Answers are grounded in this paper and cite their sources.
• A tail-attached device with an accelerometer and a thermistor was applied to calves. • Behavioral and physiological data were recorded in calves with varying health status. • Different traits were found among healthy, diarrheal, and respiratory diseased calves. • A generalized disease detection model was built using sensor-derived daily features. • The model showed acceptable discrimination between healthy and diseased cases. This study aimed to clarify how behavioral and physiological parameters differ among preweaning female calves with different health statuses (healthy, diarrheal, and respiratory diseases) using a multimodal tail-attached device equipped with a 3-axis accelerometer and a thermistor. We also evaluated the applicability of this device for disease detection. The device was fitted to 329 Holstein calves (mean age: 20.7 ± 10.9 days) and recorded data for 24.2 ± 6.2 days at 3-minute intervals. During this period, calves were assessed for health status twice weekly, and only data from assessment days were analyzed (1,996 days' data: 1,778 healthy, 148 diarrheal, 63 respiratory, and 7 with both diseases). Five daily metrics were calculated: mean activity intensity (meanActivity), total lying time (sumLying), number of body position changes (sumPosition), standard deviation of the tail-raising index (sdTail), and maximum tail skin temperature (maxST). A linear mixed-effects model showed that diarrheal disease was negatively associated with meanActivity and sumPosition, and positively with sumLying and sdTail. Respiratory disease was negatively associated with meanActivity and positively with sumLying and maxST. However, substantial overlap in metrics was observed between the two diseases. Therefore, they were merged into a single category, and a generalized disease detection model was developed using a random forest algorithm. Model performance, evaluated using a holdout approach, showed an acceptable level of discriminative ability between healthy and diseased cases (AUC = 0.71). These findings offer insights into health monitoring strategies and support the potential utility of the tail-attached device for automated disease detection in preweaning calves.
No takes yet. Share an insight, caveat, or question.
Furukawa et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: