From: doug@eris.berkeley.edu (Doug Merritt)
Subject: Re: Ultimate input hardware
Date: Tue, 12 Feb 91 16:56:53 GMT
Message-ID: <1991Feb12.165653.9655@agate.berkeley.edu>
Organization: University of California, Berkeley



In article <1991Feb8.164913.7787@umiami.ir.miami.edu> rdees@umiami.ir.miami.edu 
(Matthion) writes:
>
>First, I should point out that several people (whose opinions I deeply respect)
>have come to the conclusion that such direct neural interfaces are
>[im]possible due to the intrinsic variability between people.

First, how about getting those people to give us some good bibliographic
references so that we could understand the nature of that problem?

Second, I assume that some such intrinsic variability exists (it would be
an incredible stroke of luck were it otherwise), but that's ok...I'd be
quite happy to have a direct neural interface that required (say) a
month of training to begin using. Or even a year, if that was the best
we could do. This is not at all a reason to give up on it.

Note also that interfaces that are unique to a particular individual
might have a side benefit of giving inherent casual security to that
unit.

>Although the systems are chaotic in the sense that as you go higher patterns
>emerge, attacks at the root level are still attacks on a random system.

I'm not sure what "higher" and "root" refers to here; can you clarify?

>As for looking deeper in the brain, do we want to?  As you descend into the
>brain, you move backwards in time, through all the mammals, then reptiles, then

Yes, we probably would want to. For instance, if you wanted to get a
computer image of exactly what the subject was looking at, it's quite possible
that this information is available only in deeper, older parts of the brain.
(Unless it just happens to be available on the surface part of the visual
cortex.) If you wanted to track the precise emotional state (mood) of the
subject at all times, this almost certainly would require monitoring the
deeper parts of the brain.
--
Doug Merritt            doug@eris.berkeley.edu (ucbvax!eris!doug)
                or      uunet.uu.net!crossck!dougm