Brain Capacity?

Bill Skaggs skaggs at bns.pitt.edu
Wed Mar 12 11:29:24 EST 1997


goble at kigateway.eastend.com.au (David Goble) writes:

> 	Here is a Exerise question and its answer (Structured Computer
> Organization 3rd, Prentice Hall Internation editions, Andrew s.
> Tanenbaum)
> 
> Chapter 2. Exreise 5. page 73
> 
> Q5.	Estimate the maxium storage capacity of the human brain using
> the following assumptions. All memory is coded as DNA molecules. A DNA
> molecule is a linear sequence of the four basic nucleotides: A, C, G,
> and T. From the average weight of a nucleotide, roughly 10^-20 grams
> and an average brain weight of 1500 grams, deduce the bit capacity of
> the brain for this encoding form. Note: This calculation is only an
> upper limit, because the brain contains many more cells that perform
> functions other than memory.
> 
> A5. (Total weight of brain / Total weight of nucleotide) * No. of Bits
> 	1500                  /         10^-20                      ) * 4
> 		     1.5e+023                                      * 4
> 				= 6.e+023


Notice that there is no claim here that the assumptions are
reasonable.  In fact, they're absurd in many ways.  The fact that the
answer comes out to Avogadro's number makes me suspect that this
exercise was composed with tongue slightly in cheek.  (Actually,
though, the solution given above is incorrect, because if there are
four possible nucleotides, each one only contains 2 bits of
information, not 4.)

Incidentally, for readers who don't know this story, there was a time,
back around the 1960's, when many psychologists believed that memory
is coded as RNA molecules (not DNA).  There were a number of
experiments that purported to provide evidence for this, which later
could not be replicated, and a journal called "The Wormrunner's
Digest" was devoted to experiments along these lines.  The title
arises from the procedure that many of these experiments used:  they
would teach worms to perform some simple task (i. e., curl up when a
light comes on), then grind the worms up and feed them to other worms,
and see if the cannibals were quicker to learn the task than the
original worms were (which would indicate that memory was encoded in a
way that could be transferred in ground-up tissue).  The journal
ceased publication sometime around 1970, I think.

As far as I'm concerned, it was never a plausible theory to begin
with.  Certainly an RNA-based memory system could hold a tremendous
amount of information, but how in the world could it ever be
recalled?!  The process would have to involve some kind of
transcription, and there's no way anything like that could possibly
occur in the few hundred milliseconds needed to recall something from
human memory.

	-- Bill Skaggs



More information about the Neur-sci mailing list