machine brains

ray scanlon rscanlon at wsg.net
Mon Sep 15 11:10:35 EST 1997


As a first step toward design of a machine brain, I look at the
vertebrate brain.

--- ---   

             How the Brain Comes to Think

"What is thought except a movement that is not connected to a
motor neuron."
                              Attributed to Walle Nauta

"It is remarkable that there is no comprehensive theory of how the
brain works."
                                          Per E. Roland

Abstract: I construct a comprehensive theory of how the brain
works and show how a movement comes to be disconnected from a
motor neuron. I examine the vertebrate and in particular the
mammalian brain. I describe the brain as a neural net.

                   --- Prologue ---

When the brain is explained it is seen to be simple and easily
grasped. The explanation is not earth-shaking. It may be likened
to the game of chess. To explain the game we describe the board
and the pieces, we list the rules. This completes the explanation,
one piece of paper suffices. Endless books have been written (and
more to come) about the intricacies of play, games between masters
may be quoted, but one sheet of paper holds the explanation.

In the same way the explanation of the brain is also seen to be
simple. We describe the neurons and the larger tracts, we list the
rules of axonal growth and synaptic modification. That neither of
these is completely known is of no moment, the main picture is
clear and the details shall be filled in later. With the overall
plan in mind we can decide which details are important and which
are less so.

A distraction is the relationship between mind and brain, between
the soul and the body. We leave this to the philosophers. The
brain will work exactly the same whether it is associated with a
soul or not.

My objective universe is a closed system. Energy and mass interact
with no help from soul. This simple position is easy to state,
easy to believe, but we may maintain it only with great difficulty
when we speak of the brain. Men find themselves in need of an
homunculus, a little man who sits in the center of the brain
watching a television screen and punching buttons. The more
sophisticated postulate a soul (mind) that selects from the data
proffered by the brain, manipulates this data, decides, and
forwards the decision to the brain for execution. In a physical
universe there can be no homunculus.

Any talk of biological organisms must base itself on evolution and
homeostasis. Life is reducible to matter and motion. The particu-
lars of cellular life follow as DNA begets RNA and RNA begets
proteins. By evolution I mean that DNA replicates itself imper-
fectly and is subject to mutation.  By homeostasis I mean that the
instantaneous protein surround of the DNA reacts to a fluctuating
universe to produce a constant milieu about the DNA.

We may expect no "Why?" but only "How".

As the fetus grows, a nervous system appears. The genome exists
today because in the past it constructed a brain that functioned
at birth to preserve homeostasis in a hostile universe. Organic
chemistry leads us to believe it shall do so again. The genome
wires the brain to move the organism toward food, water, and a
mate, to find a secure resting place, and to avoid the predator.
As a group, these wired drives satisfy the larger homeostasis of
protecting the DNA from an unkind world.  The molecules of each
individual neuron rearrange themselves to keep what is in effect a
running average of the activity of afferent neurons. These running
averages alter the wired responses of the brain to incoming signal
energy.

All animals begin their lives with a neural system that can
maneuver them through the milieu that previous generations of
their species experienced at birth. Some hit the ground running
while others depend on their parents. The genome constructs a
fully functioning brain that can direct the body so it may survive
in a hostile environment.

It is we who assign purpose to the brain. The DNA molecules know
nothing of goals, they survive as they survived.

               Description of the Brain

I am interested in the human brain as a part of the physical
world. The brain is a neural net (NN) composed of sensory neurons
that receive signal energy from sense organs, motor neurons that
transmit permuted signal energy to muscles, and interneurons that
connect the two and permute the signal energy.

The most primitive NN known appears in certain jellyfish, they
have sensory neurons that are connected directly to motor neurons.
Their motor output is simple, as simple as turning on an electric
light. In certain other jellyfish a more advanced NN appears,
interneurons are interposed, they connect sensory neuron to motor
neuron but also connect with each other. No further advance in
NN's is aknown.

The neurons in the human NN are segregated and clumped into
nuclei, such an organization invites analysis. It is important to
know that all vertebrate NN's share the same architecture, all
mammalian NN's share the same floor plan. Mouse, Rat, Man, it is
all one; they differ only in neuron count.

The NN is wired (by the genome) to cause the organism to approach,
to seek. That which is approached I call the good.

The NN is wired (by the genome) to evade, to retreat. That which
is evaded I call the bad.

The NN is altered by the milieu.

We do not start our investigation at the top with an NN that has
the ability to perform a mathematical demonstration. We start at
the bottom and show how a NN survived in the world and evolved
into a structure that performs such mathematical demonstrations as
an idle by-product.

               The Source of All Things

First hypothesis: The brain stem contains the circuitry that
activates the neurons in the forebrain that seek food, water, and
a mate. The brain stem also contains the circuitry involved in
eating, drinking, and the sex act. The neurons of the reticular
formation form an alarm system; they watch for the predator.

