Biomedical Engineering Reference
In-Depth Information
ARE NEURONS THAT SIMPLE?
In fact, it constitutes a most interesting case as
it applies to several neuron types in the cortex
of mammals. One may argue that the few action
potentials fired by these near-silent neurons have a
strong informative load and they are signaling the
recognition of particular combinations of inputs
bearing a strong physiological meaning. While this
is reasonable, some questions follow. If only some
input combinations produce outgoing spikes, do
neurons consider irrelevant the vast majority of
their inputs, which cause no output? Would that
mean that the fate of most presynaptic spikes is to
become useless noise in brain circuits? If so, why
bother that much to produce them anyway? And,
is the postsynaptic cell structure or the specific
assortment of electrogenic machinery genetically
assembled to make the neuron recognize these
critical combination of inputs that initiate a few
outgoing superspikes ? From the point of a working
neuron, these and many other similar questions
can be reduced to two major questions. First,
how the thousands of inputs are selected within
dendrites to end in a few outgoing spikes? And
second, is there any benefit or function in the
remaining inputs failing to produce output? The
answers have to be found in the computational
operations performed in the dendritic arboriza-
tion of the neurons. Formerly viewed as a black
box, we shall try to introduce ourselves within
the dendritic apparatus to unveil the secrets of
this miniature communication center.
The quintessential feature allowing any process-
ing of electrical signals within individual neurons
is that their narrow elongated dendrites make them
electrically non-uniform, enabling long electri-
cal distances between different parts of the same
neuron. Whatever electrical events take place
in the dendrites, they are not the same as in the
soma or the axon. These compartments are not
totally isolated and signals travel from one another
as in an electrical circuit. From early studies,
neurophysiologists concluded that neurons were
essentially input/output devices counting synaptic
inputs in their dendrites, a numeric task that trans-
form into a temporal series of action potentials or
spikes, each one triggered when a certain voltage
threshold is reached by temporal summation of
the synaptic currents at a specialized region near
the soma-axon junction. This simple working
scheme enables the transformation of myriads of
inputs scattered in a profusely branched dendritic
arbor into a temporal sequence of binary axonal
spikes that can be read by target cells in terms
of frequency. Indeed, frequency modulation of
spikes may perform with equivalent accuracy as
the fine tuning capabilities of graded communica-
tion that prevail in non neuronal cells. For a while,
researchers felt this was good enough for a single
cell, and the uncomplicated view of dendritic trees
as not-very-clever receiving black boxes settled
in for decades. This classical picture is, however,
incomplete. Such a basic scheme holds only in a
few neuron types (if any).
Typically, neurons receive several thousand
synaptic contacts, each conveying information
from as many afferent cells. In order to under-
stand how such a huge amount of inputs are in-
tegrated, it is useful to examine the output first.
Lets have a look to it. The output range across
neuron types in the Nervous System goes from
regularly firing (pacemaker-like) to almost silent
neurons. The later type fire spikes at extremely
low rates lacking any apparent temporal structure.
CHANNELS IN DENDRITES:
A CONCEPTUAL REvOLUTION
Based on modeling studies, some authors, sug-
gested that the overall firing pattern is somehow
engraved in the architecture of the dendritic tree
(Mainen et al., 1995). Indeed, some architectonic
dendritic prototypes have been preserved in spe-
cific brain nuclei spanning long periods of brain
evolution. Although this may be indicative of a
common integrative function being preserved
Search WWH ::




Custom Search