The walk and trot are both symmetrical gaits. In dogs, there are 2 phases to the walk; ipsilateral 2-legged support of the body alternates with 3-legged support of the body.1 In the trot, the body is supported by a diagonal pair of limbs that alternates with an aerial phase between alternations. The forelimbs have a primordial weight-bearing role and contribute less to propulsion than do the hind limbs on level ground.2–5 Forelimbs have been traditionally regarded merely as compliant struts during steady-state locomotion.6 Historically, forelimb research was separated into 2 areas of study, ground-reaction forces3,7–12 and 2-D descriptions of segment and joint kinematics.13–17 Fischer and Lilje18 used a marker-based system to study the kinematics in the parasagittal plane at all gaits for 327 dogs of 32 breeds and also used fluoroscopy to describe joint angles and segment motion during a walk and trot for dogs of 7 breeds and compared those findings with those of other published reports.
Unfortunately, separate consideration of GRFs and kinematics cannot be used to accurately describe joint dynamics. Inverse dynamics analysis is an engineering science methodology that combines kinetic, kinematic, and morphometric data and provides a robust approach for describing the causes of movement patterns from the perspectives of both joint and net muscle work.19 To our knowledge, the 2-D inverse dynamics of the lower portion (ie, distal to the shoulder joint and scapular fulcrum) of the canine forelimb have been described in only a few studies.20–22 In 1 study,20 the intersegmental forces, moments of force, and mechanical power of the forelimbs had consistent patterns and magnitudes in healthy similarly sized dogs at a walk. Inverse dynamics of the forelimb have also been described for dogs with a fragmented medial coronoid process21 and dogs with chronic disease of the medial coronoid process that were treated with conservative management or arthroscopic surgery.22
Limb motion during locomotion is a 3-D phenomenon; thus, abductive and adductive motions in conjunction with axial rotation need to be quantified for a comprehensive description of limb and joint dynamics. For dogs, scientific literature regarding the dynamics of the scapular fulcrum or shoulder joint during locomotion and gait-related joint mechanics and control is lacking. Forelimb segments such as the scapula and humerus are completely or partially embedded in the thorax; therefore, accurate description of their kinematics is challenging. However, the scapula plays an essential role in forelimb dynamics. The uppermost pivot point of the forelimb is located near the dorsal aspect of the scapula, and movement of the scapula accounts for > 60% of forelimb step length.18 Additionally, fundamental biomechanical differences between the walk and trot and the effect of those differences on the general behavior of the canine forelimb (eg, vaulting mechanics23 vs bouncing mechanics24) might induce changes in joint torque similar to those observed in equine forelimbs25 as well as joint work.
The purpose of the study reported here was to conduct 3-D inverse dynamics analysis of the entire forelimb (including the most proximal aspects such as the shoulder joint and scapular fulcrum) of healthy Beagles at a walk and trot. Our specific goals were to determine the extent of nonparasagittal motion in all segments (particularly the most proximal segments) of the forelimb and whether gait (walk or trot) causes changes in the 3-D patterns of segment angles, net joint torques, and powers. Additionally, we discussed published EMG activity data relative to joint torque and power data to elucidate the functional role of the major forelimb muscles.
Materials and Methods
Animals
All experiments were carried out in strict accordance with the German Animal Welfare guidelines of the states of Thuringia and Lower Saxony. Five healthy adult Beagles with ages ranging from 3 to 5 years that belonged to a research colony of the Small Animal Hospital of the University of Veterinary Medicine, Hannover, Germany, were used for the study. To be included in the study, each dog had to be healthy and free of orthopedic abnormalities as determined by results of physical and orthopedic examinations. All selected dogs were regularly trained on treadmills for various locomotion studies; therefore, they did not require acclimation to the treadmill prior to study initiation.
Experimental design
Data required for computation of inverse dynamics include segment morphometry, segment kinematics, and GRFs. Experimental assessment of each dog consisted of 2 phases; data were first collected by use of fluoroscopy followed 2 days later by use of a motion laboratory. Data were obtained by instrumentation of dogs with reflective markers, walking or trotting the dogs on a treadmill, and capturing data with fluoroscopic imaging or video cameras for subsequent analysis. Movement of the reflective markers on the overlying skin relative to the underlying bone during locomotion represents a major source of error in kinematic analyses and can lead to systematic miscalculation of bone kinematics,26,27 especially for the scapula and humerus.18 Therefore, we also conducted SR for 1 dog to highlight the difference between scapular and humeral motions computed on the basis of information obtained with the reflective markers and that obtained from fluoroscopic evaluation.
Instrumentation of dogs for data collection
Only the left forelimb of each dog was evaluated in the study. The left forelimb was instrumented with 19 reflective markers, which exceeded the number of markers required for the analysis; however, it is helpful to have dogs instrumented with extra markers in case some are occluded during locomotion. The markers were tentatively placed at specific locations to approximate the proximal and distal centers of rotation (ie, joint centers) for each limb segment. Then, the dog was held stationary in a standing position and walked on a treadmill in front of a biplanar x-ray fluoroscope,a and the positions of the anatomic markers were adjusted as necessary to match joint position and bone orientation during the stance phase (Figure 1). After the markers were confirmed to be in the correct positions, the hair at each marker site was shaved to ensure that the markers were replaced in the correct positions in the motion laboratory. Markers were attached at the dorsal margin of the scapula; ventral angle of the scapula; spine of the scapula midway between its most dorsal and ventral points (cluster of 3 markers in the form of a T [T cluster]); lateral midshaft region of the humerus (T cluster); lateral and medial epicondyles of the humerus; cranial and caudal aspects of the antebrachium midway between the elbow and carpal joints; lateral, medial, and palmar aspects of the carpal joint; lateral, medial, and palmar aspects of the metacarpophalangeal joint; and dorsal aspect of the third phalanx.
Representative photographs of the treadmill and biplanar x-ray fluoroscope system used to confirm correct positioning of reflective anatomic markers (A) and the left forelimb of a healthy adult Beagle (B) that depicts the locations at which 19 anatomic markers were placed on each of 5 similar dogs to collect 3-D kinematic and kinetic data for inverse dynamics calculations. For each dog, the markers were tentatively placed at specific locations (dorsal margin of the scapula; ventral angle of the scapula; spine of the scapula midway between its most dorsal and ventral points [cluster of 3 markers in the form of a T {T cluster}]; lateral midshaft region of the humerus [T cluster]; lateral and medial epicondyles of the humerus; cranial and caudal aspects of the antebrachium midway between the elbow and carpal joints; lateral, medial, and palmar aspects of the carpal joint; lateral, medial, and palmar aspects of the metacarpophalangeal joint; and dorsal aspect of the third phalanx. The locations for the 2 markers [medial condyle of the humerus and medial aspect of the metacarpophalangeal joint] that cannot be seen in the photograph are indicated [green arrows]) to approximate the proximal and distal centers of rotation (ie, joint centers) for each limb segment. Then, the dog was held stationary in a standing position and walked and trotted on the treadmill in front of the fluoroscope, and the positions of the anatomic markers were adjusted as necessary to match joint position and bone orientation during the stance phase. After the markers were confirmed to be in the correct positions, the hair at each marker site was shaved to ensure that the markers were replaced in the correct positions for subsequent data collection.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
Representative photographs of the treadmill and biplanar x-ray fluoroscope system used to confirm correct positioning of reflective anatomic markers (A) and the left forelimb of a healthy adult Beagle (B) that depicts the locations at which 19 anatomic markers were placed on each of 5 similar dogs to collect 3-D kinematic and kinetic data for inverse dynamics calculations. For each dog, the markers were tentatively placed at specific locations (dorsal margin of the scapula; ventral angle of the scapula; spine of the scapula midway between its most dorsal and ventral points [cluster of 3 markers in the form of a T {T cluster}]; lateral midshaft region of the humerus [T cluster]; lateral and medial epicondyles of the humerus; cranial and caudal aspects of the antebrachium midway between the elbow and carpal joints; lateral, medial, and palmar aspects of the carpal joint; lateral, medial, and palmar aspects of the metacarpophalangeal joint; and dorsal aspect of the third phalanx. The locations for the 2 markers [medial condyle of the humerus and medial aspect of the metacarpophalangeal joint] that cannot be seen in the photograph are indicated [green arrows]) to approximate the proximal and distal centers of rotation (ie, joint centers) for each limb segment. Then, the dog was held stationary in a standing position and walked and trotted on the treadmill in front of the fluoroscope, and the positions of the anatomic markers were adjusted as necessary to match joint position and bone orientation during the stance phase. After the markers were confirmed to be in the correct positions, the hair at each marker site was shaved to ensure that the markers were replaced in the correct positions for subsequent data collection.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
Representative photographs of the treadmill and biplanar x-ray fluoroscope system used to confirm correct positioning of reflective anatomic markers (A) and the left forelimb of a healthy adult Beagle (B) that depicts the locations at which 19 anatomic markers were placed on each of 5 similar dogs to collect 3-D kinematic and kinetic data for inverse dynamics calculations. For each dog, the markers were tentatively placed at specific locations (dorsal margin of the scapula; ventral angle of the scapula; spine of the scapula midway between its most dorsal and ventral points [cluster of 3 markers in the form of a T {T cluster}]; lateral midshaft region of the humerus [T cluster]; lateral and medial epicondyles of the humerus; cranial and caudal aspects of the antebrachium midway between the elbow and carpal joints; lateral, medial, and palmar aspects of the carpal joint; lateral, medial, and palmar aspects of the metacarpophalangeal joint; and dorsal aspect of the third phalanx. The locations for the 2 markers [medial condyle of the humerus and medial aspect of the metacarpophalangeal joint] that cannot be seen in the photograph are indicated [green arrows]) to approximate the proximal and distal centers of rotation (ie, joint centers) for each limb segment. Then, the dog was held stationary in a standing position and walked and trotted on the treadmill in front of the fluoroscope, and the positions of the anatomic markers were adjusted as necessary to match joint position and bone orientation during the stance phase. After the markers were confirmed to be in the correct positions, the hair at each marker site was shaved to ensure that the markers were replaced in the correct positions for subsequent data collection.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
Fluoroscopic evaluation
Each dog was held stationary in a standing position, walked (velocity, 1 m/s), and trotted (velocity, 2 m/s) on a treadmill in front of a biplanar x-ray fluoroscopea to analyze the position of the joint centers of rotation for the scapula and humerus relative to the anatomic markers and for SR (performed for only 1 dog) of the scapula and the humerus. Lateral and ventral fluoroscopic recordings were obtained during each phase (standing, walk, and trot). The fluoroscope settings used were 90 kVp and 70 mAs, with a sampling frequency of 250 Hz.
