skaggs at bns.pitt.edu
Mon Mar 17 19:25:25 EST 1997
chouse at uoguelph.ca (Chris J House) writes:
> [ . . . ]
> Secondly, the number of synapses do not relate to memory. Memory, for
> the sake of this argument, is located within the hippocampus lets say.
> Well, there is a huge portion of the brain that we will not be adding up
> synapses in if we were to only contend with the idea of 'how much memory
> does the brain have in computer terms'.
> [ . . . ]
What you say is correct, but the situation is really not as bad as you
make it sound. The thing that simplifies it is that one particular
type of synapse accounts for at least 80% of the synapses in the
brain, namely, glutamatergic synapses onto dendritic spines of
pyramidal cells (or spiny stellate cells) in the neocortex. There are
about 10^14 of these synapses in the human brain, and the total number
of synapses of all other kinds is at least an order of magnitude
smaller. All that is actually necessary, then, is to assume that this
one type of synapse is capable of storing memory -- quantitatively,
the other types just don't matter. Of course, if this assumption
turns out to be wrong, then the calculations will have to be redone.
Functionally, of course, the other types of synapses are critical.
Neglecting them for purposes of calculating the total memory of the
system is analogous to looking only at the hard disk when calculating
the memory of a PC.
I agree completely with the final paragraph you wrote: thinking about
a brain solely in terms of storage and memory is unreasonable. But so
is thinking about a computer that way! It doesn't change the fact
that storage and memory are important parameters.
More information about the Neur-sci