The objective of this study was to investigate 3D-kinematics as a method for determining if primiparous dairy cows display differences in behaviour biometrics during the estrus period as ovulation approaches in a tie-stall system. The second objective was to evaluate the accuracy of behaviour biometrics as estrus alerts. Fourteen primiparous dairy cows (n = 14) were studied as part of a split-plot over time design. 3D-kinematic assessment took place during proestrus-estrus-ovulation period, Follicular diameter, corpus luteum (CL) diameter and estradiol (E₂) concentrations served as physiological parameters to indirectly estimate the estrus period (d -1). The frequency of Pelvic tilts, Pelvic shifts left (Pelvicsl), Pelvic shifts right (Pelvicsr), Total pelvic shifts; TPelvicS), Foot strikes left (Foot strike L), Foot strikes right (Foot strike R), and Total feet strikes (TFootS) were recorded. Additionally, the frequency of Micro tail left (TailLMicro), and right (TailRMicro) movements, Middle tail left (TailLMid), and right (TailRMid) movements, Macro tail left (TailLMacro), right (TailRMacro) movements were also recorded. The overall length of estrous cycle in primiparous dairy cows in this study was 21.66 ± 3.09 (LSMeans ± SEM days). The largest Follicular diameter (LSMeans ± SEM; 17.04 ± 0.59 mm) and E₂ (17.43 ± 1.76 pg/mL) occurred 24 h before ovulation. The frequency of some behaviour biometrics increased (LSMeans ± SEM Events/5min) at d -2 including Pelvic tilts (19.75 ± 8.67 Events/5min), Pelvicsl (20.26 ± 13.64), and TPelvicS (20.82 ± 8.79) compared to baseline (d -4, Pelvic tilt; 13.10 ± 8.32, Pelvicsl; 3.71 ± 2.52, and TPelvicS; 6.34 ± 2.72). Other significant patterns observed include a decrease at d -1 in the frequency of TFootS (9.86 ± 1.98), TTailMicro (7.30 ± 3.62), TTailMid (1.82 ± 1.01), and TTailMacro (1.66 ± 1.01) movements followed by an increase in frequency at d -4 (TFootS; 14.44 ± 2.78, TTailMicro; 14.57 ± 7.23, TTailMid; 6.07 ± 3.27, and TTailMacro; 1.84 ± 1.12). The accuracy of each behaviour biometric as a potential estrus alert was analyzed using J index values with balance sensitivity – specificity (J index, Se–Sp) levels (ROC curves analysis). Feet strikes had the greatest accuracy (0.50; 0.90-0.6) followed by Pelvic tilts (0.37; 0.78-0.59), Foot strikes L (0.33; 0.44-0.89), TailLMid (0.30; 1.00–0.30), TailLMacro (0.41; 1.0–0.41) and TFootS (0.34; 0.67-0.68). Our results indicate that naturally cycling, primiparous dairy cows housed in tie-stalls exhibit behavioural fluctuations such as the estrus period approaches, which can be used as estrus alerts.

这项研究的目的是研究3D运动学，作为一种确定初发奶牛是否在发情期系统中排卵时在动情期表现出行为生物学特征差异的方法。第二个目标是评估作为发情警报的行为生物识别的准确性。作为时间划分设计的一部分，研究了14头初产奶牛（n = 14）。在发情-发情-排卵期进行3D运动学评估，卵泡直径，黄体直径（CL）和雌二醇（E ₂）浓度作为生理参数间接估算发情期（d -1）。骨盆倾斜的频率，骨盆向左移动（Pelvicsl），骨盆向右移动（Pelvicsr），总骨盆移位；TPelvicS），左脚踩动（F脚踩L），右脚踩动（F脚踩R）和总脚踩动感（TFootS）被记录。此外，还记录了Micro尾部向左（TailLMicro）和向右（TailRMicro）移动，中尾向左（TailLMid）和向右（TailRMid）移动，宏尾向左（TailLMacro），向右（TailRMacro）移动的频率。在这项研究中，初产奶牛的发情周期总长为21.66±3.09（LSMeans±SEM天）。最大卵泡直径（LSMeans±SEM ; 排卵前24小时出现17.04±0.59 mm）和E ₂（17.43±1.76 pg / mL）。一些行为生物识别的频率增加（LSMeans±SEM事件/ 5min）在d -2时，与基线（d -4，骨盆倾斜; 13.10±8.32，Pelvicsl; 3.71± 2.52和TPelvicS； 6.34±2.72）。观察到的其他显着模式包括TFootS（9.86±1.98），TTailMicro（7.30±3.62），TTailMid（1.82±1.01）和TTailMacro（1.66±1.01）的运动在d -1处降低，随后是频率增加在d -4（TFootS; 14.44±2.78，TTailMicro; 14.57±7.23，TTailMid; 6.07±3.27，和TTailMacro; 1.84±1.12）。使用具有平衡敏感性-特异性（J指数，Se–Sp的J指数）的值来分析每个行为生物特征作为潜在发情警报的准确性。）级别（ROC曲线分析）。脚底打击的准确性最高（0.50; 0.90-0.6），其次是骨盆倾斜（0.37; 0.78-0.59），脚底打击L（0.33; 0.44-0.89），TailLMid（0.30; 1.00-0.30），TailLMacro（0.41; 1.0） –0.41）和TFootS（0.34； 0.67-0.68）。我们的研究结果表明，饲养在领带架中的自然循环初生奶牛表现出行为波动，例如发情期临近，可以用作发情预警。