Cherrypicking The Transgender Brain

(Post in question)

I won’t comment on the Daphna et. al literature, because it’s still an emerging line of research that has some more philosophical and definitional questions as to what actually constitutes dimorphism (I’ve seen contradicting meanings).

Ummm… this result is in complete agreement with another hypothesis that many transwomen find uncomfortable, one made by Ray Blanchard, in which he hypothesized that late transitioning transwomen would have brain structure differences from both men and women that would NOT be sexually dimorphic; while young (“homosexual”) transsexuals would show shifts in sexually dimorphic structures toward female morphologies.  There was an earlier review of previous studies (which I also wrote a post about) that had shown that hypothesis to be supported.

There’s also recent research that shows controlling for sexuality has absolutely no effect on trans brain structures.

From the abstract:

However, controlling for individual estradiol, testosterone, or progesterone plasma levels or for subjects’ sexual orientation did not change group differences

And then in the statistical analyses section they explain their control:

To investigate whether the three sex steroid hormones (E2, T, and P4) or sexual orientation explained group differences, the named variables were added as covariates of no interest within separate ANCOVA analyses. Finally, multiple regression analyses were performed to determine the effects of hormones and sexual orientation on diffusivity maps independent of group membership, i.e., with group in addition to TIV as factor of no interest. Separate models were calculated for each of the independent variables (E2, T, P4, and sexual orientation). Because hormone values over the entire sample were non-normally distributed, they were transformed to ordinal scales based on ranks before inclusion in the analysis. Sexual orientation was entered into analyses in three different ways of coding: (1) as raw values of the Likert scale (1 for attraction toward females; 7 for attraction toward males); (2) as a spectrum from homosexuality (e.g., 1) to heterosexuality (e.g., 7) with reference to the raters’ genetic sex; and (3) as a spectrum from homosexuality to heterosexuality with reference to the raters’ gender identity. The statistical threshold was set at p < 0.05 FWE corrected, using the threshold-free cluster enhancement method to define the clusters (). Voxels showing significant differences were assigned to white matter tracts using the DTI-81 white matter atlas of the International Consortium for Human Brain Mapping as provided by the DiffeoMap software package ( Diffusivities based on tract-specific quantification for right and left CST, Fmajor, and Fminor were compared using ANOVA in SPSS. Separate models were calculated for each tract and diffusivity parameter, followed by post hoc pairwise comparisons and correction for multiple comparisons using the Bonferroni’s procedure. Analyses were run with and without TIV as covariate of no interest. Separate correlation analyses were performed to examine the association between diffusivities and age and between volumetric data [gray matter volume (GMV), white matter volume (WMV), CSF, and TIV] and age for each group.

And their results:

Here, we investigated whether sexual orientation associates with diffusivity measures. No effects on our main findings were observed when sexual orientation was regressed out in the ANCOVA design. Moreover, there was no significant effect of sexual orientation on diffusivity parameters in the regression analysis including all subjects and using group as factor of no interest.

And furthermore:
In our study, we find robust differences between investigated groups in MD, AD, and RD indicating that biological sex and gender identity both contribute to observed group differences. Moreover, the high positive correlation with adult plasma T levels (controlling for group membership) indicates that group differences cannot be explained by peripheral sex hormone plasma levels. As expected, biological males (MCs and MtF transsexuals) had higher T levels than biological females (FCs and FtM transsexuals), whereas group differences in diffusivity values showed the transition FC > FtM > MtF > MC. Furthermore, group differences were not explained by differences in sexual orientation, narrowing potential determinants for differences in diffusivity parameters to biological sex and gender identity.
As I said, I don’t particularly care about etiological debates over neurology considering the field is new enough to warrant skepticism about replicability and usefulessness, and the extent that socialization and environmental factors affect brain neurology.
But I’ll offer an alternative interpretation of the data that lesbian trans women have ‘masculinized brains’ (taking that as an axiom for the purpose of argument).
We know from some recent on the neurology of gay men and lesbian women that sexual orientation has an effect on brain neurology independent on trans status:

diffusivities and age and between volumetric data [gray matter volume (GMV), white matter volume (WMV), CSF, and TIV] and age for each group.

A 2008 study demonstrating that homosexual men and women have sex-atypical brain structures

The connectivity pattern in homosexual subjects was almost reciprocal in relation to the same-sex controls

Second, HoM, just as HeW, displayed connections with the contralateral amygdala, the anterior cingulate, the subcallosum, and the hypothalamus

Group comparisons confirmed these findings: HeW, as well as HoM, showed a greater connectivity with the contralateral amygdala and the cingulate cortex compared with both HeM and HoW.

We also found that the hemispheric ratios, as well as the patterns of amygdala connectivity, were sex atypical in homosexual subjects, with HoM exhibiting more female patterns and HoW showing more male-like features (albeit less pronounced).

(Note that this has been replicated quite a few times)

A TERF or transphobe could easily interpret this data to show that “trans women are just gay men!” and “trans men are just lesbian women”, which is exactly why the neurological argument is intrinsically flawed. But this data yields an alternative interpretation of the (contested) result that lesbian trans women have brains more similar to that of men: they should. Just like lesbian cis women have brains “more similar to that of heterosexual men” (to paraphrase), lesbian trans women do too. My conclusion is that brain structures are not organized by sex or gender, but by who one is sexually attracted to. Male/men-attracted people and female/woman-attracted people differentiate themselves. It’s internally consistent with the data, while the “male brain” / “female brain” hypothesis is not: gay men and lesbian women provide contradicting data. The issue is not one of data, but one of the paradigms we interpret the data through.

Now, this is all moot if we take the data from the study that I cited above that showed that sexual orientation had no impact on brain structures for trans individuals, which just explicates that trans brains are neurologically distinct than cis brains categorically rather than by subgroups

But either way we look at it, her argument is incorrect in some manner: paradigmatically or scientifically.

Searcy is likely to have written her article in an attempt to discount the growing evidence from transgender brain scan research that shows that the two type taxonomy for transwomen is supported.

There’s exactly one study that has shown gynephilic trans women have ‘male brains’, and there is at least one that shows the contrary. I wouldn’t be so eager to run and shout that the two type taxonomy is supported by the research

Where once older transitioning transwomen cherry picked the brain structure research in an attempt to spin it such that all transwomen had female brains.  She is spinning the science to lead us to believe that brain structure research is unimportant and should be ignored, first by saying that there is no brain structure sexual dimorphism of any consequence and then say that what differences between transgender folk and nontransfolk is unimportant anyway.

This just gets back into the debate over dimorphism, which as I said I won’t get into those weeds today, but I’d just emphasize that it’s not a simple topic. There are fundamental philosophical disagreements over what constitutes sexual dimorphism between the two camps that can’t be ignored by saying that ‘one side is supported by the evidence’ and that the other isn’t.

I believe it represents a growing fear by autogynephilic transwomen that the brain scan science will undermine their own identity as transwomen if the public were to become aware of what the evidence means.

To me it seems Brown is alluding to her own denial of lesbian trans women’s identities and womanhood, which contradicts her stated “support” of “autogynephilic transwomen”


Here Be Liars

(Post in question)

Autogynephilia in cis women does exist. Denying the facts is just harmful to the cis women in question. It’s science-denying and it prevents them from fully expressing their sexuality.

Now that was mostly satire, but what I said about the existence of autogynephilia in cis women is true.

First, there is the wonderfully creative, if fallacious, redefinition of autogynephilia, the softening of the language, more than simple euphemism, of calling it “female embodiment fantasies”.  How delightfully it allows one to then state that, of course, women naturally see themselves as female embodied as they have sexual fantasies.  See, women are autogynephilic too.  Or, as some twist it around, “Blanchard is defining normal female sexuality as a paraphilia!”

Autogynephilia is stigmatizing. It’s pathologizing. Reconceptualizing it (despite whether cis women have autogynephilia or not) as female embodiment fantasies allows trans women with supposed autogynephilia to feel more comfortable about labeling themselves.

Women are not sexually aroused by, nor become romantically enamored with, their femaleness

Here’s where Kay Brown is wrong. I don’t like the blog or the person here, but they did have a survey that found significant levels of attraction to femaleness among cis women that’s about the same/slightly higher than in trans women. Furthermore, Veale et. al 2008 used Blanchard’s scale and found that cis women do qualify as autogynephilic.

But that is what autogynephilia is… sexual arousal and/or romantic attachment to the contemplation of becoming or being female in and of itself.

So, definitionally cis women have autogynephilia

If autogynephilia were an effect of gender dysphoria and a female gender identity, we would predict several consequences from that effect to show up in the data.  We would expect that those who were the most gender dysphoric from an early age, those who are the most naturally feminine from an early age, those who transition the soonest, to report the most autogynephilia.

This ignores at least one of the Blanchardianist skeptics argument about the origin of autogynephilia: that autogynephilia is a result of being closeted and internalizing gender dysphoria for long periods of time. If an individual is open about their gender dysphoria (see: ‘early transitioners’), then they would not have the decades of denial and internalization that the closeted individuals do (see: ‘late transitioners’). Having the highest level of gender dysphoria does not necessitate that autogynephilia is developed, but rather that the internalization of gender dysphoria does. She homogenizes anti-Blanchardianist arguments into one neat and nice strawman that she tears down.

 In Nuttbrock (2009), those who had begun Hormone Replacement Therapy as teenagers only 14% reported having any autogynephilic arousal to cross-dressing compared to 82% of the gynephilic subjects (of whom only one had started HRT as a teen).

This could alternatively lead to the conclusion that transition reduces and/or prevents autogynephilia from developing. This conclusion is further supported by evidence of reduction of FEFs/AGP after transition in adults.

Further, if it is an effect of gender dysphoria and of a female gender identity, we would expect that only those who experience gender dysphoria and claim a female gender identity, to experience autogynephilia

The contention that autogynephilia is a result of both gender dysphoria and a female gender identity is a vicious strawman for a few reasons. The idea that autogynephilia is a result of gender dysphoria and a female gender identity is only posited as a single etiology, which does not preclude the possibility of other etiologies of autogynephilia in individuals of different gender identities.


Nuttbrock et. al

Many transfolk make the claim that data supporting Blanchard’s hypothesis has not been replicated

Yes, because it hasn’t. The non-monotonicity has never been replicated, neither has his findings using penile plethysmograph or the social desirability scale. I could go on, but the short is that a lot of his research has not been replicated.

Already I’ve shown how it had been replicated by both the Leavitt&Burger study and the Smith Study, which was further refined by Lawrence when she showed that the statistical signal strength is increased when one carefully sorts by sexual history.

I’ll have to go read her articles on these two, but I doubt she’s accurately representing the research.

First, the conclusion, so as not to lose my reader’s attention:  The Nuttbrock paper confirms, absolutely confirms, the Freund/Blanchard two type taxonomy for Male-To-Female (MTF) transsexuals, one that is exclusively androphilic and one that is autogynephilic. There can be no doubt now.

There are a number of gigantic issues with this claim. First off is that the study never claimed that. Nuttbrock found a monotonic association between gynephilia and autogynephilia, found substantial differences between heterosexual and bisexual respondents, and concluded;

These limitations notwithstanding, we nonetheless conclude that a classification of the MTF population, based solely on sexual orientation, is fundamentally limited

Which seems to point out that Blanchard’s typology at its core is not as useful as other models.

The second issue is that an association between autogynephilia and gynephilia (which every study on sexuality and trans people has found) is not contested. I have never met anyone who contests that, I know I don’t. The contest is over whether a correlation proves causation (it doesn’t), and whether we can coherently create a mutually exclusive typology (we can’t). Nuttbrock et. al does not do any research on those claims, which makes it problematic to make the type of conclusion Brown does from the study.

Lawrence took the Smith study data set and further sorted out those who self-reported being androphilic but had a sexual history of being attracted to women, as being non-homosexual.

Here’s the issue. A lot of homosexual men (cis gay men) have a history of attraction to women. It’s pretty difficult to coherently categorize people’s sexualities into nice neat boxes when sexuality has been shown to change over time, and that the boundaries between bisexual and homosexual are permeable (Is homoflexible subsumed under the category of bisexual or homosexual? Or is it a distinction orientation?) What Lawrence did is disingenuous to say the least.

A very important study that Brown forgets to cite here is Nuttbrock’s reply to Lawrence, which replies to some of Lawrence’s claims and points out her fundamental misunderstand of a few core methods used in the study;

Our analysis showed that age and ethnicity, in addition to homosexuality, were statistically significant predictors of transvestic fetishism. Lawrence pointed to the moderately high correlations among homosexuality, age, and ethnicity and asserted that our multivariate regression analysis of these three predictors of transvestic fetishism was flawed due to multicollinearity (Light, 1995). The‘‘moderate effect sizes’’ among these predictor variables, according to Lawrence, rendered our regression analysis invalid. This reflects a basic misunderstanding of multicollinearity

Furthermore, they detail problems with Blanchard’s typology (these quotes are out of order and miss headings and other portions of the reply);

Blanchard’s dualistic typology of homosexual versus nonhomosexual or autogynephilic gender dysphoria is one broadly defined dimension for classifying this population, but it is simply too crude to reflect the diversity of sexuality in this population and to serve as a singular basis for understanding these individuals and their sexuality across differentage groups and ethnic categories.

