. PLoS One. 2020; 15(1): e0228130. doi: 10.1371/journal.pone.0228130 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Lundblad J, Rashid M, Rhodin M, Andersen PH. Effect of transportation and social isolation on facial expressions of healthy horses. PLoS One. 2021; 16(6 June): 1–17. doi: 10.1371/journal.pone.0241532 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Young T, Creighton E, Smith T, Hosie C. A novel scale of behavioural indicators of stress for use with domestic horses. Appl Anim Behav Sci. 2012; 140(1–2): 33–43. doi: 10.1016/j.applanim.2012.05.008 [DOI] [Google Scholar]
17.Hallgren KA. Computing inter-rater reliability for observational data: an overview and tutorial. Tutor Quant Methods Psychol. 2012; 8(1): 23–34. doi: 10.20982/tqmp.08.1.p023 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Wickham H, Averick M, Bryan J, Chang W, McGowan LD, François R, et al. Welcome to the tidyverse. J Open Source Softw. 2019; 4: 1686. doi: 10.21105/joss.01686 [DOI] [Google Scholar]
19.Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Softw. 2015; 67. doi: 10.18637/jss.v067.i01 [DOI] [Google Scholar]
20.Peters SM, Bleijenberg EH, van Dierendonck MC, van der Harst JE, Spruijt BM. Characterization of anticipatory behaviour in domesticated horses (Equus caballus). Appl Anim Behav Sci. 2012; 138(1–2): 60–69. doi: 10.1016/j.applanim.2012.01.018 [DOI] [Google Scholar]
21.Mal ME, McCall CA, Newland C, Cummins KA. Evaluation of a one-trial learning apparatus to test learning ability in weanling horses. Appl Anim Behav Sci. 1993; 35(4): 305–311. [Google Scholar]
22.Christensen JW, Rundgren M, Olsson K. Training methods for horses: habituation to a frightening stimulus. Equine Vet J. 2006; 38(5): 439–443. doi: 10.2746/042516406778400574 [DOI] [PubMed] [Google Scholar]
23.Hall C, Goodwin D, Heleski C, Randle H, Waran N. Is there evidence of learned helplessness in horses? J Appl Anim Welf Sci. 2008; 11(3): 249–266. doi: 10.1080/10888700802101130 [DOI] [PubMed] [Google Scholar]
24.Fureix C, Hausberger M, Seneque E, Morisset S, Baylac M, Cornette R, et al. Geometric morphometrics as a tool for improving the comparative study of behavioural postures. Naturwissenschaften. 2011; 98(7): 583–592. doi: 10.1007/s00114-011-0803-2 [DOI] [PubMed] [Google Scholar]
25.Smiet E, Van Dierendonck MC, Sleutjens J, Menheere PPCA, van Breda E, de Boer D, et al. Effect of different head and neck positions on behaviour, heart rate variability and cortisol levels in lunged Royal Dutch Sport horses. Vet J. 2014; 202(1): 26–32. doi: 10.1016/j.tvjl.2014.07.005 [DOI] [PubMed] [Google Scholar]
26.Hall C, Randle H, Pearson G, Preshaw L, Waran N. Assessing equine emotional state. Appl Anim Behav Sci. 2018; 205: 183–193. doi: 10.1016/j.applanim.2018.03.006 [DOI] [Google Scholar]
27.Stratton R, Cogger N, Beausoleil N, Waran N, Stafford K, Steward M. Indicators of good welfare in horses: Final report. MPI Technical Paper, (2014/44). Ministry for Primary Industries, New Zealand Government, 52 p.; 2014. https://www.mpi.govt.nz/dmsdocument/5089. [Google Scholar]
28.Reefmann N, Kaszàs FB, Wechsler B, Gygax L. Ear and tail postures as indicators of emotional valence in sheep. Appl Anim Behav Sci. 2009; 118(3–4): 199–207. doi: 10.1016/j.applanim.2009.02.013 [DOI] [Google Scholar]
29.Sankey C, Richard-Yris MA, Henry S, Fureix C, Nassur F, et al. Reinforcement as a mediator of the perception of humans by horses (Equus caballus). Anim Cogn. 2010; 13(5): 753–764. doi: 10.1007/s10071-010-0326-9 [DOI] [PubMed] [Google Scholar]
30.Dyson S, Berger J, Ellis AD, Mullard J. Development of an ethogram for a pain scoring system in ridden horses and its application to determine the presence of musculoskeletal pain. J Vet Behav. 2018; 23: 47–57. doi: 10.1016/J.JVEB.2017.10.008 [DOI] [Google Scholar]
31.Wathan J, McComb K. The eyes and ears are visual indicators of attention in domestic horses. Curr Biol. 2014; 24(15): R677–R679. doi: 10.1016/j.cub.2014.06.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Roberts K, Hemmings AJ, Moore-Colyer M, Parker MO, McBride SD. Neural modulators of temperament: A multivariate approach to personality trait identification in the horse. Physiol Behav. 2016; 167: 125–131. doi: 10.1016/j.physbeh.2016.08.029 [DOI] [PubMed] [Google Scholar]
33.Merkies K, Ready C, Farkas L, Hodder A. Eye blink rates and eyelid twitches as a non-invasive measure of stress in the domestic horse. Animals. 2019; 9(8): 1–10. doi: 10.3390/ani9080562 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Des Roches ADB, Richard-Yris MA, Henry S, Ezzaouïa M, Hausberger M. Laterality and emotions: Visual laterality in the domestic horse (Equus caballus) differs with objects’ emotional value. Physiol Behav. 2008; 94: 487–490. doi: 10.1016/j.physbeh.2008.03.002 [DOI] [PubMed] [Google Scholar]
35.Leliveld LMC, Langbein J, Puppe B. The emergence of emotional lateralization: Evidence in non-human vertebrates and implications for farm animals. Appl Anim Behav Sci. 2013; 145: 1–14. doi: 10.1016/J.APPLANIM.2013.02.002 [DOI] [Google Scholar]
36.Hockenhull J, Creighton E. Training horses: Positive reinforcement, positive punishment, and ridden behavior problems. J Vet Behav. 2013; 8(4): 245–252. doi: 10.1016/j.jveb.2012.06.002 [DOI] [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0286045.r001

Decision Letter 0

Chris Rogers

16 Feb 2023

PONE-D-22-34365Horse behavior and facial movements in relation to food rewardsPLOS ONE

Dear Dr. Daros,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Apr 02 2023 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Chris Rogers

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

Additional Editor Comments:

Thank you for the submission. The reviewers have asked for some edits to improve the clarity on the experimental design. Please also see the comments regarding the use of language to describe outcomes.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This is a straightforward piece of working describing the behavioural and facial responses of horses being provided positive reinforcement through the provision of food rewards to a routine task. While few results were obtained, it nevertheless provides information in an area where little research has been done to date. The biggest comment I have to improve the paper would be to analyze the carryover effect as described below.

Introduction: well-written and succinct

Methods:

- L106 – was a lead rope used to lead the horses to the chute?

- L117 - Were the horses familiar with the testing environment prior to the experiment? Were they accustomed to entering the chute? If so, for what purpose – eg. veterinary treatment? The question is, what previous experience did they have with the chute and how might this affect their perception. This is alluded to on L344 but should be clearly defined in the methods.

- L135 – why did you offer apple and molasses treats when not all the horses preferred them (L112)?

- L146 – authors acknowledge the differences between the control and the treatment in regards to the holding time in the chute. Although perhaps not directly comparable, it seems a wise choice. How long were the video clips for the control while in the chute?

- L158 – authors comment the behaviours were collected continuously. Please indicate which behaviours were frequency counts, which were duration, and how long the duration was. Were all videos standardized to a certain length of time? This is better described for the FAUs and in table 2.

- L159 – the observer was not blind to the treatments – this should be addressed in the discussion

- L192 – did you standardize the length of time observed for each horse? As noted above, it is not clear how long the video clips were (I think they were all one minute, but this should be clearly stated). If the video clips were not standardized, this would greatly affect the results.

- L204 – did you account for any carryover effect for those horses who experienced the positive reinforcement phase the first week and then the control phase the second week? This may affect your comparisons between treatment and control. Also why did you use only the third day for comparisons of behaviours but used the second and third days for comparisons of the FAUs?

Results:

- L211-215 – does this first sentence not say the same thing as the second sentence?

- L231 – perhaps specify in the table caption that the behaviours listed in table 3 are ones that occurred infrequently

- L236 – the phrase “All but one showed differences between the phases.” is confusing – do you mean only one behaviour differed between phases? Because you only give results for lowered neck. Does this mean that none of the other behaviours differed between phases? Please clarify.

Discussion:

- L322 – perhaps amend to saying “slightly negative”

- L327 – can you speculate on why the control horses may have closed their eyes more? Was this indicative of resting behaviour (low arousal, not much going on)?

Reviewer #2: General Comments.

The purpose of this study is laudable and it is necessary that research focus on positive affects in animals, as also stated by the authors. This is however a difficult task, because the ground truth of affective states of animals is hard to achieve and that many different cognitive processes may influence the affective states.

This study aims at investigating the effect on behaviours of food rewards in horses (n=13). Horses were brought from the pasture to a test area every day for three weeks. From this area, the horses were brought individually to a chute, in a cross over design, where a food reward was given or not. Video was recorded in the test area and while in the chute after 3 days of intervention.

I am however not sure that I have understood the design correctly, new additions come here and there. Please clarify, maybe add a figure of the study design, and make sure that the abstract contains all necessary information about the study design, e.g. time in chute, the food rewarding schedule etc. Please also clarify the figure texts accordingly.

Please explain which behaviours you believe you reinforce by administering food in the chute? I am not an ethologist, but are you reinforcing something or just teaching the horses to get food? What is meant by “toggled”? Would the horses, that were habituated to get food, when in the chute, not get “frustrated” when they got no food? Can you then use the horses as their own controls, if one control group is neutral and the other is “frustrated” – negative valence affect?

Besides from the difficulty in understanding the study design, and therefore also how the data were produced, the use of “positive emotions” seems mixed up with concept of positive welfare. Explain why food should produce positive affects? Please refer to your reference 6 for terminology and revise the entire text according to published concepts.

