Digital Video Analysis Reveals Gait Parameters that Predict Performance in the Stadium Jumping Phase of Three-Day Eventing

摘要: In international equestrian sport, visual inspections assess gait changes and subtle lameness to protect the welfare of performance horses during competition. Horses competing at the international level in three-day eventing must pass two mandatory inspections (pre-competition and post-cross country) before attempting the final phase: stadium jumping (SJ). We hypothesize that by quantifying objective gait parameters for signs of post-cross-country gait changes, success in SJ performance can be predicted. We constructed our custom gait analysis pipeline using the software package DeepLabCut (DLC) and videos of horses at the trot, compiling a sample of 194 horses across three levels (CCI-2*, CCI-3*, and CCI-4*) and five venues. We previously used a Quantitative Discriminant Analysis to assign a standardized coefficient score for gait changes between the inspections. This change in gait score, based on six gait parameters (hind limb swing range, fore limb swing range, duty factor, speed, stride length, and fetlock joint range), significantly predicted the accrual of faults in the SJ phase (P= 0.0229, T-test), where a greater score demonstrated a higher chance of fault accrual. We found that individually, speed (P= 0.0039, T-test), stride length (P= 0.0058, T-test), hindlimb swing range (P= 0.0482, T-test) and duty factor (P= 0.0039, T-test) were significantly correlated to SJ finish status. While, prior to cross-country, the gait parameter of duty factor (P= 0.0044, T-test) correlated to show jumping performance. Further work investigating competition performance will encourage the use of objective measures to protect sport horse welfare, as well as …