Neuroimaging studies in people with gender incongruence

Abstract

The current review gives an overview of brain studies in transgender people. First, we describe studies into the aetiology of feelings of gender incongruence, primarily addressing the sexual differentiation hypothesis: does the brain of transgender individuals resemble that of their natal sex, or that of their experienced gender? Findings from neuroimaging studies focusing on brain structure suggest that the brain phenotypes of trans women (MtF) and trans men (FtM) differ in various ways from control men and women with feminine, masculine, demasculinized and defeminized features. The brain phenotypes of people with feelings of gender incongruence may help us to figure out whether sex differentiation of the brain is atypical in these individuals, and shed light on gender identity development. Task-related imaging studies may show whether brain activation and task performance in transgender people is sex-atypical. Second, we review studies that evaluate the effects of cross-sex hormone treatment on the brain. This type of research provides knowledge on how changes in sex hormone levels may affect brain structure and function.

Keywords:

• Gender dysphoria
• sexual differentiation
• cross-sex hormone treatment
• organizing effects
• activating effects
• gender identity development
• transgender
• transsexual

Introduction

Gender incongruence (GI) refers to the feeling that the physical characteristics of the body are not in line with the experienced gender. This may (in the case of gender dysphoria) or may not be accompanied by distress. People with gender incongruence may socially transition to another gender, or may seek treatment to align their body to their experienced gender by cross-sex hormone treatment and surgical procedures. The Diagnostic Statistical Manual for Mental Disorders provides criteria for the diagnosis gender dysphoria (American Psychaitric Association, 2013). (For changes in the nomenclature of the diagnosis see Beek, Cohen-Kettenis and Kreukels, 2015) Likewise the current International Classification of Diseases and Related Health Problems provides criteria to diagnose transsexualism (World Health Organization, 1992; see also Beek, Cohen-Kettenis and Kreukels, 2015, elsewhere in this special issue). Treatment guidelines can be found in the World Professional Association for Transgender Health’s Standards of Care (SOC) (Coleman et al., 2011, SOC7). In this article we will use gender incongruence to refer to the condition as a whole; MtF when we refer to natal men who want to align their bodies to their female gender identity, or have begun treatment to do so, or have already transitioned; and FtM when we refer to natal women who want to align their bodies to their male gender identity, have begun treatment to do so, or have already transitioned.

Causal mechanisms for feelings of gender incongruence are unknown, but biological factors are suggested to play a role. Men and women have been shown to differ in several characteristics, but the largest difference may be found in gender identity: most women feel that they are women, and most men feel that they are men. Sex differences in brain morphology, connectivity, and function are thought to underlie sex differences in behaviour, psychopathology, and cognitive performance on certain tasks. Men have a larger brain volume than women and this is only partly due to the larger body size in men (Luders & Toga, 2010). Boys and girls show differences in the development of grey and white matter volume over the course of puberty (Giedd et al., 2012), and sex differences in the ratio of brain tissue compartments have been reported (Luders & Toga, 2010). Cortical thickness is generally higher in women compared with men (Luders et al., 2006).

Subcortically, the amygdala is larger in men and has a higher density of androgen than oestrogen receptors, whereas portions of the hippocampus are larger in women, with a higher density of oestrogen than androgen receptors (Halpern, 2012). Boys and girls (8–13.3 years) already differ in structural connectivity, and differences increase during development (Ingalhalikar et al., 2014). In maturity, men have greater intra-hemispheric connectivity while, in women, inter-hemispheric connectivity predominates.

Sex differences in the performance of cognitive tasks (i.e. visuospatial and verbal fluency tasks) have been found (Halpern, 2012; Voyer et al., 1995), as well as differences in brain activation patterns during these tasks (Hugdahl et al., 2006). Sometimes performance may be similar for men and women, but activation patterns differ (Schoning et al., 2007).

Animal studies examining the role of gonadal steroids in differentiating male and female mammals at morphological, physiological and behavioural levels have been guided by thoughts arising from embryological (Jost, 1947) and behavioural (Phoenix et al., 1959) studies (Guillamon & Segovia, 1996). In early (prenatal) development, gonadal steroids direct the sexual differentiation of the brain (organizing effects of sex hormones), and later during life, circulating hormones influence the brain (activating effects of sex hormones). Hormone levels may fluctuate or change during puberty, the menstrual cycle, menopause and hormone treatment. A prominent hypothesis for the mechanism behind feelings of gender incongruence is that exposure to sex hormones during prenatal development has led to atypical sexual differentiation of the brain, with the body and genitals developing in the direction of one sex, and the brain and gender in the direction of the other sex (Swaab & Garcia-Falgueras, 2009). The time window for prenatal sexual differentiation of the genitals precedes the time window for brain sexual differentiation. The results of a series of post-mortem studies at the laboratory of Swaab were pivotal for the formulation of this hypothesis. In several hypothalamic nuclei, a sex reversal was found in volume and neuron number in male-to-female transsexuals (MtF) (Garcia-Falgueras & Swaab, 2008; Kruijver et al., 2000; Zhou et al., 1995). The sexual differentiation hypothesis is studied by evaluating whether the brains of people with gender incongruence resemble those of their natal sex or their experienced gender.

Genetic factors also affect sexual differentiation of the brain (Ngun et al., 2011). Genes located on the X and Y chromosomes are likely candidates for direct genetic effects (Arnold, 2009). Studies of sex differences in the brain should consider whether they originate from hormonal or genetic factors, or both (McCarthy et al., 2012). In addition, environmental factors such as nutrition and stress may also influence brain structure and function (Wachs et al., 2014).

Some evidence exists for genetic factors in the development of gender identity and gender incongruence. The genetic background of gender incongruence is largely unidentified, but in twins, monozygotic (MZ) twin pairs show a higher concordance of gender incongruence than dizygotic (DZ) twin pairs (Heylens et al., 2012), suggesting genetic involvement in the development of gender incongruence. Several other studies looked into polymorphisms of sex steroid-related genes as possible candidates for a role in the development of gender incongruence (Bentz et al., 2008, Fernandez et al., 2014; Hare et al., 2009; Henningsson et al., 2005).

First, we will review the literature on brain studies that have been performed in people with gender incongruence before the start of treatment. Differences in presentation and outcome are reported in people with gender incongruence with regard to their sexual orientation (Smith et al., 2005). Because different developmental trajectories may underlie these subtypes, we will provide information on sexual orientation if this is available from the papers. Second, we will review studies aiming to examine the effects of cross-sex hormone treatment on the brain, because testosterone (Höfer et al., 2013) and estradiol (Casanova et al., 2011; Resnick et al., 2009) therapies have been associated with structural and functional changes in the brain.

Brain phenotypes of people with gender incongruence

Grey matter – volumetric and cortical thickness studies

As mentioned above, the total volume of the brain is generally larger in men than in women. Before treatment, similar volumes to natal sex are found in adults (Hahn et al., 2014; Rametti et al., 2011b; Savic & Arver, 2011) and adolescents (Hoekzema et al., 2015) with gender incongruence. However, MtF (natal men with a female gender identity) had a total intracranial volume between those of male and female controls (Hahn et al., 2014).

Regional grey matter (GM) variation in MtF, who were mixed with regard to sexual orientation, was more similar to the pattern found in men than in women (Luders et al., 2009). In cortical regions in which no sex differences were found by Savic and Arver (2011), namely the right parieto-temporal junction, the right inferior frontal and the insular cortices untreated non-androphilic MtF, that is to say those who were not sexually attracted to men, had a larger GMvolume compared to both control groups.

Participants with a female gender identity (female controls and androphilic (those with a sexual attraction to men) MtF) showed larger GM volumes than both male controls and FtM in the right middle and inferior occipital gyri, the fusiform, the lingual gyri and the right inferior temporal gyrus (Simon et al., 2013). Participants with a male gender identity (male controls and FtM) had larger GM volumes than both female controls and MtF in the left pre- and postcentral gyri, left posterior cingulate, calcarine gyrus and the precuneus (Simon et al., 2013).

