Prospective categorisation of maternal care style in Thoroughbred broodmares

This is the authors' abstract. We don't add key points for this paper.

The quality of maternal care style (MCS) has long-lasting influences on a neonate’s subsequent stress response and behavioural phenotype. In humans and rodents, early post-natal stress due in part or wholly to poor MCS has been linked to hippocampal atrophy and poor stress adaptation later in life. Maternal care in horses centres around protective and nursing behaviours, with the foal most dependent on the mare’s input during the first two weeks of life. Maternal care has been characterised in feral horses by assessing these behaviours, although a robust, validated system to quantify MCS has not been established. Maternal bond and fearfulness have also been used to assess MCS by evaluating a mare’s acute stress responses to a brief separation from the foal and a novel object. The objective of the study was to test the hypothesis that early MCS in Thoroughbreds can be determined by categorising protective and nursing behaviours, validated against behavioural tests assessing the maternal bond and fearfulness. Thoroughbred broodmares ( n =36, 4−23 years of age) were recruited from one farm. Mare-foal interactions were assessed using ethogram and qualitative behavioural assessments (QBA) of videos taken during the foal's first two weeks of life. Behavioural and physiological measurements (salivary cortisol, heart rate variability) were taken from each mare during a three-minute mare-foal separation test and for 25/36 mares, during a five-minute novel object test. The ethogram and QBA results were significantly correlated ( r =0.411, p =0.013, Pearson’s correlation). From a combination of the two datasets, principal component analysis revealed three principal components (PC) responsible for 56.4% of the variance in behaviours (PC1 37.9%, PC2 10.7%, and PC3 7.8%). Using k-means clustering, two distinct clusters corresponding to high and low MCS phenotypes were identified, with ethogram (successful nursing, p =0.017; proximity to the foal when down, p =0.004) and QBA (indifference, p <0.001; fearfulness, p <0.001) behaviours significantly different between phenotypes. Probit regression confirmed the MCS phenotypes to be reliably predicted by the maternal bond and fearfulness tests with a receiver operating characteristic curve producing a high area under the curve (0.87). It was concluded that ethogram/QBA-based assessments of maternal care in Thoroughbred broodmares during the first two weeks after foaling provide robust and validated MCS phenotypes. This has applications for future studies on the effect of MCS on the stress response in equine neonates later in life, with a long-term goal to improve welfare and attenuate stress-related diseases by managing the early life environment in horses. • Maternal care style (MCS) was categorised with an ethogram and qualitative behavioural assessment. • A separation and novel object test were performed to quantify maternal bond to the foal and fearfulness. • MCS phenotype can be predicted through behavioural and physiological responses to behavioural tests. • This work has applications for investigating the effect of MCS on development and welfare long-term.