Effect of milk allowance on the suitability of automated behavioural and physiological measures as early disease indicators in calves

This study investigated the effect of milk allowance on the suitability of behavioural and physiological responses, individually and in combination, as early disease indicators. A total of 120 heifer calves were assigned into a 5 L/d or 10 L/d milk allowance treatment group. Daily health checks were conducted to determine when calves presented clinical signs of neonatal calf diarrhoea (NCD). Automated calf feeders recorded milk feeding behaviour, and infrared cameras automatically recorded eye and cheek temperatures. Accelerometers recorded lying behaviour, and water drinking behaviour was recorded using an automated water system and video observations. Respiration rate was recorded once per day by counting flank movements. Calves were used as their own controls, with data analysed relative to the day of clinical identification (d 0). Based on changes which occurred over the 6 days prior to clinical signs of disease (across d -6 to 0), typically, feeding behaviours only changed significantly for calves on the 10 L/d milk allowance with an increase in total (P < 0.001) and rewarded visits to the feeder (P = 0.013), and a decrease in milk consumption (P = 0.011). Infrared temperatures only changed significantly prior to clinical signs for calves on the 5 L/d milk allowance, with a decrease in eye (P = 0.013) and cheek (P = 0.006) temperatures. Regardless of milk allowance, lying time (P = 0.028, 5- L/d; P = 0.011, 10 L/d), number of lying bouts (P < 0.001, 5 L/d; P = 0.007, 10 L/d), and average bout duration (P < 0.001, 5 L/d; P = 0.002, 10 L/d), changed significantly prior to clinical signs. The only significant change in water drinking behaviour prior to clinical signs was an increase in total trough visit duration for calves on the 5 L/d milk allowance (P = 0.029). Respiration rate showed no significant change prior to clinical signs regardless of milk allowance. For calves on the 10 L/d and 5 L/d milk allowances, the most suitable indicator of disease was the total number of visits to the feeder (P < 0.001) and the number of lying bouts (P < 0.001), respectively. Regardless of milk allowance, combinations of feeding and lying behaviours, and additionally infrared temperatures in the case of calves on the 5 L/d milk allowance, provided the strongest composite indicators of disease. Overall, results suggest milk allowance can influence the suitability of behavioural and physiological responses as early disease indicators. Milk allowance should be considered when determining which measures will act as the optimum indicator/s when developing and implementing algorithms into automated on-farm systems for disease detection.