Monday, March 31, 2008

Learning to Learn II – Learning Can Increase the Biological Capacity to Learn

I explained in my book on memory that the hippocampus is the brain structure that promotes consolidation of (declarative) short-term memories into long-term memories. I have also reviewed studies showing that the hippocampus is the one structure in the brain that clearly receives newborn nerve cells, even in the adult. New cells can enhance the ability of the hippocampus to create lasting memories. What has not been emphasized is the importance of survival of new neurons. To be of lasting benefit, new neurons must survive beyond just being born.

Insight into the requirements for neuron survival has come in a recent study by J. R. Epp and colleagues at the University of British Columbia. They injected rats with a chemical marker for DNA that shows up in any new DNA, that is in any newly born cells. If that marker shows up in a cell, it means that that this is a new cell that has incorporated the marker along with its new DNA.

Immediately after injection of the marker, the experimenters trained the rats in a large pool of water that had a safe platform located 2 cm under the water surface where rats could learn its location from seeing cues outside of the pool (such as windows, doors, pictures on the wall, etc.). Other studies had established that learning this task is accomplished by the hippocampus. Rats were divided into groups and trained on days 1-5, 6-10, or 11-15 after injection of the DNA marker. The new-DNA marker showed up only in rats trained on days 6-10 after marker injection. This indicated that there must have been new neurons in the hippocampus of these rats that did not survive in the two groups where marker was not seen. Put another way, for new neurons to survive there is a critical period where they have to be stimulated by learning experiences. Without that stimulus, they die.

Earlier studies had shown that new neurons in rat hippocampus have a development cycle wherein 6-10 days after birth is a middle stage of development in which new neurons are rapidly sending out membrane processes in search of contacts with other neurons. When neurons make contact with targets they can survive. The stimulus of learning thus provides a stimulus for forming new synapses with other neurons, thus enabling new neurons to survive.

The data were originally pooled across all rats in each test group. However, when the data were segregated by how well rats learned (the top and bottom 50 %), it became clear that it was only the poor learners that were showing an effect on new-neuron survival by maze learning. Poor learners probably got more stimulation from the learning because their brains had to work harder at it. It wasn’t that much of a mental challenge for good learners.

We know that humans are continually producing new neurons in the hippocampus. The issue is the need to experience enough demanding learning to help these new neurons survive. The critical period for learning to influence new-neuron survival in humans is not known. So, the practical take-home message is that we need to be learning constantly, every day, so that no matter what the critical period is, we will be helping our new neurons to survive. Survival of new neurons means a greater biological capacity for learning, at least in people who are not good learners. In other words, here is a clear case where the “poor get richer.”

Source: Epp, J. D., Spritzer, M. D., and Gales, L. A. M. 2007. Hippocampus-dependent learning promotes survival of new neurons in the dentate gyrus at a specific time during cell maturation. Behavioural Neuroscience. 149: 273-285.

Thursday, March 20, 2008

Learn One Movement Skill At a Time


"Motor memory" refers to a mental model (MM) that the brain constructs from past experience. In the example given by researchers Reza Shadmehr and Thomas Brashara-Krug, when a person plans to pick up a brick, a MM of the amount of force required to pick up the brick is used to execute the action.The brain does not estimate the force as if it were a feather nor if it were a sack of cement, rather it uses its memory of what a brick weighs to create a model of how much force will be needed to pick it up.

In the studies they reported, they used a robotic arm that subjects used to manipulate objects. In learning how to use the mechanical arm, subjects had to create a MM of how to make it do what they wanted. Like other kinds of learning, the MM is consolidated with practice into long-term memory.Moreover, motor performance continues to improve, even after actual practice has stopped, indicating that the MM itself may be subconsciously rehearsed, off-line so to speak.

Motor memory processes have great applicability in everything from learning to touch-type to learning to throw a football to a moving target. The study by Shadmehr and Brashara-Krug explored the finding that a recently acquired MM (MM1) can be disrupted if a second MM (MM2) was introduced too soon after MM1.That is, a MM1 has to have enough time to consolidate, just as declarative memories do.

Also, a MM1 can interfere with learning a MM2, if there is not enough time separation between learning the two motor tasks.This was demonstrated in the present study by having 60 subjects learn how to make two conflicting movements using the robotic arm. The MM for both tasks could be learned but only if the training sessions were separated by at least 5 hours. If the interval was shorter, learning of the second MM (MM2) was impaired, as was the likelihood of consolidating the first MM.

The “take home message” of this research is that learning different movement tasks should be separated in time, lest there be interference with forming long-term memory of both tasks. My explanation is the following: Once MM1 gets consolidated (that is, after about 5 hours), the circuits that sustain its short-term representation now become available for learning a second motor memory (MM2). That is, MM1 has proactive interfering after-effects that dissipate with consolidation of the MM1 and thus no longer interfere with learning an MM2.

Athletic coaches might be well advised to ponder the application of this principle.


Shadmehr, R., and Brfashers-Krug, T. 1997. Functional stages in the formation of human long-term motor memory. J. Neuroscience. 17(1): 409-419.