This paper (below) looked at whether biological sex and “gender” are independently associated with certain aspects of brain structure. In 88 healthy adults, they found that sex is mainly associated with regional brain volumes, while “gender” is associated with cortical thickness in various prefrontal regions. However, as usual there are many problems with how this data is interpreted... Firstly, they don’t really measure “gender”. They use the Bem Sex Role Inventory (BSRI), which measures adherence of traditional masculine/feminine personality traits. Even then, there are questions around how well the BSRI actually captures these constructs. A review of 23 validation studies found that the underlying structure captured by the BSRI was more complicated than simply “masculinity” and “femininity”, and questioned whether these constructs were adequately captured by the scale. The authors then simplify this even further by creating a single “gender continuum”: Femininity − Masculinity This seems odd to me because masculinity and femininity were specifically designed as separate dimensions in the BSRI, rather than opposite ends of the same scale. Someone high in both could therefore have exactly the same “gender” score as someone low in both. So I’m not sure this can really be interpreted as a continuum of “gender”. My biggest issue, though, is their claim of “sex-specific” effects. When they split the sample by sex, several correlations between femininity and cortical thickness are significant in females but not males. But a significant correlation in females and a non-significant correlation in males does not mean the two sexes significantly differ. You need to directly test whether the relationship differs between males and females (i.e., an interaction). They actually do this, and the sex × gender interaction is not significant (p=.651). These subgroup correlations are also uncorrected for multiple comparisons, with several only just reaching p<.05. So I think describing these as “sex-specific” is quite a strong interpretation. I’m also not convinced by their brain-size correction. They divide regional volumes by overall brain volume, which effectively assumes that regional volumes scale proportionally with brain size. But we know that isn’t necessarily true. If a structure increases less than proportionally as brains get larger, smaller brains will naturally have a larger regional volume / brain volume ratio. Since females have smaller brains on average, this can make some structures look “proportionally larger” in females simply because of normal allometric scaling. So some of these apparent regional sex differences could actually be mediated by the underlying difference in brain size, rather than representing a specific effect of sex on that region. I’d much rather see brain size modelled directly (e.g., ANCOVA), alongside testing the allometric relationship. Finally, it’s cross-sectional. Even if we accept the BSRI as capturing some aspect of gendered experience, a correlation with cortical thickness doesn’t tell us whether gendered experience caused the brain difference, whether the relationship goes in the opposite direction, or whether something else influences both. Overall, I do NOT think this paper provides convincing evidence for separate effects of sex and “gender” on brain structure. At most, it shows that biological sex and scores on a questionable measure of masculine/feminine personality traits correlate with different structural measures. That’s a much narrower finding than the paper implies