Search

For centuries humans have been fascinated by the natural beauty of horses in motion and their different gaits. Gait classification (GC) is commonly performed through visual assessment and reliable, automated methods for real-time objective GC in horses are warranted. In this study, we used a full body network of wireless, high sampling-rate sensors combined with machine learning to fully automatically classify gait. Using data from 120 horses of four different domestic breeds, equipped with seven motion sensors, we included 7576 strides from eight different gaits. GC was trained using several machine-learning approaches, both from feature-extracted data and from raw sensor data. Our best GC model achieved 97% accuracy. Our technique facilitated accurate, GC that enables in-depth biomechanical studies and allows for highly accurate phenotyping of gait for genetic research and breeding. Our approach lends itself for potential use in other quadrupedal species without the need for developing gait/animal specific algorithms.

BACKGROUND: In Warmblood horses, degenerative joint disease is involved in cervical malformation and malarticulation (CVM). The degree of contribution of articular process joint (APJ) osteochondrosis (OC) is not clear.OBJECTIVES: 1) To explore presence of predilection sites for APJ OC in cervical and cranial thoracic vertebral columns of Warmblood foals and 2) to examine correlation of such a site with the predilection site of CVM.STUDY DESIGN: Case series.METHODS: Seven hundred APJ facets of C2 to T2 of 29 foals (11 months gestation to 12 months [median age 7 days; range 365 days; 95% confidence interval [95% CI] 2- 47 days]) were examined for OC and prevalence between joints, and the predilection site for CVM and the cranial cervical vertebral column were evaluated.RESULTS: 20.6% of facets revealed OC. There was no predilection site. Prevalence decreased with age up to 1 year (odds ratio (OR) 0.997; (95% CI 0.975-0.998) but not up to 5 months. Severity increased with age in all age ranges (up to 1 year OR 1.023; 95% CI 1.005-1.049; > 1 to 5 months, OR 1.203; 95% CI 1.014e+00-1.921; up to 1 month, OR 1.114; 95% CI 1.041-1.228). Highest prevalence was in cranial facets of the cervical and cervical thoracic joints and in caudal facets of the thoracic joint up to 1 year and up to 1 month (OR 0.364; 95% CI 0.170-0.745, OR 0.434; 95% CI:0.235 -0.782, OR 7.665; 95% CI: 1.615-66.553 and OR 0.400; 95% CI 0.170-0.880, OR 0.351; 95% CI 0.172-0.700, OR 5.317; 95% CI 1.098-44.344 respectively).MAIN LIMITATIONS: Two thirds of the foals were less than one month of age.CONCLUSIONS: Articular process joint OC in Warmblood foals is common and is not more prevalent at CVM predilection sites, suggesting that abnormalities of enchondral ossification may not be major contributors to CVM.

BACKGROUND: Variation in equine caudal cervical spine morphology at C6 and C7 has high prevalence in Warmblood horses and is suspected to be associated with pain in a large mixed-breed group of horses. At present no data exist on the relationship between radiographic phenotype and clinical presentation in Warmblood horses in a case-control study. OBJECTIVES: To establish the frequency of radiographically visible morphologic variation in a large group of Warmblood horses with clinical signs and compare this with a group without clinical signs. We hypothesised that occurrence of morphologic variation in the case group would not differ from the control group, indicating there is no association between clinical signs and morphologic variation. STUDY DESIGN: Retrospective case-control. METHODS: Radiographic presence or absence of morphologic variation of cervical vertebrae C6 and C7 was recorded in case (n = 245) and control horses (n = 132). Case and control groups were compared by univariable Pearson's Chi-square and multivariable logistic regression for measurement variables age, sex, breed, degenerative joint disease and morphologic variation at C6 and C7. Odds ratio and confidence intervals were obtained. A P-level of ≤0.05 was considered statistically significant. RESULTS: Morphologic variation at C6 and C7 (n = 108/377 = 28.6%; Cases 58/245 = 23.7%; Control 50/132 = 38%) was less frequent in horses with clinical signs in univariable testing (OR 0.48, 95% CI 0.3-0.8, P = 0.001). Age, sex, breed and degenerative joint disease were not retained in the final multivariable logistic regression step whereas morphologic variation remained significantly less present in horses with clinical signs. MAIN LIMITATIONS: Possible demographic differences between equine clinics. CONCLUSIONS: Morphologic variation in the caudal cervical spine was detected more frequently in horses without clinical signs. Therefore, radiographic p

For centuries humans have been fascinated by the natural beauty of horses in motion and their different gaits. Gait classification (GC) is commonly performed through visual assessment and reliable, automated methods for real-time objective GC in horses are warranted. In this study, we used a full body network of wireless, high sampling-rate sensors combined with machine learning to fully automatically classify gait. Using data from 120 horses of four different domestic breeds, equipped with seven motion sensors, we included 7576 strides from eight different gaits. GC was trained using several machine-learning approaches, both from feature-extracted data and from raw sensor data. Our best GC model achieved 97% accuracy. Our technique facilitated accurate, GC that enables in-depth biomechanical studies and allows for highly accurate phenotyping of gait for genetic research and breeding. Our approach lends itself for potential use in other quadrupedal species without the need for developing gait/animal specific algorithms., QC 20201127Correction: DOI:10.1038/s41598-021-88880-7