From: Adrian Tymes (wingcat@pacbell.net)
Date: Sun Jun 16 2002 - 23:35:56 MDT
Smigrodzki, Rafal wrote:
> Some time ago I had an idea which should help: make electrodes in the
> shape of pigtails, or helical springs.
[etc.]
Not bad. Looking forward to seeing it in lab tests
> (the
> current Dobelle implant contains a pedestal open to the outside,
> traversing the skin, which in practice could pose the risk of infection).
Which is more of a problem than most literature acknowledges. Long term
but nearly constant exposure to minute amounts of electricity will break
down the skin. There is a simple solution, used by most devices
seriously intended for long-term implantation (like pacemakers): use
induction, not conduction, to transmit power and data. No skin
breakdown and no skin penetration means no risk of infection.
> Any other down-to-the-cortex suggestions for hot-wiring our brains?
There are already silicon chips on the market which one can thread
neurons through, or implant and let neurons grow around. They can sense
neural pulses, and excite the neurons on command (which, with the proper
software - which turns out to be more difficult than a Small Matter Of
Programming; indeed, it's one of the main problems in this field that
people are trying to solve today - can mean firing the neurons on
command). I'd say they work well, but then I am biased: they seem to be
based - in *very* small part, granted - upon the results of my own
(brief) academic career. They could also stand to be much more densely
packed: neurons are single cells, on the order of a human hair in
thickness, yet the holes in these chips for neurons are rather wide.
http://www.trnmag.com/Stories/2001/120501/Nerve-chip%20link%20closer_120501.html
This becomes a problem for, say, using these as silicon retinas.
Pulling numbers out of thin air for purposes of demonstration...imagine
the human eye sees at about 2048*1024 pixels at true color (24 bits per
pixel). Now imagine being forced to see at only 32*32 black and white.
(I pulled those numbers from thin air, but on review, they actually seem
about right...)
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:14:51 MST