Sunday, October 18, 2009

Information Processing in Nerve Cells, My Own Theory

Way back when I was a kid, I think this was when I was in Junior High or about 1972, my science teacher showed us a movie about how nerve impulses traveled up a nerve fiber. I remember that they took a live but "pithed" frog (they had destroyed the brain with a needle) and strapped it's leg onto a board. They had dissected out one of the nerve fibers in the leg and connected an electrode to the nerve. That electrode was then connected to an oscilloscope (an electronic device used to look at the shape of electrical signals. I know that all sounds pretty gross but all they showed us in the movie was the leg strapped down with the electrode coming out. Anyway, they also had a device that could be made to touch the frog's foot with a specific amount of pressure. They proceeded to touch the foot and show the signal that was detected in the nerve fiber.

Now most people think of a nerve signal traveling up a nerve fiber as a singular event. They think only one pulse travels up the nerve for any one event sensed by the nerve endings. After all, all the biology textbooks ever describe is how one pulse travels up the fiber. However, when the device touched the frog's foot a series of pulses were detected by the oscilloscope. What's more, the pulses were not regular. They were separated by varying spaces like the lines on a bar code. This is very similar to what radio control hobbyists call Pulse Code Modulation, or PCM. In PCM the differences in the spacing between a set of pulses provide information to the device on the receiving end, such as the R/C airplane. In addition, the pattern of pulses detected in the frog's nerve was repeated every time the device touched it with the same pressure. However, when the device was set to touch the foot with a different pressure, a different pattern was detected, which also showed up exactly the same every time that particular amount of pressure was applied. This is again similar to PCM in that the same pattern always means the same thing to the receiving end and different patterns mean different things. Usually different patterns in PCM mean to set the flaps on an airplane to a different angle, or to turn the wheels of a car to a different degree.

Yet, another similarity with PCM is that the height or amplitude of the pulse doesn't matter. Only the pattern of the pulses. In PCM the amplitude of the pulse is ignored because it will always be different depending on how far away the radio controlled car or plane is from the controller. Paying attention only to the pattern of the pulses - which can be detected even when the signal strength is very low - makes the system much more reliable. And so it seems to be with nerve signals. The strength of the signal may diminish as the signal moves down the fiber - even though there are little repeater stations called the Nodes of Ranvier - but the pattern will always stay the same. The narrator of the film even made sure to point out that, even though one would expect more pressure to simply make the nerve impulse signal have a higher amplitude (be stronger), that aspect of the signal always stayed the same. Only the pattern changed. So, it would seem that nerve signals store information in a pulse code modulated fashion.

It has been 37 years since I saw that film and I have, obviously, never forgotten it. I have read a lot of material about how nerve signals are transmitted over the years and not once have I seen anything about this pulse code modulated nature of nerve signals. In fact, almost everything I have read seems to treat a nerve cell as some kind of binary device that is either on or off. The only information current researchers seem to believe a nerve cell contains is a single bit of binary data that is either there or it isn't. Granted, I haven't read a lot of actual research papers. Just articles in the popular press and more textbooks than I care to cite. So it is possible that real neuroscientists are well aware of the phenomenon I describe and that it just doesn't get into the popular press, not even Scientific American, because some authors deem it too complicated for the masses. But one would think I would have seen at least some mention of it in 37 years of reading.

Over all that time, I have developed a theory of sorts as to how these patterns of pulses play a vital role in how information is stored, transmitted, and processed in the brain. However, it will take quite a while to explain, so bear with me.

