Q & A: A Question About Place Preference Boxes

I have done place preference experiments in both types of boxes over the years and prefer the three-chambered box overall. In a traditional two-chambered CPP box, the animal necessarily must spend time, during the test, in one side or the other. Imagine a case where a test drug is given in one side and saline in the other. During the test, the animal spends more time on the saline-paired side. Is this a preference for saline or an aversion to the test drug? Generally, this is viewed as an aversion to the test drug, but the interpretation is sometimes unclear.

In a three-chambered CPP apparatus, imagine the same scenario. The exception is that the animal spends more time in the saline-paired chamber and equal time in the middle and drug-paired chambers. This is a preference for saline. Imagine another case where the animal spends equal time in the middle and saline-paired chamber and little time in the drug-paired chamber. This is an aversion.

I find that the resulting data are clearer when three chambers are used. Generally, the middle chamber is much smaller than the other two and the animal spends very little time in there aside from shuttling back and forth. The advantage is that the middle chamber allows one to clearly see where the animal is spending its time during the test.

My two cents.

I have only had experience using three-chambered conditioned place preference boxes purchased from two different well-known vendors. Both of these apparatuses had a smaller middle chamber, about half of the size of the pairing chambers on either end. I am inclined to agree with the comments from Robert Sorge. Another advantage of the middle/neutral chamber is the fact that it allows for a consistent starting place for each preconditioning and postconditioning test, rather than in a two-chamber apparatus where animals must start in a chamber that is being paired with a drug during conditioning training.

For me, the first challenge using three chambers was getting the mice/rats to spend less time in the middle neutral chamber, thereby necessitating more time being spent in the end/pairing chambers, which increases the sensitivity of the assay to detect smaller preferences (i.e., smaller changes in pre vs. post times). We accomplished this by increasing the brightness of the lights in the middle chamber, or decreasing the light intensity in the pairing chambers. However, this complicates the interpretation of an aversion if an animal spends more time in the saline chamber and non-equal time in the drug and middle chamber, a scenario presented by Robert Sorge, because we've created somewhat of an aversion in the middle chamber by increasing the light intensity. In this scenario it would be advantageous to include control groups that receive saline in each end chamber or receive drug in each end chamber, but of course this increases the complexity of the experimental design and interpretations. In general though for me, the animals spend very little time in the middle chamber if contextual cues are set appropriately. 

Another challenge that I've come across using the three-chambered boxes is the following scenario. Imagine you place an animal in the middle/neutral chamber during preconditioning and the animal fails to explore the end chambers, resulting in most of the time of the test spent in the middle chamber, with equal but very little time spent in the end pairing chambers. What happens then? If you are aiming to create a preference, then perhaps this scenario is ok given that the animal would leave the middle chamber to spend time in the drug-paired chamber, assuming it is rewarding. But what if the goal is to create an aversion, then the only option would be to pair the drug with the middle chamber but then the initial decision made by the animal on which end chamber to enter during the post test is completely random and could complicate the interpretation of the results. In fact, I've had this very problem with type 2 diabetic animals. My results and results from other labs indicate decreased open field activity in type 2 diabetic animals, and in CPP this translated into not exploring the end chambers during preconditioning and remaining in the neutral start chamber virtually the whole time. In this case, a two-chambered box would be desirable and a biased conditioning approach must be used. So depending on the goal of the experiment there are considerations that need to be made.

It should be noted that the boxes I've used are a rectangular prism, with the middle/neutral chamber sandwiched in between the end pairing chambers. I have seen other three-chamber boxes where the neutral chamber is offset to one side, so that in fact the pairing chambers share a wall (not see through), and the neutral chamber acts as a side hallway between the end chambers. Having gone through the process of purchasing and using CPP boxes, I would suggest buying or building a system that can accomodate both three- and two-chamber layouts. 

I tend to agree with comments from both of the above commenters. A large neutral chamber preference (at baseline) does seem to happen more often than not. We have found this to be especially true for rats (compared to mice) where animals had an equal preference for the neutral chamber and the testing chamber. To minimize the neutral chamber time but to take advantage of the 3-chamber format, we have reduced the size of the neutral chamber so that it is basically a hallway between the two larger chambers. This system allows us to place the animals during the final day of testing in the neutral chamber so that they have to make a true choice between the two testing chambers. Of note, the animals still spend ~1/5 of their time in the neutral chamber at baseline. We attempted to alter the lighting in the neutral chamber/testing chamber but were never able to set on a format that was not overly aversive. In other words, if the neutral chamber is too bright, animals may not want to cross through it due to the bright light. This would introduce a significant bias in the testing.