IUBio

question: job of a neuron (emergent behavior)

Richard Norman rsnorman at mediaone.net
Sat Sep 8 15:48:32 EST 2001


On Sun, 9 Sep 2001 01:22:14 +0800, "yan king yin" <y.k.y@(no spam
please)lycos.com> wrote:

>"Matt Jones" <jonesmat at physiology.wisc.edu>:
>> In the study of cellular automata (sometimes called artificial life)
>> there are a lot of interesting so-called "emergent" behaviors that
>> populations exhibit, such as flocking of birds, ants following
>> coherent trails, stuff like that.
>>
>> Turns out these behaviors, that are only visible at the level of the
>> whole population, actually arise from very very simple rules followed
>> by each individual. For example, "Head toward the center of mass of
>> the flock, but if another bird gets too close then change direction
>> randomly." In these systems, the entire population behavior can be
>> understood -completely- by knowing what rules the individuals follow.
>>
>> What reason do we have to think that brains don't operate in similar
>> ways, using very simple rules at the level of each neuron to determine
>> the extremely complex behavior at the top?
>
>I have been thinking that the brain has emergent behavior at the level
>of neurons. But there is a problem with this idea: The pattern of
>innervation inside the brain is specified by genes and is stereotypic,
>ie not varying among individuals. For example the corpus collosum
>connects the 2 hemispheres, and some genes that code for the
>commissure has been identified in the fly. It seems that these patterns
>of connection serve specific purposes and they are not random.
>
>If it does not matter how the hemispheres are connected, then we
>would expect to find different innervation patterns in different people,
>but the fact is they are all anatomically similar.
>
>In other words, if there is emergent behavior, the brain should look
>like chaotic, but it is not (when compared among individuals).
>
>It seems that the brain is hard-wired this way in order to maximize
>intelligence and adaptive behavior. It could also be for historical
>reasons from the phylogeny.
>
>On the other hand some aspects of innervation are activity-dependent
>(eg ocular dominance) and therefore emergent.
>
>Is the wiring of the brain also emergent -- determined by a small
>number of mechanisms?
>

The general "wiring diagram" of the brain may be coded in the genes,
but certainly not the fine structure.  Consider fingerprints, for
example.  Identical twins with the same genes still have different
fingerprints because these are determined by localized and
unpredictable (sort of "random") environmental cues at the size level
of a millimeter in the development of the finger ridges.  At the same
size level in the brain, in one cubic millimeter, there are thousands,
perhaps tens of thousands of neurons each synapsing with many
thousands of neurons in its vicinity.  Certainly the 40,000 genes in
the human genome are not going to code for specific neuronal
connections!  But the large pathways with dimensions of many
centimeters (to a specificity of a few millimeters) may be determined
genetically.

There is also a great deal of evidence to indicate that a tremendous
number of neurons form and interconnect with each other initiallly.
Later experience whittles down (sculpts) these connections and cell
numbers to a smaller but functionally important number.  Again, the
fine details may be determined by experience even though the overall
grand scheme "determined" by genetics.

>From what we know of development, a small number of rules can produce
a highly organized pattern of structure.  Since this organization
occurs because of the interaction between cells, not because of the
separate activities programmed into each separate cell, you could well
call that an example of emergent properties.




More information about the Neur-sci mailing list

Send comments to us at biosci-help [At] net.bio.net