Return to the archive index

SMUT Digest reply [was another input device . . .]

Let me offer again to assist anyone who is interested in studying
single motor unit training.  I've personally gotten as far as
reproducing several analog experiments from the late 60's and early
70's.  I've only just started playing with digitizing.
-- 
La Monte H.P. Yarroll      Home:                
   If you remember nothing else:           NIC Handle: LY
   GPL - "Just give source a chance."

Gregory Martin Pfeil <> writes:
> I don't know if Mr. Yarroll has forgotten about this, but about two 
> months ago he sent me information about single motor unit training.  
> This made it very difficult to complete school successfully.  I was 
> trying to research this as much as I could and it resulted in lower
> grades this semester.

I'm happy to hear that you are not letting grades interfere with your
education.  :-)

...
> I haven't actually experimented with this yet, so I can't say if it
> really is as easy to control as I've heard.  If it is, one can type
> up to 240 wpm.

Run to your nearest Radio Shack and buy a TL082 and a big pile of 1K
and 100K resistors--enough for two matched pair.  Until you've played
with SMUT (Single Motor Unit Training) you'll have many mistaken ideas.

...
> The band should probably be some kind of mesh, maybe lycra, to keep
> it tight while being breathable.  What does everyone think?

There is no need for a band--the MedTronic nerve stimulation
electrodes are easy to apply and very stable.  Not adhesive electrodes
introduce really nasty motion artifacts.

"R. Paul McCarty" <> writes
> If I'm understand this correctly, it sounds like you would have to learn
> to control small groups of muscles in your arm individually, rahter than
> as a group when you move your arm around, and I'm having trouble imagining
> this to be easy to do.  How long does it typically take to train 12
> units? and what is the percentage of errors? Is this just using skin
> contact electrodes or is something inserted beneath the skin?

Adhesive surface electrodes.  Most of the liturature is based on work
with intramuscular electrodes (needles in the muscle), but for this
application we don't care which muscle cells our signals come from.

I have no data on error rates--"good enough" is the best I can offer
at this point.

Training 12 SMU's takes many hours of practice.  Most people can
isolate and control their first SMU within half an hour.  After the
first one it is a lot easier.

Gregory Martin Pfeil <> writes:
> ... Your arm would not move while "typing", but 
> you wouldn't have to hold it still either.  The reason is because there 
> are a large number of moto-neurons that aren't really used.
> ... you don't have to worry about neighboring ones triggering the
> electrodes.

No, you do use all your motorneurons--that's why you have them :-).
The training process consists of learning how to SUPPRESS the
neighboring motor units.  It is possible to control SMU's while
moving, but it is a bit more difficult.

> I haven't done any experimentation myself (yet), but apparently, training 
> 3 to fire between 1 and 20 times a second (with absolute precision), 
> takes a few hours.  It probably wouldn't take much longer to control more 
> if the rate isn't focused on.  Speed can be gained over time, as with any 
> typing method.

If you can't control the rate you can't control the SMU.  Build the
apparatus--try some experiments--you will understand.

Paul Archer <> writes: 
> That is a better explaination. I am just missing two things. One is
> exactly how you would provide input. In other words, would you chord
> several different individual sites, or would you make a site fire a
> certain number of times or at a certain rate to give different input.

Encoding methods are still an open issue.

> The other question is how do you train to conciously fire these sites?
> What do you need to get started?

You MedTronic nerve-stimulation electrodes (with integral gell), a
small instrumentation amplifier (see Basmajian&Hudson below), and an
audio amp.  Oscilliscope is optional.

Play with the analog stuff before you start to worry about digital
processing and information encoding.

Gregory Martin Pfeil <> writes:
> Well, basically, they are trained with electrodes and an oscilloscope.  

You need the audio feedback--a 'scope doesn't suffice.

Gregory Martin Pfeil <> writes:
> This is actually what I was considering.  A small tattoo over each area 
> that could be lined up with the electrode.  I don't know how many people 
> would go in for that though--computers aren't at the top of everyone's 
> list of things worth mutilating themselves for.

Electrode placement affects detected wave shape, but is not critical
for training purposes--once you have trained an SMU you can fire it at
will regardless of the exact audio feedback.

Andrew Edelstein <> writes:
> Another thought: would a change in one's physique ("bulking up" or the
> reverse) change the position of those marked spots in relation to the nerves
> they mark? If the diameter of one's bicep area changes, does that throw off
> the makrs? Would seem to make a good argument for semi-perminent, rather th=
> an perminent, marks.

This isn't an issue I've had to contend with :-).  You don't generally
grow new neurons, so I would not expect radical changes in enervation
patterns.  Enervation patterns can change in cases of severe muscle or
nerve injury.

"Felix S. Gallo" <> writes:
> It'll be interesting to see, though, not least because I just had an
> electromyographical procedure done on my arm which suggests that
> there's enough bandwidth through the nerves to do quite a bit.  If the
> nerves don't get really bothered by constant electrocution, even at
> low currents.

"Electrocution" is death by electricity--the word you want is
"electrification".

In theory there is no extra current passing through your muscles.  In
practice most apparatus I've built seems to produce a bit of back
current--I use DC coupling because the signals are so small.  I have
no ill-effects to report at this point.

"R. Paul McCarty" <> writes:
> He said its measuring muscles in the upper arm (bicep).  As far as I
> know the muscles that control the fingers are all in the forearm. :)

You are of course assuming that your correspondant has mammalian
anatomy.  :-).
Gregory Martin Pfeil <> writes:
> Yes, it is a passive device.  Should be completely safe.

Just the same, I encourage you to stick with 9v transistor batteries
for your initial experiments.

Thad Starner <> writes:
> The problem with EMG is that electrode contact is difficult to
> maintain in a mobile setting.

This is true for most electrodes I've experimented with.  The second
best are disposable EKG electrodes--they adhere well and produce
almost no motion artifact.  The only problem is that the snap
connectors break loose too easily (this is a design feature).

The MedTronic nerve-stimulation electrodes are conductive rubber with
an integral gell.  This gell has the remarkable property of sticking
even better when it gets wet.  I've worn these electrodes for a
full week without electrical degradation.

They have a horizontal tube on the back into which you insert a
contact probe.  The friction fit is sufficient for fairly vigourous
activity.

Bibliography:

Muscles Alive!  John Basmajian
  -- This is the Bible of Electromyography.  It includes a
     full chapter on SMUT and has a good bibliography.

Electromyography for Experimentalists,  Gerald E. Loeb and
Carl Gans, University of Chicago Press, 1986 ISBN 0-226-49015-7
  -- If Basmajian is the Bible, this is the Catechism.  This book is
     loaded with practical advice with enough theory to hold it all
     together.  My copy is wearing out.

Atlas of Human Anatomy for the Artist, Stephen Roger Peck, Oxford
University Press, 1951, 1982 ISBN 0-19-503095-8 (pbk.)
  -- Part II of this book gives detailed information about surface
     muscles.  This is very helpful for electrode placement and SMU
     training.

From Wear-Hard Mailing list Archive (WH)
Maintained by R. Paul McCarty

Archive created with babymail