Speaking of rats… [psychology]

Operant conditioning of any mammal requires that you provide a reward for certain behavior. If you want a dog to sit on command, you give the command and reward compliance. If you do that enough times, you can shift the reward ratio away from a 1 to 1 reward and into a schedule of rewards. Maybe you give the dog a treat every other time it complies. It’ll still “sit” when you tell it to, because it likes rewards and two “sits” treat. It works even better when you reward sits in a more unpredictable pattern (thus the high persistence obtained by variable ratio schedules or “VR” in the graph above). But that’s outside the scope of this post.

For simple reward schedules, it only gets complicated when the reward is tied to a lot of effort. If you reward a rat every 100o lever presses, for example, it will work for that reward. But what you see afterward, before it begins to press feverishly the next 1000 times for the next reward is a break. A motivational lag. A pause. That’s the flat part of the Cumulative Number of Responses line you see on the “FR” or “Fixed Ratio” part of the graph. The reward signals to the rat that it has a hell of a lot of effort to put in before it gets to the same reward.

So, if we’re a little light on the posts after the 1000th one, please understand: we’re just simple mammals. Look forward to the feverish lever presses starting again in the near future.