That, my friend, is the million pound question! We don’t really know for sure but our best bet comes from computational neuroscience which tells us that memories and knowledge of all sorts, is probably “stored” in the brain by changing the connections between brain cells. These can either be strengthened — so that if one cell fires it makes it more like the other will fire too — or weakened by learning. In fact, at cellular level, this is precisely what Mariana works on. Strengthening is called “long term potentiation” (or LTP for short) and Mariana’s work aims to study how this occurs and what exactly it means for memory.
So although we’ve known about these cellular connection changes for a while, no one really knows how it works in detail. Computational neuroscientists have done the most in this respect. By using computer models of the way neurons learn, they’ve shown how it is possible to encode and retrieve information in a set of neurons. One of the cool things they’ve shown is that different brain systems learn differently. So, for instance, there was a question previously about how we learn and I’ll find the link to that and post it in a minute. But the short story is that the hippocampus learns quickly but doesn’t store information for very long whereas the rest of the neocortex learns slowly but keeps it more or less forever. The computational people were able to show why this happens in a very mechanistic fashion and for that reason, I think they’ve done more towards answering your question than anyone else.
Basically because even though it is relativly small, it contains billions of neurons (I just checked wikipedia and apparently the cerebral cortex alone conatins 15-33 billion neurons by itself. Therefore there is actually a huge capacity to store lots of information. While it doesn’t work like this, it each neuron could hold either a “yes or no” item of information, then the brain would hold 10-100 gigabytes of information – as much as hard disk on a computer.
(If anyone knows more exact numbers for this I would be very intriuged to know!)
Hi Nick,
It’s interesting to put a number on the information because 100GBs is fairly small — it’s probably many orders of magnitude less than the information content of your average 6 year old. The issue here is that neurons aren’t binary information stores — in fact, neurons are very complex, nonlinear processing devices that are so sophisticated, that for a state-of-the-art desktop computer to simulate about 1s of typical information processing in a neuron would take several days of computing. It’s difficult to over-emphasize the complexity of the processing that occurs in a single neuron.
Then there is the additional fact that the roughly 100 billion neurons in the brain make between 100 trillion and 10 quadrillion connections and these connections are the main way that information is stored. So instead of having the capacity to store 100GB of information, brains may store up to 10 million times as much information as that…
Comments
Joe commented on :
Here was an early question is that is related. This is a really hot topic — great question, christy!
/march2010brain-zone/2010/03/what-pathways-are-involved-in-learning-and-what-species-and-why-are-most-adept-to-learning/
Joe commented on :
Hi Nick,
It’s interesting to put a number on the information because 100GBs is fairly small — it’s probably many orders of magnitude less than the information content of your average 6 year old. The issue here is that neurons aren’t binary information stores — in fact, neurons are very complex, nonlinear processing devices that are so sophisticated, that for a state-of-the-art desktop computer to simulate about 1s of typical information processing in a neuron would take several days of computing. It’s difficult to over-emphasize the complexity of the processing that occurs in a single neuron.
Then there is the additional fact that the roughly 100 billion neurons in the brain make between 100 trillion and 10 quadrillion connections and these connections are the main way that information is stored. So instead of having the capacity to store 100GB of information, brains may store up to 10 million times as much information as that…