From: hsr4@vax.oxford.ac.uk
Subject: Re: VR Oneupmanship ?
Date: 1 Nov 91 19:12:56 GMT
Message-ID: <1991Nov1.191256.2559@vax.oxford.ac.uk>
Organization: Oxford University VAXcluster


In article <1991Oct26.034745.12560@milton.u.washington.edu>, dall@hfrd.dsto.
oz.au (Ian Dall) writes:

> In article <1991Oct18.193925.2368@vax.oxford.ac.uk>, hsr4@vax.oxford.ac.uk 
> writes:
>
>> [NMRI as an input device...]
> 
> I'm told that having an NMRI scan is not a pleasant experience. Evidently
> those intense electromagnetic fields can be felt!

A few years ago, when I got involved briefly with NMR (MRI as it was), none
of the human subjects reported any ill-effects; the only discomfort came from
the aluminium formers used to clamp the arm in place.

What DID happen and caused much consternation was a presentation by a 
researcher about his use of intense magnetic fields to isolate malarial
parasite-infected red blood cells (the presence of the parasite causes a
change in the chemical structure of haemoglobin, rendering the iron content
amenable to magnetisation).  This did not bode well for clinical use since
the subject might have experienced localised blood clots resulting in
fairly hairy consequences.  Eventually the facts were ascertained and the
episode was resolved (I cannot remember exactly how - I have a feeling that
the blood was passed over wire  wool in the presence of a strong magnetic
field and the affected cells remained trapped in the wool).

> How ever I had a similar idea. NMRI is used to image the structure of the
> brain based on its chemical composition. I kind of doubt that the way
> actual transmitter substances are there in enough quantity to detect
> indvidual neuron firings however. In short, I don't think you could
> detect though processes with NMRI.
> 
> What is needed is a non invasive means of measuring the *electrical*
> state of neurons. I think that this must be "in principle" doable with
> a large enough array of sensitive enough detectors. What I find mind
> boggling is the amount of computing power needed to process the output
> of the sensor array. I think that maybe the way to go is to "focus" on
> a small area of the brain which a person could consciosly (perhaps with
> training) control.
> 
> Of course, this would be a nice interface whether or not it is used for VR.

What I had in mind (and it is dependent upon further research) was something
akin to a skullcap with a matrix of copper coil pickups linked to sensor
hardware.  The controlling computer could then identify general areas of
activity in the brain of a subject while that subject is given a series of
specific tasks to perform, said tasks being associated with activity in
specific sites.  I hadn't considered identification of individual neuron
firing since it is probable that the brain's functions are not identifiable
with single neuron activity - it's more likely that multiple firings at
multiple sites achieve the effects we observe (I'm not an expert).

The cap wouldn't require direct contact with the skin or with conductive
jelly (no shaving the bonce for a VR experience, and no need to take your
shampoo into the shower...:-)).

+--------------------------------------+---------------------------------------+
| Peter G. Q. Brooks HSR4@OX.VAX.AC.UK |                                       |
| Health Services Research Unit        |                                       |
| Dept of Public Health & Primary Care |                                       |
| University of Oxford                 |                                       |
+--------------------------------------+---------------------------------------+
| +44 (0) 865 224375  8.30 am - 7.30pm |                                       |
+--------------------------------------+---------------------------------------+