Morphometry
A CT scan of the left forelimb was performed for one of the study dogs to obtain the morphometric data needed for the inverse dynamics calculations. A virtual reconstruction of the bones and segments (bones and tissues) of the forelimb was created. First, the bones and segments of the forelimb were segmented by use of the segmentation editor of a commercially available 3-D visualization and analysis software package.b The objects obtained were converted to solids with freeform modeling softwarec and exported to a computer-aided design program.d The resulting virtual reconstruction was used to obtain the mass, position of the CoM, and inertia tensor around the principal axes for each segment by linking the geometry of each segment to a material density of 1.1 g/cm3 (Table 1).
Morphometric data obtained for the left forelimb of a 13.8-kg healthy adult sexually intact male Beagle, which were subsequently used for inverse dynamics calculations.
| Segment | ||||
|---|---|---|---|---|
| Variable | Scapula | Humerus | Ulna | Carpus |
| Mass (kg) | 0.272 | 0.360 | 0.118 | 0.041 |
| Length (m) | 0.12 | 0.09 | 0.13 | 0.06 |
| Position of CoM (% of segment length) | ||||
| From proximal end | 0.54 | 0.50 | 0.51 | 0.50 |
| Medial-lateral* | 0.01 more medial | 0.00 | 0.03 more medial | 0.05 more lateral |
| Cranial-caudal (palmar-dorsal)* | 0.035 more caudal | 0.00 | 0.03 more caudal | 0.01 more palmar |
| Moment of inertia around CoM (× 10−3 km•m2) | ||||
| Longitudinal axis | 0.290 | 0.254 | 0.025 | 0.005 |
| Mediolateral axis | 0.342 | 0.435 | 0.108 | 0.014 |
| Caudocranial (dorsopalmar) axis | 0.126 | 0.329 | 0.115 | 0.013 |
Measurements were obtained from a virtual reconstruction of the forelimb constructed from a CT scan. The mass, position of the CoM, and inertia tensor around the principal axes for each segment were determined by linking the geometry of each segment to a material density of 1.1 g/cm3.
Deviation from the line between the joints on either side of the segment of interest.
SR
For the dog that underwent a CT scan of the left forelimb for morphometric analysis, x-ray reconstruction of moving morphology was performed by use of noninvasive SR as described28 to estimate the kinematics of the scapula and humerus. Ten strides were analyzed at both a walk and trot. Prior to SR, the unedited x-ray video recordings were corrected for distortion by use of a freeware program,e and the orientation of the x-ray image intensifiers relative to the subject was determined by recording a calibration object (0.2 × 0.12 × 0.12 m) with metal beads inserted at regular intervals within the biplanar field of view.28 For SR, an articulated 3-D model of the subject's skeleton (which was derived from the CT virtual reconstruction) was manually posed with a commercial 3-D animation packagef so that it matched the subject's x-ray shadow in both biplanar views. This process was repeated for key frames of the x-ray video recordings with cubic spline interpolation to produce smooth movements that closely approximated the recorded kinematics. During SR, bone models are linked to form a hierarchical chain. Anatomic coordinate systems are implemented at each joint to measure the movement of the distally adjacent bone relative to the proximal bone directly from the animation.
3D kinematic and kinetic data collection
For each dog, 3-D kinematic and 3-D kinetic data for the left forelimb were obtained simultaneously in a motion laboratory. Six infrared camerasg were strategically placed around an instrumented quad-band treadmill.h Kinematic data were collected at 100 Hz, whereas GRFs were collected at 1,000 Hz. Each dog was instrumented with anatomic markers as previously described and then walked or trotted across the treadmill. Data collection began as soon as the dog was walking or trotting smoothly and comfortably. Data were recorded for a maximum of 45 seconds. For computation, a series of at least 5 cycles (strides) were evaluated in which the dog moved steadily over the treadmill without overstepping the force plates. Data were collected in 3 dimensions (x, y and z), where x represented motion in the mediolateral plane with positive values to the left, y represented motion in the craniocaudal plane with positive values opposite to the direction of motion, and z represented motion in the longitudinal, or vertical, plane with positive values in the up direction (Figure 2).
Representative illustrations of segmental motion for the canine humerus in body frame. The motion of the humerus was determined and is depicted in 3 dimensions (mediolateral [x-axis; red], craniocaudal [y-axis; green], and longitudinal [z axis; dark blue]). The dotted arrows represent motion outside of the 2-D plane of the paper. The coordinate system is fixed at the proximal end of the segment (humerus) and rotates with the segment. Rotation around the mediolateral axis represents protraction-retraction movement or torque; protraction (pro) is characterized by a negative slope or value (light blue arrow) and retraction (re) is characterized by a positive slope (orange arrow). Rotation around the craniocaudal axis represents abduction-adduction movement or torque; abduction (ab) is characterized by a negative slope and adduction (ad) is characterized by a positive slope. Rotation around the longitudinal axis represents internal-external movement or torque; internal rotation (int) is characterized by a negative slope and external rotation (ext) is characterized by a positive slope.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
Representative illustrations of segmental motion for the canine humerus in body frame. The motion of the humerus was determined and is depicted in 3 dimensions (mediolateral [x-axis; red], craniocaudal [y-axis; green], and longitudinal [z axis; dark blue]). The dotted arrows represent motion outside of the 2-D plane of the paper. The coordinate system is fixed at the proximal end of the segment (humerus) and rotates with the segment. Rotation around the mediolateral axis represents protraction-retraction movement or torque; protraction (pro) is characterized by a negative slope or value (light blue arrow) and retraction (re) is characterized by a positive slope (orange arrow). Rotation around the craniocaudal axis represents abduction-adduction movement or torque; abduction (ab) is characterized by a negative slope and adduction (ad) is characterized by a positive slope. Rotation around the longitudinal axis represents internal-external movement or torque; internal rotation (int) is characterized by a negative slope and external rotation (ext) is characterized by a positive slope.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
Representative illustrations of segmental motion for the canine humerus in body frame. The motion of the humerus was determined and is depicted in 3 dimensions (mediolateral [x-axis; red], craniocaudal [y-axis; green], and longitudinal [z axis; dark blue]). The dotted arrows represent motion outside of the 2-D plane of the paper. The coordinate system is fixed at the proximal end of the segment (humerus) and rotates with the segment. Rotation around the mediolateral axis represents protraction-retraction movement or torque; protraction (pro) is characterized by a negative slope or value (light blue arrow) and retraction (re) is characterized by a positive slope (orange arrow). Rotation around the craniocaudal axis represents abduction-adduction movement or torque; abduction (ab) is characterized by a negative slope and adduction (ad) is characterized by a positive slope. Rotation around the longitudinal axis represents internal-external movement or torque; internal rotation (int) is characterized by a negative slope and external rotation (ext) is characterized by a positive slope.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
Inverse dynamics
Synchronized 3-D kinematics, kinetics, and segmental properties were processed and combined in a customized program.i The 3-D coordinates of marker trajectories were smoothed by use of a Butterworth fourth-order low-pass filter with a cutoff frequency of 6 Hz. To obtain 3-D angular kinematics, a Cardan sequence of 3 rotations around the x, y and z axes was used as described.29 The transformation matrix for the Cardan sequence was as follows:
The angles θ1, θ2 θ3 were obtained from the trigonometric relations present in that matrix. Those angles and their time derivatives were expressed in global coordinates. For the purpose of this study, we were interested in segmental (body-frame) kinematics and dynamics. To transform velocities expressed in laboratory frame into body frame, the following transformation matrix was used:
where the angular accelerations in body frame (αx, αy, and αz) are simply the time derivatives of the respective velocities (ωx, ωy, and ωz). Then, the force (resp. torque) around the proximal end of each segment was estimated by use of the Newton-Euler equations of motion in the CoM as follows:
where m is the mass of the segment, I is the principal moments of inertia of the segment around the CoM, R is the reaction force, M is the torque in the proximal (p) and distal (d) ends of the segment, and l is the lever arm of the forces relative to the CoM. To compute the power in a joint, the velocity in the laboratory frame of the proximally adjacent segment (i + 1) was transformed into the body frame of the distally adjacent segment (i) as described previously. For example, the power (P) in a joint for the x-axis was calculated as follows: 
. Note that the superscript indicates the frame in which data were evaluated.
Results
Dogs
The mean ± SD body weight of the 5 dogs was 15.9 ± 2.3 kg. During the walk, the mean ± SD speed, number of steps, and step frequency were 0.98 ± 0.04 m/s, 17.8 ± 5.7 steps, and 1.62 ± 0.07 Hz, respectively. During the trot, the mean ± SD speed, number of steps, and step frequency were 2.2 ± 0.22 m/s, 28 ± 4.3 steps, and 2.55 ± 0.11 Hz, respectively. During fluoroscopic data collection for the dog that was used for SR, the speed, number of steps, and step frequency were 1 m/s, 10 steps, and 1.61 Hz, respectively, during the walk and 2 m/s, 10 steps, and 2.52 Hz, respectively during the trot.
Protraction-retraction dynamics
Despite efforts to record kinematic data with the anatomic markers as close as possible to real bone motion, there was a substantial difference between fluoroscopy-based and marker-based measurements for scapular and shoulder joint motion. Therefore, for the scapula and humerus, the first value reported is the marker-based measurement and the value presented in parentheses is the fluoroscopy-based measurement.
The angle, torque, and power for each forelimb bone segment (carpus, ulna, humerus, and scapula) in the mediolateral (ie, protraction-retraction motion; x axis) axis during a walk and trot were summarized (Figure 3) for the study population. At TD, the scapula was in a protracted position of −32° (−41°) during a walk and −32° (−39°) during a trot. Then, the scapula was retracted throughout 80% of the stance phase achieving a vertical position of −7° (2°), which was maintained until just before TO, when it began to protract in preparation for the swing phase. At TO, the scapula was still in a slightly protracted position (range, −5° to −10° for both the walk and trot), and continued to protract during most of the swing phase until just before TD when it began to retract.
Mean ± SD curves for segment angle, torque, and power for each segment of the left forelimb around the mediolateral (x) axis throughout a stride cycle at a walk (black) and trot (red) for 5 healthy adult Beagles. In the graphs, the solid black and red lines represent the means and the gray and red shaded areas represent the SD for the means at a walk and trot, respectively. For the segment angle graphs for the humerus and scapula, the solid and dashed cyan lines depict SR data determined for 1 of the 5 dogs and represent the mean of 10 steps for that dog at a walk and trot, respectively. Each segment angle, torque, and power was computed at the proximal end of the given segment and presented in a body coordinate system (body frame). For all graphs, values of 0 and 100 indicate TD and solid vertical lines indicate TO. For the segment angle graphs, 0° indicates that the segment is parallel to the vertical axis, negative values indicate protraction, and positive values indicate retraction. All torque and power values were normalized to body weight. For the torque graphs, positive values indicate net retractor torques and negative values indicate net protractor torques. A retractor torque flexes the shoulder joint, extends the elbow joint, and flexes the carpal joint. For the power graphs, positive values indicate energy generation and negative values indicate energy absorption. The color-coded illustrations at the bottom indicate torque and power at (from right to left) TD, 50% of the stance phase, TO, and 50% of the swing phase during a walk and trot. The arrows indicate the direction of torque for each segment of the limb. The color of the arrows indicates power; green represents energy generation, red represents energy absorption, and gray represents power values close to 0. The proportion of each arrow that is colored red or green represents the amount of torque at that segment (an arrow with just the tip colored red or green indicates torque values from 0 to < 0.15 Nm/kg, an arrow that is approximately 25% red or green indicates torque values from 0.15 to < 0.30 Nm/kg, and an arrow that is approximately 50% red or green indicates torque values ≥ 0.30 Nm/kg); torque values near 0 are not shown.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
Mean ± SD curves for segment angle, torque, and power for each segment of the left forelimb around the mediolateral (x) axis throughout a stride cycle at a walk (black) and trot (red) for 5 healthy adult Beagles. In the graphs, the solid black and red lines represent the means and the gray and red shaded areas represent the SD for the means at a walk and trot, respectively. For the segment angle graphs for the humerus and scapula, the solid and dashed cyan lines depict SR data determined for 1 of the 5 dogs and represent the mean of 10 steps for that dog at a walk and trot, respectively. Each segment angle, torque, and power was computed at the proximal end of the given segment and presented in a body coordinate system (body frame). For all graphs, values of 0 and 100 indicate TD and solid vertical lines indicate TO. For the segment angle graphs, 0° indicates that the segment is parallel to the vertical axis, negative values indicate protraction, and positive values indicate retraction. All torque and power values were normalized to body weight. For the torque graphs, positive values indicate net retractor torques and negative values indicate net protractor torques. A retractor torque flexes the shoulder joint, extends the elbow joint, and flexes the carpal joint. For the power graphs, positive values indicate energy generation and negative values indicate energy absorption. The color-coded illustrations at the bottom indicate torque and power at (from right to left) TD, 50% of the stance phase, TO, and 50% of the swing phase during a walk and trot. The arrows indicate the direction of torque for each segment of the limb. The color of the arrows indicates power; green represents energy generation, red represents energy absorption, and gray represents power values close to 0. The proportion of each arrow that is colored red or green represents the amount of torque at that segment (an arrow with just the tip colored red or green indicates torque values from 0 to < 0.15 Nm/kg, an arrow that is approximately 25% red or green indicates torque values from 0.15 to < 0.30 Nm/kg, and an arrow that is approximately 50% red or green indicates torque values ≥ 0.30 Nm/kg); torque values near 0 are not shown.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
Mean ± SD curves for segment angle, torque, and power for each segment of the left forelimb around the mediolateral (x) axis throughout a stride cycle at a walk (black) and trot (red) for 5 healthy adult Beagles. In the graphs, the solid black and red lines represent the means and the gray and red shaded areas represent the SD for the means at a walk and trot, respectively. For the segment angle graphs for the humerus and scapula, the solid and dashed cyan lines depict SR data determined for 1 of the 5 dogs and represent the mean of 10 steps for that dog at a walk and trot, respectively. Each segment angle, torque, and power was computed at the proximal end of the given segment and presented in a body coordinate system (body frame). For all graphs, values of 0 and 100 indicate TD and solid vertical lines indicate TO. For the segment angle graphs, 0° indicates that the segment is parallel to the vertical axis, negative values indicate protraction, and positive values indicate retraction. All torque and power values were normalized to body weight. For the torque graphs, positive values indicate net retractor torques and negative values indicate net protractor torques. A retractor torque flexes the shoulder joint, extends the elbow joint, and flexes the carpal joint. For the power graphs, positive values indicate energy generation and negative values indicate energy absorption. The color-coded illustrations at the bottom indicate torque and power at (from right to left) TD, 50% of the stance phase, TO, and 50% of the swing phase during a walk and trot. The arrows indicate the direction of torque for each segment of the limb. The color of the arrows indicates power; green represents energy generation, red represents energy absorption, and gray represents power values close to 0. The proportion of each arrow that is colored red or green represents the amount of torque at that segment (an arrow with just the tip colored red or green indicates torque values from 0 to < 0.15 Nm/kg, an arrow that is approximately 25% red or green indicates torque values from 0.15 to < 0.30 Nm/kg, and an arrow that is approximately 50% red or green indicates torque values ≥ 0.30 Nm/kg); torque values near 0 are not shown.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
At the beginning of the stance phase (TD), there was a net retraction torque acting on the scapular fulcrum, followed by a protraction torque, which continued for approximately the first two-thirds of the stance phase (Figure 3). The maximum mean protractive torque for the scapular fulcrum during the stance phase was 2.3 times that when dogs were trotting, compared with when dogs were walking. The scapular fulcrum generated energy at the beginning and end of the stance phase but absorbed energy during most of the stance phase, which indicated lengthening or eccentric contraction of the protraction muscles. The maximum mean energy absorption for the scapular fulcrum during a trot was 4 times that during a walk. During the first third of the swing phase, the scapular fulcrum had a net protraction torque and positive net power, which indicated concentric contraction of the protraction muscles. During the last two-thirds of the swing phase, the scapular fulcrum entered into retraction torque, and energy was absorbed to slow protraction and initiate retraction of the scapula prior to TD.
The humerus was in a retracted position (mean segment angle, 18° [29°] during a walk and 23° [28°] during trot) at TD. It continued to be retracted during the stance phase, achieving a mean peak segment angle of 63° (77°) during a walk and 59° (73°) during a trot immediately before TO. The humerus began to protract at or shortly before TO during a walk and trot, respectively, and continued to protract throughout the swing phase until just before TD when it began to retract (Figure 3). Initiation of humerus retraction was accompanied by a net retractor torque that rapidly changed to a protractor torque during the early portion of the stance phase despite continued retraction of the humerus. Retractor torque at the proximal end of the humerus tends to flex the shoulder joint. Mean peak protractor torque for the humerus during a trot was 2.5 times that during a walk. During the walk, the humerus generated energy only at the beginning of the stance phase and absorbed energy for the rest of the stance phase, and the negative power values indicated that the protractor (extensor) muscles were working eccentrically. During the first part of the swing phase of the walk, the humerus had a net protractor torque and was in a state of energy absorption. As the swing phase continued and the humerus entered into retractor torque, energy continued to be absorbed to slow the protraction of the humerus. Just before TD, the concentric work of the flexor muscles initiated retraction of the humerus.
The ulna was in a protracted position of approximately −22° at TD. It was retracted throughout the stance phase and achieved a maximum mean segment angle of 34° during a walk and 28° during a trot. The ulna began to be protracted shortly before TO and continued to be protracted for the first 80% and 90% of the swing phase during the trot and walk, respectively (Figure 3). During the last stages of the swing phase, the ulna was retracted in preparation for TD, although it did not achieve a net retracted position until after TD. During the first half of the stance phase, the ulna underwent a net retractor torque that tended to extend the elbow joint. During the last half of the stance phase and early portion of the swing phase, the ulna underwent net protraction torque; the maximum mean protractor torque observed during the trot was twice that observed during the walk. At TD, the ulna had a net negative power, which quickly changed to a net positive power during the first half of the stance phase, indicating that the extensor muscles were working to prevent collapse of the elbow joint. During the last half of the stance phase, energy was first absorbed and then generated, and the shape of the power curve was similar for both the walk and trot. During the first part of the swing phase the ulna had a net protractor torque and net positive (albeit declining) power, which helped to flex the elbow joint. During the latter stages of the swing phase, the ulna underwent retractor torque and energy was generated to retract and prepare the bone for its antigravity work at TD.
The carpus was in a protracted position (mean segment angle, −46° for the walk and −38° for the trot) at TD. It was retracted throughout the stance phase and into the early stages of the swing phase (Figure 3). The maximum mean segment angle achieved by the carpus during retraction was 76° for the walk and 95° for the trot. Following peak retraction, the carpus was rapidly protracted for approximately 80% of the swing phase, achieving a maximum mean protraction that was maintained for the last tenth of the swing phase and first portion of the stance phase. The carpus underwent retractor torque throughout the stance phase, which resulted in palmar flexion of the carpal joint. The maximum mean retractor torque for the carpus during the stance phase for the trot was 2.2 times that for the walk. The carpus had a net negative power at TD, but energy was generated throughout most of the stance phase, which indicated eccentric contraction of the retractor muscles, and the carpus had a net positive power at TO. The maximum mean power achieved during the stance phase for the trot was 11 times that for the walk. Both the torque and power of the carpus remained close to zero throughout the swing phase during the walk and trot.
Abduction-adduction dynamics
The angle, torque, and power for each forelimb segment in the craniocaudal (ie, abduction-adduction motion; y axis) axis during a walk and trot were summarized (Figure 4). An adductor torque causes adduction or prevents abduction, and an abductor torque causes abduction or prevents adduction (Figure 2). All forelimb segments had a similar mean torque pattern for both gaits; there was a net abductor torque throughout the stance phase, whereas both torque and power remained approximately 0 during the swing phase.
Mean ± SD curves for segment angle, torque, and power for each segment of the left forelimb around the craniocaudal (y) axis throughout a stride cycle at a walk (black) and trot (red) for the dogs of Figure 3. For the segment angle graphs, negative values indicate abduction and positive values indicate adduction. For the torque graphs, negative values indicate abductor torque and positive values indicate adductor torque. See Figure 3 for remainder of key.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
Mean ± SD curves for segment angle, torque, and power for each segment of the left forelimb around the craniocaudal (y) axis throughout a stride cycle at a walk (black) and trot (red) for the dogs of Figure 3. For the segment angle graphs, negative values indicate abduction and positive values indicate adduction. For the torque graphs, negative values indicate abductor torque and positive values indicate adductor torque. See Figure 3 for remainder of key.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
Mean ± SD curves for segment angle, torque, and power for each segment of the left forelimb around the craniocaudal (y) axis throughout a stride cycle at a walk (black) and trot (red) for the dogs of Figure 3. For the segment angle graphs, negative values indicate abduction and positive values indicate adduction. For the torque graphs, negative values indicate abductor torque and positive values indicate adductor torque. See Figure 3 for remainder of key.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
The segment angle curves for the scapula were similar for both gaits; however, the scapula was more abducted during a walk than during a trot. The scapula was in an abducted position at TD and became abducted further during the stance phase. After TO, the scapula was adducted until just before TD, at which point it began to be abducted again. Maximum mean abductor torque for the scapula during a trot was 1.3 times that during a walk. During the stance phase, the pattern of the power curve for the scapula during a walk was opposite that during a trot. During a walk, energy was generated during the first half of the stance phase and absorbed during the last half, whereas during a trot, energy was absorbed during the first half of the stance phase and generated during the second half.
The humerus was in an abducted position (mean segment angle, −10° [−5°] during a walk and −10° [−15°] during a trot) at TD. It remained in an abducted position with small oscillations (generally toward further abduction) throughout the stance phase. The humerus underwent further abduction during the first half of the swing phase and adduction during the second half (Figure 4). During the stance phase, maximum mean adductor torque for the humerus at a trot was 1.4 times that at a walk. The humerus had a net negative power throughout the stance phase, which indicated eccentric contraction of the abductor muscles for the shoulder joint. The maximum mean energy absorbed by the humerus at a trot was 4 times that at a walk.
The segment angle, torque, and power curves for the ulna at a walk closely mirrored those at a trot (Figure 4). The ulna was in an adducted position (mean segment angle, 20°) at TD. It became adducted further during the first half of the stance phase and was abducted during the second half yet always maintained a net adducted position. Power for the ulna oscillated between net positive and net negative values throughout the stance phase, which indicated that the abductor muscles of the elbow joint controlled the adduction of that joint.
The kinematic patterns for the carpus were also similar between a walk and trot. The carpus was in an abducted state (mean segment angle, −7°) at TD. It was then adducted throughout the stance phase and first third of the swing phase (Figure 4). The carpus was abducted during the last two-thirds of the swing phase until just before TD. It had a net negative power throughout most of the stance phase, which indicated that the abductors of the carpal joint worked eccentrically against adduction of the carpus.
Internal-external rotation dynamics
The angle, torque, and power for each forelimb segment in the longitudinal (internal-external rotation; z axis) axis were summarized (Figure 5). For all segments, both torque and power remained close to 0 throughout the swing phase. The scapula was in an internally rotated position at TD; it continued to be internally rotated for the first part of the stance phase before undergoing external rotation. The transition to external rotation occurred approximately halfway through the stance phase at a trot and just before TO at a walk. The scapula underwent further external rotation and then internal rotation during the swing phase. The transition to internal rotation occurred approximately halfway through the swing phase at a trot and about two-thirds through the swing phase at a walk. The scapula had a net internal rotation torque during the first half of the stance phase and a net external rotation torque during the last half. The scapula had a net positive energy at TD and during the early stages of the stance phase; then, the net power values became negative and back to positive for the remainder of the stance phase. That energy pattern indicated that the external rotator muscles of the shoulder joint were working eccentrically (negative values) at first to slow the internal rotation of the scapula and then concentrically (positive values) to externally rotate the scapula during the late stages of the stance phase and early stages of the swing phase. Approximately halfway through the swing phase, the scapula began to undergo internal rotation torque and generate energy, which indicated concentric contraction of the internal rotator muscles of the shoulder joint.
Mean ± SD curves for segment angle, torque, and power for each segment of the left forelimb around the longitudinal (vertical; z) axis throughout a stride cycle at a walk (black) and trot (red) for the dogs of Figure 3. For the segment angle graphs, 0° indicates that the segment is in alignment with the laboratory frame (a global coordinate systemg), negative values indicate internal rotation, and positive values indicate external rotation. For the torque graphs, negative values indicate external rotation torques and positive values indicate internal rotation torques. See Figure 3 for remainder of key.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
Mean ± SD curves for segment angle, torque, and power for each segment of the left forelimb around the longitudinal (vertical; z) axis throughout a stride cycle at a walk (black) and trot (red) for the dogs of Figure 3. For the segment angle graphs, 0° indicates that the segment is in alignment with the laboratory frame (a global coordinate systemg), negative values indicate internal rotation, and positive values indicate external rotation. For the torque graphs, negative values indicate external rotation torques and positive values indicate internal rotation torques. See Figure 3 for remainder of key.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
Mean ± SD curves for segment angle, torque, and power for each segment of the left forelimb around the longitudinal (vertical; z) axis throughout a stride cycle at a walk (black) and trot (red) for the dogs of Figure 3. For the segment angle graphs, 0° indicates that the segment is in alignment with the laboratory frame (a global coordinate systemg), negative values indicate internal rotation, and positive values indicate external rotation. For the torque graphs, negative values indicate external rotation torques and positive values indicate internal rotation torques. See Figure 3 for remainder of key.
Citation: American Journal of Veterinary Research 78, 7; 10.2460/ajvr.78.7.804
The humerus was in an internally rotated position (mean segment angle, −28° [−34°] for the walk and −36° [−25°] for the trot) at TD. During the stance phase, the humerus underwent both external and internal rotation at a walk but underwent only external rotation during the trot (Figure 5). During the swing phase, the humerus again underwent both external and internal rotation at a walk, whereas it underwent only internal rotation at a trot. The humerus had a net external rotation torque throughout the stance phase, and that torque was similar between a walk and trot. Energy was generated during most of the stance phase, which indicated concentric contraction of the external rotator muscles of the humerus.
The kinematic patterns for the ulna and elbow joint were similar between a walk and trot. The ulna was in an internally rotated position (mean segment angle, −5°) at TD and remained fairly stationary for the first two-thirds of the stance phase (Figure 5). Then, it was externally rotated for the last third of the stance phase and first third of the swing phase before being internally rotated throughout the remainder of the swing phase. The ulna underwent a net external rotation torque throughout the stance phase, and the maximum mean external rotation torque at a trot was 1.4 times that at a walk. The ulna had a net positive power for a short period after TD. Energy was absorbed for approximately the first half of the stance phase and generated for the last half, which corresponded to external rotation followed by internal rotation of the ulna.
The carpus was in an externally rotated position (mean segment angle, 20° at a walk and 8° at a trot) at TD. It remained fairly stationary for the first two-thirds of the stance phase and then was internally rotated for the remainder of the stance phase and first third of the swing phase (Figure 5). During the swing phase, the carpus underwent alternating periods of internal and external rotation, which occurred more frequently during a trot than during a walk. However, the carpus remained in an externally rotated position throughout the entire stride (stance and swing phases) at both a walk and trot. The carpus underwent net internal rotation torque throughout most of the stance phase, and the maximum mean internal rotation torque for the carpus during a trot was 4 times that during a walk. During a walk, power values for the carpus remained near zero with slight oscillations throughout the stance phase, whereas during a trot, energy was generated (increasing power values) for the first half of the stance phase and absorbed (decreasing power values) during the last half. This indicated that, during the stance phase, the internal rotator muscles of the carpus first underwent eccentric contraction to prevent extensive external rotation of that segment then concentric contraction to rotate the carpus internally.
Discussion
To our knowledge, the present study was the first to describe the 3-D inverse dynamics of the entire canine forelimb during a walk and trot. Headrick et al30 were the first to describe 3-D inverse dynamic analysis of the canine hind limb. Three types of data are necessary for computation of inverse dynamics: limb segment morphometry, kinematics, and GRFs. In the present study, morphometric data were obtained from a virtual reconstruction of the forelimb, which was based on the CT scan of 1 of the study dogs. Segment kinematics and GRFs were obtained simultaneously for all 5 study dogs by use of a reflective marker–based (infrared), motion-capture system and an instrumented quad-band treadmill. The effect of the treadmill on collection of kinematic data is negligible; results of another study31 in which the 3-D motion data collected from dogs traveling overground and on a treadmill indicate that the kinematic data obtained by the 2 data-capture methods did not differ significantly in the sagittal plane and only small differences were observed for abduction-adduction rotation and long-axis (longitudinal) rotation. Prior to collection of kinematic data, high-speed x-ray fluoroscopy was used to optimize marker position and to obtain SR data for 1 dog.
In contrast to other studies30,32 that projected torque and power to anatomic planes, the results of the present study were reported in the coordinate system (body frame) for each segment evaluated. We were particularly interested in whether the proximal segments of the canine forelimb (shoulder joint and scapular fulcrum) move beyond the parasagittal plane during the stance and swing phases of the stride, which could not be addressed by projecting joint angles onto anatomic planes. Measurement of torque in the anatomic plane includes cross talk of at least 2 components measured in the body frame (eg, a combination of protraction-retraction motion with abduction-adduction motion). That cross talk increases as the limb segments become more crouched. The other objective of this study was to compare the 3-D inverse dynamics of the canine forelimb between a walk and trot.
Results of the present study indicated that the proximal segments (scapula and humerus) of the canine forelimb do not move only in the parasagittal plane. The scapula was internally rotated and abducted during the stance phase and externally rotated and adducted during the swing phase (ie, the distal portion of the scapula moved out of the mediolateral plane). The scapula remained more parasagittal during a trot than during a walk, and during a stride, it transitioned from internal to external rotation sooner at a walk than at a trot. Important differences in the kinematics of the humerus were observed between the walk and trot. During a trot, the protraction-retraction movement was smoother and more similar to a half-sine pulse, the abduction-adduction movement was larger during the swing phase, and the humerus was more externally rotated than during a walk. In contrast, no gait-related kinematic differences were observed for the ulna. Meanwhile, during a trot, the carpus underwent greater retraction after TO and less external rotation than during a walk.
