I have a new project on the brain and I’m very excited about it. It’s a collaboration with an old friend of mine from grad school: Stephane Gaskin. For his dissertation research at McGill, Stephane took advantage of a nifty little paradigm called conditioned cue preference, or CCP. The procedures for CCP buid on the clever work of some early 1900 scholars, Blodgett first, then Toleman and Honzik. In their work, they discovered that rats learned about their environment, in this case a maze with blind alleys and such, even if not explicitly reinforced with food for completing it. They figured this out by first letting rats roam the maze freely over several days without reinforcing them with food for arriving at a particular destination in the maze. Accordingly these rats, unlike rats that were being reinforced every day, entered a lot of blind alleys. As soon as the researchers began reinforcing the previously non-reinforced rats, though, they showed a steep learning curve, as good as the rats that were reinforced since day 1. These findings reveal that the non-reinforced rats were learning about the maze, but their learning was hidden, or latent, until the reinforcers were introduced.
Back to CCP. The procedures here tap into this latent learning and Gaskin’s work at McGill with Norm White, is a very interesting dissection of the neural basis for what they call “pure spatial learning” and reinforced spatial learning. In CCP, rats are placed on a radial arm maze for 3 consecutive days. On each day, all the arms but 2 are blocked and the 2 unblocked arms are adjacent to one another. Rats can explore them freely and only those rats that spend equivalent amounts of time in both continue to the next stage. After this period of non-reinforced maze experience, the rats now undergo training. Over about 8 consecutive days, each rats is restricted to just one of the two arms in a session. On half the days, they are in one arm and get access to an utter smorgasbord: 50 fruit loops!! On the other half of the days, they are in the other arm with nothing. Over the course of this training stage, rat placement in one arm or the other alternates over the days. Then comes the test. Rats are placed on the radial arm maze and, once again, have free access to the two arms. Lo and behold: rats show a significant preference for the arm of the maze that they got all those fruit loops in. Not surprising, you are thinking. Except, you do not see this preference if the rats do not have a chance to explore these arms during the 3 day pre-exposure prior to the training days! Latent learning. The ability for rats to distinguish between the two arms, which share an overlapping cue set, depends on the non-reinforced learning that is going on during the pre-exposure.
This tale now branches into two directions that I find quite fascinating. One direction was pursued rigorously by Gaskin and White and described in detail in a review article recently accepted for publication in Hippocampus. In this review, they describe the results of several experiments on the contributions of the entorhinal cortex, hippocampus and amygdala in the “pure” spatial learning and non-reinforced learning. Preview: the hippocampus (to be fair, dorsal hippocampus) is needed for the reinforced portion, and not for the non-reinforced portion. This work is elegant and worth the read.
As for the other direction that fascinates me: “pure” learning? Pure? I can’t quite get past the word. According to Gaskin, they mean “pure” in that the learning does not arise from typical, obvious reinforcers. That’s a sweet and vague definition for such a bold word. What it leads me to think about is other situations of learning that are not “typically” or “obviously” reinforced. Like the novelty preference test of memory. Gaskin, too, has been led down this road, and so a collaboration was born.
The novelty preference test rests on the tendency of most normal rats to display a bias in the exploration of novel items in familiar environments. Typically, rats are presented with two objects in an field that they have been habituated too. The objects are identical and the rat may explore them freely. After a retention delay of minutes or hours, the rat is returned to the open field and finds there one of the objects it had previously explored and a new object it has never seen before. Most normal rats will spend more time exploring the novel object. Babies do this too. And you. Underlying this preference for novelty is the memory for the previously encountered objects. Thus, this test is used to assess memory for objects. One interesting point here is that this test is quite different from other oft-used tests of object memory that relied on reinforcing the rat to make choices about objects, like object discriminations and delayed non-matching or matching to sample. Based on Gaskin’s work, I’m, of course, very curious about whether there is differential contributions of brain structures depending on which test you use. But then, an even more intriguing point, is whether exploring objects is really non-reinforced or “pure” learning. Clearly something is motivating rats to investigate objects and to spend more time with novel ones, just like there must be something motivating rats to explore a maze even when there’s no food to be had. Focusing just on the objects though, Gaskin and I plan to use CPP to figure out more about object investigation.