I spy with my little eye: Humans follow a horse-specific visual search pattern when examining facial expressions of horses with different levels of pain

Abstract

In daily interaction with horses, humans primarily rely on facial expression as a non-verbal equine cue for emotional information. Difficulties in correctly recognizing these signals might arise due to the species-specificity of facial cues, possibly leading to diminished equine welfare and health. This study aimed to explore human visual search patterns when assessing equine facial expressions indicative of various pain levels, utilizing eye-tracking technology. Hundred and eight individuals (N = 108), classified into three groups (affinity with horses (N = 60), pet owners with no affinity with horses (N = 32), and individuals with no affinity with animals (N = 16)) participated in the study; with their eye movements recorded using eye tracking glasses they evaluated four photos of horses with different levels of pain. Error score, calculated by comparing participant scores to Gold Standard Visual Analogue Score levels and fixation metrics (number of fixations and duration of fixations) were analysed across the four photos, participant group and Areas of Interest (AOIs): eyes, ears, nostrils, and mouth. Statistical analysis utilized linear mixed models. Highlighting the critical role of the eyes as key indicators of pain, findings showed that the eyes played a significant role in assessing equine emotional states, as all groups focused on them for a longer time and more frequently compared to other facial features. Also, participants showed a consistent pattern in how they looked at a horse's face, first focusing on the eyes, then the ears, and finally the nose/mouth region, indicating a horse-specific pattern. Moderate pain was assessed with similar accuracy across all groups, indicating that these signals are broadly recognizable. Nevertheless, non-equestrians faced challenges with recognizing the absence of pain, possibly highlighting the role of experience in interpreting subtle equine expressions. The study's limitations, such as variability in assessment conditions may have impacted findings. Future work could further investigate why humans follow this visual search pattern and whether they recognize the significance of a horse's ears. Additionally, emphasis should be placed on developing targeted training interventions to improve equine pain recognition, possibly benefiting equine welfare and health.