At odds with this strong prediction, in Blanchard’s studies, and our study as well, some of the homosexual MtFs reported transvestic fetishism and, in contrast, some of the non-homosexual MfFs did not do so. The cases not predicted by Blanchard’s theory have been assumed to reflect a combination of measurement errors whereby some homosexual MtFs over-report autogynephilia while non-homosexual MtFs under-report autogynephilia (Blanchard, Clemmensen, & Steiner, 1985; Blanchard, Racinsky, & Steiner, 1986). Some under-reporting of this phenomenon may indeed occur in clinic-based studies (such as Blanchard’s) but invoking this particular combination of measurement errors in a broad-based community sample (such as ours) is highly speculative. A more productive approach would be to better understand aspects of sexuality in this population that are at odds with autogynephilia theory

Lawrence, following Blanchard, claimed that a dichotomous measurement of sexual orientation (homosexual vs. nonhomosexual) is a basic divide that fully classifies and sufficiently describes this population. This broad dualistic typology necessarily assumes that differences across categories of non homosexuals(heterosexual,bisexual,andasexual)arenotempirically
and theoretically significant. We found that bisexual MtFs reported statistically significant lower levels of lifetime transvestic fetishism and (if they reported it during adolescence) they were more likely to ‘‘age out’’ of it during post-adolescence. Rather than ignoring these differences, as Lawrence chooses to do, perhaps we should attempt to better understand them

We attempted to replicate Blanchard’s (1992) intriguing finding regarding a non-monotonic association between a continuous measurement of gynephilia and autogynephilia (including transvestic fetishism). Lawrence complained that we misread Blanchard’s original article regarding a hypothesized non-linear (inverted U) association between gradations of gynephilia and transvestic fetishism (as one indicator of
autogynephilia). Lawrence was indeed correct that Blanchard clearly distinguished between transvestic fetishism and autogynephilia and predicted that autogynephilia (not transvestic fetishism) would show a curvilinear association with gynephilia. Blanchard’s (1992) empirical analysis nonetheless showed that the level of transvestic fetishism, like the specific measurements of autogynephilia, dipped significantly at the highest level of gynephilia (Fig. 2). We attempted to replicate the non-monotonic association between transvestic fetishism and a continuous measurement of gynephilia

Ironically, it seems that Lawrence fundamentally misunderstands Nuttbrock, basic concepts in statistical analysis and possibly Blanchard himself.

Nuttbrock, et al.  followed the Smith example and used self-report, but with Blanchard’s original four categories, sorting into exclusively androphilic, bisexual, exclusively gynephilic, and asexual.  I must emphasize, this was self-reported sexual identity, not actual sexual history

And I must emphasize that sexual orientation is a political category not deterministically decided by past action, but contemporary engagement in sexual intercourse as well as relevant factors like current attraction and identification with the category. To classify androphilic

Another important difference between the Nuttbrock study and the others is that while all of the other subjects were from a gender reassignment clinic seeking somatic feminization, those in the Nuttbrock study were obtained through advertisements and direct contact, in the community at large

Which is an important factor to consider when assessing the validity of Blanchard’s typology. If it only applies to a specific subpopulation, then it does not explain a large part of the trans community.

This means that potentially, an important personality type may have been missed, as it has been remarked that asexual transsexuals are typically schizotypal (fancy word for non-social, shy, loner)

It’s not clear what she means by “important personality type” here.

 It also means that many of the subjects are not strongly motivated toward somatic feminization, instead simply identifying as “transgender” or “gender-queer”.  Indeed, 28% of the subjects are not even taking feminizing hormones.

The fact that 28% of the subjects were not taking feminizing hormones could easily be explained by a number of factors; lack of the diagnoses required to obtain the hormones, needing more money before being able to afford hormones, still being closeted, and so on.

It is important when reviewing the above data, that as in all sociological studies of taxa, because we have not yet found a perfect instrument by which to sort the taxa, and do not have a perfect instrument to detect autogynephilia, we are only able to statistically tease out the two types.  After all, we are asking people to be self-reflective, honest, and accurate, about something that is very personal and as yet poorly understood.  But about the existence of the two types, there is no doubt.

I think this is where some theory on what taxon/typologies are (we should note that taxon are primarily used in the biological sciences, while the terminology of typologies is used in the social sciences, with some important differences) is going to be useful. But that’s for another day.