Richeson John T, Lawrence Ty E, White Brad J
Department of Agricultural Sciences, West Texas A&M University, Canyon, TX.
Department of Clinical Sciences, Kansas State University, Manhattan, KS.
Transl Anim Sci. 2018 Feb 23;2(2):223-229. doi: 10.1093/tas/txy004. eCollection 2018 Apr.
For decades, we have relied upon visual observation of animal behavior to define clinical disease, assist in breeding selection, and predict growth performance. Limitations of visual monitoring of cattle behavior include training of personnel, subjectivity, and brevity. In addition, extensive time and labor is required to visually monitor behavior in large numbers of animals, and the prey instinct of cattle to disguise abnormal behaviors in the presence of a human evaluator is problematic. More recently, cattle behavior has been quantified objectively and continuously using advanced technologies to assess animal welfare, indicate lameness or disease, and detect estrus in both production and research settings. The current review will summarize three methodologies for quantification of cattle behavior with focus on U.S. beef production systems; 1) three-axis accelerometers that quantify physical behavior, 2) systems that document feeding and watering behavior via radio frequency, and 3) triangulation or global positioning systems to determine location and movement of cattle within a pen or pasture. Furthermore, advances in Wi-Fi and radio frequency technology have allowed many of these systems to operate remotely and in real-time and efforts are underway to develop commercial applications that may allow early detection of respiratory or other cattle diseases in the production environment. Current challenges with commercial application of technology for early disease detection include establishment of an appropriate algorithm to ensure maximum sensitivity and specificity, reliable and repeatable data collection in harsh environments, cost:benefit, and integration with traditional methodology for clinical diagnosis. Advanced technologies have also allowed cattle researchers to determine temporal variance in behavior or variability between experimental treatments. However, these data sets are typically very large and challenges exist regarding statistical analysis and reporting.
几十年来，我们一直依靠对动物行为的视觉观察来定义临床疾病、辅助育种选择并预测生长性能。对牛行为进行视觉监测的局限性包括人员培训、主观性和短暂性。此外，对大量动物的行为进行视觉监测需要大量时间和人力，而且牛在人类评估者面前伪装异常行为的猎物本能也存在问题。最近，利用先进技术对牛的行为进行了客观且持续的量化，以评估动物福利、指示跛行或疾病，并在生产和研究环境中检测发情情况。本综述将总结三种量化牛行为的方法，重点关注美国牛肉生产系统；1）量化身体行为的三轴加速度计，2）通过射频记录采食和饮水行为的系统，3）用于确定牛在围栏或牧场内位置和移动的三角测量或全球定位系统。此外，Wi-Fi和射频技术的进步使许多此类系统能够远程实时运行，并且正在努力开发商业应用，以便在生产环境中早期检测呼吸道或其他牛病。目前，用于早期疾病检测的技术在商业应用方面面临的挑战包括建立适当的算法以确保最大的敏感性和特异性、在恶劣环境中进行可靠且可重复的数据收集、成本效益以及与传统临床诊断方法的整合。先进技术还使牛研究人员能够确定行为的时间变化或实验处理之间的变异性。然而，这些数据集通常非常大，在统计分析和报告方面存在挑战。