6. Yeates JW, Main DC. Assessment of positive welfare: a review. Vet J. 2008; 175(3):

389 293–300. doi: 10.1016/j.tvjl.2007.05.009.

Please state hypotheses clearly. I can see that they pop up here and there, but please collect them, it greatly helps the understanding of your statistics.

Specific comments.

Line 15: Positive reinforcement can elicit positive emotions in animals. Please remove this statement from the abstract. At the best this statement remains controversial and should not be in an abstract. You could instead state that food rewards are believed to have a positive valence in horses.

Line 58 and onwards: Not true, please include e.g. the works by Lea Lansade et al., for example Facial expression and oxytocin as possible markers of positive emotions in horses. DOI 10.1038/s41598-018-32993-z

Line 120 and onwards: … The horse was left in the chute without physical contact for 1 min. During this period, both cameras (Sony and Canon) were on, and the researcher observed the horses for selected behaviours live. Was it possible to observe the entire horse while in the chute?

Line 156: During the restraining period, several behaviours were recorded live and by video. I am sorry if I didn’t get the design correctly, but was this not the case also for the baseline horses? And the single observer was not blinded? Why?

Line 162: Video clips from the last two days of the positive reinforcement and control phases (totalling 52 video clips) were analysed to extract data regarding the facial action units (FAUs) of each horse. How long were the clips for FACS annotation and how did you select them? I guess that not all of the video provided footage from a good angle, as the horse moved its head? Please therefore state which visibility codes you used or explain how you handled data when AUs were not possible to score. Again, regarding the design: you did only record for facial activity in the chute, not in the test area?

Line 165: The start and end of each assessed FAU was manually annotated by a researcher watching on a… Was it the same FACS coder for all videos? Or how many different coders were engaged? Which training did coders have – certification as FACS reader?

Line 170: Therefore, the frequency of each FAU was assessed but several FAU were also selected

(i.e., the ones that are expected to be activated for longer periods) to be measured as state events.

In FACS, there are no point events. Start and stop is recorded for each AU, indeed the observation of both the onset and offset if necessary to define a specific AU. For example: the eye blink, closure and half blink cannot be properly discriminated without determination of the duration. Further, your table contains a number of Action Descriptors, which are not considered AUs, they are composed of less defined muscle movements and may have another neurophysiological background than the 17 AUs. I think it may be valuable to have also the ADs in your study, but please describe what is what. Please revise table, text and statistics accordingly.

Line184: The ICC was 0.73 (95% CI: 185 0.63–0.80); ranging from good to excellent agreement. Please use the Wexler’s agreement, which is recommended for FACS analysis.

Line 302: see earlier comment for line 58.

Line 261: Table 4: Estimated means for the positively reinforced (PR) and control phases from mixed linear regression models of each facial action unit (FAU) evaluated in 13 mares. From the perspective of facial expressions, it is highly unlikely that one single action unit should be associated to a certain affective state. This is at least not the case for human emotions, or the affective components of pain or stress.

As mentioned earlier, the behaviour is the bottom line of all cognitive and other brain processes, learning, memory. It is unlikely that the 13 mares, being from 4 of age up to 22, would be cognitively equal. Therefore, I think it could be rewarding to look at individual horses and their responses to food. While some might be unaffected you might see characteristic changes in other, and this might be related to rank, age, horsonality etc. You have only 13 horses so I think this would be doable. You already did a huge job to collect the data

Line 304: Facial movements. You have chosen to discuss individual AU/ADs. This may not give real meaning as facial expressions, since you have no co-occurrence calculated. There is no single AU/AD where the frequency or duration can be associated as a marker for a certain affective state. At least this is not the case for humans. No facial expression consist of only one AU, which gives meaning if we believe that facial expressions are used for communication.

Line 327: ….horses had higher frequencies of closing the eyelid for more than half a second. I am confused: did you state earlier that you recorded eye blink as point events? Did you still measure duration? Please clarify further in the Material and Methods section.

I think the conclusion is sound, but wonder what you had expected – what would you have interpreted as indicative of positive valence, regarding facial expressions?

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Pia Haubro Andersen

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2023 Jun 12;18(6):e0286045. doi: 10.1371/journal.pone.0286045.r002

Author response to Decision Letter 0

5 Apr 2023

Dear editor and reviewers. Thank you for your time assessing our manuscript. The detailed responses to all comments can be found below or in the rebuttal letter attached in the submission system.

Authors' response to the comments:

AU: Thank you for taking the time to review and to suggest improvements in our manuscript.

Introduction: well-written and succinct

AU: Thanks!

Methods:

- L106 – was a lead rope used to lead the horses to the chute?

AU: L106-107 “… they were taken individually by a lead rope to the chute”.

AU: We now state the activities they were used to in the facility. L119-122 “All horses were familiar with the testing environment prior to the experiment and were accustomed to entering the chute for veterinary treatment and daily reproductive management, but the individual experiences of each horse and their perceptions based on personality are unknown by the authors.”. We left the information about their potential perception to the discussion as we were not able to assess it before the study.

- L135 – why did you offer apple and molasses treats when not all the horses preferred them (L112)?

AU: Because some animals may have their preferred food, some like them all and some may like it varied we opted for offering all food items to them. We added more to this section, see L139-141 “All food items were hand-fed one at a time starting with the carrot, apple, and then the molasses treat. If the horse refused any item, it was put aside and the next piece was offered. The rationale was to offer the horses a chance to eat or pass as they wish”.

AU: Fixed. L150-151 “In the control group, no waiting period was applied, so the 60-sec video clip started immediately after the horse was closed in the chute”.

AU: Fixed. L167-171 “Behaviors described were collected continuously for 60-sec and annotated… …Pawing, kicking, rear and defecation/urination were recorded according to the frequency they occurred, while standing, resting and neck position and tail movement were recorded according to the duration”.

- L159 – the observer was not blind to the treatments – this should be addressed in the discussion

AU: Added to the limitations of our study. L386 “Blinded observers should be considered to reduce potential bias”.

AU: Information added. L154-155 “The observation time was standardized for each horse - i.e., a synchronized 60-sec video clip in both cameras (Sony and Canon).”.

AU: Thank you for this comment. We found no carryover effects as highlighted on L223-225 “The order of treatments (control/reinforced) in the experimental phase was included in all models at first, but it did not affect the results (P > 0.1)”. Regarding the use of the third day we now added the rationale for it in L219-221 Only the third day of the control and positively reinforced phases was used for comparison between data collected in the holding area, as that is the day we were most certain that the mare was habituated the experimental phase she was in.”

In previously analysis, we also used the second last day. Same results, so to keep it simple we decided to use only one day. We did not apply this same rationale for the EquiFACS analysis. Facial movements are much more subjective than the other behaviours measured, taken into consideration that the mean of samples is closer to the truth than the value of a single sample we have taken this rationale for this analysis, though we understand that this is more complex and may cause some confusion at first. Furthermore, the EquiFACS analysis we were looking at the immediate effect of having a food reward, so habituation to the experimental phase should influence less the results.

Results:

- L211-215 – does this first sentence not say the same thing as the second sentence?

AU: Sorry for the misunderstanding. First sentence refers to the comparison between the reinforced and the baseline phase. L235-239 now reads as follows: “It was expected that during the positively reinforced days, horses would arrive at the management facility before those that were under the control phase. However, this hypothesis was not confirmed because horses did not significantly change their arrival order (t = 0.44; P = 0.66) from the baseline to the positively reinforced days. When comparing the positive reinforcement with the control phase, no differences were detected in the order of arrival at the holding facility (P = 0.80)”.

- L231 – perhaps specify in the table caption that the behaviours listed in table 3 are ones that occurred infrequently

AU: Done. L256-257 “Table 3. Total events of infrequently occurring body behaviors of all 13 mares observed during all sessions of 60-sec interval in a restraining chute”.

AU: Sorry for the lack of clarity. L261 now states “Only one behavior showed difference between the phases”.

Discussion:

- L322 – perhaps amend to saying “slightly negative”

AU: Done. L352.

- L327 – can you speculate on why the control horses may have closed their eyes more? Was this indicative of resting behaviour (low arousal, not much going on)?

AU: Agreed. Speculation added L360-363 “We found that during the control phase, horses had higher frequencies of eye closure (AU143) compared to the reinforced phase (a movement different from blinking or half blinking). It is possible that the control horses closed their eyes more as a result of being in a more relaxed state or engaging in resting behavior”.

Reviewer #2: General Comments.

AU: Thank you for your time in reviewing our manuscript. We tried our best in improving the clarity of our methods. Please find below our detailed responses.

AU: Details to the abstract have been added and a new figure added to the main manuscript. See L15, 19, 21, 22. and Figure 1.

Please explain which behaviours you believe you reinforce by administering food in the chute? I am not an ethologist, but are you reinforcing something or just teaching the horses to get food?

AU: Added. L141-145 “The food items were offered immediately after the front gate of the restraining chute was closed and the horse was not pushing against the gates; thus, reinforcing the short sequence of behaviors that happens after the chute has been closed and the food offering. We highlight that the person closing the chute was not the same as the one offering the food item. This procedure and timing were applied consistently throughout the study.”.

What is meant by “toggled”?

AU: Toggled was a mistake and this section has been rephrased.

Would the horses, that were habituated to get food, when in the chute, not get “frustrated” when they got no food? Can you then use the horses as their own controls, if one control group is neutral and the other is “frustrated” – negative valence affect?

AU: Thank you for this comment. When planning the experiment, we had to decide between having this issue of potential frustration of some animals or run an experiment that would be potentially confounded by time (i.e. all mares in the reinforced phase in the same week and control on another). We opted for the former as, at least, we could try to control for the effects of different emotions during the control phase using analytics method (i.e., including the order effect in our models). We highlight this now on L264: “We found no evidence of order effect of groups in any of our models”.

6. Yeates JW, Main DC. Assessment of positive welfare: a review. Vet J. 2008; 175(3):

389 293–300. doi: 10.1016/j.tvjl.2007.05.009.

AU: Thank you for the suggested literature. The manuscript was revised and the use of "positive emotions" was replaced by "positive affects" and statements that food produces positive affects have been removed.

Please state hypotheses clearly. I can see that they pop up here and there, but please collect them, it greatly helps the understanding of your statistics.

AU: Added. L56-62 “Horses are expected to change their behavior before entering the chute (e.g., spending more time close to the gate that leads to the horse chute) while body postures in the chute are expected to be associated with behaviors of expectancy (less resting, neck above withers) during the positively reinforced phase compared with the non-reinforced phase. Only few studies have assessed facial movements of horses in positive emotional contexts. Therefore, we only predict that horses will show differences in facial movements between the positive and non-reinforced phases.”.

AU: Done. L15: “Food rewards are believed to have a positive valence in horses”

AU: Reference has been added L38-41: “Animal behavioral responses studies can provide insights into how an animal feels under a given circumstance [4] and have been successfully applied to study positive affective states in animals [5, 6] including in horses [7].”.

AU: Detail added. L168 “… was always present being possible to observe the entire horse while in the chute”.

AU: Sorry for the lack of clarity. In this paragraph we are specifically talking about the behaviour in the holding area, which is the pen were horse are used to wait before getting into the area were the restraining chute is (details in Figure 2). We now make this clear on L161-162 “Holding area behavioral data were assessed through video analysis by a single observer”.

The only fully trained observer for this analysis was also the researcher responsible for conducting the Project, thus it was impossible to blind. We address this limitation in our discussion. L385-386: “Blinded observers should be considered to reduce potential bias and the FAUs should be analyzed in co-occurrence”.

AU: L179-183 “Each video was watched multiple times as only one FAU was observed and was rewound or slowed down when in doubt, being scored only when visible. Our objective was to evaluate the occurrence of all AUs during the 60-sec, however, in the moments when it was not possible to score AUs, no data was generated. We emphasize that most of the videos provided good visibility and good angle as the horses remained still”.

Yes. L184-185: “Facial recordings were only possible in the chute as the camera recording the holding area behavior was set to capture images from the whole herd”.

AU: All videos were watched by the same self-trained (using the EquiFACS material) observer. To assess if this observer was reliable we measured interobserver reliability against another independently self-trained observer. Specified in L178-179: “The start and end of each assessed FAU was manually annotated by the same FACS coder watching on a computer”.

Line 170: Therefore, the frequency of each FAU was assessed but several FAU were also selected

(i.e., the ones that are expected to be activated for longer periods) to be measured as state events.

AU: Sorry. We understand that in FACS there are no point events and that start and stop is recorded for each AU. Declaring AUs as one-off events has been removed. However, we also evaluated the activation frequency of each AU. We reviewed this section as described in L186-191: “The categorization of each FAU was done following the EquiFACS methodology [12], though we adapted the measurement of FAUs that are short lasting to measure frequency only (eye closure, blink, half blink, lower lip depressor, chin raiser and lip presser). FAUs that lasted longer were assessed as duration and frequency. The time required to correctly classify each FAU was always accounted for. In Table 2, we inform which FAUs were observed, specifying use the codes AU, AD, AUH and EAD, as described in Wathan et al., 2015. The duration and frequency of each FAU were analyzed separately”.

Line184: The ICC was 0.73 (95% CI: 185 0.63–0.80); ranging from good to excellent agreement. Please use the Wexler’s agreement, which is recommended for FACS analysis.AU: We considered using Wexler’s however ICC is a more robust statistical measure as it takes into account not only the ratio of agreement between observers but also the variance in the data, and account for likelihood of agreeing by chance. ICC is a well established method for interobserver reliability see Hallgren, 2012 (Tutor Quant Methods Psychol. 2012 ; 8(1): 23–34.).

Line 302: see earlier comment for line 58.

AU: Context added to the discussion. L330-336 “To date facial expressions in horses have mostly been described in the context of painful or stressful experiences. However, certain features of the face have already been identified in a positive context [7, 27], although these studies have not did not use EquiFACS methodology. Lansade et al. (2018) found that analyses of facial movements during grooming revealed significant differences between groups. However, a negative stimulus (spraying a mild perfume close to the horse) induced a more marked change in facial behaviors compared to positive (grooming) and neutral (observer close to the horse, without physical contact) contexts [27]”.

AU: Agreed. "Facial Expression" has been completely removed from the manuscript. We highlighted this in our discussion. See L336-341: “We emphasize that our study did not aim to evaluate the facial expression of horses, however, we presented the activation response (duration and frequency) of each FAU separately. Isolated facial movements can also be promising to identify a prominent action (either in duration or frequency) in situations of positive affective valence. We understand that only together could they reflect a “facial expression”, therefore, in future studies, the FAUs should be analyzed in co-occurrence”.

AU: We really appreciate this comment. In fact, we just finished a trial conducted by LGC (first author) where we have more tests to assess personality and their association with other behavioural processes (e.g., attention and memory bias). In earlier versions of our analysis we included age in our models and none were significant (We now mention this in L224-225). Unfortunately, we still do not have rank and personality data ready to re-run our analysis. For the facial movements we did not test this covariate as the data collection was planned to be conducted within individuals.