At whole-brain level, adolescents with gender incongruence differed in GM volume from adolescents sharing their gender identity, and not from adolescents sharing their natal sex, arguing against a sex-atypical differentiation of the brain in gender incongruence with respect to this measurement (Hoekzema et al., 2015). However, when examining GM volumes within sexually dimorphic structures using region of interest analyses (regions in which sex differences are found), subtle deviations from the natal sex were observed in MtF adolescents in the direction of adolescents sharing their gender identity.

MtF showed higher cortical thickness compared to men in the control group in sensorimotor areas in the left hemisphere and right orbital, temporal and parietal areas (Luders et al., 2012). This study did not include a female control group, and the sexual orientation of the participants was mixed. A Spanish cortical thickness (CTh) study that included a male and a female control group found similar CTh in androphilic MtF and female controls, and increased CTh compared with male controls in the orbito-frontal, insular and medial occipital regions of the right hemisphere (Zubiaurre-Elorza et al., 2013). The CTh of FtM was similar to control women, but FtM, unlike control women, showed (1) increased CTh compared with control men in the left parieto-temporal cortex, and (2) no difference from male controls in the prefrontal orbital region.

With respect to subcortical structures, the GM volume of the right putamen of MtF was in the female range (larger than in men) in one of the earlier studies (Luders et al., 2009). However, in Savic & Arver’s study (Savic & Arver, 2011), untreated MtF had a relatively smaller putamen than both male and female controls. In yet another study, no differences were found in the right putamen between untreated MtF and male and female controls (Zubiaurre-Elorza et al., 2013). These three studies applied different techniques and their samples differ in sexual orientation. Savic & Arver studied gynephilic MtF, the Spanish study included androphilic MtF and Luders et al.’s paper had a mixed sample.

FtM, like male controls, showed a larger volume of the putamen than female controls (Zubiaurre-Elorza et al., 2013).

White matter studies

Fractional anisotropy (FA) is a measure of white matter microstructure. Measured via diffusion tensor imaging (DTI), men have greater FA values than women (Rametti et al., 2011a, 2011b). Before hormonal intervention, androphilic MtF with feelings of gender incongruence that began in childhood appeared to have a white matter microstructure pattern that differs statistically from male as well as female controls (Rametti et al., 2011b). Their values lie in between those of male and female controls in several brain fascicles in the right hemisphere, except for one, where they present a masculine pattern.

FtM FA values are significantly greater in several fascicles than those belonging to female controls, but similar to those of male controls, thereby showing a masculinized pattern. However, their corticospinal tract is defeminized; that is, their FA values lie between those of male and female controls, and are significantly different from each of these two groups (Rametti et al., 2011a).

Kranz et al. (2014b) also studied white matter microstructure by DTI in MtF, FtM, control men and control women. They found widespread, significant differences in mean diffusivity between groups in almost all white matter tracts, but no differences in FA values. Significantly increased mean diffusivity (MD) values were found in MtF compared to control men, and significantly decreased MD values in FtM compared to control women. MD values (and axial and radial diffusivity) were associated with plasma testosterone levels. The participants in this study were mixed with regard to sexual orientation. Controlling for sexual orientation did not result in changes in the findings.

Connectivity profiles

Hahn and colleagues (2015) studied structural connectivity networks in transgender people. For MtF, they found a decreased hemispheric connectivity ratio in subcortical/limbic regions when compared to male and female controls, which seemed to be driven by an increased inter-hemispheric lobar connectivity. FtM showed decreased intra-hemispheric connectivity between the right subcortical/limbic and right frontal and temporal lobes compared with male and female controls and MtF. The differences between MtF and FtM in nature and direction of brain connectivity suggest that feelings of gender incongruence are ‘accompanied by pronounced but distinct structural signatures for FtM and MtF’ (Hahn et al., 2015). They argue that it is important to identify unique features in future studies in transgender people. They do not report on sexual orientation in this study, but presumably these are the same subjects as the other study from the same group (Kranz et al., 2014b), who were mixed with regard to sexual orientation. The same group has addressed asymmetry in the serotonin transporter system (Kranz et al., 2014a) because sex differences have been reported in neurotransmitter systems (Cosgrove et al., 2007). Men show a strong rightward serotonin transporter asymmetry in the midcingulate cortex, which is absent in both women and MtF.

Regional cerebral blood flow (rCBF) of a small sample of FtM showed a significant decrease in rCBF in the left anterior cingulate cortex, and a significant increase in the right insula in FtM compared with female controls (Nawata et al., 2010).

Lin et al. (2014) hypothesized that three key regions of the body representation network (primary somatosensory cortex, parietal lobe and insula) would show a higher degree of centrality in untreated transgender people compared with controls; the degree of centrality is an index of the functional importance of a node in a neural network. They did indeed find a higher degree of centrality in the bilateral parietal lobe and the somatosensory cortex in homosexual transgender people (MtF and FtM combined) compared to controls (male and female combined). The authors did not distinguish between MtF and FtM.

Another study by the same group investigated functional connectivity in transgender people when they were viewing erotic and non-erotic interactions of male–female couples (Ku et al., 2013). Unlike control men and women, transgender people (again MtF and FtM combined) presented an increased functional connectivity between the ventral tegmental area (VTA) and the anterior cingulate cortex subregions.

The functional connectivity profile of an untreated FtM (who also had polycystic ovary syndrome) was more similar to female controls than to male controls (Santarnecchi et al., 2012). This study examined resting state fMRI, which means the participants were not involved in a task or activity while their brains were imaged.

Task-related imaging studies

Men and women show differences in hypothalamic activation in response to certain chemosignals such as androstadienone (Savic et al., 2001). A sex difference in the hypothalamic response is already apparent in pre-pubertal children (Burke et al., 2014). Gynephilic MtF adults show similarities with control women in hypothalamic activation while smelling odorous steroids (Berglund et al., 2008). Adolescents with gender incongruence showed a response to androstadienone that was similar to their experienced gender (Burke et al., 2014). For children, the findings were less clear: boys with GI responded in a similar way to the control boys, whereas girls with GI showed neither a male nor a female pattern of activation in the hypothalamus.

Sex differences in (sub)cortical activation patterns in response to erotic stimuli have been established (Stoleru et al., 2012). We have already seen above that people with gender incongruence show differences in their connectivity profiles while watching erotic interactions (Ku et al., 2013). Brain activation patterns while viewing erotic videos in MtF (mixed with regard to sexual orientation) were found to be similar to control women (Gizewski et al., 2009).

With regard to emotional processing, well-acknowledged as showing differences between the sexes, FtM who were gonadally suppressed showed less activation compared to control women in the right superior temporal lobe during processing of positive affective images (Soleman et al., 2014). However, this difference was unrelated to hormonal levels. It may be that the FtM were already different from the female control group before the start of GnRH analogues and that the observed group differences were due to prenatal hormonal influences rather than circulating hormones.

Having established that men and women differ in voice gender perception (Junger et al., 2013), the same group studied neural activation during perception of male and female voices in MtF with various sexual orientations compared to control men and women (Junger et al., 2014). They did not find differences between untreated and treated MtF, nor between sexual orientation groups. MtF differed from both control groups when listening to male versus female voices, ‘supporting the notion of an intermediate position between men and women.’ (Junger et al., 2014, p. 7).

Men generally have less trouble with visuospatial tasks, whereas women generally outperform men in verbal fluency tasks. These sex-typical cognitive abilities have been studied in people with gender incongruence to determine whether they show performance and activation patterns like their natal sex or experienced gender. Sex differences in brain activation during mental rotation have been shown: men show predominantly parietal activation, while women additionally show more inferior frontal activation (Hugdahl et al., 2006). In a mental rotation study, MtF differed from controls of their natal sex in brain activation during this visuospatial task: control men showed greater activation in the left parietal region, while untreated and hormone-treated MtF exhibited stronger activation in the temporal-occipital regions than control men (Schöning et al., 2010). The sexual orientation of this latter group was unknown. Long-term administration of oestrogens to MtF produced a decrease in activation of the parietal cortex that correlated negatively with the number of months of hormonal treatment (Carillo et al., 2010).

