Does the anticipatory behaviour of chickens communicate reward quality?

Abstract

The anticipatory behaviour of animals has been credited with enabling scientists to more closely infer what an animal wants. From a welfare perspective, this knowledge could improve how we care for animals under our management, as information about how animals prioritise rewarding items may guide how we allocate resources effectively. Our goal was to determine if behaviour in anticipation of different types of reward was differentially expressed. We investigated whether certain behaviours were characteristic of anticipation of both food and non-food rewards, and whether signals indicating rewards led to increased activity levels. Twelve laying hens experienced a Pavlovian conditioning paradigm using sound cues to signal the availability of two different food rewards (mealworms, normal food), one non-food reward (a container of mixed soil and sand substrate suitable for foraging and dustbathing (Dusty substrate)) and a sound-neutral event, which was signalled by a sound, but no reward was given. A muted-neutral treatment (no reward and no sound cue) controlled for any specific behaviour as a result of the sound cues. Behavioural responses and the number of transitions between behaviours were measured during a 15s anticipatory period, before birds accessed rewards in an adjoining compartment by pushing through a door. These responses and latency to access the rewards were analysed using linear and generalised linear mixed models. Differences in pushing and pecking at the door (frequency: Dusty substrate 4.87a, Mealworm 3.18b, Normal Food 2.23b, Sound Neutral 0.30c, Muted Neutral 0.03d, χ2(4)=228.99, p<0.001), standing (not walking) (duration (s): Sound Neutral 9.92c, Muted Neutral 7.49bc, Normal Food 7.39bc, Mealworm 7.05b, Dusty substrate 3.06a, χ2(4)=36.28, p<0.001), reflected the perceived value of the rewards, with birds appearing to be more motivated to access the Dusty substrate compared with the food rewards. Rewarded sound cues elicited increased transitions between behaviours, compared with neutral events (Dusty substrate 10.16a, Mealworm 10.13a, Normal Food 9.22ab, Sound Neutral 7.89bc, Muted Neutral 6.43c, χ2(4)=72.05, p<0.001). The sound-neutral treatment induced increased head movements, previously associated with anticipation of rewards (duration (s): Sound Neutral 1.58b, Muted Neutral 0.58ab, Normal Food 0.48a, Mealworm 0.27a, Dusty substrate 0.00a, χ2(4)=25.56, p<0.001). Latency to access rewards conveyed the relative value of rewards (Dusty substrate 7.30a, Mealworm 10.06ab, Normal Food 16.53b, χ2(2)=10.88, p=0.004). Our experiment indicates that, under certain conditions, hens increase their activity levels (behavioural responses and transitions) in anticipation of rewards. Importantly, we demonstrate that this response is not food specific, but rather a general response to both food and non-food rewards. This outcome extends our knowledge of reward-related anticipatory behaviour, and of how hens rank rewards of contrasting incentive value, which may have implications for the methods and environments applied to improve the welfare of laying hens in managed systems.