I believe that the patterns of pulses have complex interactions when they reach the thicker body of the nerve cell. Now remember that a nerve fiber is not a solid object. It is actually a tube made up of the cell wall of the nerve cell, surrounded by the myelin sheath, which acts like an insulator of sorts. The "pulse" traveling down the nerve actually travels along the cell wall. And the "pulse" is really a chain of chemical reactions that propagate down the cell wall. The reaction at one point triggers off the reaction at the next point. That is how the pulse travels down the fiber. After the reaction moves down the fiber, the cell wall resets itself to get ready for the next pulse. This all happens pretty darn fast. The nerve body is just a continuation of that cell wall, that tube. The body is more like a widening of the tube than a solid body with a solid wire attached to it. Imagine for a moment that the cell body is just a slight widening of the tube that is the cell wall. Now imagine the series of pulses traveling along the narrow tube that is the nerve fiber. Each pulse is like a ring of chemical reactions propagating down the surface of the tube like a wave. One could use the analogy of a very, very long, narrow tank of water with a series of waves traveling down it from one end to the other. However, in the case of the nerve pulses, the wave is more square shaped instead of rounded on the top like a water wave. Anyway, when the series of nerve pulses reach the thickened part of the tube we are temporarily imagining our cell body to be, they simply spread out to fill the wider space and then squeeze back together when the space narrows again and travel along the rest of the fiber unchanged.

Now imagine that we insert a vertical card in the wider space, placing it parallel to the direction of travel of the waves/pulses. The water waves would just go on either side of the card and continue on unchanged, right? OK, here is where things start to get a little complicated: Imagine instead that the cell body is actually much wider than the fiber, which is actually the case, so just remember a basic diagram of a nerve cell. (I hope to draw my own diagrams later. I just have to get this out of my head while I am on a roll.) In order to represent this in our water tank analogy we could imagine the cell body as simply a large tank of water. Unfortunately that is not how the actual cell works. Remember, the pulse only travels along the cell wall, not through the middle of the cell body the way a wave would travel right through the middle of a simple tank of water. Therefore, in order to continue our analogy we must imagine that the long narrow tank representing the nerve fiber splits into two long narrow tanks. Each side of this split represents a strip of the cell wall of the body of the nerve cell. The water waves traveling up the two different paths in the tanks is like the nerve pulses traveling along two different sides of the cell walls as they get farther apart where the nerve fiber meets the cell body. Further imagine that each of these diverging paths are the same length and come back together later. The waves in the tank would simply meet back up in perfect alignment and continue on down the rest of the tank just as it did when all we did is place a card into the tank.

Next, imagine a slightly different scenario. Imagine that one of the paths representing the cell wall of the body of our nerve cell was longer than the other. As the waves traveled down the two paths and then met up again, they would not be in alignment, now would they? One would be slightly delayed because the waves had to travel farther. This is the crux of my theory. When the waves meet on the other side some of them cancel each other out because they are shifted and some of them reinforce each other because they align perfectly. Transfer this analogy back to the actual cell wall of the nerve cell. As the rings of chemical reactions, our series of pulses, travels along the nerve fiber and reaches the cell body, those rings spread out and travel along the surface of the cell body, wrapping around it like stripes painted around a vase. Now, if our cell body were perfectly symmetrical, then the rings would meet perfectly on the other side of the cell body and continue on unchanged. But cell body's aren't perfectly symmetrical are they? They have all kinds of odd shapes and they have extra nerve fibers sticking out here and there. So the rings of chain reactions comprising our nerve pulses don't just go into the the cell body like it was a big transistor. They wrap around the cell body and get deformed as they travel along the oddly shaped surface, meeting up on the other side and canceling or reinforcing each other sort of like water waves.

I believe it is this modification of the nerve pulses as they wrap around a nerve cell that does a major portion of the processing within the brain. A portion that has been largely ignored by researchers who create an overly simple model of the nerve cell when they assume that it acts like a simple on/off switch. I believe that the very shape of the nerve cell body, as well as the exact locations of where nerve fibers enter and leave the cell, has a strong effect on what the series of nerve pulses coming out of the cell will look like. This, along with a complex set of thresholds of various neurotransmitters which can vary from one location to the next on the surface of the cell, creates a signal modification processor vastly more complex than a simple on/off switch. In addition, consider that these pulses are not pure waves like water waves so they don't cancel each other out in the same ways. It takes a short but finite time for the cell wall to recover and be ready for another signal to pass. Thus, a pulse can actually cancel out another pulse that follows too closely even if they don't exactly overlap. Also, a week pulse that may not be strong enough to continue triggering off the chemical chain reactions may set up a cell wall in such a manner that a second weak pulse passing over the same spot will be suddenly reinforced even though there is no reinforcing pulse occurring at the same place at the exact same time. Finally, all of these factors are being continuously, dynamically modified in response to all the pulses that have passed over the cell as well as many other factors such as brain chemistry.