Because the results of the present study were presented in body frame, the kinematic data cannot be directly compared with that of other studies.16,20 However, when the protraction-retraction data were converted to joint flexion-extension angles, the results of this study were consistent with the findings of those other studies.16,20 To our knowledge, the present study was the first to report the 3-D kinematics and abduction-adduction movement for the scapula in conjunction with the internal-external rotation for the shoulder, elbow, and carpal joints. In this study, the amplitude of the protraction-retraction measurements for the scapula and humerus determined by fluoroscopy differed substantially from those determined by use of the anatomic markers, a phenomenon that has been previously described.18 This indicated that the kinematics of the humerus and scapula should be interpreted with caution when data are collected exclusively by use of external anatomic markers. The extent to which the discrepancy between fluoroscopy-based and marker-based measurements affects estimation of torque and power was beyond the scope and facility capabilities of the present study and warrants further investigation. To our knowledge, a facility with the ability to measure GRFs on a split treadmill while simultaneously obtaining high-speed fluoroscopic imaging does not currently exist.
The torque and power results calculated for the left forelimb of the dogs of the present study were similar to those reported by investigators of another study20 that involved healthy dogs. However, in that other study,20 data were estimated only for the sagittal plane with dogs at a walk. The elbow joint transitioned from retraction (joint extension) to protraction (joint flexion) at approximately 60% of the stance phase in the present study, compared with at 90% of the stance phase in that other study.20 This discrepancy may be the result of methodological differences between the 2 studies or differences in the segment lengths or posture of the study populations. For example, during a stride, the changes in the flexor torque of the elbow joint of rats33 are similar to those described for the dogs of that other study,20 whereas the changes in the flexor torque of the elbow joint of horses25 during both a walk and trot are similar to those for the dogs of the present study. The maximum mean torque values reported for the dogs of that other study20 were 30% to 50% greater than those for the dogs of the present study, despite the fact that the torque values in both studies were normalized to body weight. Thus, aside from the factors already mentioned, it is possible that the observed discrepancies between the present study and other study20 were the result of mass-scaling factors or breed-related differences. The dogs of that other study20 were described as mixed-breed dogs but were not further characterized, and the effect of breed on forelimb torque requires further investigation. The power pattern for the carpus also differed between the dogs of the present study and those of that other study.20 In the present study, the carpus had a net negative power only after TO, whereas in the other study,20 the carpus had a net negative power during the last stages of the stance phase before TO. Additionally, in the present study, the carpus had a net negative power during the first third of the stance phase because of the short protraction (dorsiflexion) of the carpal joint as it becomes load bearing. This power pattern for the carpus (eg, energy absorption [negative power] during the early stages of the stance phase followed by energy generation [positive power] during the late stages of the stance phase) observed for the dogs of the present study was similar to that described for rats.33
The fundamental biomechanical differences between the walk and trot can be observed at the CoM and in the behavior of the limbs (ie, vaulting mechanics23 vs bouncing mechanics24). In the present study, those differences were reflected in torque and power values that were frequently greater at a trot than a walk, which was indicative of increased work in the joints when dogs were trotting versus walking. Although the magnitude of the torque and power oscillations varied between the walk and trot, the patterns (direction of changes) for torque and power were similar between the 2 gaits, which indicated that the sequence of limb motions and control of the joints remain fairly consistent between the walk and trot. This was consistent with results of studies34–36 published in the early 1970s, which indicate that the activity patterns of various limb muscles of quadrupeds do not differ significantly between a walk and trot and vary only slightly during a gallop. The activity pattern describes the initiation and duration of muscle activation. Possible electromechanical delays in forelimb muscle activity patterns were not evaluated for the dogs of the present study. However, differences were observed in the intensity of muscle activity between a walk and trot, which was an indirect measure of the number of activated muscle fibers. Slow locomotion can be more tiring than rapid locomotion because the inherent energy storage mechanisms of muscles can be used more efficiently during rapid locomotion than during slow locomotion. Thus, it appears gait is a function of the speed at which the limbs are rotated and their interaction with the environment. Therefore, the walk and trot are just 2 different oscillation modes of the same limb properties. Results of simple models like the bipedal spring-mass model37 and pronograde virtual pivot model,38 which reduce the action of limb muscles and soft tissues to 1 or 2 variables such as limb stiffness and damping, indicate that walking, grounded running, and a gait that is a combination of walking and grounded running are all produced by the same limb motions with the only difference being the amount of energy in the system (i.e., transitioning among those 3 gaits occurs by simply increasing or decreasing the speed).39
It has been stated that inverse dynamic analysis is inherently unable to characterize the activity of individual muscles,40 which is correct in principle, but it is possible to validate crosswise EMG data and joint torques. Although the torque estimated by inverse dynamic analysis is a collective measure of the net action of all muscles and passive structures that span a joint, when we compared the forelimb torque values of the present study with EMG results for canine forelimbs reported in other studies,6,14,34,35,41 we found that there was generally good correlation between torque and muscle activation patterns. As expected, the correlation between torque and muscle activation patterns was weaker for muscles with multiple tasks than for muscles with limited tasks. Additionally, there were discrepancies in the timing and duration of activation for certain muscles among various studies,6,14,34,35,41 and both activation patterns are discussed when necessary. Results of an EMG-mapping study in rats indicate that use of only 1 or too few electrodes may fail to record sequential activation in muscles.42,43 Comparison of EMG data with torques might identify knowledge gaps in the existing literature.
Retraction of the scapula is produced mainly by the action of the cervicis part of the serratus ventralis muscle.34,41 In the present study, energy was absorbed (negative power values) by the scapula throughout most of the swing phase, which indicated eccentric muscle contraction, and was only generated (positive power values), which indicated concentric muscle contraction, just prior to TD (Figure 3). At TD, the scapula was in a net protraction position, but during the early stages of the stance phase, it underwent a short period of active retraction followed by further retraction induced by the action of gravity and momentum. That activity counteracted the eccentric work of the teres major muscle, thoracic part of the serratus ventralis muscle, long head of the triceps brachii muscle, omotransversarius muscle, and thoracic part of the trapezius muscle. Eccentric contractions are more powerful and cheaper (require less energy) than concentric contractions. In fact, energy absorption by eccentric contractions can be up to 15 times that of concentric contractions.44 In locomotion, energy absorption is necessary for deceleration, downhill movement, jumping, and dissipation and damping of high-impact forces, which would otherwise damage soft tissue structures. The contractions of the retractor muscles of the scapula become concentric before TO and during the early stages of the swing phase.
The activation patterns for the thoracic part of the trapezius muscle reported by Tokuriki35 agreed better with the torque values for the scapula calculated during a trot for the dogs of the present study than did those reported by Carrier et al.6 Similarly, the activation patterns for the thoracic part of the serratus ventralis muscle reported by Tokuriki35 agreed better with the torque values of the present study than did those reported in another study41 by Carrier et al. The discrepancies in the activation patterns for the trapezius and serratus ventralis muscles among those studies6,35,41 might have been the result of differences in electrode positioning or cross talk. Unfortunately, Tokuriki35 did not describe the position of the electrodes.
The abductor torques for the scapular fulcrum and shoulder, elbow, and carpal joints observed in the present study can only be produced by muscles that span the proximal aspect of the scapula from medial origins such as the thoracic and cervicis divisions of the rhomboideus muscle and cervical part of the trapezius muscle. However, good agreement between the torque values calculated during the stance phase for the dogs of the present study and activation patterns for the thoracic part of the trapezius muscle,6,34 thoracic34,35 and cervicis34,41 parts of the serratus ventralis muscle, and teres major muscle suggests that those muscles may also contribute to forelimb torque.
Because the origin of the cervicis part of the serratus ventralis muscle is located cranial and medial to its insertion on the scapula, concentric contraction of that muscle retracts and internally rotates the scapula and accounts for the internal rotation torque observed for the scapula at the end of the swing phase and beginning of the stance phase. Indeed, both timing and magnitude of torque observed for the scapula in the present study matched well with the activation patterns for the cervicis part of the serratus ventralis muscle described in other studies.34,41 The external rotation torque that the scapula underwent during the last half of the stance phase and early stages of the swing phase was most likely caused by concentric contraction of the thoracic part of the serratus ventralis muscle and teres major muscle because the activation patterns for those 2 muscles34,35,41 agreed well with the torque patterns. However, the torque patterns observed during a trot agreed better with the activation patterns for the thoracic part of the serratus ventralis muscle35 than did those observed during a walk.
The supraspinatus muscle and acromial part of the deltoideus muscle work primarily to stabilize the shoulder joint. The superficial pectoral and brachiocephalicus muscles are the primary protractors of the humerus, and both of those muscles work eccentrically or isometrically against gravity, especially during a walk.6,34,35 The activation patterns for the superficial pectoral muscle reported by Carrier et al6 matched the torque patterns observed for the humerus in the present study better than those reported by Tokuriki.35 During the latter stages of the swing phase, the humerus underwent a retractor torque that slowed and then retracted the humerus just before (trot) or after (walk) TD. The primary muscles that retract the humerus are the latissimus dorsi muscle, pectoralis profundus muscle, long head of the triceps brachii muscle, and teres major muscle. The activation patterns for the latissimus dorsi and pectoralis profundus muscles reported by Carrier et al6 perfectly matched the retractor torque of the humerus during the swing phase. Thus, during the swing phase, the latissimus dorsi and pectoralis profundus muscles actively underwent eccentric contraction (ie, lengthening with negative power values), which might have slowed the forward motion and initiated retraction of the limb prior to TD. The activation and function of the latissimus dorsi and pectoralis profundus muscles in dogs appear to be similar to those for horses during a trot.45 The retraction torque patterns of the humerus observed for the dogs of the present study did not match the activation patterns reported by Tokuriki35 and Goslow et al14 as well as they did the activation patterns reported by Carrier et al.6 Following TD, the long head of the triceps brachii muscle and the teres major muscle induce a retractor torque on the shoulder joint. However, it appears that activation of the long head of the triceps brachii muscle is primarily responsible for extension of the elbow joint,14 whereas activation of the teres major muscle during the stance phase controls or minimizes joint extension.34,35
Owing to the lack of or low activity of the main protractor and retractor muscles of the humerus during the stance phase, some investigators6 assumed that the GRFs should point toward the shoulder joint, and therefore the shoulder joint moment should be close to 0. They6 postulated that the forelimbs mainly work axially (ie, as struts at the shoulder joints). The torque results of the present study do not support that postulate because the forelimb underwent a short period of low retractor (flexor) torque followed by an important protractor (extensor) torque during the stance phase, which was similar to the forelimb torque pattern described for another population of dogs20 and other quadrupeds such as horses,25,46,47 therian mammals (pika, cui, and tupaia),48 and rats.33 Thus, it appears more likely that the sagittal projection of GRFs point near or toward the shoulder joint only at the beginning of the stance phase. During the first 5% to 15% of the stance phase, weak activation of the superficial pectoral and brachiocephalicus muscles in conjunction with the action of the supraspinatus muscle stabilizes the shoulder joint. The subsequently stronger activation of the superficial pectoral muscle likely accounts for the protractor torques that were estimated during the inverse dynamics analysis of the present study. Additionally, the lack of or low activity of muscles spanning the shoulder joint during the first half of the stance phase suggests that passive structures such as tendons and ligaments have an important role in the generation of torque. It is clear that the forelimbs of dogs have a greater potential for the storage and recovery of elastic energy than do the hind limbs; all 3 of the major distal joints (elbow, carpal, and metacarpophalangeal joints) of the forelimb have the capacity for substantial storage of elastic energy, whereas in the hind limb, only the tibiotarsal joint has the capacity for substantial storage of elastic energy.49 A detailed musculoskeletal model is needed to determine the relative contributions of muscle work and elastic energy storage. The external rotation torque that the humerus underwent during the stance phase, first portion of the swing phase, and just prior to TD was produced by the concentric contraction of the scapular part of the deltoideus muscle during the first half of the stance phase (Figure 5).
Following TD, the ulna undergoes retractor torque that is produced by eccentric and then concentric contraction of all 4 heads (long, lateral, medial, and accessory) of the triceps brachii muscle. At the middle of the stance phase, the torque changes to flex the elbow joint subsequent to the action of the biceps brachii and brachialis muscles. The action of the brachialis muscle first help the biceps brachii muscle to flex the elbow joint at the end of the stance phase and then controls the joint angle during the early stages of the swing phase.34,35 When the protractor muscles are switched off, gravity and presumably passive structures help to slow and brake protraction of the ulna at no cost in terms of energy. At the end of the swing phase, the extensors of the elbow joint are activated again to brake and retract the ulna prior to TD in preparation for its anti-gravity function.14 The elbow and shoulder joints both have the capacity to store and recover elastic energy, but the balance between energy storage and recovery differs between a trot and a gallop. For the elbow joint, although there is the potential for > 96% energy recovery during a trot, < 60% of the positive work results from storage of elastic energy during a gallop.49 That energy pattern is opposite that for the shoulder joint. The shoulder joint has the potential to store and recover 49% of the elastic energy during a gallop but only 38% of elastic energy during a trot.49
We believe the biceps brachii also contributes to the external rotation of the ulna; however, when the torque and muscle activation patterns were compared, we noticed that they did not match at the beginning of the stance phase. The reason for this, we believe, is that the external rotation torque acting at the shoulder joint is transferred to the elbow joint at that point of the stride.
The flexor digitorum profundus and flexor carpi ulnaris muscles produce a net retractor torque at the carpal joint throughout the stance phase. During the first third of the stance phase, the tendon and muscle units lengthen and store energy, which is added to that subsequently generated by the eccentric contraction that first extends the carpal joint and then impulses the limb upwards and forward. This is somewhat analogous to a catapult mechanism, in that energy is stored in elastic tissue fairly slowly during limb loading and then released quickly at TO resulting in protraction of the limb.50 Additionally, the mass of a muscle can be reduced if most of its work is through elastic storage. Minimizing the mass of the distal limb segments results in a disproportionate decrease in the internal work of locomotion (ie, a decrease in the energy required to swing the limbs back and forth51). It is estimated that up to 97% of the positive work performed by the carpal and tibiotarsal joints during a trot results from the recovery of elastic energy that was stored during the first half of the stance phase.49
After TO, the extensor digitorum communis (digital extensors II, III, IV, and V) muscle and extensor carpi radialis muscle brake the retraction of the carpus and cause it to protract, and during the second half of the swing phase, the retractor muscles of the carpus are activated. Co-contraction of the flexor and extensor muscles of the carpus controls the carpal joint angle prior to TD, which may reduce the torque-generating capacity of those muscles and stiffness and stability to the joint. That co-contraction also induces compressive joint forces by pulling the carpal bones together. Compressive forces can reduce the net shear or tensile loads on the joint that contribute to injury.25,52 In the present study, we did not identify any muscles that might cause internal-external rotation of the carpus.
Acknowledgments
Supported by Biologische Heilmittel Heel GmbH.
The authors thank Prof. Dr. Ingo Nolte for making available the dogs and allowing access to his motion laboratory for this experiment; Rommy Petersohn, Alexandra Anders, Lisa Dargel, Patrick Arnold, Julia Wildau, and Ben Derwel for technical assistance; Anvar Jakupov for assistance with estimation of moments of inertia; and Prof. Heiko Wagner and Dr. Brandon Kilbourne for assistance with manuscript preparation.
ABBREVIATIONS
| CoM | Center of mass |
| EMG | Electromyography |
| GRF | Ground reaction force |
| SR | Scientific rotoscoping |
| TD | Toe down |
| TO | Toe off |
Footnotes
Neurostar, Siemens, Erlangen, Germany.
Amira 3-D visualization and analysis software, Visualization Sciences Group Inc, Burlington, Mass.
Rhinoceros, McNeel Europe, Barcelona, Spain.
Inventor, Autodesk, San Rafael, Calif.
Matlab routine for x-ray reconstruction of moving morphology, Brown University, Providence, RI.
Maya 2014, Autodesk, San Rafael, Calif.
Vicon, Oxford Metrics, Oxford, England.
Force Plate FP4060-08, Bertec Corp, Columbus, Ohio.
Matlab, Mathworks, Natick, Mass.
References
1. Muybridge E. Animal locomotion: an electrophotographic investigation of consecutive phases of animal movements. 1872–1875. Philadelphia: University of Pennsylvania, 1887.
2. Griffon DJ, McLaughlin RM Jr, Roush JK. Vertical ground reaction force redistribution during experimentally induced shoulder lameness in dogs. Vet Comp Orthop Traumatol 1994; 7: 23–26.
3. Lee DV, Bertram JE, Todhunter RJ. Acceleration and balance in trotting dogs. J Exp Biol 1999; 202: 3565–3573.
4. Bertram JE, Lee DV, Case HN, et al. Comparison of the trotting gaits of Labrador Retrievers and Greyhounds. Am J Vet Res 2000; 61: 832–838.
5. Williams SB, Wilson AM, Daynes J, et al. Functional anatomy and muscle moment arms of the thoracic limb of an elite sprinting athlete: the racing Greyhound (Canis familiaris). J Anat 2008; 213: 373–382.
6. Carrier DR, Deban SM, Fischbein T. Locomotor function of forelimb protractor and retractor muscles of dogs: evidence of strut-like behavior at the shoulder. J Exp Biol 2008; 211: 150–162.
7. Budsberg SC, Verstraete MC, Soutas-Little RW. Force plate analysis of the walking gait in healthy dogs. Am J Vet Res 1987; 48: 915–918.
8. Riggs CM, DeCamp CE, Soutas-Little RW, et al. Effects of subject velocity on force plate–measured ground reaction forces in healthy Greyhounds at the trot. Am J Vet Res 1993; 54: 1523–1526.
9. Brebner NS, Moens NM, Runciman JR. Evaluation of a treadmill with integrated force plates for kinetic gait analysis of sound and lame dogs at a trot. Vet Comp Orthop Traumatol 2006; 19: 205–212.
10. Fanchon L, Valette JP, Sanaa M, et al. The measurement of ground reaction force in dogs trotting on a treadmill: an investigation of habituation. Vet Comp Orthop Traumatol 2006; 19: 81–86.
11. Bockstahler BA, Skalicky M, Peham C, et al. Reliability of ground reaction forces measured on a treadmill system in healthy dogs. Vet J 2007; 173: 373–378.
12. Abdelhadi J, Wefstaedt P, Galindo-Zamora V, et al. Load redistribution in walking and trotting Beagles with induced forelimb lameness. Am J Vet Res 2013; 74: 34–39.
13. Arshavsky YI, Kots YM, Orlovskii G, et al. Investigation of the biomechanics of running by the dog. Biophysics (Oxf) 1965; 10: 737–746.
14. Goslow GE Jr, Seeherman HJ, Taylor CR, et al. Electrical activity and relative length changes of dog limb muscles as a function of speed and gait. J Exp Biol 1981; 94: 15–42.
15. Brown CM, Dalzell B. Dog locomotion and gait analysis. Wheat Ridge, Colo: Hoflin Publishing Ltd, 1986; 160.
16. Hottinger HA, DeCamp CE, Olivier NB, et al. Noninvasive kinematic analysis of the walk in healthy large-breed dogs. Am J Vet Res 1996; 57: 381–388.
17. Carrier DR, Gregersen CS, Silverton NA. Dynamic gearing in running dogs. J Exp Biol 1998; 201: 3185–3195.
18. Fischer MS, Lilje KE. Dogs in motion. Dortmund, Germany: VDH Service GmbH, 2011; 1–208.
19. Bresler B, Frankel J. The forces and moments in the leg during level walking. Trans Asme 1950; 72: 25–35.
20. Nielsen C, Stover SM, Schulz KS, et al. Two-dimensional link-segment model of the forelimb of dogs at a walk. Am J Vet Res 2003; 64: 609–617.
21. Burton NJ, Dobney JA, Owen MR, et al. Joint angle, moment and power compensations in dogs with fragmented medial coronoid process. Vet Comp Orthop Traumatol 2008; 21: 110–118.
22. Burton NJ, Owen MR, Kirk LS, et al. Conservative versus arthroscopic management for medial coronoid process disease in dogs: a prospective gait evaluation. Vet Surg 2011; 40: 972–980.
23. Cavagna GA, Heglund NC, Taylor CR. Mechanical work in terrestrial locomotion: two basic mechanisms for minimizing energy expenditure. Am J Physiol 1977; 233:R243–R261.
24. Blickhan R. The spring-mass model for running and hopping. J Biomech 1989; 22: 1217–1227.
25. Harrison SM, Whitton RC, King M, et al. Forelimb muscle activity during equine locomotion. J Exp Biol 2012; 215: 2980–2991.
26. Reinschmidt C, van den Bogert AJ, Nigg BM, et al. Effect of skin movement on the analysis of skeletal knee joint motion during running. J Biomech 1997; 30: 729–732.
27. Günther M, Sholukha VA, Kessler D, et al. Dealing with skin motion and wobbling masses in inverse dynamics. J Mech Med Biol 2003; 3: 309–335.
28. Brainerd EL, Baier DB, Gatesy SM, et al. X-ray reconstruction of moving morphology (XROMM): precision, accuracy and applications in comparative biomechanics research. J Exp Zool A Ecol Genet Physiol 2010; 313: 262–279.
29. Winter DA. Biomechanics and motor control of human movement. 4th ed. Hoboken, NJ: John Wiley and Sons, 2009; 1–384.
30. Headrick JF, Zhang S, Millard RP, et al. Use of an inverse dynamics method to describe the motion of the canine pelvic limb in three dimensions. Am J Vet Res 2014; 75: 544–553.
31. Torres BT, Moëns NM, Al-Nadaf S, et al. Comparison of overground and treadmill-based gaits of dogs. Am J Vet Res 2013; 74: 535–541.
32. Eng JJ, Winter DA. Kinetic analysis of the lower limbs during walking: what information can be gained from a three-dimensional model? J Biomech 1995; 28: 753–758.
33. Andrada E, Mämpel J, Schmidt A, et al. From biomechanics of rats' inclined locomotion to a climbing robot. Int J Des Nat Ecodyn 2013; 8: 191–212.
34. Tokuriki M. Electromyographic and joint-mechanical studies in quadrupedal locomotion. I. Walk. Nippon Juigaku Zasshi 1973; 35: 433–436.
35. Tokuriki M. Electromyographic and joint-mechanical studies in quadrupedal locomotion. II. Trot. Nippon Juigaku Zasshi 1973; 35: 525–533.
36. Tokuriki M. Electromyographic and joint-mechanical studies in quadrupedal locomotion. III. Gallop. Nippon Juigaku Zasshi 1974; 36: 121–132.
37. Geyer H, Seyfarth A, Blickhan R. Compliant leg behaviour explains basic dynamics of walking and running. Proc Biol Sci 2006; 273: 2861–2867.
38. Andrada E, Rode C, Sutedja Y, et al. Trunk orientation causes asymmetries in leg function in small bird terrestrial locomotion. Proc Biol Sci 2014; 281: 1797.
39. Andrada E, Haase D, Sutedja Y, et al. Mixed gaits in small avian terrestrial locomotion. Sci Rep 2015; 5: 13636.
40. Griffon DJ. Canine gait analysis: a decade of computer assisted technology. Vet J 2008; 178: 159–160.
41. Carrier DR, Deban SM, Fischbein T. Locomotor function of the pectoral girdle ‘muscular sling’ in trotting dogs. J Exp Biol 2006; 209: 2224–2237.
42. Fischer MS. Kinematics, EMG, and inverse dynamics of the therian forelimb—a synthetic approach. Zool Anz 1999; 238: 41–54.
43. Fischer MS. Locomotory organs of mammals: new mechanics and feed-back pathways but conservative central control. Zoology 2001; 103: 230–239.
44. McFadyen BJ, Winter DA. An integrated biomechanical analysis of normal stair ascent and descent. J Biomech 1988; 21: 733–744.
45. Payne RC, Veenman P, Wilson AM. The role of the extrinsic thoracic limb muscles in equine locomotion. J Anat 2005; 206: 193–204.
46. Colborne GR, Lanovaz JL, Sprigings EJ, et al. Forelimb joint moments and power during the walking stance phase of horses. Am J Vet Res 1998; 59: 609–614.
47. Clayton HM, Lanovaz JL, Schamhardt HC, et al. Net joint moments and powers in the equine forelimb during the stance phase of the trot. Equine Vet J 1998; 30: 384–389.
48. Witte H, Biltzinger J, Hackert R, et al. Torque patterns of the limbs of small therian mammals during locomotion on flat ground. J Exp Biol 2002; 205: 1339–1353.
49. Gregersen CS, Silverton NA, Carrier DR. External work and potential for elastic storage at the limb joints of running dogs. J Exp Biol 1998; 201: 3197–3210.
50. Wilson AM, Watson JC, Lichtwark GA. Biomechanics: a catapult action for rapid limb protraction. Nature 2003; 421: 35–36.
51. Hildebrand M, Hurley JP. Energy of the oscillating legs of a fast-moving cheetah, pronghorn, jackrabbit, and elephant. J Morphol 1985; 184: 23–31.
52. Ackland DC, Pandy MG. Lines of action and stabilizing potential of the shoulder musculature. J Anat 2009; 215: 184–197.
