The following is based on a seminar given at the University of Florida, Department of Psychology, on October 31, 2002.
Thumbnail images may be clicked for enlargement; use your browser's "back" button to return.
First, a bit of introduction. My principal qualification for work
related to psychiatry is my doctoral dissertation
on crustacean neuroanatomy.
Quick quiz: Who is the most famous crustacean neuroanatomist
of all time? You do know this person's name. Here's a hint.
His reputation has undergone a steep decline over the last couple
decades (but not because anything was wrong with his crustacean neuroanatomy,
which he completed in 1881). [Click here for answer.]
More recently, with several other researchers scattered around the world, I have been trying to promote interest in the role that genetic organization plays in evolutionary adaptability. In particular, I have become impressed by the evolutionary function of microsatellite DNA. You may be aware of the several hereditary neurological disorders, such as fragile-X and Huntington's disease, which are caused by expansion of triplet CAG repeats. You may not know that CAG repeats are found all over the genome, or that other kinds of repeats are also common but have received much less attention.
The most curious feature of microsatellite repeats is that the number of repeats can exert a quantitative effect on gene regulation, such that mutations affecting repeat number can provide a prolific source of genetic variation. In brief, microsats can encourage and facilitate variation in quantitative traits.
Probably the best-characterized microsatellite, in terms of evolutionary function, is a hexanucleotide repeat in the period gene of Drosophila, where the number of repeats adjusts the thermal sensitivity of the fly's biological clock.
Perhaps more relevant to psychiatry, we have a 48-nucleotide repeat in a human dopamine-receptor gene (D4DR), where variation in the number of repeats has been correlated with variation in "novelty-seeking" (Dean Hamer, Living with Our Genes, 1998).
There is even some suggestion that bipolar disorder may show the pattern of inheritance, called "genetic anticipation", which is characteristic of the triplet-repeat diseases. The only studies I have seen so far report a failure to find variation in triplet repeats that is correlated with bipolar, but unfortunately these studies have not considered that many other DNA motifs might be involved in repeat-associated gene regulation.
The above digression is dated. New information on repeat function continues to appear.
Anyway, returning to today's topic... I'm supposed to be talking about the rate of perceptual alternation in binocular rivalry.
Three years ago, I had not even heard of binocular rivalry. [If you want to learn about binocular rivalry in relation to bipolar disorder from someone with more experience, go to Jack Pettigrew's webpage.] My ignorance is no measure of other's education, but I really have no reason to expect that you are any more familiar with binocular rivalry than I had been. If you are already familiar with binocular rivalry, skip ahead. Otherwise, here's a brief introduction.
Normally, both of your eyes look together at the same visual field. Your brain then combines the two images into a single, unified perception of the world "out there". Binocular rivalry emerges when each eye views a different visual scene.This not a normal situation, but it's pretty easy to contrive.
One convenient approach is to hold a small cardboard tube in front of one eye and the palm of your hand in front of the other eye.
The resulting perception will probably start out looking like a "hole" in the palm of your hand.
If the view through "hole" is not too interesting, after a few moments the "hole" may appear to "fill in" with a view of your palm. If you continue to look in this way, the views will shift back and forth, at somewhat irregular intervals, between "hole" and "palm". This is the phenomenon of perceptual alternation. [There are several other circumstances which produce an experience of perceptual rivalry (more).]
I first encountered binocular rivaly as a topic of scientific interest while reading an autobiographical sketch by Jack Pettigrew in the July 1999 newsletter of the International Society for Neuroethology. I already knew something of Jack's work on the evolutionary implications of the neuroanatomy of the auditory system in bats. Here in this essay I learned that Jack's interest in interhemispheric switching had led him to study binocular rivalry in people.
I also learned that Jack had fortuitously noticed that the rate of perceptual alternation during binocular rivalry was markedly slower in some subjects with bipolar disorder. Binocular rivalry is indeed a curious phenomenon, but the world is full of curious phenomena. However, I had recently developed a very personal interest in bipolar disorder.
So at this point I wanted to see for myself, to measure perceptual alternation in a few people and check out Jack's results. But I quickly learned that research with human subjects has complications that I had never encountered while working with lobsters and flies. While awaiting approval from the Human Subjects Committee for some exploratory observations, I learned at a meeting of NAMI that NAMI's research affiliate, the Brain & Behavior Research Foundation (formerly NARSAD), offered research grants for investigators wishing to embark on new studies, with a very efficient application and review process. So, while awaiting Human Subjects approval I sent a proposal to NARSAD. NARSAD was very generous, and I have been able to see many more subjects than I had initially expected.
So now we're ready to begin. The biggest effort in this project has been recruiting subjects, for which I was ably assisted by Jeannie Killian, who quit a secure job in order to join this project.
Actually
measuring binocular rivalry is fairly easy and cheap. The setup consists
of an ordinary personal computer, software (which was provided to me by
Keith
White), and a set of cunning stereoscopic
glasses with rapid liquid crystal shutters for lenses.
Here
is the view that a subject sees through the glasses. The spot in the
middle of the screen changes rapidly back and forth -- 60 times per second
--between vertical and horizontal bars. The shutters in the glasses
also switch on and off 60 times per second, coordinated with the display
by the tiny red light on top of the computer.
The
result of this is that one eye is always looking at vertical bars while
the other eye is always looking at horizontal bars. Thus we have binocular
rivarly.
For most subjects, the perception that results from this setup alternates between vertical and horizontal bars, sometimes with a combined (cross-hatched) pattern during the switchovers.
About ten percent of our subjects did not experience rivalry. They would see just one of the two percepts, unchanging. In most cases this had a straightforward explanation. Either the subject had amblyopia and his brain was ignoring input from one eye, or the subject's two eyes did not focus together at the same distance (sometimes because of Lasek surgery for presbyopia) and her brain was ignoring input from the out-of-focus eye. Binocular rivalry only emerges if neither of the mismatched images is more salient than the other (e.g., sharper, brighter, more contrasty, more motion, etc.)
After
all that, and two years, we have data. Our principal result was a
complete surprise. The rate of perceptual alternation shows a distinct
decline with the age of the subjects.
Note that we should not presume that this "age effect" is causally related to aging per se. This is a cross-sectional study. Subjects who differ in age may also differ in other ways unrelated to age, most notably in such things as diet during childhood, exposure to various illnesses, life-style choices, etc. We did not plan this study as a "fishing expedition" to discover what variables might correlate with perceptual alternation rate, so we did not subject our volunteers to an extensive questionnaire. All we recorded were basics: gender, race, age, and family history of severe mental illness.
However, the differences between these groups are quite small relative to the variation within the groups.
I have not subjected this data to sophisticated statistical analysis, in part because I am inexperienced with statistics and in part because proper analysis is confounded by a weakness in our experimental design. Any competent statistics course will emphasize the importance of proper experimental design. Unfortunately, our experimental design has problems. The biggest problem is that we did not anticipate that age would turn out to be a significant confounding variable.
Furthermore, although the line on this graph plots a linear regression with age, we have no expectation for how alternation rate may be correlated with age. There is no reason to expect the relationship to be linear. In fact, the data suggest that most of the decline occurs after age 40.
It is treacherous to try to correct for a confounding variable when the sample populations were not randomly sampled with respect to that variable.
[Because we did not design our study to control for an age effect, samples are not balanced for age. Our control sample happens to be weighted toward younger subjects. I suspect that this is because older (retirement-age) subjects were more strongly motivated to volunteer if they were themselves bipolar or had close relatives with bipolar disorder.]
There are several other reservations that might be mentioned.
Assignment of subjects to groups has some unavoidable uncertainty. We are not qualified to perform any psychiatric evaluation ourselves. We asked every subject whether they had received any psychiatric diagnosis, and we requested a response from the subject's health care provider indicating any psychiatric diagnosis. But most subjects in our control group had never received any psychiatric evaluation. So there's a small chance that a few of our control subjects may be undiagnosed. Furthermore, based on the prevalence of bipolar disorder in the general population, a small percentage of the control subjects may develop bipolar disorder at some time in the future.
For subjects who indicated that they had been diagnosed with "bipolar disorder", I have here included only those whose providers confirmed that diagnosis. But I have heard enough stories about changing diagnoses to have a certain skepticism about this data. Whether or not a particular subject has been diagnosed as bipolar may depend on which of several local psychiatrists provided the diagnosis, and at what point in the subject's life.
Inclusion of subjects in the category of "having a parent, child, or sibling with bipolar disorder" was based entirely on self-reporting. Not all families share accurate information about any medical matters, especially mental illness. And, rather obviously, even if family history is accurately known, what we would like to have is information about the families' genetic potential for bipolar disorder rather than how this potential has been realized in a small sample of actual relatives.
Finally, measurement of perceptual alternation is inherently very noisy. I do not yet know how consistent binocular measurements may be when taken repeatedly at different times or under varying circumstances.
Curiously, subjects who reported psychiatric diagnoses other than bipolar disorder also showed a decline in alternation rate similar to that shown by subjects with bipolar. (Sorry, I don't have a slide for that data.)
So. That's we where are now. My major conclusions, with interpretation, are the following:
