Evaluation of a sensor-based system of motion analysis for detection and quantification of forelimb and hind limb lameness in horses.

Abstract

To compare a sensor-based accelerometer-gyroscopic (A-G) system with a video-based motion analysis system (VMAS) technique for detection and quantification of lameness in horses. 8 adult horses. 2 horses were evaluated once, 2 had navicular disease and were evaluated before and after nerve blocks, and 4 had 2 levels of shoe-induced lameness, alternatively, in each of 4 limbs. Horses were instrumented with an accelerometer transducer on the head and pelvis, a gyroscopic transducer on the right forelimb and hind feet, and a receiver-transmitter. Signals from the A-G system were collected simultaneously with those from the VMAS for collection of head, pelvis, and right feet positions with horses trotting on a treadmill. Lameness was detected with an algorithm that quantified lameness as asymmetry of head and pelvic movements. Comparisons between the A-G and VMAS systems were made by use of correlation and agreement (kappa value) analyses. Correlation between the A-G and VMAS systems for quantification of lameness was linear and high (r2 = 0.9544 and 0.8235 for forelimb and hind limb, respectively). Quantification of hind limb lameness with the A-G system was higher than measured via VMAS. Agreement between the 2 methods for detection of lameness was excellent (kappa = 0.76) for the forelimb and good (kappa = 0.56) for the hind limb. The A-G system detected and quantified forelimb and hind limb lameness in horses trotting on the treadmill. Because the data are collected wirelessly, this system might be used to objectively evaluate lameness in the field.