[From Bruce Abbott (951007.1110 EST)]
Bill Powers (951007.0600 MDT) --
Bruce Abbott (951006.2045 EST)
I would argue that there are times when the single-subject approach
cannot be applied and yet it would be of value to collect data
anyway. An example would be studies aimed toward understanding the
roles of various brain systems which manipulate levels of brain
biochemicals in experimental animals. I once participated in a
study that looked at the effect of alpha-MSH, a peptide hormone, on
separation distress in young chicks. ...
Question 1: How did they determine that it was "distress" that was being
affected?
The loud, regular, incessant peeping, coupled (in natural circumstances)
with wild dashing to and fro strikes the human observer as the very picture
of distress. It strikingly parallels the similar behavior of very young
children who have become separated from their parents at the mall. But do I
actually know what the chick is experiencing? No, of course not.
Question 2: If it was distress being affected, how did they determine
that the distress concerned "separation?"
"Distress vocalizations" (DVs) begin immediately upon separation from the
mother/brood and cease as soon as contact is restored.
So it is established that peeping and separation from the brood are
inversely related. How was it determined that peeping is related to
"distress" or "anxiety"? If the subjects had been human beings, you
could have asked them about their internal states, but chicks can't
report how they are feeling so there would seem to be no way to test the
hypothesis about what is being experienced.
Research in this area began with the discovery of imprinting in fowl.
Immediately after hatching, the chick shows no particular inclination to
follow its mother or broodmates around, nor does it appear at all disturbed
when removed from the mother and brood. Within a couple of hours, however,
a radical change occurs: the chick will now follow its mother wherever she
goes, maintaining a particular distance from her (or another chick) just as
in the CROWD demo. (I think we've all seen chicks of this age strung out
behind a mother hen like beads on a string.) If separated from her, the
chick races to catch up (if she is in sight) or (if she is not) begins to
emit DVs at high intensity. The function of these peeps, which are readily
distinguishable from the peeping chicks ordinarily produce when not
separated, is evidently to call the mother hen's attention to the chick's
plight and bring her to the chick.
Imprinting occurs when the chick is exposed to the appropriate conditions
during the first couple of hours of life. Those conditions include a nearby
moving object of more-or-less appropriate size being visible to the chick;
it also helps if the object emits sounds at least somewhat like the clucking
of the hen.
The more general term for imprinting is "attachment." Attachment is found
not only in certain birds but also in mammals up to and including humans.
In young human children the distress and anxiety that often follows
separation from the parents is probably something all of us have distinct
memories of. The parallels between these human responses and those of many
other animals strongly suggest an evolutionary continuity.
Question 3: how was it determined that the effect of alpha-MSH is
specific to the effect of the mirrors on vocalizations? I deduce from
your desciption that when alpha-MSH was administered, the experimental
group continued peeping, whereas the control group did not. Why was it
not concluded that alpha-MSG stimulates peeping, or stimulates anxiety,
or reduces the capacity to recognize other chicks, or increases the
desire to be with other chicks beyond what can be satisfied by proximity
to the number seen to be present, or causes hallucinations of other
conditions that also usually lead to peeping (such as hunger)?
My description of the study was given to make a point about between-subjects
methodology and hypothesis testing; for this reason I provided a simplified
description that omitted some important controls. The chicks were tested
both in mirrored boxes and in unmirrored boxes. If MSH stimulated peeping
or anxiety, it should increase the rate of DVs in the unmirrored boxes above
the rate found for control subjects. It did not. As for other
possibilities, some have been ruled out by other tests; others remain to be
investigated. Of course, it is impossible to rule out all possible
alternative explanations; the data simply may be taken to support some
reasonable interpretations.
Question 4: of the total number of chicks in the group that showed the
effect, how many did not show it? The reason I ask is that if the effect
of alpha-MSG were indeed specific, direct, and simply a matter of
biochemistry, every experimental animal should have shown the same
effect to the same degree, within some small scatter.
This is where I would have preferred a single-subject approach, which would
have allowed me to vary the MSH dosage during continuous observation of the
chick and under both separation and non-separation conditions. Such a study
would have been difficult (although not impossible) to conduct as it would
require implanting a catheter in the chicks' brain and running a line back
to a microinjection pump whose rate could be controlled. The line itself
may have disturbed the chick's behavior.
In the study as conducted, there was a chick or two that did not show the
effect. The reason probably has to do with the difficulty of getting the
MSH into the correct location in the brain. To assess the reliability of
the injection procedure, tests were conducted with dye (in a pevious study
using another drug). In chicks that showed no effect of the drug,
examination of the chick's brain showed that the needle had missed its target.
The reason I ask is that if the effect
of alpha-MSG were indeed specific, direct, and simply a matter of
biochemistry, every experimental animal should have shown the same
effect to the same degree, within some small scatter.
Not so. Every chick's brain as a bit different from the brain of every
other chick, in the same way that no two faces are identical. How much
effect MSH may have on a given chick will depend on such variables as the
extent to which separation produces "distress" and the density of MSH
receptors in the critical areas of the chick's brain. Do you expect every
person in a tracking task to produce the same figure for the control system
gain? Of course not, and for the same reason one would not expect MSH to
have an identical magnitude of effect on every chick.
Question 5: What is the probability that when alpha-MSG is injected into
the ventricles of a particular chick, that chick will continue to peep
at the rate that indicates suppression of the mirror effect when placed
in the mirror chamber? This number should be obtainable from the
statistical record, or from the raw data.
To attempt an answer to this question I need to provide additional data on
the rate of DVs for the same chick in both the plain and mirrored boxes.
Here are the data for the both groups of chicks:
MSH Control
Plain Mirrors Change Plain Mirrors Change
467 382 - 85 464 24 -440
310 559 +249 162 32 -130
354 416 + 62 0 9 + 9
542 490 - 52 505 90 -415
230 187 - 43 528 158 -370
345 444 + 99 515 417 - 98
469 282 -187 552 38 -514
---------------------- ----------------------
Mean 388.1 394.3 6.1 389.4 109.7 -279.7
The third chick in the control group did not show any DVs in the plain box;
I would discard this one from the analysis, although this was not done in
the actual study. To compute the requested probability, we need to
establish some criterion for the effectiveness of MSH in reducing the mirror
effect. We need to convert a continuous variable to an artificial YES/NO
dichotomy. If I arbitrarily use a reduction of 100 DVs as indicating a
mirror effect, then in the MSH group the probability that an MSH-treated
animal would show the mirror effect, estimated from these data, is 1/7 =
0.14; for a control animal it is 5/7 = 0.71. Other criteria will yield
different values. Larger samples would give more stable estimates; the
question you ask is a population question. The point I was making in my
post is that the question about whether MSH has an effect on the rate of DVs
is not a population question; this is a different question from yours.
By the way, as you are aware, Wayne Hershberger's elegant study of chicks'
following behavior investigated this behavior from a PCT viewpoint (although
they were attempting to "follow" food rather than the mother hen). I think
this demonstrates the feasibility of investigating attachment behaviors
(including following and distress vocalization) from a control systems
perspective. From this viewpoint certain questions arise as to what MSH is
doing to the control system regulating the chick's distance from its mother
(or the brood). It would appear that a perception normally capable of
eliminating DVs (because it eliminates the error in perceived distance to
other chicks) does not have this effect (or at least not to the same extent)
when MSH is present. Is this because the chick no longer recognizes the
visual input (unlikely on other evidence), or is it the case that the MSH
prevents the visual input from canceling the error (more likely); that is,
despite the visual feedback of non-separation, the chick continues to
experience separation anxiety. It is probable that actual contact with the
other chicks may be necessary to quell the anxiety when MSH is present.
Regards,
Bruce