Table 1.
Dog characteristics and the number of recorded strides.
Fig 1.
A: schematic drawing of a walking dog, with the stick diagram representing the analyzed segments and markers. B: analyzed segments with joint and segment elevation angles. C: mean speed (+SD, n = 6). D: cycle duration. E: duration of backward (relative to the body) limb excursion. For galloping (asymmetrical gait), both leading and trailing limb values were computed.
Fig 2.
Forelimb and hindlimb endpoint (EP) path.
A: stick diagram of a single stride (limb segment movements relative to hip and scapula). EPx and EPy denote horizontal and vertical endpoint excursions, respectively. B and C: mean values (+SD, n = 6) of EPx and EPy excursions for the hindlimb and forelimb.
Fig 3.
A: stick diagrams of a single cycle and limb contact patterns of one dog in walk, trot, gallop and swim. Black bars indicate the time of backward limb excursion of each leg: hindlimb on the recording camera side (HL), ipsilateral forelimb (FL), contralateral hindlimb (HL_contr) and contralateral forelimb (FL_contr). PL–phase lag between limbs determined as the relative timing of the limb cycle onset (relative to HL) expressed as a percentage of the gait cycle. B: mean PL values (+SD, n = 6) for all limbs and gaits.
Fig 4.
A: ensemble-averaged waveforms (±SD, n = 6) of the limb joint (knee, ankle, elbow, wrist) and segment elevation (thigh, shank, foot, scapula, upper arm, lower arm, hand) angles in the sagittal plane (for galloping, the waveforms for the leading HL are shown). Data are plotted versus normalized gait cycle. Note a similar temporal sequence of minima of the elevation angles across gaits for each limb (schematically marked by green arrows). The dashed vertical lines in each subplot represents the onset of forward limb movement. B: range of angular motion (mean+SD) for HL and FL in different gaits.
Fig 5.
Temporal characteristics of the HL and FL elevation angles.
A: timing of minima (triangle) of the elevation angles (±SD). Y axis: t—thigh, s—shank, f–foot, u–upper arm, l–lower arm, h–hand. B: Characteristics of the first harmonic of HL and FL segment elevation angles. Upper panels—percent of variance accounted for by the first harmonic. Lower panels—phase of the first harmonics (circle) (zero refers to the cosine function). Note the phase-lead of the thigh segment in HL and of the lower arm segment in FL in all gaits (consistent with the temporal sequence of minima in the elevation angles, panel A).
Table 2.
Correlations between angular waveforms during trot, gallop and swim with those during walk.
Fig 6.
Planar covariation of segment elevation angles in different gaits.
A: covariation of limb segment elevation angles during walking, trotting, galloping (trailing limb) and swimming computed for ensemble-averaged elevation angles. The best-fitting plane is shown by grids. B: 95% confidence cones characterizing spatial distribution of the normal to the covariation plane between dogs in each gait and for each limb (left panel–HL, right panel–FL) in the 3D space defined by the elevation angles (see Methods). The foot (and hand for FL) semiaxis is positive, and the shank and thigh (lower and upper arm for FL) semiaxes are negative. Note overlapping of confidence cones in different gaits. C: direction cosines (u3t, u3s, u3s for HL and u3ua, u3la, u3h for FL) of the normal to the covariation plane (u3 vector) for all individual strides and all dogs versus locomotion speed. D: percent of variance (+SD) accounted for by u3 (PV3). Asterisks denote significant differences.
Fig 7.
Planar covariation of forelimb segment elevation angles using different tri-segmental FL models.
A: 'upper arm—lower arm–hand’ model. B: ‘scapula—upper arm—lower arm’ model. The same format as in the right panel of Fig 6A. Note roughly similar orientation of the covariation plane between these two models of FL (A and B) and different orientation compared to the HL model (Fig 6A).
Fig 8.
Basic patterns of ipsilateral EMG profiles of 23 extrinsic FL and HL muscles during canine locomotion.
A: stride-normalized (with respect to touchdown of each limb), averaged and low-pass EMG data during walk, trot and gallop of twelve mixed-breed dogs taken from [59]. In dark gray—HL, from bottom to top (1–12): m. tensor fasciae latae, m. semitendinosus, m. semimembranosus, m. sartorius cranial, m. sartorius caudal, m. rectus femoris, m. gracilis, m. gluteus superficialis, m. gluteus medius, m. biceps femoris cranial, m. biceps femoris caudal, m. adductor magnus. In light gray–FL, from bottom to top (1–11): m. trapezius p. thoracica, m. trapezius p. cervicalis, m. serratus ventralis thoracis, m. serratus ventralis cervicis, m. rhomboideus, m. pectoralis superfic. transversus, m. pectoralis profundus, m. pectoralis superf. descendens, m. omotransversarius, m. latissimus dorsi, m. clediobrachialis. B: schematic of motor modules. Simulated example of muscle activity profiles as weighted sum of basic patterns (pj): mi = ∑jpj(t)⋅wij + residual. The outputs of the first (green), second (blue), third (red) and forth (violet) modules are summed together to generate overall muscle activation (black envelope). C: basic activation patterns of HL (black) and FL (gray) muscles during walk, trot and gallop. EMG data were decomposed into basic activation patterns using non-negative matrix factorization. For gallop, the patterns for the leading (solid lines) and trailing (dotted lines) limbs were superimposed. The four basic patterns account for ∼90% of variance in each limb and gait and are each characterized by a relatively narrow peak of activation (Gaussian-like) at a particular phase of the cycle (with respect to touchdown of each limb). Components are designated in chronological order. D: the timing of the main peaks of 4 basic patterns across gaits and limbs. Zero timing for HL and FL refers to touchdown of HL and FL, respectively. E: locomotion motor program as a sequence of activation pulses [52,54]. The schematic diagram of activation pulse timings corresponds to those of walk. TD = touchdown. Note different phases of basic muscle activity in HL and FL for each gait.
Fig 9.
Pattern of inter-segmental coordination in human (A) and canine (B) gait.
Covariation of thigh, shank, and foot elevation angles is shown during walking at a natural speed. Gait loops of each species are represented with respect to the best-fitting planes (grids). Note different covariation plane orientation in human and dog locomotion, attributable to different phase relationships between limb segment oscillations. Data from human were adapted from [92].