All of this is going on in just one single cell. Multiply that by all the connections between all the cells and you have really got your hands full. Think of all the other nerve endings going into the body of the cell and all the synapses receiving additional series of pulses from other nerve cells. All of these additional sequences of pulses are being added to the complex pattern of pulses washing over the surface of the nerve cell, affecting the pattern of pulses in various ways. They may directly reinforce or cancel another pulse, or they may influence some chemical threshold on the cell wall of our nerve so that it will influence the propagation of waves that pass by later. It is known that signals from some nerves can suppress or enhance the reactivity of other nerves. All of this together adds up to a great deal of processing that is done in real time by all the molecules making up every tiny discrete piece of the cell wall, all of those pieces acting together perform yet another vast level of processing.

As with all my other ideas, I am not an expert in this field. There may be research that directly contradicts what I have hypothesized here. But then I don't really have time to do all that research. That would be a PhD thesis all in itself. I am just tossing this idea out there for others to mull over and use as they see fit. Believe it or not I still have something more important to do than figure out how the brain works. That something is DEMML. If I can get that going then there will be millions more people out there with the capability to figure out how the brain works and lots of other important things.

The contents of this post is Copyright © 2009 by Grant Sheridan Robertson.
Again, anyone can use this idea for research. Just be honest about where you got the idea.


  1. What you are describing is a well known phenomenon called synaptic integration. No neuroscientist thinks a neuron is as simple as you describe it. In fact, a single synapse is usually not powerful enough to elicit an action potential in the postsynaptic neuron. Instead, excitatory and inhibitory postsynaptic potentials (EPSPs and IPSPs, respectively) must integrate before an AP can occur. This integration can happen spatially, as you suggest in your post, but it also occurs temporally. Look up "axon hillock" or "synaptic integration" if you want to learn more about it.

  2. Hey, Anonymous, thanks so much for the feedback. I did find, when I did my research for this ( paper, that it can take many synapses and overcoming a complicated threshold mechanism to start an action potential. What I haven't been able to sus out is whether that action potential (AP) occurs simultaneously across the entire cell body, or is localized on the dendrite. If the latter, then does this "spatial integration," as you call it, come into play across the cell body as other dendrites also reach AP and send their own pattern of pulses down to the cell body asynchronously? What then is the final determiner as to what signal gets sent down the axon to other cells? Recent research now shows that other cells may influence the pulse even after it is heading down the axon, so that certainly complicates things. Please note, I did take temporal integration into consideration. I just did not know the correct terminology.

    I will definitely look up "axon hillock" and "synaptic integration."

    I am curious, though. If no neuroscientist thinks of neurons in the simple model I describe, then why do almost ALL of them describe neurons using that same simple model. Do they think that no one could possibly understand the truth or are they just trying to keep the secrets to themselves? ;^)

  3. I enjoyed reading your article thanks!

    I would like to ad that neuroscience is a medical profession where the related literature is full of dry medical related old latin words and phrases. Your text is more related to the engineering profession which is more pedagogical and logical to read. Regarding the content I would assume that neuroscience literature does have similar understanding. Although I find your metaphor compare about the various wave forms interesting and would like to see the neuroscience version of the same.

  4. Thank you, Anonymous from Jan 10.

    Being a technical person at heart and not knowing all the neuroscience terminology, I wrote the best description I could. In addition, I tend to write for a general audience here, even though I am writing about technical things. I believe reducing the barriers to learning, thinking, and sharing is important.

  5. Technics - even Physics - are called "reductional" for Biology. Physicians try to use laws, but Biologians use all as it is needed to survive - therefor much more complicated.

    If you like to use a model - leave the wave model. I learned it too so. Use the Electrostatic Potential Model.

    I am also only an Informatician coming from Electronics and Physics. I understand the theory of "energy transport" in our nerves as moving of electical loaded ions (parts of molecules). They obay the physical rule of attracting and pushing off different or same loaded ions. All meshured pulses are in nature positive or negative loaded ions.

    I hope I could help You!

    Look also to