A Brief Sketch of the Brain
What's the scale of things here?
The following lengths (from Posner
p. 305) give approximate sizes for structures in the nervous system:
0.001 mm: synapses (tip of a connection between neurons)
0.1 mm: neurons (brain cell)
1 mm: local circuits (small networks of cells)
10 mm: maps (spatially organized topographic maps)
100 mm: systems (e.g., the visual system)
1000 mm: the central nervous system (including spinal cord)
How many things are we talking about?
Short answer: a LOT.
Long answer: Per cubic millimeter (mm^3), there are about 10^5 neurons
and 10^9 synapses. For you computer types, that's about 2^16 neurons
and 2^30 synapses. It's estimated that there are around 10^12 (or 2^40)
neurons and 10^15 (or 2^50) synapses in the nervous system. A typical
brain cell receives inputs from thousands of other cells, and the
influence of each connection is 1-5% of threshold -- that is, only 1-5%
of what the cell needs to respond.
How fast does the brain work?
Not very fast by computer standards. An action potential (i.e., nerve
impulse) lasts about 1 ms (millisecond). Axons, the long output
connection from a cell, come in two basic types: myelinated and
unmyelinated. Myelinated axons have an extra layer of
"insulation," which allows the action potential to travel about 10 to
100 meters per second. Unmyelinated axons are slower, transmitting at
only about 1 meter per second. When the signal reaches the end of the
line, it has to cross the synapse to influence the next cell; this
process takes about 5 ms. The effect can last from a millisecond to
many minutes, depending mostly on the type of the synapse. [Posner p.
How are memories stored in the brain?
This is one of the great questions of neuroscience, and research has
nearly converged on an answer. Short-term memories -- those which
last a few minutes or hours -- may be stored in a variety of ways,
including protein activation and inactivation within neurons, or
simply cycles of neural activity. More important to uploading is the
issue of long-term memory. It appears that long-term memories are
stored by structural changes in neural processes. These changes
include the number of branches a neural process makes and the number
and efficacy of synapses. Byrne et
present an excellent overview of recent work documenting these changes
in Aplysia (a marine mollusc). There is currently no
reason to believe that memories in humans operate by a different
mechanism. [It may also be speculated that these same morphological
characteristics are responsible for other aspects of personality,
though this has yet to be demonstrated.]
OK, so neurons and connections, is that it?
Well, there's more in the brain than that. There are also various forms
of "glia" cells (meaning "glue"), which serve mostly support functions.
The most important of these, from a processing standpoint, are the
astrocytes; astrocytes may affect information processing by regulating
the neuronal microenvironment, by slow signaling mechanisms (e.g.,
calcium waves), or by other poorly understood processes.
Also, there are all sorts of
hormones and other chemical interactions in the brain which definitely
do have an effect on the mind; these chemical interactions would need to
be simulated for the upload to be accurate.
Facts and Figures are available from the
University of Washington.
brainfacts.html . . . . . . . .
2/20/98 . . . . . . .