Observation

The reticular formation in the brain stem, the size of a little
finger in man, is the prime alarm. The genetic residue of the
original pre-vertebrate brain, it retains its pre-eminence. This
primal brain monitors the environment for any evidence of the
approach of a predator. All sensory input (except olfactory) falls
indiscriminately on neurons of the reticular formation; they
habituate to the environment. This puzzles neurophysiologists; how
is a neuron to integrate these diverse afferents except as noise?
Noise it is, but it is noise with a level and the level is the
message. Evolution has written in the genome that any change may
be followed by sudden death. Upon such a disturbance the reticular
formation halts motor output, alerts the forebrain, and directs
sensory organs toward the source. If it is asleep, alerting the
forebrain wakes the mammal. The forebrain evaluates the threat and
evades the predator. The DNA knows nothing of death, but it
survives because it produces a structure that is attentive.

Teleological insert

We create the object: quarks, sub-atomic particles, atoms,
molecules, synapses, neurons, columns, tracts, ganglia, neural
centers, nuclei, brains. Having created the object, we say that
this object does things for a reason. If I look at the neuron, I
say that it sums the activity of afferent synapses. If I look at
the reticular formation, I see the neurons acting as a group. I
say they have a function, a purpose. The molecules involved say
that they are obeying the fundamental laws of physics.

end insert

Second hypothesis: The neurons of the reticular formation turn off
the forebrain during periods in which there is more chance of
being eaten than of eating.

Conjecture: The reticular formation contains two groups of neurons
that make up a clock. This clock gradually phase-locks with the
sun during the first year of life. Diurnally a rostral group
dominates and puts the forebrain to sleep and, alternately, a
caudally located group dominates and wakes the forebrain. These
two groups of neurons form a flip-flop; they alternately inhibit
each other.  

Teleological insert


All the neurons are alive and fully functioning always. I see a
neuron as dynamically active along a scale that runs from bursting
to quiescent, the neuron sees only afferent synapses. When I say a
neuron, or several neurons, "activates" another, I speak of its
function as part of a brain. The conductor of an orchestra and a
member of the audience are both alive but somehow we wish to say
the conductor is working and the auditor is not. We wish to say
that the conductor is creating music although he plays no instru-
ment. We wish to continue this distinction even if the auditor
makes small movements in time with the music.

It would be a basic error, however, to think that a neuron turns
on and off like an electric light. What does happen is that the
molecules rearrange themselves to alter the probability that a
neuron will fire during a particular interval. Exocytosis is the
release of a packet of neurotransmitter following the arrival of
an axonal pulse. It occurs with probability. We do not know how
the molecules alter this probability.

The neurons of the brain stem make the neurons of the forebrain
ineffective by stopping the inhibition of the neurons of the
reticular nucleus of the thalamus that in turn then inhibit the
neurons of the lateral and medial geniculate bodies and of the
ventral posterior nucleus of the thalamus. The reticular formation
disconnects the forebrain by cutting off its sensory input. 

I say the reticular formation 'turns off' the forebrain. That
sleep accompanies this 'turning off' is irrelevant. We have ruled
out all questions involving awareness other than noting them.

end insert

Observation

The forebrain, as activated by the brain stem, evades the preda-
tor, seeks food, water, a mate, and finds a secure resting place.
The brain stem handles eating, drinking, and the sex act. At other
times, the mammal retreats to its burrow and the reticular
formation turns off the forebrain to immobilize the body and
conserve energy. This is the foundation of a foraging strategy
that resides in the genome. The DNA that balanced its energy
equation survived.

All the things I think of as living, as wanting to get about in
the world, originate in my brain stem. In humans the forebrain has
so flowered that it apparently functions by itself, but this is
not true. Without an active constellation of neurons in the brain
stem there is nothing.

When the reticular formation turns on the neurons of the forebrain
by lowering their thresholds and by inhibiting the reticular
nucleus of the thalamus, the neurons of the thalamus and the basal
ganglia open to the universe. Signal energy now passes through the
thalamus, filters through the neocortex, the striatum, and the
globus pallidus, being permuted into a motor program that is fine
tuned by the cerebellum.  After passing through the ventral
anterior and ventral lateral nuclei of the thalamus and the pre-
motor and motor areas of the neocortex, this motor program drives
the muscles.

Teleological insert

A motor program is a sequence of activated neurons that we see as
a process. We see it headed for muscles that when energized will
perform an action that has meaning to us. I liken it to the
perforated paper roll that drives a player piano. The music lies
in the manner in which the holes are sequenced. Similarly, the
motor program exists only as a sequence of activated neurons.

end insert

Man sleeps in remembrance of his evolutionary past when he was not
nearly so high on the food chain. These two functions of the
reticular formation, watching for the predator and providing the
master foraging strategy, are essential to life. Take them away
and the species disappears. We can survive without metaphysics but
not without the reticular formation.

The thalamus acts as a peripheral unit, sometimes on (awake) and
sometimes off (sleeping). The neurons of the reticular formation
control the action. When they call for movement, there is move-
ment; when they call for sleep, there is sleep.

(Note: These are the first two of six hypotheses.)

ray

-- 

email: rscanlon at wsg.net

If you are interested in how the brain works, visit
http://www.wsg.net/~rscanlon/brain.html







More information about the Neur-sci mailing list