In Men and Women’s Brains Aren’t Really That Different, Gina Rippon provides an excerpt from her book on gender differences in the brain. Unfortunately, the piece reads more like an op-ed than a form of sci-com. It could be that in other areas of the book, there is substance which justifies her position, but there is absolutely nothing in the article which does so. There are also two issues that I have with the claim. The first issue is a semantic one. Scientists need to become better at differentiating sex and gender. The second issue is in actual substantiation and a contradiction between the science that does exist and her position.
The Semantic Issue
From a semantic perspective, it seems that Gina is referring to potential sex based differences in the brain, rather than gender based differences. I’ve written extensively on the concept of gender, both here on Medium and elsewhere. Gender is what I refer to as an eidos: a culturally defined role which is related to, but distinct from, underlying biological features.
Sex is biological. Gender is cultural. While the two are related, they are not the same, and while I admit that I do sometimes use gender terms and sex terms in place of one another, I do try to make sure that I avoid this error, as all science communicators should. Indeed, science communicators are apt to point out the distinction between sex and gender, when there is pseudo-scientific or anti-scientific rhetoric used to attack transgender people and other groups, but we often fail to recognize the distinction between sex and gender, in our own work, thus clouding the issue and spreading the misunderstanding further.
The Science
The second issue is that science does not seem to justify her position. While it may be true that the science on sex based differences is not as robust as some might think, that’s not enough to argue that there is no inherent sex based difference in how the brain functions. There are a number of reasons to think that the brains of females would be, on average, different from the brains of males. These differences could take multiple forms. Gross size of regions of the brain are of little interest. On that matter, I agree with Gina. Functional differences are what matters. These differences can be permanent or transient.
Impact of Hormones on Functional Connectivity
The greatest reason to think that such differences between sexes exists is that males and females have different levels and cycles of hormones, and hormones seem to impact brain function. For instance, according to Lu et al. 2019, changes in sex hormones during perimenopause apparently changes functional connectivity in the brain. Additionally, Hahn et al. 2016 noted that high doses of testosterone caused significant functional changes in adult brains in female-to-male transsexuals. This finding is inconsistent with the idea that sex dimorphisms are due to differences in enculturation and how the brain adapts to expectations. The adult brain of a male-to-female transsexual is unlikely to undergo such rapid and precise changes in expectation simply because they are starting transition. Given that hormones have such a large impact on functional connectivity, and there are significant differences in hormone levels between males and females, it becomes very difficult to argue that such differences are almost entirely due to conditioning.
For Gina’s position to be robust, she would have to respond to these issues. Either she would have to provide justification for the claim that hormones do not impact brain connectivity, or that somehow, even with this theory in place, it is conditioning, and not hormonal differences, that override and result in the dimorphism in brain function, and that without this conditioning, somehow the difference in hormones between males and females does not result in dimorphism.
Fetal Research
But to further investigate the matter, it is important to conduct research on neurological distinctions that arise during fetal development. If sex based differences are due to later enculturation, they should be less likely to arise during fetal development. One such study is Sex differences in functional connectivity during fetal brain development, conducted by Wheelock et al. 2019. Gina actually provides criticism of this paper, and while there are some valid concerns, most of the criticism is not exactly valid. Of interest are the sample differences between the male and female participants, the testing of a large number of potential regions of interest, and the use of non-parametric tests due to a lack of apparent normality.
The sample sizes were unequal between male and female participants, and ages varied considerably, and during that time, cortical development varies considerably. If the age distribution of females was significantly different from the age distribution of males, in the study, the age difference could account for apparent sex based differences. However, the researchers specifically checked to see if there was a difference in gestational age at the time of the scan. The results of the test can be found in Table 1 — Demographic Information, and the results do not indicate a difference between the two populations.
Additionally, Gina mentioned that there are a lot of potential regions of interest. And on this matter, I agree. If a researcher performs enough hypothesis tests, there will be some that are successful. It’s a simple matter of chance. Therefore a followup study needs to be conducted to further justify that these regions of interest are not simply false positives, but actual differences in development.
Gina also commented on the use of parametric tests and their low statistical power. However, this argument really makes no sense. A hypothesis test is a statistical form of proof by contradiction. The null hypothesis is simply an initial assumption that is made, with the intent to try to find a result that contradicts it. The simpler the hypothesis — and non-parametric tests make fewer assumptions than parametric tests — the less likely the hypothesis is to be falsified.
In essence, this idea is the flip side to Occam’s razor: if a “simpler” explanation is more likely to be true, all else being equal, then it is less likely to be falsified. Therefore, it is more difficult to reject the null hypothesis. In the case of this study, the null hypothesis is equality, and so when using a non-parametric test, as is required because there wasn’t sufficient justification to assume normality, it would be more difficult for the researchers to conclude that the regions of interest are unequal.
Additionally, Gina mentioned that non-parametric tests cannot test for multivariate influences. While it is true that tests like the Spearman rank correlation is univariate, there are multivariate non-parametric tests available. Since Gina is concerned about the lack of attention to other variables, perhaps the solution is to see if the raw data is available and perform such a multivariate non-parametric test.
While there are some issues with a key study which indicates that sex based differences in the brain start to emerge in-utero, this kind of study does not exist in a vacuum. Alone, the study does not provide much justification. But given that we see various sex based differences, in fetal development, and in childhood, and in adolescence, across various cultures, the compass needle is pushed far away from the theory that the sexes are neurologically the same. It’s true that it would be more powerful, if it were repeated. If the same regions of interest were tested and similar results were found, then it would be justification that these hits were not just due to random chance.
Yes; more research should be done, as is generally the case in science. And perhaps it is reasonable to dismiss the claim that there are sex based differences in the brain. However, when it comes to claims that there is no difference, the claim is simply unjustified. Moreover, the compass needle does point away from that theory, especially when coupling the findings of fetal studies with studies that indicate that hormones have a significant impact on functional connectivity in the brain.