AU: Agreed. That’s the reason we refrain to mention facial expressions. See earlier comment for line L336-341: “We emphasize that our study did not aim to evaluate the facial expression of horses, however, we presented the activation response (duration and frequency) of each FAU separately. Isolated facial movements can also be promising to identify a prominent action (either in duration or frequency) in situations of positive affective valence. We understand that only together could they reflect a “facial expression”, therefore, in future studies, the FAUs should be analyzed in co-occurrence”.

AU: Sorry for the lack of clarity, we do not measure the duration. The rationale for this is in L186-189 and reads as follows: “The categorization of each FAU was done following the EquiFACS methodology [12], though we adapted the measurement of FAUs that are short lasting to measure frequency only (eye closure, blink, half blink, lower lip depressor, chin raiser and lip presser). FAUs that lasted longer were assessed as duration and frequency. The time required to correctly classify each FAU was always accounted for.”. We also rewrote part of the discussion that now reads (L359-361) now reads as follows: “We found that during the control phase, horses had higher frequencies of eye closure (AU143) compared to the reinforced phase (a movement different from blinking or half blinking)”.

I think the conclusion is sound, but wonder what you had expected – what would you have interpreted as indicative of positive valence, regarding facial expressions?

AU: Unfortunately, we didn't have a clear hypothesis regarding facial movements. We address this in L60-62 “Therefore, we only predict that horses will show differences in facial movements between the positive and non-reinforced phases”. Although we found no clearly evident facial movements with food rewards, we can say that positive reinforcement affected the facial movements of group-housed mares.

Attachment

Submitted filename: Response to Reviewers_RRD_RRD_RRD.docx

Click here for additional data file.^{(88.5KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0286045.r003

Decision Letter 1

Chris Rogers

9 May 2023

Horse behavior and facial movements in relation to food rewards

PONE-D-22-34365R1

Dear Dr. Daros,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Chris Rogers

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Thank you for the revised manuscript and attention to the reviewers comments.

Reviewers' comments:

PLoS One. doi: 10.1371/journal.pone.0286045.r004

Acceptance letter

Chris Rogers

2 Jun 2023

PONE-D-22-34365R1

Horse behavior and facial movements in relation to food rewards

Dear Dr. Daros:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Chris Rogers

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Attachment

Submitted filename: Response to Reviewers_RRD_RRD_RRD.docx

Click here for additional data file.^{(88.5KB, docx)}

Data Availability Statement

The original data, R scripts, and output can be found at URL: (https://doi.org/10.6084/m9.figshare.21714233.v1) DOI: (10.6084/m9.figshare.21714233.v1).

[pone.0286045.ref001] 1.Mellor DJ. Updating animal welfare thinking: Moving beyond the “Five Freedoms” towards “a Life Worth Living”. Animals. 2016; 6(3): 21. doi: 10.3390/ani6030021 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0286045.ref002] 2.Boissy A, Manteuffel G, Jensen MB, Moe RO, Spruijt B, Keeling LJ, et al. Assessment of positive emotions in animals to improve their welfare. Physiol Behav. 2007; 92(3): 375–397. doi: 10.1016/j.physbeh.2007.02.003 [DOI] [PubMed] [Google Scholar]

[pone.0286045.ref003] 3.Rault JL, Hintze S, Camerlink I, Yee JR. Positive welfare and the like: Distinct views and a proposed framework. Front Vet Sci. 2020; 7: 370. doi: 10.3389/fvets.2020.00370 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0286045.ref004] 4.Weary DM, Droege P, Braithwaite VA. Behavioral Evidence of Felt Emotions: Approaches, Inferences, and Refinements. Adv Study Behav. 2017; 49: 27–48. doi: 10.1016/bs.asb.2017.02.002 [DOI] [Google Scholar]

[pone.0286045.ref005] 5.van der Harst JE, Spruijt BM. Tools to measure and improve animal welfare: Reward-related behaviour. Anim Welf. 2007; 16(Suppl.): 67–73. [Google Scholar]

[pone.0286045.ref006] 6.Yeates JW, Main DC. Assessment of positive welfare: a review. Vet J. 2008; 175(3): 293–300. doi: 10.1016/j.tvjl.2007.05.009 [DOI] [PubMed] [Google Scholar]

[pone.0286045.ref007] 7.Lansade L, Nowak R, Lainé AL, Leterrier C, Bonneau C, Parias C, et al. Facial expression and oxytocin as possible markers of positive emotions in horses. Scientific reports. 2018; 8(1), 14680. doi: 10.1038/s41598-018-32993-z [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0286045.ref008] 8.Laule GE, Bloomsmith MA, Schapiro SJ. The use of positive reinforcement training techniques to enhance the care, management, and welfare of primates in the laboratory. J Appl Anim Welf Sci. 2003; 6(3): 163–173. doi: 10.1207/S15327604JAWS0603_02 [DOI] [PubMed] [Google Scholar]

[pone.0286045.ref009] 9.Lambeth SP, Hau J, Perlman JE, Martino M, Schapiro SJ. Positive reinforcement training affects hematologic and serum chemistry values in captive chimpanzees (Pan troglodytes). Am J Primatol. 2006; 68(3): 245–256. doi: 10.1002/ajp.20148 [DOI] [PubMed] [Google Scholar]

[pone.0286045.ref010] 10.Hendriksen P, Elmgreen K, Ladewig J. Trailer-loading of horses: Is there a difference between positive and negative reinforcement concerning effectiveness and stress-related signs?. J Vet Behav. 2011; 6(5): 261–266. doi: 10.1016/j.jveb.2011.02.007 [DOI] [Google Scholar]

[pone.0286045.ref011] 11.Freymond SB, Briefer EF, Zollinger A, Gindrat-von Allmen Y, Wyss C, Bachmann I. Behaviour of horses in a judgment bias test associated with positive or negative reinforcement. Appl Anim Behav Sci. 2014; 158: 34–45. doi: 10.1016/j.applanim.2014.06.006 [DOI] [Google Scholar]

[pone.0286045.ref012] 12.Dalla Costa E, Minero M, Lebelt D, Stucke D, Canali E, Leach MC. Development of the Horse Grimace Scale (HGS) as a pain assessment tool in horses undergoing routine castration. PLoS One. 2014; 9(3): 1–10. doi: 10.1371/journal.pone.0092281 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0286045.ref013] 13.Wathan J, Burrows AM, Waller BM, McComb K. EquiFACS: The equine facial action coding system. PLoS One. 2015; 10(8): 1–35. doi: 10.1371/journal.pone.0131738 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0286045.ref014] 14.Trindade PHE, Hartmann E, Keeling LJ, Andersen PH, Ferraz GDC, Paranhos da Costa MJR. Effect of work on body language of ranch horses in Brazil. PLoS One. 2020; 15(1): e0228130. doi: 10.1371/journal.pone.0228130 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0286045.ref015] 15.Lundblad J, Rashid M, Rhodin M, Andersen PH. Effect of transportation and social isolation on facial expressions of healthy horses. PLoS One. 2021; 16(6 June): 1–17. doi: 10.1371/journal.pone.0241532 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0286045.ref016] 16.Young T, Creighton E, Smith T, Hosie C. A novel scale of behavioural indicators of stress for use with domestic horses. Appl Anim Behav Sci. 2012; 140(1–2): 33–43. doi: 10.1016/j.applanim.2012.05.008 [DOI] [Google Scholar]

[pone.0286045.ref017] 17.Hallgren KA. Computing inter-rater reliability for observational data: an overview and tutorial. Tutor Quant Methods Psychol. 2012; 8(1): 23–34. doi: 10.20982/tqmp.08.1.p023 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0286045.ref018] 18.Wickham H, Averick M, Bryan J, Chang W, McGowan LD, François R, et al. Welcome to the tidyverse. J Open Source Softw. 2019; 4: 1686. doi: 10.21105/joss.01686 [DOI] [Google Scholar]

[pone.0286045.ref019] 19.Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Softw. 2015; 67. doi: 10.18637/jss.v067.i01 [DOI] [Google Scholar]

[pone.0286045.ref020] 20.Peters SM, Bleijenberg EH, van Dierendonck MC, van der Harst JE, Spruijt BM. Characterization of anticipatory behaviour in domesticated horses (Equus caballus). Appl Anim Behav Sci. 2012; 138(1–2): 60–69. doi: 10.1016/j.applanim.2012.01.018 [DOI] [Google Scholar]

[pone.0286045.ref021] 21.Mal ME, McCall CA, Newland C, Cummins KA. Evaluation of a one-trial learning apparatus to test learning ability in weanling horses. Appl Anim Behav Sci. 1993; 35(4): 305–311. [Google Scholar]

[pone.0286045.ref022] 22.Christensen JW, Rundgren M, Olsson K. Training methods for horses: habituation to a frightening stimulus. Equine Vet J. 2006; 38(5): 439–443. doi: 10.2746/042516406778400574 [DOI] [PubMed] [Google Scholar]

[pone.0286045.ref023] 23.Hall C, Goodwin D, Heleski C, Randle H, Waran N. Is there evidence of learned helplessness in horses? J Appl Anim Welf Sci. 2008; 11(3): 249–266. doi: 10.1080/10888700802101130 [DOI] [PubMed] [Google Scholar]

[pone.0286045.ref024] 24.Fureix C, Hausberger M, Seneque E, Morisset S, Baylac M, Cornette R, et al. Geometric morphometrics as a tool for improving the comparative study of behavioural postures. Naturwissenschaften. 2011; 98(7): 583–592. doi: 10.1007/s00114-011-0803-2 [DOI] [PubMed] [Google Scholar]

[pone.0286045.ref025] 25.Smiet E, Van Dierendonck MC, Sleutjens J, Menheere PPCA, van Breda E, de Boer D, et al. Effect of different head and neck positions on behaviour, heart rate variability and cortisol levels in lunged Royal Dutch Sport horses. Vet J. 2014; 202(1): 26–32. doi: 10.1016/j.tvjl.2014.07.005 [DOI] [PubMed] [Google Scholar]

[pone.0286045.ref026] 26.Hall C, Randle H, Pearson G, Preshaw L, Waran N. Assessing equine emotional state. Appl Anim Behav Sci. 2018; 205: 183–193. doi: 10.1016/j.applanim.2018.03.006 [DOI] [Google Scholar]

[pone.0286045.ref027] 27.Stratton R, Cogger N, Beausoleil N, Waran N, Stafford K, Steward M. Indicators of good welfare in horses: Final report. MPI Technical Paper, (2014/44). Ministry for Primary Industries, New Zealand Government, 52 p.; 2014. https://www.mpi.govt.nz/dmsdocument/5089. [Google Scholar]

[pone.0286045.ref028] 28.Reefmann N, Kaszàs FB, Wechsler B, Gygax L. Ear and tail postures as indicators of emotional valence in sheep. Appl Anim Behav Sci. 2009; 118(3–4): 199–207. doi: 10.1016/j.applanim.2009.02.013 [DOI] [Google Scholar]

[pone.0286045.ref029] 29.Sankey C, Richard-Yris MA, Henry S, Fureix C, Nassur F, et al. Reinforcement as a mediator of the perception of humans by horses (Equus caballus). Anim Cogn. 2010; 13(5): 753–764. doi: 10.1007/s10071-010-0326-9 [DOI] [PubMed] [Google Scholar]

[pone.0286045.ref030] 30.Dyson S, Berger J, Ellis AD, Mullard J. Development of an ethogram for a pain scoring system in ridden horses and its application to determine the presence of musculoskeletal pain. J Vet Behav. 2018; 23: 47–57. doi: 10.1016/J.JVEB.2017.10.008 [DOI] [Google Scholar]

[pone.0286045.ref031] 31.Wathan J, McComb K. The eyes and ears are visual indicators of attention in domestic horses. Curr Biol. 2014; 24(15): R677–R679. doi: 10.1016/j.cub.2014.06.023 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0286045.ref032] 32.Roberts K, Hemmings AJ, Moore-Colyer M, Parker MO, McBride SD. Neural modulators of temperament: A multivariate approach to personality trait identification in the horse. Physiol Behav. 2016; 167: 125–131. doi: 10.1016/j.physbeh.2016.08.029 [DOI] [PubMed] [Google Scholar]

[pone.0286045.ref033] 33.Merkies K, Ready C, Farkas L, Hodder A. Eye blink rates and eyelid twitches as a non-invasive measure of stress in the domestic horse. Animals. 2019; 9(8): 1–10. doi: 10.3390/ani9080562 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0286045.ref034] 34.Des Roches ADB, Richard-Yris MA, Henry S, Ezzaouïa M, Hausberger M. Laterality and emotions: Visual laterality in the domestic horse (Equus caballus) differs with objects’ emotional value. Physiol Behav. 2008; 94: 487–490. doi: 10.1016/j.physbeh.2008.03.002 [DOI] [PubMed] [Google Scholar]

[pone.0286045.ref035] 35.Leliveld LMC, Langbein J, Puppe B. The emergence of emotional lateralization: Evidence in non-human vertebrates and implications for farm animals. Appl Anim Behav Sci. 2013; 145: 1–14. doi: 10.1016/J.APPLANIM.2013.02.002 [DOI] [Google Scholar]

[pone.0286045.ref036] 36.Hockenhull J, Creighton E. Training horses: Positive reinforcement, positive punishment, and ridden behavior problems. J Vet Behav. 2013; 8(4): 245–252. doi: 10.1016/j.jveb.2012.06.002 [DOI] [Google Scholar]

PERMALINK

Horse behavior and facial movements in relation to food rewards

Laize G Carmo

Laís C Werner

Pedro V Michelotto Jr

Ruan R Daros

Roles

Abstract

1. Introduction

2. Materials and Methods

2.1 Animal handling

Fig 1. Diagram of the study design.

Fig 2. Horse management facility.

2.2 Data collection setup

Fig 3. Video camera and observer positioning for capturing body and facial behaviors of horses while in the restraining chute.

2.3 Habituation phase

2.4 Experimental phase

2.5 Behavioral analysis

Table 1. Behaviors assessed during the restraining period in a horse chute.

Table 2. Facial action units (FAUs) assessed in the study using EquiFACS: The equine facial action coding system.

2.6 Statistical analysis

3. Results

3.1 Holding area behavior

3.2 Behavior in the chute

Table 3. Total events of infrequently occurring body behaviors of all 13 mares observed during all sessions of 60-sec interval in a restraining chute.

3.3 Facial action unit—EquiFACS

Table 4. Estimated means for the positively reinforced (PR) and control phases from mixed linear regression models of each facial action unit (FAU) evaluated in 13 mares.

4. Discussion

4.1 Behavior in the holding area

4.2 Behavior in the chute

4.3 Facial action units

5. Conclusions

Acknowledgments

Data Availability

Funding Statement

References

Decision Letter 0

Chris Rogers

Roles

Author response to Decision Letter 0

Decision Letter 1

Chris Rogers

Roles

Acceptance letter

Chris Rogers

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases