Gait analysis in swine, sheep, and goats after neurologic injury: a literature review

Jacob W. Sveum, Raveena R. Mishra, Taylor L. Marti, Jalon M. Jones, Daniel J. Hellenbrand, Amgad S. Hanna

Abstract Medical research on neurologic ailments requires representative animal models to validate treatments before they are translated to human clinical trials. Rodents are the predominant animal model used in neurological research despite limited anatomic and physiologic similarities to humans. As a result, functional testing designed to assess locomotor recovery after neurologic impairment is well established in rodent models. Comparatively, larger, more clinically relevant models have not been as well studied. To achieve similar locomotor testing standardization in larger animals, the models must be accessible to a wide array of researchers. Non-human primates are the most relevant animal model for translational research, however ethical and financial barriers limit their accessibility. This review focuses on swine, sheep, and goats as large animal alternatives for transitional studies between rodents and non-human primates. The objective of this review is to compare motor testing and data collection methods used in swine, sheep, and goats to encourage testing standardization in these larger animal models. The PubMed database was analyzed by searching combinations of swine, sheep, and goats, neurologic injuries, and functional assessments. Findings were categorized by animal model, data collection method, and assessment design. Swine and sheep were used in the majority of the studies, while only two studies were found using goats. The functional assessments included open pen analysis, treadmill walking, and guided free walking. Data collection methods included subjective behavioral rating scales and objective tools such as pressure-sensitive mats and image-based analysis software. Overall, swine and sheep were well-suited for a variety of assessment designs, with treadmill walking and guided free walking offering the most consistency across multiple trials. Data collection methods varied, but image-based gait analysis software provided the most robust analysis. Future studies should be conducted to standardize functional testing methods after neurologic impairment in large animals.

Key Words: functional testing; gait analysis; goats; large animals; neurologic injury; sheep; spinal cord injury; swine

Introduction

The majority of medical research is performed on smaller animals such as rats and mice. Approximately 111.5 million mice and rats are used in the United States of America for medical research each year, and they make up around 99.3% of animal subjects used (Carbone, 2021). Rats and mice are affordable, require little space to grow, and can rapidly produce offspring. Their gestation period is approximately 19–21 days and they have short life spans, making them easy and efficient experimental models (Bryda, 2013). Despite these benefits, the study results are not always applicable to humans due to limited similarities in anatomy and physiology. Translational studies between rats and humans are particularly difficult when studying nervous system ailments due to significant differences in spinal cord size, subarachnoid space, and cerebrospinal fluid flow rate (Kim et al., 2018). Numerous functional testing methods have been designed for smaller animal models and are primarily grouped into pen analysis and forced walking.

In pen analysis, the test subject walks around normally in an open field environment and is assessed using the Basso, Beattie, and Bresnahan (BBB) Scale and the Basso Mouse Scale (BMS) for rat and mouse models, respectively. The BBB scale is used to evaluate the functional recovery of rats following spinal cord injury (SCI), in which the rat receives a numerical score to describe its range of motion and limb usage as it walks around an open field (Basso et al., 1995). Similarly, the BMS scale is used to measure the gait of a mouse (Pomeshchik et al., 2015).

Alternatively, forced walking methods are also used to analyze gait in small animals. One method of forced walking, automated CatWalk gait analysis, has been used in several studies involving rat models (Bozkurt et al., 2008; Timotius et al., 2019). Timotius et al. (2019) described a system in which a plexiglass sheet serves as a walkway and a camera is placed below the plexiglass. In the ladder rung walking technique, the animals are trained to walk on the metal rungs before data collection, and video recordings are used to determine the animals’ motor skills and limb coordination (Riek-Burchardt et al., 2004; Metz and Whishaw, 2009). Lastly, the rotarod test focuses on motor coordination before and after injury. The rotarod is a rotating cylinder which the animal must run on to avoid falling (Farr et al., 2006; Navarro et al., 2012; Imani et al., 2016; Sato et al., 2018). For a more thorough assessment, combinations of tests can be used to analyze motor function.

Although the BBB/BMS scales, CatWalk treadmill, ladder rung test, and rotarod test are effective ways to evaluate motor function in small animals, it can be difficult or even unrealistic to implement these methods in larger animals (Russell et al., 2011). For better translation to humans, larger animals with similar anatomy need to be prioritized; thus, using functional testing methods specific to large animal models is imperative.

This literature review aims to identify and assess existing locomotor testing methods used in large animals with the goal of encouraging testing standardization. In selecting non-rodent animal models for review, anatomical, financial, and ethical factors were considered. Non-human primates are the most representative animal model for translation to humans due to their genetic, biological, and physiological similarities (Nardone et al., 2017), however, they were excluded from this review due to significant financial and ethical constraints that limit the accessibility of this model. The purpose of this review is to investigate larger animal models that can be used to validate foundational results observed in rodents before proceeding to test in nonhuman primates or humans. Further, smaller non-rodent species that do not approximate the size of adult humans such as cats and rabbits were also excluded. Dogs were not reviewed due to their widespread domestication and ethical barriers that may impede testing standardization. These ethical concerns were further validated with 2019 legislation that restricts the use of dogs in the U.S. Department of Veterans Affairs laboratory testing to scenarios when a scientific question can only be met using dogs. Per the ruling, researchers must use other animals even when a scientific question could be answered with fewer dogs than a comparable animal, such as swine (National Academies of Sciences, 2020). Large animal models, such as swine, sheep, and goats, may be preferable due to their anatomical features, pharmacologic responses, and overall size when compared to humans.

Literature Search Strategy

The online database PubMed was used to retrieve articles that were used in this review of gait analysis in large animals. Searches were performed between August 2021 and September 2022 using the format (x) AND (y), where group (x) included (swine), (pig), (sheep), (horse), and (goat), while group (y) included one or more from: (gait), (brachial plexus), (treadmill), (MATLAB), (spinal cord injury), (brain injury), (nerve injury), and (sensors). Large animals were included if their motor function was studied following neurologic injury. Non-human primates were excluded from the review due to ethical and financial considerations mentioned previously. Dogs were also excluded from this review due to the widespread domestication of these animals and higher ethical considerations that may impede testing standardization. Horses were initially included as part of the literature review; however, to our knowledge, all gait analysis studies in horses involve nonnervous system ailments and were therefore excluded. In addition, central nervous system or peripheral nervous system injury-related studies involving the use of small animal models such as rats, rabbits, and cats were excluded due to the small anatomic size compared to humans. Finally, if a large animal met the inclusion criteria and therefore had been functionally studied after nervous system impairment, non-nervous system-related articles were included for that animal if they presented information relevant to gait analysis.

Results

Of the functional testing studies reviewed, swine and sheep were the predominant animal models used, while only two studies were found using goats (Figure 1). Swine studies used a variety of assessment types, while sheep and goat studies primarily utilized pen analysis. Data collection methods were structured within subjective and objective frameworks and this delineation was used for comparative analysis across each animal model.

Figure 1｜Charts depicting the animal models and functional assessment types used in literature for swine, sheep, and goats with nervous system ailments.

Swine

Multiple studies have used swine to measure functional recovery outcomes after an injury to the nervous system (Lee et al., 2013; Kinder et al., 2019; Benasson et al., 2020). Many of these studies have been successful in using pen analysis, food-guided walking, and forced walking techniques such as treadmill walking (Streijger et al., 2015; Hrycushko et al., 2019; Cerro et al., 2021). Compared to smaller animal models, some breeds of miniature swine, such as the Yucatan minipig or the Wisconsin Miniature Swine, resemble humans in size with weights similar to a human adult when fully grown. Minipigs, which have been previously used in gait analysis studies (Streijger et al., 2015; Kim et al., 2019; Benasson et al., 2020; Boakye et al., 2020; Hrycushko et al., 2020), weigh 25–50 kg when 4 months old and 68–91 kg when mature (Hanna et al., 2022). Alternatively, full-size breeds, such as Landrace pigs, can weigh around 50 kg when they are 5–6 months old and have been used for brain injury research involving gait analysis (Kinder et al., 2019; Kaiser et al., 2020).

The body and limb lengths of swine and humans are comparable, indicating that the forces involved in ambulation are also similar to those in humans (Webb et al., 2018). Specifically, the upper cervical vertebrae are nearly identical between swine and humans (Sheng et al., 2016). The location of dorsal and ventral horns as well as the lengths and cross-sectional area of the lumbosacral spinal cord are very similar (Sheng et al., 2010; Toossi et al., 2021). The brachial plexus of Wisconsin Miniature Swine and humans have very similar size and origin (Hanna et al., 2022) and the vagus nerves of both species have multiple fascicles and similar fibrous tissues (Stakenborg et al., 2020). Swine may be advantageous as models for human nervous system ailments because piglets have a similar pattern for brain development when compared to human adolescents (Kinder et al., 2019). In addition, the brains of swine myelinate similarly to human brains and have comparable biochemistry (Schomberg et al., 2017; Cerro et al., 2021). The surface of a pig’s brain is more similar to the human gyrencephalic neocortex when compared to that of a rat, which is lissencephalic (Lind et al., 2007; Weber-Levine et al., 2022). Rat brains contain less than 10% white matter while both human and swine brains can contain more than 60% white matter (Webb et al., 2018). Several studies that utilize gait analysis of swine involve SCI (Lee et al., 2013; Benasson et al., 2020; Fadeev et al., 2020), ischemic stroke, and brain injury, as organized in Table 1.

Table 1 ｜Summary of the studies included in this review about gait analysis in swine, sheep, and goats after neurologic injury

Subjective data collection

One subjective measurement of the gait of a pig is called the Porcine Thoracic Injury Behavior Scale (PTIBS), which can be used to assess the functional recovery of pig models after SCI. Like the range of 0–21 for BBB scores (Basso et al., 1995), the PTIBS contains values from 1, indicating no movement in the pig’s hindlimb, to 10, indicating that the pig has a normal gait and full range of motion in the hindlimbs (Fadeev et al., 2020). A score between 1–3 is assigned if the hindlimb is ‘dragging’ on the floor, 4–6 indicates that the pig can ‘step’ with the hindlimbs, and 7–10 indicates that the pig can ambulate (Lee et al., 2013; Kim et al., 2019; Fadeev et al., 2020). The PTIBS scoring system can be used in pen analysis, guided free walking, or treadmill walking assessment designs.

Because the determination of a PTIBS score involves qualitative observations and may be more subjective, Lee et al. (2013) confirmed the reproducibility of these PTIBS scores between multiple investigators. This was done by having some investigators assign PTIBS scores to the pigs in all the sample videos at the beginning and end of 3 weeks and determining whether the scores were reproduced. In addition, several investigators assigned scores to the same video and compared the scores. Lee et al. (2013) also validated the PTIBS using histologic analysis and found the score to be closely correlated with the extent of white and gray matter sparing at the SCI site.

PTIBS scores are only used to assess the hindlimb function of pigs, and thus, can only be practical in functional testing of nervous system injuries that affect the posterior portion of the pig but leave the anterior portion functional. Additionally, hindlimb assessments may not account for compensation for healthy forelimbs. Instead of using the PTIBS, Cerro et al. (2021) designed two scales to measure functional recovery in 2-month-old large, white male pigs after cervical SCI. The first, the Individual Limb Motor Scale (ILMS), scores each limb based on its movement and ability to bear weight. A score of 1 indicates a completely inactive limb while 8 indicates normal use. The second, the Quadruped Position Global Scale (QPGS), scores the ability of the pig to stand up. Ranging from 1 to 6, the scores indicate the “axial muscle tone” with a score of 6 being able to stand up for 5 minutes or more. The conjunction of these scales allowed Cerro et al. to track the changes in gait even with greater paralysis in the upper limbs (Cerro et al., 2021).

Objective data collection

Several research groups have used a pressure-sensitive mat and trained the pigs to walk across it during testing. The mat contains sensors that can detect changes in pressure providing data on all four limbs of the pig. Pigs naturally place 40% of their body weight on the hindlimbs and 60% on their forelimbs (Kaiser et al., 2020), which may be important depending upon the downstream effects of the nervous system ailments on the individual limbs. Kinder et al. (2019) used a 6.10 m × 0.61 m mat that contained 23,040 total sensors and was sensitive to the pressure from the pigs’ hooves. To force the pigs to trot on the mat, training was required before the trials (Sneed et al., 2021) and food was used as an incentive for them to travel from one end to the other (Kinder et al., 2019). Kinder et al. (2019) repeated the trials until they obtained four or five consistent trials or until they reached their upper limit of 20 minutes spent on a single pig’s trials. A successful trial consisted of at least 3 gait cycles in which the pig’s walking speed had 10% or less variability.

An image-based gait analysis can also be set up using cameras to record the pigs’ gait cycles. Fadeev et al. (2020) used a single camera for recording, while Streijger et al. (2016b) used two cameras to laterally capture the pigs’ gaits. A successful recording included 5 gait cycles and was 10 to 15 seconds long (Fadeev et al., 2020). The recordings were used to objectively measure joint kinematics and assign PTIBS scores. Fadeev et al. (2020), Streijger et al. (2016b), and Cerro et al. (2021) placed markers on the joints of the pigs, and the changes in position were used to calculate the angles of each joint using Kinovea software 0.8.25, Dartfish Software, or WINanalyze V. 2.2 software. Benasson et al. (2020) guided the pig in a line using food and utilized eight cameras to capture 18 distinct marks that were placed upon the joints of the limbs. Six marks were placed on each hindlimb, and three on each forelimb.A custom script “(MATLAB R2019b, The MathWorks, USA)” calculated each parameter based on the changes in the position of each marker (Benasson et al., 2020).

Assessment designs

Pen analysis, treadmill walking, and guided free walking assessment designs have all been used with the swine model. Both objective kinematic measurements and the subjective assignment of PTIBS, ILMS, or QPGS scores can be used to analyze gait within the different assessment designs. Several studies have used both measurement types whereas others employed only a single type.

Pen analysis

Observing swine inan open pen can be an effective means for assessing functional abnormalities after injury. Pigs typically have a “symmetrical fourbeat gait” and alternate between using two or three limbs for support while walking (von Wachenfelt et al., 2008). When exploring the tolerance of Yutacan minipig spinal cords after radiation therapy, Hrycushko et al. (2019, 2020) allowed the pigs to walk freely in their pen and observed them for 52 weeks or until they underwent a change in gait. To determine a change in gait, Hrycushko et al. (2019) checked for asymmetric gait and any other visual indication that the pig was unwell or unable to care for itself. Thus, the symmetry of the pig’s gait can be a clear indication of whether the pig’s ability to move one or more limbs may be impaired and should be further studied.

The floor conditions are also an important factor in open pen analysis as inadequate conditions may lead to unwanted changes in gait. For example, the pig can slip when walking on floors with low friction while abrasive floors can wear out their hooves (von Wachenfelt et al., 2008). Kaiser et al. (2020) used rubber mats to avoid slipping. Additionally, the floor material and cleanliness should remain constant to avoid variability and extraneous changes in gait unrelated to the study. When the floor was clean in the study by von Wachenfelt et al. (2008), the pigs had less stance time as well as longer, faster strides. There should also be plenty of space for the pig to explore within the enclosed pen. Scheulin et al. (2021) used a pen that had an area of 2.7 m2for the pig to easily ambulate. Lastly, white curtains can be hung around the enclosed pen during each trial to avoid distractions (Kaiser et al., 2020). When analyzing the pig’s gait within the pen, a digital camera can be used to capture limb and body movements. These video recordings can then be analyzed by software or subjective assessment. Very few nervous system injury-related studies have analyzed pigs walking freely in their pen, and specific equipment set-ups are not well established. Webb et al. (2018) performed gait analysis in an open field, and although the camera set-up was not given, they did measure several parameters including distance traveled, velocity, cadence, stride length, and relative pressure.

Treadmill walking

An alternative to allowing the pig to walk freely is to analyze its gait on a treadmill, which typically requires rigorous training beforehand. A study by Streijger et al. (2016b) required at least 3 weeks of daily training before inflicting SCI. They allowed two pigs to walk together with food rewards to increase initial comfortability and then trained each pig separately until it could walk for 6 minutes at a speed of 1.5 m/h (~0.67 m/s). At this point, a harness was placed around their midsection to simulate the weight support system used after SCI (Streijger et al., 2016b). In the study performed by Fadeev et al. (2020), swine were trained on the treadmill 2 weeks prior to surgery and the speed was set between 0.3 and 0.4 m/s.

Both studies began treadmill trials 2 weeks after SCI (Fadeev et al., 2020) and continued testing 8 to 12 weeks post-injury (Streijger et al., 2016a, b). Fadeev et al. (2020) wrapped a support band around the midsection to allow only 5% to 20% of the weight to fall on the hindlimbs. On the other hand, Streijger et al. (2016b) allowed 100% of the pig’s weight to be held by the band for the initial post-injury trials. Boakye et al. (2020) also conditioned the pigs to associate an audible click with food, incentivizing them to walk toward a target with their snout and perform desired tasks such as walking on a treadmill.

Guided free walking

Instead of allowing pigs to roam freely in the pen, several investigators have trained pigs to walk in a straight line at a constant speed using food or other positive reinforcement techniques. Benasson et al. (2020) and Streijger et al. (2021) independently explored this method of gait analysis of female Yutacan miniature pigs that underwent a surgically-induced SCI. Streijger et al. (2021) familiarized the pigs with the facility for 1 week and the researchers themselves for 4 days. Then, the pigs practiced walking across a rubber mat for 1 hour each day for 5 days. In a later study, Streijger et al. (2021) trained the pigs for 2 weeks before SCI to desensitize them to human handling and allow them to walk on a rubber mat calmly. Food was used to incentivize them to walk and PTIBS scores were used to rate the swine’s gaits (Streijger et al., 2016a, 2021). Several cameras, set at different angles, captured a video for analysis and assignment of a subjective score (Streijger et al., 2015).

Sheep

Sheep were the second most frequently used large animal model to test locomotor function after neurologic injury and they are the most common animal model in myelomeningocele (MMC) studies (Yamashiro et al., 2020). Sheep were also used in SCI studies and their gait was assessed post-injury by forced treadmill walking (Safayi et al., 2015; Wilson et al., 2017). Some studies did not involve any injury to the sheep but still analyzed gait using techniques such as guided free walking (Agostinho et al., 2012; Faria et al., 2014) or treadmill walking (Diogo et al., 2021).

The anatomical characteristics of sheep make them a viable model for translating results to humans. Scientists at Iowa State University selected sheep for functional testing because the size and morphology of the adult ovine spinal cord are similar to those of humans (Safayi et al., 2015). In addition to spinal cord similarities, the cervical vertebrae of sheep and humans have remarkably similar functions despite their differences in appearance (Cain and Fraser, 1995). The diameter of the transverse foramen of cervical vertebrae is slightly smaller but very similar in size. Transverse foramen was measured to be 3–6 mm in the ovine models, and 4–7 mm in humans (Cain and Fraser, 1995). The atlas and axis of the ovine model do have different structures, as the vertebrae in sheep are taller (Kandziora et al., 2001). Despite this structural difference, the atlantoaxial articulation that allows for the rotation of the head side to side functions the same as it does in humans (Cain and Fraser, 1995).

A systemic literature review conducted by Sheng et al. (2010) measured morphometric components of vertebrae in many animal models such as sheep, porcine, calf, baboon, and deer. There is no ideal animal model for the human spine, but different models have more anatomical similarities to specific regions of the human spine (Sheng et al., 2010). The sheep model was determined to be a viable model for research involving the gross structure of the thoracic and lumbar spine (Wilke et al., 1997). The size and morphology of the adult ovine spinal cord are similar to those of humans (Gibson-Corley et al., 2012) and therefore, a mature sheep model could be chosen forin vivoexperimental use (Sheng et al., 2010; Gibson-Corley et al., 2012).

Subjective data collection

Some sheep studies use an observational rating scale such as the Thoracic Injury Behavior Scale (TIBS) to track the recovery of the subject after the injury. The TIBS system is a slightly modified version of the PTIBS system mentioned earlier. The scale is broad, with a range from 1 to 10. A score of 1 is assigned to an animal with no hindlimb movement and a score of 10 indicates normal walking (Wilson et al., 2017).

In a study done by Wilson et al. (2017) at the University of Iowa, the TIBS score was used to evaluate the recovery of nine sheep each day for 16 days after the SCI while walking on a treadmill. The final TIBS score was determined 30 days after the injury to ensure the effects of the injury were stabilized. Six sheep had their spinal cords contused with a weight drop height of 10 cm and three sheep had their spinal cords contused from a 7.5 cm height. The researchers aimed to obtain a cohort of sheep with final TIBS scores ranging from 7 to 9, indicating a mild SCI that would induce hindlimb muscle spasticity but still allow for balance and ambulation (Wilson et al., 2017).

For the six 10 cm weight drop sheep, the TIBS scores varied from 2 to 9 with a mean of 4 on post-injury day 1. The day 30 scores for these sheep ranged from 8 to 10 with a mean of 9. Comparatively, the three 7.5 cm weight drop sheep were measured for 16 days post-injury and had similar TIBS scores ranging from 8 to 10 on post-injury day 16. Wilson noted that while the consistency of these results is encouraging, this scale is still quite broad, and a higher level of detail could improve post-injury analysis (Wilson et al., 2017).Another type of subjective evaluation used in the ovine model is the locomotor rating scale (Brown et al., 2015; Galganski et al., 2020; Yamashiro et al., 2020). The locomotor rating scale for sheep was developed from the original rodent BBB scale to address the many anatomical differences between the two animals.

The scale was developed based on the behavior of 20 sheep. An MMC defect was surgically created in fetal lambs, which accounted for 15 of the sheep. A spontaneous MMC defect is a central nervous system defect where the neural tube does not fully close during the development of the spinal cord (Wang et al., 2015). In brief, each ewe underwent a survival laparotomy and hysterotomy at a gestational age of approximately 75 days (Brown et al., 2015). Four of the lambs did not undergo any surgical procedures and served as control subjects. The final sheep was a pregnant ewe that underwent the same procedures as the other 15 ewes but did not have a defect creation at the gestational age of 75 days to control for any possible effects of surgery on the lamb’s motor function (Brown et al., 2015).

The locomotor rating scale is divided into seven categories of assessment: hindlimb movement, stance with help, spontaneous hindlimb weight support, spontaneous standing, stepping, coordination, and hindlimb clearance. An overall grade is assigned based on the combined performance in all seven categories, with a grade of 0 indicating complete paraplegia and 15 indicating normal ambulation.

The hindlimb movement category assesses the hip, knee, and ankle joints of each hindlimb with a score of 0 to 2 for each joint, with 0 indicating no movement and 2 indicating extensive movement. A maximum score of 12 is achieved if all three joints on both hindlimbs have extensive movement.

The stance with help category assesses whether the sheep can stand on its own or stands with help and is scored as yes or no. The spontaneous hindlimb weight support category assesses the ability to lift the hips off the ground and support full body weight. Spontaneous standing is rated as yes or no based on the ability to complete a full sitting-to-standing movement.

If the sheep can stand (assisted or spontaneously), the ability to ambulate is assessed. Stepping is divided into three categories based on the number of continuous steps taken by the hindlimbs: 0, 1–4, and > 5 steps. Coordination of the forelimbs and hindlimbs is assessed as no coordination, occasional (< 50% of the time), or frequent (≥ 50% of the time) coordination, with a coordinated step being one forelimb step for every hindlimb step at a constant speed.

Finally, the clearance test assesses the ability to walk over a 4-inch by 4-inch wooden bar in an open field and is scored as pass or fail. A final grade is assigned based on the combined performance score in these seven categories. For a more detailed table of the sheep locomotor rating scale, see Brown et al. (2015).

Objective data collection

Image-based gait analysis has been used alongside treadmill walking with sheep because the video-based observations provide subtle measures of the effects of SCI on gait (Safayi et al., 2015). In a study designed by Safayi et al. (2016), sheep were fitted with reflective limb markers and recorded while ambulating on a treadmill. Each trial was recorded using six infrared cameras positioned around the treadmill (Safayi et al., 2016). Diogo et al. (2021) replicated this design using three cameras to capture the entire sheep as it ambulated. Reflective infrared markers were placed on each of the four hooves, the hock (tarsal/tarsus) joints, and upper and lower segments surrounding the hock joint. The sheep were then recorded using the Vicon Motion system (Reddy et al., 2018). All the reflective markers created a model that was used to analyze the gait pattern of the subject (Faria et al., 2014). Diogo et al. (2021) used similar markers to create a 3D reconstruction of the limb using a Direct Linear Transformation algorithm.

The “ovine model,” developed by Safayi et al. (2015), explains the kinematic parameters of thoracic and pelvic limb gait patterns, as well as temporospatial measures of hock joint angle in the sagittal plane in clinically healthy sheep using image-based gait analysis. A custom MATLAB program was created to calculate quantitative data produced by the ovine model. Two general categories of data were analyzed: kinematics of the hock joint, and interlimb coordination (Safayi et al., 2015). Interlimb coordination was evaluated by the stride, stance, and swing measurements in the thoracic and pelvic limbs of healthy sheep, as well as the hoof elevation in the swing phase and hoof velocity. The hock joint angle values were evaluated by hock joint flexion and extension during the stance and swing phases (Safayi et al., 2015). This combination of video recording, motion tracking, and kinematic analysis was the most standardized and comprehensive data collection method found in the sheep model.

Pressure-sensitive mats can also be utilized to measure limb-specific gait characteristics quantitatively. In a study by Agostinho, et al., sheep were guided along a pressure mat and measurements related to the gait cycle and stride were calculated. In addition, they measured the vertical force and impulse that the limbs placed on the sensors. This was combined with three recording cameras and markers glued onto the skin to measure the kinematics of the sheep’s gait more comprehensively (Agostinho et al., 2012).

Assessment designs

Sheep were assessed through three different methods: treadmill walking, guided free walking, and open field testing. Treadmill walking tests with sheep require an acclimation period because sheep are natural herd animals, making them averse to walking in an isolated setting (King et al., 2012). Therefore, separating them from one another and forcing them to walk in a straight line takes time for acclimation. Alternatively, the locomotor function can be evaluated through pen analysis, which does not require an acclimation period.

Treadmill walking

Several groups have used treadmills as part of their gait analysis to test the functionality of sheep limbs. Before beginning data collection, most groups trained the sheep for six to ten sessions over two to three weeks (Safayi et al., 2015; Wilson et al., 2017; Diogo et al., 2021). Diogo et al. (2021) used a halter to guide the sheep onto the treadmill and the hooves were trimmed before training. The goal of this training procedure was to get the sheep to walk at a constant speed of 3.75 km/h for 5-minute trials (Wilson et al., 2017). Diogo et al. (2021) found that healthy sheep can walk on a treadmill at a speed of 1.2 m/s. On average, it took about 30–45 minutes for each sheep to become accustomed to walking on the treadmill (Safayi et al., 2015). Because sheep tend to huddle together for protective reasons, Diogo et al. (2021) required the use of the same handler to train all the sheep to walk on the treadmill to avoid any negative repercussions of being separated from the rest of the sheep. In addition, if a sheep showed signs of stress, the subject’s trial would be stopped until it was calm and then the trial would start again (Safayi et al., 2015).

After the implementation of the injury, the subjects participated in another round of testing to look for statistically significant changes in the measurements that were recorded during the treadmill analysis before the SCI. In a study by Safayi et al. (2015), there were significant changes in the mean angular velocity of the hock joint stance to swing ratio. This indicates that the program they created was effective in yielding intricate differences in results, which is important for an experiment working with slight changes in gait measurement.

Pen analysis

Pen analysis, also called open-field testing, involves observational assessment of the animal while it roams freely in a defined area. Pen analysis was the predominant assessment design in the sheep model and was often coupled with the locomotor rating scale (Brown et al., 2015; Kabagambe et al., 2018; Joyeux et al., 2019). In the MMC study by Brown et al. (2015) mentioned earlier, open field testing was initiated shortly after lamb birth, and then 24 hours after they were allowed to acclimate to the open field environment. Each trial was recorded on a video camera for approximately 5 minutes. Locomotor patterns of the sheep were observed and recorded to collect data for the development of the locomotor rating scale described earlier. Lambs capable of spontaneous standing and/or ambulation were allowed to move without interference. If unable to stand spontaneously, lambs were first observed, then assisted to stand by lifting the lamb into a standing position with all limbs in contact with the ground. Any ability to stand unassisted or ambulate after being assisted to stand was observed without interference. Lambs unable to stand even after assistance were evaluated by assessing movement in each joint of each hindlimb (Brown et al., 2015).

Guided free walking

In guided free walking, sheep can be guided in a straight-line using food as positive reinforcement or previous training to walk along a pathway (Agostinho et al., 2012; Faria et al., 2014). A halter is used to assist with guiding the sheep. Agostinho, et al. (2012) found that it required training twice per day for three weeks before data collection. Forty total trials were collected for each animal and were considered valid if the sheep walked between 1.1–1.3 m/s and accelerated between –0.15 to 0.15 m/s. The trial could have no head movement or halter pulling, and all four limbs had to contact the walking surface. Gait analysis was then performed by Walkway 7.0 (Tekscan) software (Agostinho et al., 2012).

Goats

To our knowledge, there are very few studies that have tested goats for locomotor function following a neurologic injury. Still, goats are a viable research model due to their similarities in overall and vertebral canal size when compared to humans. Additionally, the goat vertebral column has a uniform and regular shape that simplifies surgical procedures and spinal manipulation (Cao et al., 2014). The human and goat spine share similarities, most notably in the structure and biomechanics of both the lumbar and thoracic vertebrae (Qiu et al., 2015). The goat spinal canal’s depth and width are quite comparable to a human’s thoracic spinal cord region (Roels et al., 2022). Goats’ larger anatomy more closely resembles that of an adult human in comparison to other smaller animal models (Qiu et al., 2015). In this review, goat studies included SCI and pure observational analysis without injury.

Subjective data collection

In a study by Cao et al. (2014), a Tarlov motor function grade test was performed for behavioral analysis of the goats after a balloon-induced SCI. This test examines the goat’s motor function of their hindlimbs after SCI. It was performed 24 hours before surgery and a week following the surgery in the study by Cao et al. (2014), and 5 days, 2 weeks, 4 weeks, and 8 weeks post-surgery in a study by Jiang et al. (2017). The goats were scored according to their locomotive ability by blinded assessors to eliminate any potential biases. Goats were assigned a numerical value of 0–5 according to the following scale: 0 = paralysis of hindlimbs, 1 = Inability to stand but ability to move hindlimbs, 2 = inability to walk but the ability to stand on hindlimbs, 3 = inability to run but the ability to walk, 4 = inability to jump but ability to run, and 5 = ability to jump and normal gait pattern (Cao et al., 2014; Jiang et al., 2017).

Objective data collection

Pressure mats have been used to assess gait characteristics as a goat walks across an array of sensors. Determining how the goat distributes weight across its limbs can help assess the functionality of the individual limbs. The use of pressure-sensitive mats is a time-efficient method, as one can evaluate several limbs in the same trial (Rifkin et al., 2019). Rifkin et al. (2019) used the Walkway Pressure Mapping System (Tekscan Inc., South Boston, MA) with a mat that was 87.1 cm × 36.9 cm and a sensor density of 1.4 sensors/cm2. Along with the data collected from the pressure mats, each trial was recorded using a digital video camera. Video frames were aligned with pressure mat data to calculate specific gait characteristics. Gait was measured in time, number, and distance and involved the stance and stride. Force and impulse were also measured for each limb (Rifkin et al., 2019).

Assessment designs

Due to the limited number of studies utilizing the goat model for functional analysis, guided free walking was the only assessment design used in the goat model.

Guided free walking

Like swine, goats can also be guided to walk along a straight pathway. In a study by Rifkin et al. (2019), goats were required to walk along a set path of pressure-sensitive mats that were placed over concrete flooring. A 2.5 to 5 cm layer of wood shavings was also spread over the mat. The researchers created an alleyway with walls to allow the goat to walk freely and ensure that each contact of the hoof with the mat will be measured by the pressuresensitive mat. The alleyway also discouraged goats from turning around and walking in the opposite direction. In addition, they also used a halter to lead the goat at a walking pace along the mat, which encouraged the goat to not stop during the trial. However, there was no pressure applied to the halter to avoid forcing any ambulation and to allow consistency between goats. They rejected any data where the goat hesitated or changed their gait or pace of gait for any reason. The goats had no difficulty walking in the alleyway and were trained quickly (Rifkin et al., 2019).

Discussion

Rodents are the most commonly used animal models for motor function testing after neurological ailments due to their affordability, short gestational length, low upkeep, and overall ease of use. This has led to standardized testing procedures and data collection methods, such as the BBB scale in the rat model. However, the anatomical differences between rodents and humans limit the utility of study results in translational research. Large animals, such as swine, sheep, and goats, are more anatomically similar to humans but have not been studied to the same extent as rodents. Thus, there is a need for standardization in large animal testing models with neurologic impairment.

Animal model comparison

In comparing the swine, sheep, and goat animal models, both anatomical characteristics and assessment compliance must be considered. In all three models, the relative size and morphology of the spinal cord were more similar to humans when compared to smaller rodent models (Cao et al., 2014; Safayi et al., 2015; Schomberg et al., 2017). There were also no appreciable anatomical differences between the three models as it pertains to gait analysis that made one superior to the others. In general, these animals were well suited for gait analysis due to their larger limb length, spinal cord morphology, and body weight which may better simulate the biomechanical forces of ambulation experienced by adult humans (Webb et al., 2018).

It should be noted that as animal weights begin approximating that of humans, the associated costs of housing, husbandry, and veterinary resources tend to increase as well. For example, in the case of SCI, post-injury assistance may be required for eating, drinking, and mobility to avoid pressure sores. Antibiotics and pain medications may also be required under the guidance of a veterinary team (Weber-Levine et al., 2022). These handling challenges could offset some of the benefits of using larger animals and have caused some researchers to utilize animals at younger ages due to their lighter weights (Lee et al., 2013). We acknowledge that access to proper facilities and veterinary resources can also be limiting factors. However, in the case of preclinical therapies that have been effective in rodent models, larger animals still present a valuable opportunity for further pre-clinical validation.

In testing, swine performed well in pen analysis, guided free walking, and treadmill walking, particularly when food is used as a guide. Comparatively, sheep demonstrated sufficient study compliance when subjected to isolated treadmill walking despite their natural herd tendencies (King et al., 2012; Safayi et al., 2015; Wilson et al., 2017; Diogo et al., 2021). Once sufficiently trained, sheep assume a regular gait pattern that has little variation in their stride mechanics, making them a good candidate for gait analysis (Safayi et al., 2015). Goats did not exhibit any difficulty in a guided walking format when being led by a trainer, but no examples of treadmill or open field studies could be found.

Selecting an appropriate large animal model ultimately depends on the housing and veterinary resources available to the researchers, as well as the aims of the study. For example, swine have well-established testing protocols for SCI, while sheep were predominantly used in MMC defect studies. Future studies may replicate or modify these frameworks based on the goals of the researcher.

Assessment design comparison

Pen or open field analysis, treadmill walking, and guided free walking are the predominant assessment designs in large animal functional studies (Figure 1). Each method has differing degrees of variable control and preparatory animal training required to perform the assessment.

Pen and open field analyses allow the animal to behave naturally and require little to no training or acclimation period. Both subjective scoring and objective data collection methods can be used with an open field model, however, no established camera/sensor setup was found for image-based gait analysis. This may be due to the high variability in animal behavior and walking speeds in an open-field environment. Despite the difficulty in using objective assessment, gait symmetry can be assessed subjectively and has been shown to indicate functional status in swine (Hrycushko et al., 2020). Floor conditions within the pen can also influence the animal’s gait and should be carefully controlled in each assessment iteration (von Wachenfelt et al., 2008).

Treadmill walking offers the highest degree of study control but also the largest amount of training required. The treadmill allows the researcher to precisely vary the speed while incorporating subjective and objective data collection methods such as cameras and motion sensors for gait analysis. Training for treadmill walking can be variable for both herd and non-herd animals. Swine required rigorous training for 2–3 weeks while sheep required 6–10 sessions over 2 weeks to acclimate to the treadmill (Streijger et al., 2016b; Diogo et al., 2021). No studies were found involving goats walking on a treadmill.

Guided free walking allows the animal to walk naturally in a straight line while being guided by food, halter straps, or previous training. This method appears to work best with swine, but it was not studied extensively in sheep or goats. At least 2 weeks of training were required for acclimation to the facility and testing surface for swine (Streijger et al., 2021). Comparatively, goats were trained quickly with a halter strap and wood chip path, though an exact time frame was not given (Rifkin et al., 2019). No examples of guided free walking were found in sheep models.

Assessment design selection may depend on the degree of neurologic injury and may therefore change throughout the study as the animal regains motor function. For example, an open field analysis may be more appropriate immediately after injury when an animal has a limited motor function. A subjective behavioral scale can be used to assess the animal’s gross motor recovery in the open field before moving to more complex motor assessments such as treadmill walking. Animal compliance, training time, and study aims should all be considered when choosing the most appropriate assessment design.

Data collection comparison

Subjective analyses such as the PTIBS, TIBS, locomotor rating scale, and Tarlov Motor Scale have been translated from the rodent BBB scale, creating a potential method of comparison between the small and large animal studies. PTIBS and TIBS are limited by their primary assessment of the hindlimbs without considering possible compensation by the forelimbs. Despite this limitation, Lee et al. (2013) found the PTIBS to be closely correlated to the extent of white and gray matter sparing at the SCI site.

The ILMS/QPGS addresses the issue of forelimb compensation by individually scoring each limb’s ability to move and bear weight. Further, the ILMS/QPGS evaluates the ability to stand up, though this has only been used in the swine model. The ILMS/QPGS was used in only one study within this review, but the results were useful in tracking gross motor recovery up to 90 days after SCI in swine. The ILMS/QPGS was able to differentiate the swine with either unilateral or bilateral spinal cord lesions and was also useful in assessing animal readiness for quantitative gait analysis.

The locomotor rating scale, which was adapted from the BBB scale, evaluates sheep based on their hindlimb joint movement, ability to bear weight and stand, and ability to walk with forelimb-hindlimb coordination. The locomotor rating scale also addresses all four limbs and incorporates coordination when walking. In practice, the scale was able to discriminate between a wide range of neurological recovery and demonstrated reliability with low standard deviations between examiner scores. However, this scale was only used to assess fetal lambs with MMC defects in an open pen analysis and therefore, has only been proven effective in this narrow use case (Brown et al., 2015; Wang et al., 2015; Galganski et al., 2020).

Finally, the Tarlov motor test is used in goat models to assess hindlimb function, standing ability, and walk/run/jump ability but again does not address forelimb compensation. This test was only used in two spinal cord ballon-compression studies, however, the motor deficits represented by the Tarlov scores were highly correlated with spinal cord conduction deficits when somatosensory evoked potentials were measured post-injury. More studies are needed to further validate the predictive value of the Tarlov test, but the initial results are promising (Cao et al., 2014).

With each of the observational methods, a precise injury characterization can be difficult when multiple degrees of injury severity may present with similar functional capacity. In these scenarios, it is likely best to supplement the observational scale with a quantitative analysis that provides more granularity to the injury classification. However, the observational scales still serve an important role as an efficient means for monitoring gross injury recovery. It should be noted that these behavioral scores are subject to the individual assessor, creating inherent variability. Score reproducibility should be checked by having all assessors score sample videos before and after testing, as well as having multiple assessors score the same videos when possible (Lee et al., 2013).

In summary, researchers should select an appropriate behavioral assessment based on the animal and injury models used. These studies present promising results to support testing standardization within their respective animal and injury models.

Objective data collection models utilized pressure mats, video recording, and image-tracking software to analyze large animal gait after neurological injury. Each objective model can be used in combination with other subjective analyses but vary in technical complexity.

Pressure mats were used in the swine and goat models and integrated well with guided free walking analysis as the animal can acclimate to both modalities simultaneously. There was not a standardized mat size, number of sensors, or software used across the studies, but each study incorporated a live or recorded visual analysis component. Pressure mats do not require additional training for the animal and provide data for all four limbs, which may otherwise be omitted in some subjective scoring methods that only evaluate the hindlimbs. It should be noted that pressure mats are ineffective if the animal is dragging its legs, which is often seen immediately following SCI.

The image-based gait analysis model, developed by Safayi et al. (2015) for sheep, tracks gait using a motion tracking system, along with infrared cameras and reflectors attached to the animal’s joints. The raw 3D positioning data is exported to an external software program for analysis (i.e., MATLAB, WINanalyze, Dartfish, Kinovea). This model provides the most detailed assessment of hock joint kinematics and interlimb coordination but requires the most equipment and technical expertise. Despite technical complexity, the image-based gait analysis model was replicated multiple times (Faria et al., 2014; Wilson et al., 2017; Reddy et al., 2018; Diogo et al., 2021).

The selection of an objective assessment is dependent on the facilities, resources, and technical expertise of the research team. The objective models avoid the potential for assessor bias that exists in behavioral assessments, so it is recommended that a combination of the two assessment types is used.

Conclusion

To improve treatments and functional recovery outcomes in patients with neurological ailments, larger, more anatomically similar animal models with standardized testing methods are needed. While all the reviewed animal models are viable options, swine and sheep possess the ideal combination of anatomical similarity and testing compliance in a variety of assessments. Data collection methods for functional recovery should include both a subjective and objective component, similar to well-established small animal assessment designs in rodents. The objective data collection and assessment design in large animals are dependent on the resources available to the researcher. Additional large animal testing is needed to improve and standardize the current functional testing models for neurological ailments.

Author contributions:Literature search: JWS, RRM, TLM, and JMJ. Analysis and interpretation of data: JWS, RRM, TLM, and JMJ. Drafting of the manuscript and critical revision of the manuscript: JWS, RRM, TLM, JMJ, DJH, and ASH. All authors had full access to all the data in the manuscript and take responsibility for the integrity of the data and the accuracy of the data analysis, and approved the final manuscript.

Conflicts of interest:The authors declare no conflict of interest.

Data availability statement:No additional data are available.

Open access statement:This is an open access journal, and articles are distributed under the terms of the Creative Commons AttributionNonCommercial-ShareAlike 4.0 License, which allows others to remix, tweak, and build upon the work non-commercially, as long as appropriate credit is given and the new creations are licensed under the identical terms.