Friday, May 11, 2007

Thoughts on Information VI (Man is a Creature of Habit)

First of all, let me apologize to any readers for being so long in getting this post up. It did not flow so easily as the first five. If anyone is, by chance, reading this for the first time, the prior posts on information can be found at:

Some Thoughts on Information I
Some Thoughts on Information II
Some Thoughts on Information III
Some Thoughts on Information IV
Some Thoughts on Information V

I would like to address both Mr. Putnam’s and the Growelry’s thoughts on “information” after I have finished this post of the “information” thread. There is much to ponder in their posts and, while it intersects with what is below, there are some ideas here that I would like to develop in commenting on their discussion. In addition, I will try to use the word “data” for where I previously used “information.”

What I hope to do in this post is to finish up the data transfer from the eye to the brain as we have discussed for the last 5 posts in the series and get to the heart of the matter (or brain of the matter, if you will) to discuss how some of the workings of the human brain might be viewed in terms of the molecular mechanisms of nerve conduction and transfer.

Please let me reiterate that these are all personal observations and do not carry the weight of the academy, though I have tried to document as best one can the mechanisms such as nerve pulse generation in the eye. Perforce, the discussion of the action of the brain is much harder to document since there is much disagreement on how we actually “think.” (a much more loaded word than “information”).

Let me also state early on that the hidden purpose of this exercise has been to eventually address the question of free will. I know that sounds extraordinarily presumptuous, but decisions deemed to be made by a person exercising his or her "free will" are made in the brain, and that is what we are examining. Just think of how much rests on the belief that humans have the facility of Free Will. From this belief flows all laws and responsibility, in particular the concept of evil. It is the absolute underpinning of our modern Society. With that, we will cut to the chase.

Eye to Brain
To follow up from the previous posts, the path data takes from the eye to the brain can be seen in the following diagram.

It is interesting to note that not only does some processing occur in the retina, as postulated before, but most assuredly there is processing of the data in the lateral geniculate nucleus (LGN) before traveling to the occipital cortex.

In the occipital cortex, there are a number of areas of interest where data is further processed and then sent on to other areas of the brain. The principal one is V1:

In the following diagram of the posterior brain (from here)there are two “streams” arising out of the visual cortex. The first is the inferior or ventral stream. It is associated with recognition and in some way with long term memory. Since memory is the basis of “thinking” we will return to this point. The upper or dorsal stream is associated with motion and control of the eyes or arms.

Thus, there is radiation to the motor cortex, and also to areas known for memory location (e.g. hypocampus and the amygdala.).

At this point, then, we have tracked data from the eye to the brain. In order to move further, I would like to propose a simple model. Though we “see” the world in all its glory, data that strikes the eye is actually contained in a vast (and I am speaking vast here) spatial and temporal array of photons. In two dimensions, and in an much simplified view, the data appears as the following;

When these data strike the retina they are transduced into a spatial and temporal data stream that is carried to the brain via action potentials down the optic nerve bundle. At both the level of the retina and at the lateral geniculate nucleus, there is initial processing of the data. We will propose a simple model of processing in a minute.

At some point we should review the purpose of this data. Clearly, it is to determine the actions of the organism. Data are essential for an organism to find food, ingest food, reproduce and avoid destruction. Quite simply, an organism gathers data, compares it to stored information, and then makes a decision based on that comparison. I see an elephant, I run. I see a Big Mac, I eat. I see Anna Nicole Smith, I puke. Man is a creature of habit. While man’s behavior is complex in the extreme (to us, maybe not to an alien) it may still be reduced to the laws of chemistry and physics.

Just as we discussed isolation tanks (and I have just finished reading "Altered States" by Paddy Chayefsky, who, also wrote "Network" with the great line: "I'm mad as hell and I'm not going to take this any more") it is my contention that man's action are entirely driven by external stimuli.

In past posts we have reviewed the electro-chemical processes that convert photons into action potentials. Importantly, we have observed that these pulses from the eye to the brain along the optic nerve (there are 1.2million "nerves" (fibers) in the optic nerve) are an either/or phenomena and that the presence or absence of an action potential depends, in the end, on the difference of one molecule of G protein or neurotransmitter. Recently I discovered that this is termed "The Principal of Bivalence."

We will never unravel the complexity of the signal from the eye to the brain and that is certainly not our purpose here. What we want to do is show that the restrictions imposed by the either/or nature of the transmission of pulses place restrictions on how we think.

The key cell in the transfer process and, of course, in the brain is the neuron.

The brain is composed of approximately 100 billion neurons plus a number of other, supportive cells. The neurons have about 100 trillion synapses. All neurons are not completely alike but their function is similar in many respects. The nerve cell is composed of the nucleus and cytoplasm with multiple connections to other nerve cells via the dendrites and with a long process called the axon with connects with the dendtrites of other cells. As we have shown before (neurotransmitters and the action potential), the action potential is generated in the nerve cell by the change of the permeability of the cell membrane to ions which is, in turn, caused by the binding of neurotransmitters.

The major cell in the cerebral cortex is the pyramidal neuron:
A pyramidal cell (or pyramidal neuron, or projection neuron) is a multipolar neuron located in the hippocampus and cerebral cortex. These cells have a triangularly shaped soma, or cell body, a single apical dendrite extending towards the pial surface, multiple basal dendrites, and a single axon. Pyramidal neurons compose approximately 80% of the neurons of the cortex, and release glutamate as their neurotransmitter, making them the major excitatory component of the cortex.

Here is a slide show on neurons:

How the brain works on a biochemical level is not very complicated except for, possibly, short and long term memory. Action potentials arrive from other nerve cells (e.g. a fiber of the optic nerve) at the soma, or body, of the nerve cell. The level of the potential is summed and, depending on the threshold, a new potential is generated. The site of the summing is at the axon hillock.

Here I quote from the Wikipedia reference:
The Axon Hillock is the anatomical part of a neuron that connects the cell body called soma (biology) to the axon. It is attributed as the place where Inhibitory Postsynaptic Potentials (IPSPs) and Excitatory Postsynaptic Potentials (EPSPs) from numerous synaptic inputs on the dendrites or cell body summate.

It is electrophysiologically equivalent to the 'initial segment where the summated membrane potential reaches the triggering threshold, an action potential propagates through the rest of the axon (and "backwards" towards the dendrites as seen in backpropagation). The triggering is due to positive feedback between highly crowded voltage gated sodium channels, which are present at the critical density at the axon hillock (and nodes of ranvier) but not in the soma.
Thus, the brain is composed of neural circuits where outgoing impulses are generated depending on the incoming pulses at the level of the neuron. Again, remember that this is an either/or phenomenon because of the biochemical nature of the action potential

Neural circuits are, of course, very complex. Here are just ten out of the 100,000,000,000 nerve cells in the brain:

Neural circuits are similar to electrical circuits in that there is potentials (generated by batteries, e.g.) and conductors. The neural "batteries" are gated, in that they either generate a potential or not. I am unsure whether the strength of the action potential varies. I am inclined to think that it doesn't but I need to research this in neurophysiology.

We are all familiar with electrical circuits. Here is a mundane but interesting example. A hobbyist has built a gizmo that detects when trains are approaching a crossing in a model train set.
It has pulse width modulation so it can be used to control the speed of motors, not just on/off. So far I have only used mine to control my Remote Control Level Crossing, an infrared sensor detects a train approaching the crossing and automatically lowers the barriers. When the train has passed, it raises them again.
. Here is copy of the circuit:

Notice the similarities to the 10 neurons above.

One of the interesting thing that comes out of thinking about this is how on earth did the brain evolve? If there are 100 trillion connections, each one, of course could not have be the result of evolution. Furthermore, this is done with only 25,000 genes! I will leave this now since it certainly deserves much discussion. One final thought though is that the brain is not predesigned (as ID contends) but is the result of evolution. Successful evolution always leaves an out; unsuccessful evolution (the dodo bird) doesn’t. That is, unless we are at the DoDo end of evolution, the brain can go much much further even biologically. Unfortunately for us, it will take millions of years.

Back to a discussion of neural networks:

The complexity of neural networks is daunting. However, just as one does not need to know the direction vector and the kinetic energy of every molecule in a gas to know the behavior of the collection, there should be ways of simplifying neural circuits so that they can be better understood.

Let me simplify the scheme of photons hitting the retina even further. I do not know whether you have read “Flatland” by Edwin Abbott. In it, a man must function in two dimensions, as a square (pentagons are superior, triangles are serfs.) However, he does dream about a one dimensional world called “LineLand.” In following data from the eye to the brain and beyond, it is useful to simplify to the utmost. Thus, consider the scenario of a single photon (and its opposite, no photon) as the ultimate carrier of data in “LineLand.”

This impulse, on reaching the brain of the individual in Lineland must be compared to something in order to trigger a response. This something is, of course, memory. In Lineland, every photon hitting the eye triggers a response just as every photon that doesn't triggers a no response. This seems trivial in the extreme but it is the absolute basis of behavior. Behavior as reflex. Behavior as habit, if you will.

