Tuesday, April 10, 2007

Learning described at the cellular level: Finding from our laboratory in Lund



Today I want draw the attention to some very exciting discoveries from our neurophysiology laboratory here in Lund. Because I have just started working in the lab I cannot take any credit for the work, though I have been spending a lot of time lately, trying to develop the experimental setup further. Instead it is Dan Anders Jirenhed, Fredrik Bengtsson, and Germund Hesslow who have worked for several years to achieve the experimental setup that allowed the findings that you can read about here.

The findings have already received quite a lot of attention from places with more readers than my blog. See for instance the article in New Scientist, or if you understand Swedish you can listen or read about the discoveries on Sveriges Radio homepage. If you are interested in the technicalities I recommend that you read the original article which was published in the Journal of Neuroscience.

What has been found is a perfect correlate between behavioral learning and cellular behavior. To explain what this means you first have to know what classical conditioning is. To understand what classical conditioning is, imagine that you sit in a comfy chair (doesn't really matter whether it is comfy or not). Suddenly you hear a loud tone, and immediately after that you get a puff of air in your eye. If you have a normal brain you will blink when your eye is hit by the air. Now imagine that this occurs over and over again, first the tone, then the unpleasant air puff. Again, if you have a normal brain it will eventually realize that "aha, if I blink when I hear the tone, I can avoid that nasty puff of air in my eye, I think I will do that".


Though you may be conscious of the association between the tone and the air puff, this type of learning does not take place in the cortex of our brain, but rather in the cerebellum. Take away the cerebellum and this type of learning is severely impaired if not entirely abolished. In our laboratory we used tiny winy electrodes to measure the activity of single purkinje cells (see picture) in the cerebellum while presenting tones and air puffs. Purkinje cells are a type of neuron located in the cerebellar cortex (near the surface), which because of their morphology appear to be particularly good candidates for the learning that occurs during classical conditioning.


So in essence we could see what happened to the activity in this single purkinje cell when we were applying a classical conditioning paradigm (presenting tones and air puffs). If you just leave them alone, purkinje cells will fire action potentials at a rate of about 60Hz (60 times every second), however, amazingly during the learning paradigm described above the purkinje cells started to show a very distinct pause just prior to the presentation of the air puff (see picture). Change the time lag between the tone and the air puff, and the timing of the pause in the purkinje cell will change so that it is always perfectly synchronized with the air puff. Eliminate the association between the tone and the air puff by presenting only the tone without the air puff and the pause will disappear. Re-establish the association between the tone and the air puff and the pause in the purkinje cell will rapidly re-appear. In essence there is a perfect agreement between the behavior of the animal and the behavior of the purkinje cell.



So why is this so important? Well first of all it is just super interesting, I mean it appears that we can start to describe learning in terms of patterns of neuronal firing. This also means that we will soon be able to see what happens to learning during different treatments. One question that I find interesting is what would happen to the activity in the purkinje cell if we squirt some narcotics (e.g. Cannabis) onto the cell. Will the learning be damaged in some way. There are probably millions of follow up questions that we can now start to explore. All this because of the work done here in Lund, Sweden.


No comments: