REVIEWS AND OVERVIEWS

Mechanisms of Psychiatric Illness

Transgender Research in the 21st Century: A Selective

Critical Review From a Neurocognitive Perspective
Sven C. Mueller, Ph.D., Griet De Cuypere, M.D., Ph.D., Guy T’Sjoen, M.D.

Gender dysphoria describes the psychological distress Forty percent of transgender persons endorse suicidality, and
caused by identifying with the sex opposite to the one the rate of self-injurious behavior and suicide are markedly
assigned at birth. In recent years, much progress has been higher than in the general population. Individual factors
made in characterizing the needs of transgender persons contributing to mental health in transgender persons include
wishing to transition to their preferred gender, thus helping to community attitudes, societal acceptance, and posttransition
optimize care. This critical review of the literature examines physical attractiveness. Neurobiologically, whereas structural
their common mental health issues, several individual risk MRI data are thus far inconsistent, functional MRI evidence in
factors for psychiatric comorbidity, and current research on trans persons suggests changes in some brain areas con-
the underlying neurobiology. Prevalence rates of persons cerned with olfaction and voice perception consistent with
identifying as transgender and seeking help with transition sexual identification, but here too, a definitive picture has yet
have been rising steeply since 2000 across Western countries; to emerge. Mental health clinicians, together with other
the current U.S. estimate is 0.6%. Anxiety and depression are health specialists, have an increasing role in the assessment
frequently observed both before and after transition, and treatment of gender dysphoria in transgender individuals.
although there is some decrease afterward. