Fait Non Accompli: Social Structures Influence Studies of Sex Differences

Can you explain what the correlation is between our brains, sexual orientation and gender? -Tim Rymel

There are more than thirteen ways of looking at brain organization in humans, but all of them are sideways. -Rebecca M. Jordan-Young, Ph.D. (from Brainstorm: The Flaws in the Science of Sex Differences)

Earlier this month we learned that a Google employee wrote a manifesto that, among other things, reiterated the idea that there are intrinsic biological differences between the abilities, interests, and ambitions of males and females. The document uses scientific sounding language to support its arguments and thereby deflect the PC police. But free expression and scientific expression are not the same thing. Science has its own politics of language limited to the confines of hypothesis, experiment, and result. The manifesto attempts to sidestep a lot of scientific language on the topic of sex differences in order to reclaim a freedom of expression not given to science.

I bring this up because last year we received the following question from a reader: “Can you explain what the correlation is between our brains, sexual orientation and gender?”

It is challenging to write a succinct but adequate answer to this question in a single blog post. So I have been sitting on my response for a while. But reading the manifesto made me realize the importance of bringing attention to why questions about the biological basis of gender norms are so difficult to answer. Below is my attempt to address one aspect of Tim’s question: structural differences between the brains of females and males.

—

Tim, thank you so much for your question. Sex is one of those areas where it’s hard to disentangle biology from heuristics. In her book “Sexing the Body,” Dr. Anne Fausto-Sterling notes that when it comes to the relationship between the brain and biological sex, “scientists do not simply read nature to find truths to apply to the social world. Instead, they use truths taken from our social relationships to structure, read and interpret the natural.” Within this infinite loop of society’s structures informing science, which in turn informs how we structure society, even evidence of small differences gets exaggerated to the extent that it pigeonholes and restricts individual groups. 

Adding to the fray, given the many ways in which one can measure the brain and behavior, studies of structural differences in the brain are apt to be misinterpreted, whether or not they take sex into account. See the sidebar below for a few things to keep in mind when reading about structural differences in the brain. For one, it’s important to know the exact scale of difference in order to avoid exaggerated conclusions. Secondly, it’s important to know what is being measured directly and what is being extrapolated. In this post, I focus on the handful of things we know about sex/sexual orientation differences and brain structure with some certainty and the ways in which researchers are prone to overreach their conclusions. For simplicity, I have limited my answer to the following:

1. Human studies. There have been myriad studies on sex differences in animal brains, but I won’t mention them here. The best model for the human is other humans. (Unless there’s a specific animal we should aspire to emulate.)
2. Brain structure. The interaction of genes and sex is fascinating and surprising and the reason I got into biology in the first place. But I’ll avoid speaking of genes here. I’ll also avoid speaking of behavioral studies. Measuring brain structure is messy; adding the variable of cultural reinforcement of sex-typical behaviors only makes the task murkier.

CPU and CPMe

The moment you were born you became a socialized being in which your individual biology — the fullness of your molecular self — was affected by your environment, including the way people looked at you, talked to you, held you, and reacted to you. All of these actions impacted the microstructures within your ever-changing brain, and your perceived gender impacted those actions.

With that in mind, let’s try a thought experiment: Think of yourself as a sexless superhero from a distant galaxy, and you’ve just arrived on earth. You want to take human form, but you notice that humans seem to be divided into two general categories and you’re not sure which to become. You ask me what's different about the CPUs of these two different specimens, so I show you whole brains taken from 1000 different people: 500 male brains and 500 female brains. You come back to me exasperated, because, despite your best efforts, you can't figure out how to sort these brains into two groups. There's too much variability and complexity to base your division on anything consistent. You can arrange them by size, smallest to largest, but then you have a range rather than a dichotomy, a linear line rather than two curves.

In reality, I can’t pull 1000 brains out of a hat. But human scientists are trying to solve a very similar problem and still encounter the same challenges as your hypothetical sexless self: Is there something fundamentally different between the structure and organization of male and female brains?

Science, Like You and Me, is Flawed

In her book Delusions of Gender, Cordelia Fine emphasizes that we should be cautious when looking for a potential link between sex differences in the brain and sex differences in behavior. Here are five things she says to keep in mind:

1. Brains are more alike than different. Findings of no difference rarely, if ever, get published. As you can imagine, evidence of difference gets more attention, even when we don’t understand what that difference actually means for behavior, learning, and mental ability. Yet, trivial suggestions of difference are used to support things like separate schools and curricula for males and females.
2. Brain does not equal mind. There is no one-to-one correlation between a region of the brain and a way of thinking. When performing a certain mental task, such as addition, different brains may have different ways of arriving at the same conclusion. If Region A is activated more in a female brain, and Region B is activated more in a male brain, we cannot conclude that females and males think differently. We can only say that the females tended to rely more on one part of the brain, while males tended to rely on another part of the brain. Another way of looking at this is to imagine that women have more cells in Region A than males do, and males have more cells in Region B than females do. As a result, the female brain compensates for its smaller Region B by activating Region A, and the male brain compensates for its smaller Region A by activating Region B.
3. Neural activity does not imply primary involvement. Just because a region of the brain happens to be observed and activated during a mental task does not mean that it is the only or even the primary part of the brain responsible for performing said task.
4. Snapshots of the living brain are not indelible. Our brains change with experience. The way a brain appears at a specific moment in time is not hardwired. Socialization can and does change the brain. Moreover, much of what we know about brain structure has been gathered from studies of cadavers. This is problematic, because it necessarily misses the nuances of personality, preference and behavior.
5. Most studies of sex differences do not correct for multiple comparisons. If you roll a dice three times, you’re more likely to roll a four than if you roll the dice only once. But your chances of rolling a four are still 1/6 each time, not 1/2, as it may seem from the three dice roles. In many fields, researchers correct for this inflated chance by using a statistical test that corrects for multiple comparisons, or multiple “rolls of the dice.” In studies of sexually dimorphic traits, the sample sizes are often too small to correct for the inflation of multiple comparisons. So the chances of getting a result that shows a difference is often inflated.

Subdivided We Stand

Madhura Ingalhalikar and colleagues at the University of Pennsylvania got close to 1000 brains. They analyzed 949, to be exact, representing 428 males and 521 females. Employing diffusion tensor imaging, a type of magnetic resonance imaging (MRI) that uses the diffusion of molecules in the brain to visualize its architecture, they examined differences in the structural connectivity of the brains of male and female youths between the ages of 8 and 22. Rather than measuring smaller regions within the brain, they looked at the brain as two units: the cerebrum (top/front) and the cerebellum (bottom/back), each divided into two sides or hemispheres. They found that in the cerebrum, males had more connections between the hemispheres, while females had more connections within the hemispheres. The trend was reversed in the cerebellum. These results should be interpreted with caution, however: variability within a sex was likely large, and the researches did not appear to account for brain size, which can affect the precision of the MRI. Moreover, a similar study of 439 individual brains found no differences between males and females. 

Despite these limitations, Ingalhailikar and colleagues could not resist the temptation to claim that their results suggest “that male brains are structured to facilitate connectivity between perception and coordinated action, whereas female brains are designed to facilitate communication between analytical and intuitive processing,” even though they show no data to support a link between connectivity and cognitive ability. Without behavioral and cognitive measurements, we cannot make any claims about the role of structural connectivity in behavior and cognition.

Daphna Joel and colleagues at Tel-Aviv University took a slightly different approach: they looked at differences within a single brain. Rather than measuring the whole brain as one unit, they subdivided it into multiple regions, assigning male-end and female-end qualities to each region. Examining the MRIs of over 1400 brains, each subdivided into multiple regions, they found that most brains exhibit a mosaic of male-leaning and female-leaning structural characteristics regardless of biological sex, suggesting that typical male brains are similar to typical female brains. 

