To train a starling so that it will eat from a hopper whenever you operate the hopper, then to approach the hopper when you turn on a cue light.
One of the first steps in a behavioral experiment is to train the subject to obtain food or water from a remote-controller hopper or magazine. Eventually we will use the food or water as a reinforcement for performing the correct behavior in an operant discrimination task. This conditioning is called magazine training or autoshaping, and it is a well-studied form of classical conditioning.
As you’ll recall from your reading, classical conditioning involves an association between a neutral stimulus (the conditioned stimulus) and a stimulus with innate value (the unconditioned stimulus). In this exercise, the US is the starling chow in the hopper, and the conditioned response that you are trying to evoke is the starling approaching the hopper. What are the CS and the UR?
Autoshaping also has elements of operant conditioning, because the food acts as a reinforcer for whatever the animal was doing before the hopper came up. We will use this facet of the learning to train the birds to peck a key, and then to peck a series of keys to initiate a song playback and then indicate what category the song belongs to. At first, though, it’s important to avoid reinforcing any particular behavior, because once the animal starts to associate some behavior with getting fed, it will stop engaging in other, exploratory behaviors, and you’ll have a more difficult time later on. For example, if you only operated the hopper when the bird was standing next to it, you’d notice that it would spend more and more time in that part of the cage. Thus, it’s essential in this exercise to vary many parameters of the context when you operate the hopper to avoid “stamping in” problematic behaviors.
The basic procedure is to operate the hopper at random intervals and when the bird is at different places in the cage. Each time you operate the hopper (a trial), you’ll record the amount of time it takes the bird to approach the hopper and eat from it. You will also need to record the distance the bird is from the hopper and any environmental events that could explain sudden changes in behavior (loud noises, sudden movements, etc).
If your bird completes hopper training before the end of class, you can attempt to train the animal with a light as a secondary cue. The goal here is to direct the conditioned response towards a manipulandum, which in this case is a port that can detect when the bird pecks at it.
There are some interesting questions you can answer once you’re through with this training. What happens if you use a different color cue? A cue in a different location? Do you notice anything different about the CR to the cue light versus the CR to the hopper alone?
Next week you will train the birds to discriminate between songs from different individuals. To perform this task, the birds will need to learn to initiate trials by pecking the response keys, a classic operant behavior. You may have noticed that your bird’s CR to the cue light involved approach to the cue light rather than the hopper. At some point, it will peck the cue and we will immediately reinforce that response. Over the week ahead, your instructor and TAs will finish shaping your starlings for next week’s experiment, using a completely automated procedure.