LEVAY CRITIQUE

*Neuroscience or Nonsense*

Written by Sandy Zetlan, Ph.D.Neurosciences
zetlan@wiscmacc (Bitnet)
zetlan@wisc.macc.edu (Internet)
Women's Studies Program
University of Wisconsin
209 North Brooks St
Madison, WI 53706
INTRODUCTION
In August of 1991 Simon LeVay published an article in the interdisciplinary
journal Science (A difference in hypothalamic structure between heterosexual
and homosexual men. Science 253:1034-1037, 1991). LeVay's conclusion from this
work, was that *sexual orientation has a biological substrate*. This research
study met with rapid and voluminous accolade in the media and little critique
from the scientific community. The article is short and fairly accessible, and
I recommend that anyone interested in this topic read it in its entirety.
Much of the following analysis, and my understanding of this subject, has been
formed by discussions with William Byne MD,PhD, a current psychiatric resident
and neuroscientist at Columbia University, New York City and former doctoral
student of Dr. Ruth Bleier, Marsha Segerberg PhD Neurosciences, Postdoctoral
Fellow in Physiology at the University of Arizona, Tucson, and Mariamne Whatle
PhD, Chair of the Women's Studies Program at the University of Wisconsin,
Madison. However, none of them should be held at all accountable for my
ramblings! I review LeVay's background assumptions and methodology, and offer a
critique of both. The following discussion is unreferenced. Quotes are between
*stars*. Unlike LeVay, I will use GM for males in the homosexual male group
(because it is clearer) and, like LeVay, I use M for males in the heterosexual
male group.
SECTION ONE :What did LeVay do?
Using autopsied brain material, Simon LeVay measured the volume that different
cell groups takes up within the hypothalamus. He reported that the volume of
one group of cells, called INAH3 (the third interstitial nucleus of the
anterior hypothalamus) was on average 2x larger in the M subject group than in
the homosexual male (GM) or female subject (F) groups. He found no differences
in nearby groups designated INAH 1,2 or 4. LeVay's conclusion from this work,
was that *sexual orientation has a biological substrate*. The results have
been interpreted as evidence that the reason that GM sexual behavior is *more
like* female behavior than M behavior is due to corresponding differences in
INAH3. ***FYI: The human hypothalamus is about the size of a pea, and the INAH3
is about the size of a grain of sand.***
SECTION TWO: I suggest that there are major methodological flaws in this
study which invalidate LeVay's conclusions. In addition, LeVay, and other
researchers, make highly contestable assumptions regarding the biological basis
of human sexual orientation that are based primarily on inappropriate animal
models. I will first critique the methodology and then the background
assumptions.
SECTION THREE: What are some methodological problems in the LeVay study?
1. LeVay obtained inadequate information on the sexual background of his
subjects. The sexual orientation of the 6 females was not known. The sexual
orientation of the males in M was PRESUMED, based on the numerical
preponderance of males in the population. It was not clear on what basis men in
G were so labeled. LeVay never defines the criteria for sexual orientation.
What constitutes sexual experience? How many times did you have to do IT? Does
fantasy *count*? What if one's sexual orientation changed over time? That
LeVay never defines the criteria for sexual orientation is a major problem,
since separating groups with clearly defined sexual orientation is the basis
for this study. However, since sexual orientation seems to be an elusive blend
of physical, mental and emotional properties, a uniform definition that is
applicable to each subject within each group may be an unachievable goal.
2. LeVay obtained inadequate information on the medical background of his
subjects.. All GM (n=19) died of AIDS related opportunistic diseases. Six M
subjects died of complications of AIDS, and 10 of other causes. Males
contracting AIDS from sexual exposure, IV needle use, or blood transfusions may
incur different opportunistic diseases and treatments, as well as varying
exposure time before detection (perhaps as a result of differential interaction
with the health care system), any of which might affect the INAH3 volume.
3. Brain material was processed a variable length of time (one to two weeks).
While resulting differential artifacts may or may not be negligible, nervous
tissue can undergo distortion during fixation, and all tissue should be
processed identically, if possible.
4. Even if his three groups were actually representative of people with
different sexual orientations, it would be impossible to discern if INAH3
volume differences are the *cause* of certain sexual behaviors or the *result*
of certain sexual behaviors or other unknown and uncontrolled variables. The
brain is incredibly plastic, and can be physically altered by life experience,
even in adults.
5. In light of the volume of research which suggests that sexual orientation
appears greatly influenced by social factors, it seems naive to expect that
there would be a direct relationship between hypothalamic structure and sexual
behavior. This position is supported by the overlap of INAH3 volumes between
the subjects of all three groups. In other words, given the size of a
particular INAH3, it would be impossible to predict into which sexual
orientation group a subject would fall.
SECTION FOUR: Background information: Hormones, behavior, and brain.
Rodents have provided much of the information used to construct the current
biological model of human sexuality. Normally male rats (and male humans) are
exposed to high levels of androgens secreted from their own testes during
development, while female rats (and female humans) are not. In RATS, androgens
(eg: testosterone) affect sex differences in reproductive motor behavior and
in hypothalamic structure and function. There is no evidence that androgens
have these effects in humans. The arguments are explained in greater detail
below.
a) To a large extent, male and female rats do have different stereotypic
reproductive behaviors which can be altered by varying levels of androgens
during development. Male rats tend to mount other rats more than do female
rats. Females rats tend to display lordosis (back arched) behavior more than
male rats and allow themselves to be mounted at certain times in their cycle.
Male rats in which androgen levels were decreased experimentally during
development tend to show decreased mounting frequency and allow themselves to
be mounted more frequently (so called male *homosexual* behavior in rats).
Female rats exposed to increased androgen levels will tend to increase mounting
frequency (so called female *homosexual* behavior in rats). These effects of
androgen levels on reproductive posture and motor behavior in rats form the
basis of theories on human homosexuality.
b) In rats, exposure to androgens during fetal development alters the ability
of the hypothalamus to regulate the release of the hormone LH. In response to
natural ovarian or administered estrogen, the adult female rat hypothalamus
signals the the pituitary gland to release leutinizing hormone (LH is the
hormone that causes ovulation in female rats). Adult male rats who have been
exposed to androgens secreted from their own testes during development cannot
release a surge of LH in response to administered estrogen. In addition, female
rats exposed to experimentally induced high levels of androgen during their
development will not have an LH surge in response to estrogen. In other words,
the after exposure to high levels of androgens during fetal development, the
hypothalamus no longer has the ability to mediate the positive feedback of
estrogen. Androgen exposure has a permanent effect on the hypothalamus. Some
researchers say that *brain* has been permanently *organized, masculinized or
defeminized* by androgens. I like to say disorganized. Of course the word
*brain* , in this instance, means a region of the hypothalamus, probably
serving a specific hormonal function.
c) Parts of the rat hypothalamus are structurally different in males and
females. These areas are called the sexually dimorphic nuclei, or SDN. The
functions of these areas are not known.
SECTION FIVE: Why shouldn't LeVay (or anyone else) extrapolate an androgen
theory of rodent sexual behavior to models of sexuality in humans? Because...
1. Applying a rodent model of sexual behavior to human sexual orientation
requires several highly contestable assumptions.
a. That reproductively linked posturing behavior (mounting and lordosis) is
somehow representative of sexual orientation in rats AND in humans.
b. That human male heterosexuality is characterized by *masculine or
male-typical* behavior (eg:mounting), and that human female heterosexuality is
characterized by *feminine or female-typical* behavior (eg: lordosis) during
sexual encounters.
c. That human male homosexuality is characterized by * feminine or
female-typical* behavior (eg: lordosis and lack of mounting) and human female
homosexuality is characterized by *masculine or male-typical* behavior (eg:
mounting and lack of lordosis), sexual encounters. In addition, I object to
the terminology *masculinization* and *feminization* as applied to to the
brain since they give the misleading impression that stereotypical masculine
and feminine human behaviors are already known to be a consequence of brain
structure and function.
2. There is evidence that, unlike in rats, the non-human primate and human
hypothalamus is NOT permanently altered by exposure to androgens during
development. In other words, there is no corresponding evidence in humans for a
*male* and *female* brain. (Aside: Nevertheless, differential levels of
androgen exposure to the brain during development are also often cited as the
basis of any presumed cognitive sex differences in humans.)
3. There are no replicated studies in humans which correlate level of
androgens and sexual orientation. High levels of testosterone during
development in human females and low levels of testosterone in human males do
NOT correspond with an incidence of homosexuality.
4. Attempts to replicate reported sex dimorphic structural differences in the
human hypothalamus have not been successful. There is no evidence in humans
that an area of the hypothalamus directs sexual activity. There is not evidence
that INAH3 has a function similar to sex dimorphic areas in the rat
hypothalamus, which may or may not be correlated with sexual activity in the
rat.
SECTION SIX: Summary of major arguments against LeVay's conclusions and
assumptions.
1. LeVay's method of separating male subjects into heterosexal and homosexual
groups is highly contestable.
2. The length of exposure to AIDS, type of opportunistic infection or other
disease acquired, or type of medicative or other treatment was not reported
for any subject.
3. The impact of disease state, treatment, or length of exposure to AIDS on
the INAH3 volume is unknown.
4. Sex and sexuality differences research is currrently hampered by poor
reproducibity of results. Until LeVay's work is improved and replicated, it
stands unsupported.
5. Androgen exposure during development has not been shown to affect human
brain function or structure.
6. Androgen exposure during development has not been shown to correlate with
later sexual orientation.
7. Androgen levels in adults have not been shown to correlate with self
reported adult sexual orientation.
8. Rat motor patterns of copulation cannot be extrapolated to human sexual
orientation or sexual activity.
9. The ability of the hypothalamus to mediate positive feedback of estrogen
is blocked by early androgen exposure in rats but not in primates. That is,
there may be *male and female brains* in rats, but this has not been shown to
be true in humans.
10. The function of the sexually dimorphic nucleus (SDN) of the rat brain is
not known and there is no convincing evidence that the SDN in rats corresponds
structurally to the INAH3 in humans. Thus, there is no evidence that sexual
orientation in humans has a *biological substrate* similar to that suggested
for sexual posturing in rats.
Therefore, as far as I can tell, LeVay has not found a biological substrate
for sexual orientation. All LeVay has reported is that in groups of people with
unknown medical and sexual histories there is a significant difference in the
size of a structure whose function is not known.
SECTION SEVEN: addendum
The assumptions LeVay makes in his work are those made by many other
researchers in the field. The critiques presented here are those that should
be applied to any scientific study. I attribute the exorbitant amount of press
coverage to the idea that the media has finally found something even sexier
than biological theories on cognitive differences between the sexes. I
attribute the exorbitant amount of latitude allowed LeVay by the scientific
community to the belief, as expressed by one scientist, that such results were
*not unexpected*.