Assessment of Gait Data Collection Part I of III Parts: camera separation, position, and minimum number with a neurologically intact female domestic short-hair feline

STUDY PURPOSE: Reproducibility of functional outcome measures remains a significant issue in pre-clinical studies of injury and disease models that affect stepping movements, such as spinal cord injury. The goal of Part I of this study is to test if separation of infrared cameras positioned 60-80° apart provide the most accurate and consistent gait analysis measurements and to determine the minimum number of cameras needed for reliable 3D coordinate marker data collections. Part II and III will address data interpolation of coordinate marker trajectories and best methods for knee triangulation, respectively. DATA COLLECTED: There were 2–10 infrared motion capture cameras (4 T-160 narrow lens and 6 T-40S wide lens cameras; Vicon Motion Systems, Oxford, United Kingdom) positioned around a space with dimensions 6m x 6m x 2.5m. Camera positions (x, y, z) and orientations (x, y, z, real) were recorded using an electronic calibration device with 5 pairs of infrared LEDs positioned in the center of the capture volume. Static experiments were conducted with two T-40S cameras positioned at the same height as a goniometer set to 90°, simulating a knee joint. The goniometer was equipped with 6 mm reflective markers, and 70 frames were recorded at 100 Hz, which is the duration of a typical step cycle while walking. The cameras were on tripods and positioned at angles of 20°, 40°, 60°, 80°, or 100° relative to each other to determine optimal camera spacing. Dynamic experiments were conducted with 2–10 cameras and one purpose bred adult female feline conditioned with food reward to perform comfortable slow to moderate walking speeds on an overground (OG; 0.65 m/s; 7 steps) walkway and bipedal stepping on a treadmill (TM; 0.5 m/s; 1 step). Hindlimb steps started at the first frame of stance, which begins with touchdown (TD) of the hindlimb paw as it contacts the walking surface. Toe-off (TOff) began at the first frame of swing after the propulsion state at the end of stance. The last frame of a step occurred when the swing phase was concluded, which is the last frame before the TD frame of the next step. Reflective markers were used to represent bony landmarks for the iliac crest (IC), greater trochanter (GT), distal fibula ~2.5 cm from the lateral malleolus (vector marker, V), lateral malleolus (LM), calcaneus (Calc), and base of the second (2M) and fifth metatarsals (5M). The V and LM markers were placed on the lower arm of the goniometer, segment 1 representing the lower leg (tibia and fibula). The GT was placed on the upper arm, segment 2, representing the thigh (femur). For dynamic experiments, markers were placed bilaterally on the hindlimbs. Marker coordinates (x, y, z) were recorded at 100 Hz for static and dynamic experiments and Vicon Nexus software was used for gait analysis and data interpolation when data gaps occurred. When all cameras were recording the 7 steps on OG, anywhere from 2–10 cameras record markers during any frame. Less than 2 cameras caused a gap that was then interpolated using the spline, pattern, or rigid body tools in Vicon Nexus (see part II for details on gap filling). There were two 1-frame gaps that occurred during the 7 left OG steps in the GT and LM that were filled with the spline tool in Vicon Nexus. When all cameras were recording the single right hindlimb TM step, there were no gaps, but when just cameras 8 and 9 were used, there were 3 gaps (1,1, and 12 frames) filled in the LM and one 14-frame gap in the V marker. When just cameras 9 and 10 were recording, there were 7 gaps filled in the V, LM, and 5M markers, which ranged from 1–5 frames in length. All marker coordinate data was smoothed with a 4th order low-pass Butterworth filter (10 Hz cut off frequency). DATA USAGE NOTES: