ATTENTION BIAS: A NEW TOOL FOR WELFARE ASSESSMENT IN CAPTIVE RHESUS MACAQUES

Abstract

Attention bias (AB) describes a tendency to differentially allocate attention towards one of two or more emotional stimuli. In humans these biases reliably map onto physiological and self-reported measures of affect. AB tasks have been shown to detect shifts in emotional state and have been proposed as a novel method of animal welfare assessment. This PhD aimed to determine which factors might influence AB in rhesus macaques (Macaca mulatta) through triangulation of these cognitive data with behavioural observations, physiological measures (salivary cortisol) and key genetic polymorphisms related to oxytocin, serotonin, dopamine, and cortisol. Key factors of interest were condition (baseline and post-stressor) sex, age, and time of day. AB trials were conducted with 61 (45 female, 16 male) adult rhesus macaques (Macaca mulatta) using an automated computer operated apparatus with threat-neutral conspecific face stimuli presented on screens. Duration of looking at these stimuli was recorded. Two looking time measures were used throughout the thesis: duration looking at the threat face stimulus (THR), total duration looking at the threat and neutral face stimuli (TL). AB trials were conducted before and after the macaques’ annual veterinary health check, which is thought to be acutely stressful. A total of 640 AB trials were conducted. The main findings were the relationship between AB measures and the interaction between condition and sex. Female macaques became less attentive to social information from baseline to post-stressor, while male macaques became more attentive. Further, an association between AB measures and time of day was revealed. This thesis demonstrated that the inclusion of pedigree (relatedness) data is vital when conducting genetic analysis to avoid type I errors. Without pedigree data, six genotypes had a significant association with the AB measures; however, with pedigree data only one statistically significant association was found. The cognitive, behavioural, and physiological results suggested that the veterinary health check may be too mild a stressor for use in future AB studies. The use of a more stressful event or procedure may be more informative while the AB measure is studied and developed. This project has shown AB to be a promising tool for welfare assessment, highlights some important influencing variables and should act as a guide for further research.