question: job of a neuron

Matt Jones jonesmat at physiology.wisc.edu
Mon Sep 3 13:53:03 EST 2001

remove_this!helbrecht at gmx.net ("Wolfram") wrote in message news:<9mroi3$3pcpu$1 at ID-40201.news.dfncis.de>...
> Hello everyone,
> in the field of neuroscience i am layman. I have no clinical
> background nor nothing else. Just a person (a bit) interested in
> the field.
> Everywhere on the web i can read pages about _how_ neurons work,
> _how_ they interact with others, _how_ the whole brain works (some
> pages assert this).
> Has yet someone figured out _what_ a neuron does, i.e. what its
> task is?

Boy, didn't -this- question become controversial all of a sudden!

I see a lot of people don't think it's very meaningful to ask what a
single neuron does, and draw the analogy with asking what a particular
wire or electronic component does.

Me, I think it's a perfectly reasonable question. For example, what
does a wire do? It does the following (ideally):

1) Conducts current with zero resistance.
2) Acts as an isopotential element (i.e., causes anything connected to
its two ends to be at the same potential.
3) In non-ideal cases, it may get hot, or even give off light. This
can sometimes be useful.

How 'bout a transistor?

1) It acts as a variable conductance, allowing current to flow if
certain conditions are met at its base (or gate).

An op-amp?

1) It produces an output that rises until its two inputs attain
equivalent potentials.

That doesn't seem so hard. Why not a neuron?

1) It takes some number of inputs, each of which occurs at a
particular time and  has a particular shape and a particular size, and
performs -some- sort of operation on those inputs. Then it produces
-some- sort of output (including the case of no output at all).

That's -what- a neuron does, at least in the most general sense. The
details include such worries as -what sort- or operation, and -what
sort- of output. These details are currently in dispute among
neurophysiologists. For example, some proposed operations are:

1) Summing, or integrating, or integrating in a leaky fashion. This
operation is most often supposed to operate on the membrane potential
(i.e., summing IPSPs and EPSPs), but could just as easily work on
membrane current or conductance.

2) Detecting coincidences.

3) Multiplication (actually, division).

4) Nonlinear operations, such as detecting a threshold-crossing in the

Some of the proposed outputs are:

1) A certain rate of spiking.

2) A certain pattern of spiking.

3) In the case of both of the above, as well as for nonspiking
neurons, a release of some number of packets of neurotransmitter.
These can be either "excitatory" or "inhibitory", but in either case
they contribute either to electrical or chemical changes in
postsynaptic cells.

So these are a brief, definitely nonexhaustive list of things that a
single neuron is -capable- of doing. Just as with a wire, or an op
amp, they can act as building blocks for more complicated tasks.
Nonetheless, whatever the task, or whatever the context, the final
circuitry is made up of things that do some combination of the above.

I think it makes perfect sense to ask about this. If we didn't -know-
the answer to this question for a wire, how much success would we have
figuring out how a toaster works? If we can't answer it for a neuron,
how are we ever going to figure out how the -brain- works?



More information about the Neur-sci mailing list

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