How does memory work? What changes in the brain when we learn something?
We don't know for sure. But two outstanding
papers have just provided an important piece of the puzzle, using a truly amazing technique which allowed them to examine the brain of a living, breathing mouse under the microscope.
The approach uses mice genetically engineered such that some of their neurons contain yellow fluorescent protein (YFP). You may have already heard of the cute glowing mice who have green fluorescent protein (GFP) in all their cells. In these YFP-H mice, only some of their neurons are fluorescent.
Two-photon microscopy uses a focused laser beam to image fluorescent tissue. The authors of these papers were able to image the brain (the cortex) after surgically thinning - but not penetrating - the mice's skulls. The bone over the brain area in question was removed until it was just 20 micrometers thick. The brain itself was not interfered with in any way, which is what makes this method so remarkable. Generally, when you put a brain under a microscope, you've had to cut slices off it first.
Using this transcranial two-photon microscopy, these two teams of researchers (Xu et al from Santa Cruz and Yang et al from New York) were able to directly observe the neural changes that took place following motor skill learning. Adolescent and adult mice were trained on a difficult movement task, such as the "rotarod", in which the animal has to avoid falling off a constantly rotating metal rod. With a few day's practice, most of the mouse got better at the tasks.
Both of the papers report that the skill learning was associated with the formation of new dendritic spines in the motor cortex. The image below shows the kind of data we're talking about: this is a single neuron, and the little blobs above and below it are individual dendritic spines, or outgrowths, of the cell. The top image shows the cell before training, and the bottom image is the same cell 24 hours later, after skill learning. Several new dendritic spines have grown. Almost certainly, these spines have formed synapses with another cell.
The results of these studies show that training increases the amount of new dendritic spine formation in the motor cortex, compared to control conditions in which there is no skill learning, and that many of the new spines persist for months. Learning also seems to be associated with the removal of some already existing spines, so the overall number of spines in the brain remains roughly constant.
Overall, this is a pretty amazing set of results, and it suggests that the learning of new skills is associated not only with changes in the "strength" of existing synapses between neurons, but actually with the growth of entirely new synapses. New brain cells are not generated in the adult brain except in a couple of very specific areas, but it seems that experience causes the reshaping of existing cells.
There are lots of unanswered questions - such as whether the same process underlies other forms of learning as well as motor skill training, what triggers the formation of new dendritic spines, and how the process works in humans. But this is a very exciting first step.
Xu, T., Yu, X., Perlik, A., Tobin, W., Zweig, J., Tennant, K., Jones, T., & Zuo, Y. (2009). Rapid formation and selective stabilization of synapses for enduring motor memories Nature DOI: 10.1038/nature08389
Yang, G., Pan, F., & Gan, W. (2009). Stably maintained dendritic spines are associated with lifelong memories Nature DOI: 10.1038/nature08577