Unfortunately, human memory, or even nematode worm memory (302 neurons) is not well understood.

It occurred to me to question in this respect what is going on in the retina of the eye and the LGN where nerve signals are processed even before they get to the brain. I suspect there is no "comparison" with stored memory there but most likely a type of filtering of the data. Processing that is entirely under genetic control. Sort of like behavior in the nematode.

An ongoing theory has memory as as a hologram, i.e. that the memory itself is spread out over many neurons and is like the visual hologram which appears as a diffraction pattern on film. As attractive as this hypothesis is, it seems difficult to reconcile with the physical reality of what we have presented above. If, indeed, a pattern of neural signals is the pattern of currents in a neural network, it would make sense only if the memory that they are compared to is itself such a pattern.

However, there are many who believe that there is permanent change in the nerve cells in the brain that accompanies long term memory. Any change like this would most likely be at the gene level where genes were activated or suppressed depending on the potential of the cell. If the memory is long term, i.e. not erasable (though we do "forget") it would needs permanent.

An infinitude of research is going on in this area and I don't pretend to be conversant in it in the least. Let me cite the abstract of one paper:
Vision, emotion and memory:from neurophysiology to computation
Edmund T. Rolls*
Department of Experimental Psychology, Oxford University, South Parks Road, Oxford OX1 3UD, England, UK
The inferior temporal visual cortex provides invariant representations of objects. The computations underlying this can be understood in the framework of a hierarchical series of competitive networks in the ventral visual stream which learn invariant representations by using a short-term memory trace to extract properties of the visual input that are invariant over short periods and are thus statistically likely to be from the same object.

Summary (again):
Information from the eye arrives in the brain a a patteren of digital information in both “space” and time. The spacial information depends upon which of the 1.2 million nerves in the optic nerve carries an action potential and the time factor depends on the frequency of firing. There is a limit on the latter dependent on the latency of the action potential and, perhaps, the distance from the eye to the brain. The latter is probably not significant but it could be (need to get the speed of a nerve impulse in a mylinated nerve).

While, as we have mentioned, there has been pre processing of information prior to arrival in the brain, there is probably no actual processing of information between the retina and the occipital cortex.

Once in the occipital cortex the signals are further processed, i.e. undergo modification by neural circuts, and radiate to areas that are known to contain memory.

Once at the memory "site" the impulses are compared to memory, whatever that is, and a signal is sent to the motor area to "do something."

In the end, the workings of the brain are totally dependent on the electrochemical reactions of the neuron which are bivalent. That is, Either/Or.

Free Will:
Let me give an example. Most of us remember this awful picture as one of the strongest to come out of the Vietnam War:

If asked, one would be inclined to say that this assasian was practicing free will. One would say that this action is premeditated (i.e. he is not reacting in self defense), and that he is not being compelled to murder a defenseless prisioner.

However, this man is not in an isolation tank. He is receiving mostly visual, photon input that exists in patterns. Those patterns are being preprocessed in the retina and the LGN and are travelling to the occipital cortex where they are further processed. Once there, they radiate to areas of long term memory. If this was a reflex, like raising your hand to avoid a blow (i.e. something the prisoner might have done if he was not bound) the comparison here might go immediately to the motor cortex.

In this case, there are most likely signals to the prefrontal cortex. In the prefrontal cortex is the executive center which:
The so-called executive functions of the frontal lobes involve the ability to recognize future consequences resulting from current actions, to choose between good and bad actions (or better and best), override and suppress unacceptable social responses, and determine similarities and differences between things or events.
I would like to postulate at this point that what transpires in the executive center is exactly like what transpires in any other neural circut. That just as in Flatland, there is a decision made on a signal neuron basis. That the command to pull the trigger is an either/or decision based on the threshold reached in a signal cell with the concentration of a signal molecule or neurotransmitter.

Either the man pulls the trigger or he doesn't. While the actual pulling of the trigger involves many muscles and commands, the actual decision between pulling and not pulling rests on the potential (or, more likely the activation or inactivation of genomic DNA) in a single cell. This is due to the "The Principal of Bivalence."

I am very weary after trying to put this together so at this point I am going to stop and return to the discussion of the ramifications of this hypothesis in the next post.


David said...

Your blog is very interesting and with good links.

Will put a link on the NYT forum Human Origins.

Dr. C said...

Thanks, David. Join in the discussion if you want.