The underlying question is easy to state and hard to answer: which part of the brain reflects biological sex, which part sexual orientation and which part gender identity? An international research team has tried to separate those three pieces using magnetic resonance imaging in 140 Thai adults, and their conclusion is that cortical structure tracks gender identity rather than orientation. The work is published in the journal Neuroscience.
Earlier studies had already observed that the cortex of transfeminine people looked more like that of cisgender women than that of cisgender men. But almost all of those studies came from clinic populations in Europe and the Americas and carried a familiar confound: gender identity was mixed up with sexual orientation. This study tackles both at once, in an Asian population and with participants recruited outside the clinic.
Four groups, one scanner
The authors compared heterosexual and gay cisgender men, heterosexual cisgender women, and transfeminine people attracted to men, known in Thailand as sao praphet song (“the second kind of woman”). Among the latter, some had used gender-affirming hormones on their own and others had never taken them, which allows the effect of identity and the effect of hormone therapy to be observed separately. All participants were aged 18 to 44 and went through the same MRI protocol to measure cortical thickness, surface area and overall brain size.
Sexual orientation does not explain the pattern
The first result is a process of elimination: the cortical thickness of gay men resembled that of heterosexual men. In other words, sexual orientation did not organise the anatomical differences observed. That strengthens the idea that what was being measured had to do with something else.
The second result points to identity. People with a feminine identity — cisgender women and sao praphet song who had never used hormones — showed greater cortical thickness in frontal and parietal regions. The pattern was not explained by sex assigned at birth, by orientation, or by having attended a clinic. It appeared just the same in transgender people who had not started any treatment.
Hormones leave a trace too
Among the transgender participants, the use of gender-affirming hormones was associated with a smaller brain and thinner cortices. The authors read this as an effect of hormone therapy on brain tissue and highlight a relevant particularity: many of them were self-medicating, a widespread and under-studied practice that makes it hard to know exactly which doses and durations were involved. This finding is consistent with previous clinic-based studies that had already described similar reductions after starting treatment.
What it means and what it does not
The study is cross-sectional: it photographs a single moment and cannot prove cause and effect. A thicker or thinner cortex is neither better nor worse, and says nothing about anyone’s cognitive ability; it is a structural measure with wide variation between individuals. The sample is also entirely Thai, and the transgender group consists of people attracted to men, so the results cannot be automatically extrapolated to all transgender identities or all cultural contexts.
Even so, the work adds two valuable things. The first is generalisation: the cortical phenotype described in Europe and America reappears in a very different population, suggesting it is not an artefact of clinical recruitment. The second is the separation of variables: by comparing identity, orientation, sex and hormone use within the same design, the data point to gender identity as the factor most closely related to cortical structure. The neurobiology of identity remains a field with many open questions, but every study that isolates one variable narrows down where to look for answers.
Source: Cortical structure and brain size in Thai transfeminine individuals with and without gender-affirming hormone use compared with cisgender heterosexual men, heterosexual women, and gay men. Skorska MN, Folkierska-Żukowska M, Rodkong A, et al. Neuroscience 2026;614:373-384. DOI: 10.1016/j.neuroscience.2026.07.051.
Article published on dosier.es. Image generated with artificial intelligence (marta ai → marta.syf.es).
