Perception of the environment by farm animals is fundamental for expression of behaviour and of their adaptation to different environmental conditions. From a breeding-environmental perspective, perception becomes increasingly important when a production system negatively impacts on animals such that their normal expression of behaviour is compromised. Therefore, research on the perceptual abilities of farm animals is of crucial importance to understand the animalenvironment relationship. This review is focused on research related to sensory perception of farm animals. It should be stressed that the world of animal senses is very difficult to explore, we have limited knowledge of the complexity of the animal’s ability to perceive and process environmental stimuli.

As visual attention is an intrinsic part of social relationships, and because relationships are built on a succession of interactions, their establishment involves learning and attention. The emotional, rewarding or punishing, content can modulate selective attention. In horses, the use of positive/negative reinforcement during training determines short and long-term human-horse relationships. In a recent study in horses, where either food or withers’ grooming were used as a reward, it appeared that only the food-rewarded horses learned the task and show better relationship with humans. In the present study, we hypothesized that this differential effect of grooming/food rewards on learning performances could be due to attentional processes. Monitoring, gazes and behaviors directed towards the trainer revealed that the use of a food reward (FR) as positive reinforcement increased horses’ selective attention towards their trainer. Conversely, horses trained with grooming reward (GR) expressed more inattentive responses and did not show a decrease of “agitated” behavior. However, individual plotting of attention versus rate of learning performances revealed a complex pattern. Thus, while all FR horses showed a “window” of attention related to faster learning performances, GR horses’ pattern followed an almost normal curve where the extreme animals (i.e. highest and lowest attention) had the slowest learning performances. On the other hand, learning was influenced by attention: at the end of training, the more attentive horses had also better learning performances. This study, based on horses, contributes to the general debates on the place of attentional processes at the interface of emotion and cognition and open new lines of thought about individual sensitivities (only individuals can tell what an appropriate reward is), attentional processes and learning.