chicks with prism glasses

[From Bill Powers (951204.1320 MST)]

Martin Taylor 951204 10:40 --

     I seem to remember a Scientific American article in which it was
     shown that chicks who were fitted with prism spectacles did not
     learn to compensate for the offset between where they pecked and
     where the grain was. Is this now not believed to be so? Or is it
     true only of chicks and not of other bird species or ages?

If a chick wearing prism spectacles strikes off-center, this is a clear
demonstration that the chick is accurately controlling a perception. If
it were just kinesthetically emitting a peck, it would continue to
strike the same place, not an offset place. The only real question is
whether the control of apparent target position occurs only prior to the
peck, or whether it continues during the peck.

In order to execute a peck, the chick must orient its head so the visual
image of the target is in a standard place in the visual field, probably
the center (not of the retina, but of the inner mapping). With the
spectacles on, the chick orients itself in what we can see is the wrong
direction, but in the direction that looks right to it. It then must
initiate the peck in the direction that keeps the target stationary. The
prism spectacles, however, create a situation where moving the head
toward the target using the existing organization (along the body's
centerline) will cause the target to move to one side during the peck,
at first slowly and then more and more rapidly as the head approaches
the target.

If there were continuous closed loop control based on the visual image,
the head would accelerate sideways toward the direction the target
appears to be moving. However, a peck occurs very rapidly and the
greatest acceleration occurs near the end of the downward stroke, so
this acceleration might not go very far toward correcting the error
before the strike is complete. To see if there is really active control
during the strike, the path of the head and beak would have to be
measured accurately. Force sensors under the feet might reveal the
reaction forces due to any sideward acceleration. Also, the displacement
of the apparent target from the initial point of view of the chick's
eyes could be computed. If the average miss-distance corresponded
exactly to the optical displacement, we could probably conclude that
there was no attempt at error correction during a peck. If the
displacement is less than the computed displacement, some attempt at
error correction occurred.

However, careful measurement might be necessary, because (depending on
the speed of the peck) the amount of correction actually accomplished
might be quite small. Normally, only a minor amount of correction would
be required, if the initial orientation were actually aligned with the
kinesthetically-controlled direction of pecking, so the control systems
would not normally have to make large corrections during the peck.

If the chick (like a frog or a praying mantis) relies completely on the
initial centering of the visual image, and simply strikes along the
body's centerline, then any control of the striking position would have
to be done over many pecks, by a higher-order sampling control system.
The reference condition for the orienting system would have to be
changed to "target such-and-such a distance to the side." From my
memory of these experiments, however, the chicks do not learn to strike
to one side of the apparent target to get the food even over many pecks.
This tells us that they lack such a higher-order control system, and
possibly it also tells us that there is nothing at that higher level to
reorganize.

In human beings reaching as rapidly as they can toward a target, a
similar effect occurs. However, if the target is shifted at the start of
the movement, there is a clear attempt to control based on the visual
error, but it is not very successful when the movement is made at
maximum speed. At slower speeds, control is continuous. I have a paper
on this somewhere in a box, done by a couple of Italians.

Chicks can obviously orient themselves to center a target in the visual
field; disturbing that perception results in exactly the expected action
to compensate for it. They can't compensate for a lateral visual
disturbance during a peck, but whether that is due to the fact that they
can peck only at one very fast speed which is too fast for the existing
control system, or whether the visual control system cuts off during the
peck (the way the human eye blanks during a saccade), I don't know. Does
anyone?

···

-----------------------------------------------------------------------
Best,

Bill P.

[Martin Taylor 951205 11:40]

Bill Powers (951204.1320 MST)

Martin Taylor 951204 10:40 --

    I seem to remember a Scientific American article in which it was
    shown that chicks who were fitted with prism spectacles did not
    learn to compensate for the offset between where they pecked and
    where the grain was.

If a chick wearing prism spectacles strikes off-center, this is a clear
demonstration that the chick is accurately controlling a perception.

A nice discussion indeed, of the pecking process!

But point of the Scientific American article, and of my posting, was
unrelated to the process of pecking. It was that the chicks "did not learn
to compensate", whereas humans, kittens, and perhaps other mammals do learn
to compensate in similar circumstances. Initially, when a person puts on
prism spectacles, hand movements are misdirected and may, as you point out,
not be correctible in the time of one attempted pick-up of an object. But
after quite a short time, things start to look more or less normal again,
and one's movements regain their external-world accuracy, more or less.

Chicks don't seem to learn, which suggests that the related control system
hierarchy is not easily reorganized, as compared to ours. They seem to
be born with it, or to learn it in the short time before being fitted
with the spectacles (I don't know whether they were given any such
opportunity).

Anyway, as I said, it's probably a side issue, and I don't know whether
even the facts are still thought to be true, because all I know of it is
what I read in that long-ago Scientific American article. It would be nice
to get an update from chick-peck-researcher, but not very important in
respect of elucidating HPCT.

Martin

Martin & Bill P.
        I believe that it has been found that a chick's pecking will adapt
to prismatic displacement if the displacement comes on gradually so that
the chick senses the slight displacement of the kernel with its beak.
Regards, Wayne

Wayne A. Hershberger Email: wah@niu.edu
Department of Psychology Fax: (815) 753-8088
Northern Illinois University Phone: (815) 753-7097
DeKalb, IL 60115