# Prediction of Nocturnal Foaling Using Ventral Tail Base Surface Temperature Recorded by a Wearable Device Attached to the Mare’s Tail

## TL;DR

This study developed a method to predict when a horse will give birth at night by measuring the temperature at the base of its tail using a wearable device.

## Contribution

The study is the first to propose an algorithm for predicting nocturnal foaling using ventral tail base surface temperature.

## Key findings

- The foaling prediction model achieved a sensitivity of 68.2 to 81.8% and precision of 51.4 to 62.5%.
- The algorithm uses ventral tail base surface temperature recorded by a thermistor-equipped wearable device.
- The model was validated using 147 days of data from 22 pregnant mares.

## Abstract

It is desirable for a horse manager to attend to the mare at the time of foaling in order to assist in fetal delivery and prevent complications. However, the gestation length of horses is highly variable, and most mares give birth during the night. These characteristics of equine parturition make it difficult to attend to the foaling mare. It is known that a mare’s body temperature drops before parturition, but no research has applied this thermal change to the prediction of foaling yet. In this study, the ventral tail base surface temperature was recorded by a tail-attached device equipped with a thermistor in pregnant mares kept in an outdoor paddock all day. The objective of the present study was to make an algorithm for predicting nocturnal foaling (18:00 to 6:00) and to verify its accuracy. The foaling prediction model was validated using 147 days of data recorded from 22 mares. When the threshold was set at −0.2 to −0.3 °C, sensitivity was 68.2 to 81.8% and precision was 51.4 to 62.5%. Although the accuracy might be insufficient for clinical application, to our knowledge, this present study is the first to propose an algorithm for predicting nocturnal foaling at a specific time using ventral tail base surface temperature.

It is known that a mare’s body temperature drops before parturition, but no research has yet applied this thermal change to the prediction of foaling. In this study, the ventral tail base surface temperature (VTB-ST) was recorded by a tail-attached device equipped with a thermistor in pregnant mares kept in an outdoor paddock all day. The objective of the present study was to make an algorithm for predicting nocturnal foaling (18:00 to 6:00) and to verify the accuracy of the algorithm. Prediction of nocturnal foaling was performed at 15:00 every day. The foaling prediction model was validated using 147 days of data recorded from 22 mares. The sensitivity of the foaling prediction model proposed in this study was 68.2 to 81.8% and the precision was 51.4 to 62.5%. To our knowledge, the present study is the first one to establish an algorithm for predicting nocturnal foaling at a specific time interval using VTB-ST. Further study will be necessary to improve the foaling prediction model, as the accuracy of the algorithm proposed in this study was considered to be insufficient for practical use in stud farms.