Memory – Storage – Your Memory

Friday, December 5th, 2008 at 8:00 am.
by pre.

Last loop around the spiral we dived right into systems, tricks and mnemonics, ways to achieve almost immediate gain in pushing information into your memory such that you can recall it. We hope by now you have a rich loci-map full of details, strange events and objects, which you’re wondering through often to remind you of, well, anything and everything. You should have letters associated with each digit and, at least a few times, you’ll have remembered a number (perhaps a bank PIN?) using that mnemonic system. If not, why not? Are you trying? Are you listening to the meditations? Meditating along?

This loop around we’re aiming for more understanding of how your memory works, which should help you to understand why the tricks in the last lap around the spiral work, and indeed how the other techniques you’ll be learning this time work.

How does your memory work?

We don’t know.

Well, that’s not quite true. We have some pretty good clues, some detailed proposals. We understand at least some of the components of the memory system. The models presented here are roughly right, even though many of the details are still not understood.

Detail

You likely already know that your brain contains a few neurons. Well, more than a few. Around a hundred billion of ’em in fact. That’s 100,000,000,000. Nearly Fifty a second for your entire life (not to mention the new ones you’re growing each day). They’re all shaped a little like trees. Each one of those has a few (well, okay, mostly over 1,000) dendrites, like the leaves of the tree and an axon, like the trunk which branches at the end to a lot of root-like Axon Terminals too. So inside your brain there are around a hundred trillion (100,000,000,000,000) places where a dendrite (the neuron’s ‘leaves’) and an axon terminal (the neuron’s ‘roots’) meet. If you wanted to count ’em all you’d have to count fifty thousand connections for every second of your life or else you’d end up dead before you finished.

The physiology of these connections, called synapses, is fairly well understood, in more detail than we can possibly go into here. A wave of concentrated ions travels from the tips of the dendrite/leaves down the axon/trunk and into the terminals/roots. This causes tiny molecules known as neurotransmitters to be released at the end of that axon terminal. These drift the unimaginably small distance between there and the next neuron’s dendrite in almost no time at all and there they are attracted to proteins in the wall of a dendrite/leaf in the next neural cell. These proteins in the cell wall are called receptors and when they hit that receptor they encourage yet another wave of concentrated ions to travel along the axon of the next neuron.

Neurons are built so that it takes a certain amount of it’s receptors collecting neurotransmitters before that wave of ion concentration moves down the axon to the next neuron. And this amount, the amount required so that one neuron will pass on a signal to the others it’s wired to, can change.

There are many systems which contribute to this process. Most notably, when a neuron fires lots, it changes so that it will be more easily fired in future. This happens in a few different ways, which together are called long term potentiation, and that allows a neuron to change radically. As this process happens more often (or more extremely), the change in the neuron grows more radically and irrevocably.

This means that the more a connection is used, the lower that threshold will be, the more easily it’ll pass on that signal. Long Term Potentiation, therefore, is the key to how memory works, how learning works. It works by changing the likelihood that one neuron will respond to signals from the thousands of neurons connected to it.

Structure

Wiring up these axons between neurons randomly, even allowing that they can grow connections together and form useful bonds, wouldn’t really be terribly useful. Random neurons, even neurons which adapt their synapse connection strengths, do not make a memory system. Thanks to ingenious experimentation science has also developed a broader understanding of how networks of learning neurons are organized to produce the whole memory system.

Information from the senses is collated, analysed, processed and then stored temporarily in a sensory memory. When we say temporarily, we mean less than half a second. After that the next load of data from the senses is coming in and needs the room.

The information representing things which we are paying attention to in sensory memory can be moved, presumably along thousands and thousands of neuron’s axons, to short term memory. Short term memory has axon pathways which loop back into itself, so that if you attend to an item in short term memory you can keep it there indefinitely, but as soon as your attention fades the memory will begin to degrade too. Short term memory is often assumed to work as though you are repeating something over and over to yourself. Like a phone number, which is about as much as people can fit into their short term memories. If you keep repeating it you may be able to keep it in memory, but if you pay attention to something else it quickly fades.

When we talked about the peg system during the last circuit of the spiral you learned that it’s easier to remember things which you can chunk, so it’s easier to remember 0800-ENCODE than the numerical phone number which that represents. This is because the short term memory focuses on chunks like this. It holds a few of them in a working scratch-pad, for quick and easy access.

From short term memory some information can be carried deeper into the memory system, presumably along yet more thousands of axons, and into the dendrites of an even more wide-spread and numerous system of neurons known as the “Long Term Memory”. Unlike the previously described types of memory, long term memory can last for a hundred years or more, a whole lifetime.

On the level of an individual neuron, this could be achieved by wiring the neuron’s own axons and dendrites together. When it fires once, it’s output feeds back into it’s own input so it fires again, driving that process on and on. The event of it firing once makes it fire more often, increasing the amount of long term potentiation from even a single excitation. Indeed we do find neurons with this structure in the cortex.

Factors which help to determine what information passes from short term memory into long term memory include emotional arousal, attention, concentration, importance, repetition and relevance. All these things help to push the neurons involved over their limit and into that changed state where long term potentiation changes that neuron permanently, pushing the information deep into long-term storage.

Excuses

We have glossed over many, many details in this extremely brief description of your memory system. The best understanding science has to offer is still incomplete, but the systems we have described are better understood than this cursory glance would indicate. The worlds best understanding of these processes explain many features of human memory, and yet also miss enormous amounts of the detail. Our brief description performs even worse. If you read every page we have linked to here, and all the pages that they link to, you’ll begin to have a rough grasp of the rough grasp that our current science has to offer. If you listen to a university level course on these matters you’ll still not grasp all that the real experts do. And even they have gaps in their knowledge.

Yet this understanding is still deep enough that you can begin to use it to help you use your mind more skilfully, to learn to use your memory more efficiently.

Next week we’ll talk about how a long term memory consolidates, and what you can do to help it to do so.

2 Responses to “Memory – Storage – Your Memory”

  1. Sara Says:

    I thought this was a very interesting article.

    The question that came up for me is if the neuron is changed permanently by having information pushed into long term memory then how do we forget things? Do they get overwritten by more recent, more intense experiences? Which in turn change the neuron permanently?

  2. pre Says:

    The processes involved in forgetting are less well understood than those involved in learning. It’s clear that at least sometimes when you ‘forget’ something, in fact it’s still there and you’re just having trouble retrieving it. It can pop into your head days later, for example.

    Likely many of the changes to a neuron can be reversed, so it removes receptors from the dendrites for example and thus becomes less responsive to excitation. So it is certainly possible that sometimes we forget simply because the connection strengths reduce below a level from which you can filter a signal from the noise.

    Since memories are stored in a network of neurons, the configuration could indeed come to mean something else over time, which would at least confuse if not always erase the traces of patterns stored there.

    A very large amount of the answer is likely to be failure to retrieve rather than full on forgetting though I think. You have to be able to get enough prompts to excite the pattern enough that it’ll fire fully. Aiming to talk more (and indeed learn more) about ways you may be able to help this in the next lap around next year when I write about Memory Retrieval.

Leave a Reply

You must be logged in to post a comment.