Despite these findings, many researchers still insist on classifying brains as distinctly male or female. Not to be outdone, one such group offered a detailed critique of Joel’s work, in which they use her methods to detect — or fail to detect — differences in the facial morphologies of three primate species. These guys might say that, according to Joel’s approach, the typical monkey face is typical of other monkey faces.

Sexual Attraction: INAH3

To date, there has been only one replicated finding demonstrating a structural difference between the brains of heterosexual and homosexual men, independent of age and brain size. The difference is in a region of the hypothalamus called the “third interstitial nucleus of the anterior hypothalamus” or INAH3. Analogous structures in rhesus monkeys, cats, dogs, and sheep have been shown to play a role in male mating behavior. True to its name in rats — the “sexually dimorphic nucleus of the preoptic area” (SDN-POA) — this region is thought to regulate sex-typical behaviors and is larger in male rats than female rats.

In 1990, Simon LeVay and colleagues working at The Salk Institute in La Jolla, California examined the brains of cadavers from homosexual and heterosexual men. They found that homosexual males had smaller INAH3 volumes, on average, compared with heterosexual males. Even LeVay, however, recognized that there were notable exceptions and wide variability in his data. In her review of LeVay’s work, Elizabeth A. Wilson points out that, for instance, the second largest INAH3 volume measured was in the brain of a homosexual man, a finding that informed LeVay’s conclusion that sexual orientation is not the only factor contributing to IHAH3 size.

At the time of the LeVay study, the INAH3 had been known for a few years to be sexually dimorphic. Three independent studies have now demonstrated that the volume of the INAH3 relative to brain weight is smaller, on average, in females than in males. It should be telling, however, that as little as three studies lend enough certainty to this sex difference that the human INAH3 has since acquired the name “sexually dimorphic nucleus.” 

The root of the sexual dimorphism, however, may be different from that of the sexual orientation dimorphism: In women, the INAH3 was found to have fewer neurons than that of men. By contrast, in homosexual males compared with heterosexual males, neuron numbers were not counted, and the INAH3 was described as simply having a smaller volume relative to brain weight. Thus, despite there being a measurable difference between both males and females and heterosexual and homosexual men, the cause or origin of those differences in neural density may vary. Where one is a difference in number of neurons, the other is a difference in how dense they are. 

If you’re starting to wonder what this all means, you’re thinking like a scientist: “When do structural differences imply functional differences? Which measurements are important?”

Gender Identity

Let’s consider two more regions of the brain that have shown structural differences between females and males.

BSTc

One region of the brain found to be consistently sexually dimorphic is another region of the hypothalamus called the central subdivision of the bed nucleus of the stria terminalis, or BSTc. Located at the mid-section of the stria terminalis, a collection of nerve fibers connecting the amygdala with the thalamus and the hypothalamus, the BSTc is a classic sexually dimorphic structure in many mammals. In humans, the difference is not observable until adulthood: The average volume of the BSTc in adult males is significantly larger than the average volume of the BSTc in adult females when corrected for overall brain size. The size difference is thought to be attributable to the number of neurons that produce somatostatin, a hormone that inhibits the release of growth hormone from the pituitary.

The BSTc may be dimorphic between cis males and trans females as well, though this finding has only been reported once and has not been replicated by any independent groups. In seven trans female subjects -- six taking estrogen and one not taking estrogen -- the BSTc had female-typical numbers of neurons. Likewise, a trans male had male-typical numbers of BSTc neurons. But in order to make claims about brain differences based on biology, the brains of trans females taking estrogen would have to be compared with those of cis-males also taking estrogen. In fact, cis males taking hormone replacement therapy for other reasons had numbers of neurons typical of their gender identity, suggesting that the volume of the BSTc correlates with gender identity, not hormone levels.

The Controversial Corpus Callosum

The corpus callosum is a major tract of nerve fibers connecting the left and right cerebral hemispheres. This expansive structure of white matter has been the focus of controversial findings for almost four decades but has nonetheless informed the conventional wisdom of scientist and lay circles alike — that women, on average, have a thicker or more bulbous corpus callosum than men do.

Figure 1 from Allen et al. (1991), Sex Differences in the Human Corpus Callosum, showing a representative MRI scan of a male brain (A) and a female brain (B) as well as schematics of how they measured the sizes of different parts of the corpus callosum (C-F). Although they found a difference in the "bulbosity" of the splenium (S) between males and females, observers were unable to correctly categorize images on their own at a rate higher than would occur by chance.

In three dimensions, the corpus callosum resembles a bird with large, complex wings, "almost as wide as they are long." According to Fausto-Sterling, however, “scientists don’t measure, divide, probe, dispute, and ogle the corpus callosum per se, but rather a slice taken at its center.” Or they focus on the splenium, a region of the corpus callosum thought to be involved in vision, or the isthmus, a portion thought to be involved in language. No one study takes a full-on comprehensive look at its overall size, structure and organization. But meta-analyses of myriad studies give us an idea of what such a comprehensive analysis would show: a difference according to some measurements and no difference according to other measurements. For a case in point, a study early on showed a difference in the splenium shape but not isthmus size in adults, findings that are still controversial, while there were no significant differences in the corpus callosa of children. And subjective categorization based on the observer's intuition alone was not correct often enough to be considered better than chance.

The idea that women have a larger corpus callosum, though consistently debated, lives on in our collective mindset. It may be a disappointment to pop scientists that the reason for the difference could be related to something as simple as brain weight: the most reproducible difference between male and female brains is overall brain size. Males, on average, have larger brains, so the size of the corpus callosum relative to brain weight is smaller in men, while the absolute size is no different between men and women.

The hemisphere hypothesis

One reason why the sexy myth of the sexualized corpus callosum gained so much popularity is that it conveniently substantiates our collective beliefs about how women and men think. You may know this as “ the hemisphere hypthothesis,” or the belief that females think more globally and men think more locally. This was the idea that Ingalhailikar et al. above were so tempted to echo. 

In her book, Pink Brain Blue Brain: How Small Differences Grow into Troublesome Gaps and What We Can Do About It, neuroscientist Lise Eliot suggests that the reason for the myth’s longevity can be traced back to a 1995 study from the Shaywitz lab at Yale. The wife-husband team of Sally and Bennett Shaywitz used functional MRI to show that for a rhyming task, women tended to activate both frontal lobes while men activated only the left frontal lobe. Despite the low number of subjects (nineteen women and nineteen men), the finding that women process language more bilaterally than men do fit so well with the idea that women have a small advantage over men in verbal skills.

At least twenty-six similar studies, however, failed to definitively replicate the Shaywitz finding, and when combined, the data show no significant difference in bilateral language processing between men and women. But another study from 1995 did report that women have a higher density of neurons in a language center of the cortex called the planum temporale, though only nine brains were included in the study, and it has yet to be replicated. 

You Are Vast

Biology is fraught with human error. Science relies not just on the collection of data but also on its interpretation. To some extent, interpretation is always couched in a belief system. In his book Whistling Vivaldi, the psychologist Claude M. Steel asserts, “We simply are not, and cannot be, all knowing and completely objective. Our understandings and views of the world are partial, and reflect the circumstances of our particular lives. This is where a discipline like science comes in. It doesn’t purge us of bias. But it extends what we can see and understand, while constraining bias.” World views are inescapable. Fortunately, the scientific method works to keep them in check. 

Until we can can completely remove bias from our scientific understanding of males and females, I propose that rather than thinking about our brains and bodies fitting into discrete categories, let us think about the infinite combinations of qualities our biological selves bestow. That is not to ignore the ways in which social categories work against certain groups, but to recognize, and thus challenge, the ways in which those categories inhibit a person from realizing her or his potential.