From Tom Bourbon [931021.0840]
In Message Wed, 20 Oct 1993 18:34:55
Mark William Olson <mwog8061@UXA.CSO.UIUC.EDU> writes about physiological
work by Apostolos Georgopoulos and his associates.
This message is more important than the last one I sent about A.G.'s work.
A.G has found that individual neurons file preferentially and broadly for
movements in particular directions. No big surprise there. The
interesting thing is that if we look at the firing patterns of these same
neurons when the arm must move an upright lever with biases of force in
some random direction, the neurons do NOT fire in such a way that would
indicate a vector analysis of the lever's torque; the neurons still fire
for the desired direction of movement. In other words, the neurons that
are most activated are NOT the ones that one would expect if one were to
figure out what direction one must move to get to point A given the torque
of the lever. The neurons which were most active before (without the
lever) are the neurons most active with the lever.
Sounds like an excellent candidate for a reference signal if I ever saw
one. These neurons are in primary motor cortex--I don't remember which layer.
Mark, Georgopoulos has some interesting data, and he seems "close" to
characterizing them as what we call reference signals. In part due to his
use of colorful graphics, his work frequently makes the cover of _Science_.
The article closest to the work you describe is by A.G., J. Ashe, N.
Smyrnis, and M. Taira, 1992, The motor cortex and the coding of force,
Science, 256, 1692-1695.
Selecting from the abstract for that article, "The relation of cellular
activity in the motor cortex to the direction of two-dimensional isometric
force was investigated under dynamic conditions in monkeys. A task was
designed so that three force variables were dissociated: the force exerted
by the subject, the net force, and the change in force." Each of those
three variables was resolved further; for example, net force = subject
force + force bias. "Force bias" was what we would call an environmental
disturbance -- an arbitrary force in one of several other directions.
"Recordings of neuronal activity in the motor cortex revealed that the
activity of single cells was directionally tuned and that this tuning was
invariant across different conditions of bias force. Cell activity was not
related to the direction of force exerted by the subject, which changed
drastically as the bias force changed." In other words, to produce the
intended movement, a monkey uses different force vectors that depend on the
direction of a disturbance (bias force). When that happens, the activity of
cells in the motor cortex, which traditionally are assumed to command
specific muscle actions, is *not* related to the direction of force exerted
by the monkey (when it overcame the effect of the disturbance), but to the
direction of the *net* force (subject force + bias force). The result? The
monkey moved a hand in the specified direction. " . . . the direction of
net force, the direction of force change, and the visually instructed
direction all remained quite invariant and congruent and could be the
directional variabnles, alone or in combination, to ehich cell activity
might relate." That does sound something like a "reference signal" doesn't
it, Mark?
But if A.G.'s work is to merge with PCT, there are some elements missing,
especially on the perceptual side. For example, the paper concludes with
speculations about how static and dynamic components of the subject's force
might be sensed independently and then combined in the nervous system. It
would be interesting to see if those components were not independent at all,
but were "sorted out" solely in terms of comparison with a reference signal
for direction. And it would be nice to see what happens when an animal
controls continuously and smoothly, rather than with "force pulses" like the
ones in the paper I cited.
Is there any chance you can work on a problem like that at UIUC? I believe
it is one of the two most promising areas to look for PCT-like phenomena at
the levels of anatomy and physiology, as they relate to observable control
behavior. (The other area is the work on Aplysia and other "simple"
creatures with well-mapped nervous systems.) Go for the big one, Mark!
Until later,
Tom Bourbon
Department of Neurosurgry
University of Texas Medical School-Houston Phone: 713-792-5760
6431 Fannin, Suite 7.138 Fax: 713-794-5084
Houston, TX 77030 USA tbourbon@heart.med.uth.tmc.edu