A Bird's-Eye View of Regulatory, Animal Care, and Training Considerations Regarding Avian Flight Research

Abstract

A thorough understanding of how animals fly is a central goal of many scientific disciplines. Birds are a commonly used model organism for flight research. The success of this model requires studying healthy and naturally flying birds in a laboratory setting. This use of a nontraditional laboratory animal species presents unique challenges to animal care staff and researchers alike. Here we review regulatory, animal care, and training considerations associated with avian flight research.

Figure 1.

Space allowances for bird species commonly used in flight research. Space allowances are taken from US guidelines where available. When unavailable, European guidelines are provided.

Figure 2.

Low-cost custom night-vision system for veterinary and research procedures in the dark, which calms birds and stops them from associating the procedure with the handler. Birds can be habituated to this procedure, to reduce stress further. The system consists of a first-person view goggle for flying drones, to which an action camera (for example, GoPro) is connected directly through existing communication ports and cables. To enable the camera to see infrared light, the infrared filter is custom-removed by an infrared camera company. To flood the room with infrared light, commercial infrared power LED devices are used. The complete system costs between USD$500 and USD$800.

Figure 3.

Sample workflow for avian research positive reinforcement training. (A) Timeline of major training steps toward experimental readiness according to the party responsible for task completion. Breeders are responsible for handfeeding of the baby birds until they can eat on their own and for regular handfeeding until shipment of the bird to the research facility, to habituate the bird to human presence. Once at the research husbandry facility, researchers continue handfeeding and play common noises to the birds to continue habituation. Any time a new stimulus is introduced to the birds, a habituation period is necessary. Bridging can overlap with the habituation period and involves a reduction in reward size to a single seed, presented directly after the bridge (often a clicker is used). Once the bridge has been established, training of desired experimental behavior can begin, for example the step-up behavior illustrated. (B) Important behaviors taught during experiment-specific training that allow for ease of transport and voluntary flight performance. Step-up: bird steps onto a perch placed in front of and slightly above the perch on which it is stand. The bird remains on the perch until brought to a new perch, where they voluntarily step-off. Target stick: bird travels to and touches beak to target at end of target stick. The icon demonstrates an approximation toward the full behavior, because a bird often has to fly to get to target stick location. Target stick can be used as a substitute for pointing when more accurate positioning is necessary or for animals that do not respond to pointing. Pointing: bird flies to a perch that researcher points at. Carrybox: bird either flies into a transport box or is brought into one by using a step-up perch. A carrybox is used to transport birds between rooms and buildings. A carrybox is transparent and has 2 large openings on opposing sides, to train birds to go into the carrybox. The openings ensure that birds do not feel locked in during the first approximations and therefore will go in. The final training step is gradually covering the carrybox for transport. (C) Chaining links individually trained cues to accomplish a set of tasks. The loop shown demonstrates a typical chain used during experimental flight research. Because birds are social animals, at least one other bird is within auditory or visual contact with the experimental bird. However, visual contact is prevented by use of a cage cover when the experimental bird has access to the free space outside of the cages to prevent distraction or undesired flight toward the other bird. The same is true for training in the husbandry space whenever cage doors need to be open during training.

Figure 4.

Reward strategy affects precision of behavior reinforcement. This graph displays an example reward schedule for behavior reinforcement, where increase in the y axis represents the behavior to reward. Time T represents a single training session and should be limited to no longer than 2 min. Time ΔT represents breaks between training sessions. Bridging reinforces a small timeframe of the behavior, allowing for precise behavior reinforcement, without delays from procuring reward (as indicated by bridge–reward pair 1). However, this precise and narrow window of reinforcement means that slight mistiming of the bridge could reinforce incorrect behavior (bridge–reward pair 2). Lumping refers to rewarding the animal after they have performed a desired behavior; this practice lacks specific indication of what part of the behavior was actually desired; therefore lumping rewards a longer period of behavior, thus potentially reinforcing undesired behavior in addition to desired behavior, as indicated by reward 4. However, lumping can be useful for rewarding longer periods of generally good behavior where there is no specific peak in behavior, such as reward 3.

Figure 5.

Shaping is a crucial tool for training behaviors, because in general training a behavior requires intermediate steps. The first training step should be a behavior the bird will naturally offer in its repertoire; the subsequent behavioral steps can be built off the initial behavior in a stepwise fashion as the animal naturally varies its behavior until a particular behavior is reinforced, and so on. The steps need to be small enough to ensure a high rate of reinforcement (food rewards) can be maintained to keep the animal engaged. The example shows typical intermediate steps to shape the appropriate behavior for target stick touching, which is an essential husbandry behavior to get animals to move voluntarily to an indicated position. 1: Habituate the bird to the presence of the target stick. Reward the bird for responding correctly to other cues while the trainer holds the target stick within sight of the bird but outside of the cage. 2: The bird is rewarded for not moving away from the target stick when it is placed near the bird. When rewarding, it is key to feed for position, so offer the treat between the bird and the target stick so that the bird voluntary approaches the target stick to get food. Don't force the bird to be close to the target stick by offering the food too close near the target stick. If the bird moves away from the target stick in any way or indicates distress through raised feathers, then withdraw the stimulus to the amount necessary to stop the signs of distress before trying again. As a note, the speed with which the target stick can be presented should increase incrementally also, although an increase in speed and a decrease in distance should not be paired in the same training session (that is, only ask for one additional step in the behavior at a time). 3: Reward the bird for looking toward the target stick when it is placed within the cage. 4: Reward the bird for moving toward the target stick. 5: Reward the bird for moving toward and then touching the target stick with its beak. 6: Reward the bird for stepping from one perch to another nearby perch to touch the target stick. 7: Reward the bird for flying from one perch to another to touch the target stick. Then increase the distance between perches; in other steps, increase the height (angle) between the 2 perches. Only move to the next training step when the success ratio equals or exceeds 80% at or beyond 5 trials; if needed, take a step back. The criterion should be 80% success only, to ensure fluent transitions between each training approximation toward the goal behavior.