So... what is reversal learning anyway?

Reversal learning, in mammals and especially in humans, is a particular process of the Ventral Pre-frontal Cortex and the Ventral Striatum areas of the brain; it shares a part of the machinery that takes care of decision-making and emotional empathy (L. Clark,* R. Cools, and T.W. Robbins, 2003), essentially it's an incredibly important process that allows us to adapt to varying environments. With this, it's pretty easy to guess just how imperative it is for our survival that this type of learning evolved but, so that we can place it into perspective, lets imagine a situation where our ability to carry out reversal learning fails. 
Envision learning how to catch a ball, however, you learned to catch it with just your left hand and now every time a ball is thrown you can catch it with ease. One day, while working with a chain saw, you lost your left hand, if you didn't have the ability to carry out reversal learning you'd never be able to catch a ball again, you'd be impulsively trying to catch the ball with a ghost left hand. Why? What's impeding me from learning how to do it with my right hand? The answer is: "your memories". That is because the “Reversal Learning Paradigm” involves “being capable of effectively overcoming established associations and learning new ones based on feedback (Xue, G., et al., 2013)", in other words, being able to suppress negative transfer effects (Bernhard Komischke, et. al. 2002). This feedback or negative transfer effect refers to the memories already acquired before losing your hand, every time that a ball is thrown at you the only thing you'll remember is how to catch with your left hand. If you were able to do reversal learning, what would happen is that the feedback, in this case the memory for catching the ball with your left hand, would be silenced or, more specifically, eliciting its recall (the ability to bring forth a memory) would be inhibited. This way your brain is free to acquire a new association that shares certain cues with the previous one, like learning how to catch a ball with the right hand.
Thus, the concept of reversal learning essentially boils down to learning something and then learning the opposite of the concept learned, learning the reverse of what was just learned. This type of learning can be assessed in different kinds of subjects, such as animal models that use: rats honeybees, ants, horses, etc., by using various paradigms. One of the classic reversal learning tasks utilizes a reversed preference paradigm. This consists of 2 stimuli, one which will be rewarded, in this case with something like food or a drug, and another which will not be. Once the subject has learned which of these is the “correct” stimulus (meaning the one for which they receive a reward upon response), then the contingency is reversed. Now, the subject will be rewarded for response to the opposite stimulus, which means they must reverse learn that the original stimulus was the one which they obtained a reward for or in a way forget about it.

This image was obtained from: http://www.campdeninstruments.com/product_list.asp?SubCatID2=410
Here we can see how the reversal paradigm could work. In this case, the rat would receive a food reward  for pressing the "correct" and then it would change so that the rat would only obtain a reward for response to the opposite image.

Recently research on reversal learning has allowed us to understand, among many other things, how honeybees at the subspecies (race) level are affected by making choices as to which flowers they will visit, a very valuable piece of information since we have Colony Collapse Disorder (CCD) going on. Furthermore, this type of question has allowed us to use drugs such as MDMA to help us understand the mechanisms of decision-making and empathy, although, there's still leagues of information we can't even begin to fathom.