Verbal fluency during brain activation was studied in adolescents with GI and controls (Soleman et al., 2013). Control boys differed from girls and showed more activation in the right Rolandic operculum during phonetic fluency. Imaging data did not reveal significant differences between the adolescents with GI and controls in neuronal activation, although there was a trend of linear increase within the Rolandic operculum from girls to FtM to MtF to boys.

Executive functioning is still developing during adolescence and is therefore studied in adolescents with gender incongruence who are being treated with puberty-suppressing medication. Brain activation levels of untreated adolescents with GD fell between the two control groups in the areas that showed significant sex differences in the controls (Staphorsius et al., 2015). Hence, untreated MtF and FtM had a closer resemblance to each other compared to control men and control women, and no sex differences were found in adolescents with GI. Interestingly, the MtF who were treated with puberty suppression showed greater activation than the FtM on puberty suppression in the same region which was more active in control boys compared to control girls, indicating sex-typical brain activations. The gonadotrophin‐releasing hormone agonists (GnRHa)-treated adolescents with GI even appeared to have exaggerated sex-typical activation of the ROIs (Staphorsius et al., 2015). Unexpectedly, puberty suppression seemed to make some aspects of brain functioning more in accordance with the natal sex.

Effects of cross-sex hormone treatment

The brain is sensitive to physiological sex hormone changes during the menstrual cycle (Osserwaarde et al., 2013) and pregnancy (Oatridge et al., 2002), as well as during estradiol (Casanova et al., 2011; Resnick et al., 2009) and androgen (Höfer et al., 2013) therapies. It is surprising that, after more than half a century of cross-sex hormone treatment of transsexuals, only a few recent works have addressed this aspect of clinical importance. For people who receive cross-sex hormones as part of their gender-confirming therapy, it is important to know how this treatment affects their brains. In addition, these studies help us to clarify the effects of androgens and oestrogens on the brain.

The three works we could find in the literature used a longitudinal design and studied transsexuals before and after several months of hormone treatment. Oestrogen plus anti-androgen treatment is associated with a decrease in brain volume in MtF ‘towards female proportions’ (Hulshoff Pol et al., 2006). Moreover, the ventricles were seen to increase in volume. After six months of cross-sex hormone treatment, a general decrease in cortical thickness as well as cortical and subcortical volume was found, and (probably as a consequence) an expansion in ventricle volume in MtF (Zubiaurre-Elorza et al., 2014)

In FtM, testosterone treatment resulted in increased total brain and hypothalamus volumes (Hulshoff Pol et al., 2006), and increases in cortical thickness, cortical volume and subcortical volume were observed in FtM treated with cross-sex hormones (Zubiaurre-Elorza et al., 2014). Changes in CTh in parietal and occipital regions of the left hemisphere seemed to be correlated with the increments in the serum testosterone and free testosterone index.

The effects of testosterone treatment on white matter microstructure were also studied in FtM (Rametti et al., 2012). After seven months of treatment, FA values increased in two fascicles compared to pre-treatment values. These increments in FA were predicted by the free testosterone index before the testosterone treatment: the higher the testosterone index before hormonal treatment, the higher the increases in FA values in these fascicles under androgenization.

It has been suggested that the increases in cortical thickness observed in FtM under androgenization are due to the anabolic effects of testosterone. The decrease in cortical thickness and expansion of the ventricles found in MtF might be due to the suppression of the normal anabolic effect of testosterone on the brain due to the administration of antiandrogens, plus probably the deleterious effects of estradiol (Zubiaurre-Elorza et al., 2014). Indeed, post-menopausal substitutive estradiol therapy produces decreases in GM (Casanova et al., 2011; Resnick et al., 2009).

Discussion

We have reviewed the literature on brain imaging studies in vivo in people with gender incongruence. To the best of our knowledge, we have covered the entire structural and functional literature. We would like to stress that the main constraints we found were not only that there are still only a few studies on this subject, but also that techniques, design and samples are very diverse across studies. To reach a clear picture of the brains of people with gender incongruence, this should be overcome in future studies by avoiding the use of mixed samples with respect to age, onset age of feelings of gender incongruence and sexual orientation.

Nevertheless, a brain phenotype becomes apparent from DTI and CTh in early onset androphilic MtF and gynephilic FtM. Their gross morphology is similar to their natal sex, but white matter microstructure is demasculinized in androphilic MtF and masculinized in gynephilic FtM. Moreover, androphilic MtF and gynephilic FtM present with a feminine cortical thickness, but they differ from control men in various regions of the cortex (Guillamon, 2014; Guillamon et al., submitted). It is clear that there is not a complete sex reversal in brain structures in people with gender incongruence. This latter point is also supported by the GM study in adolescents with GI, which shows that their volumes are, at a whole brain level, in line with their natal sex (Hoekzema et al., 2015).

Currently, studies have not yet reported on androphilic FtM. Gynephilic MtF seem to differ from androphilic MtF, although direct comparisons have not yet been performed. A study that included only gynephilic MtF did not find any indications for atypical sexual differentiation in this group (Savic & Arver, 2011)

A difference in brain phenotype of people with GI compared to natal sex controls in various brain measures suggests a sex-atypical development of the brain. However, it remains unclear whether these changes originate from prenatal organization alone. Knowledge of the development of the brain during adolescence (Giedd et al., 2012), and the importance of puberty in the clinical presentation of GI (Steensma et al., 2013), suggest that this period is pivotal in understanding the development of GI. Recent work that found subtle deviations in GM volume (Hoekzema et al., 2015), and brain activation during executive functioning from their natal sex (Staphorsius et al., 2015), as well as a response to a pheromone-like substance that was similar to their experienced gender in transgender adolescents (Burke et al., 2014), underscores the need to determine the timing and nature of sex-atypical organization.

Reviewing the literature on the effects of cross-sex hormone treatment, the brain is astonishingly understudied compared to other body systems and organs (Gooren et al., 2015), despite the fact that the brain is abundant with androgen (Puy et al., 1995) and oestrogen (Osterlund et al., 2000) receptors. If the brain is already sensitive to differences in endogenous levels of sex hormones during the menstrual cycle (Protopopescu et al., 2008) and puberty (Peper et al., 2009), one could imagine that the administration of pharmacological doses of these hormones will have an effect as well. Indeed, there are studies that show important effects on the brain of cross-sex hormone treatment. After no more than seven months of cross-sex hormone treatment, we see decreased intracranial brain volume in MtF compared to controls (Hulshoff Pol et al., 2006), a generalized decrease in CTh and an expansion of the ventricles (Zubiaurre-Elorza et al., 2014). After androgenization, an increase in intracranial volume and hypothalamus volumes (Hulshoff Pol et al., 2006) and increases in CTh have been found in FtM (Zubiaurre-Elorza et al., 2014). Suppression of androgens and/or the deleterious effects of estradiol may induce the decreases in MtF, whereas increases in FtM may be the result of anabolic effects of testosterone (Guillamon, 2014, WPATH Bangkok; Guillamon et al., submitted; Zubiaurre-Elorza et al., 2014). These ideas are, at least with respect to estradiol, supported by the detrimental effects on the morphology of the brain of post-menopausal substitution therapy (Casanova et al., 2011; Resnick et al., 2009).

Although the number of studies examining the brain of people with GI is still low, they have taught us that brain phenotypes for FtM and MtF seem to exist, and provided evidence for the role of prenatal organization of the brain in the development of gender incongruence. Future studies should focus on different developmental trajectories (persisters versus desisters; early versus late onset of GI feelings) and should also examine the role of sexual orientation. Not all children with GI become adolescents or adults with GI, see Ristori & Steensma elsewhere in this special issue, and not all adults with GI have been children with GI. Some have an onset of GI feelings in early childhood, whereas others may have an onset of these feelings during or after puberty (Nieder et al., 2011).

With regard to the effects of the cross-sex hormone treatment on the morphology of the brain, animal models may help us to disentangle these effects. To determine whether puberty suppression and cross-sex hormones have detrimental effects on the brain we should continue to evaluate the short and long-term effects on brain measures of current treatment regimens.

Declaration of interest

A.G. is supported by a Spanish grant from the Ministerio de Economia y Competitividad (PSI2014-58004-P). The authors report no further conflicts of interest. The authors alone are responsible for the content and writing of the paper.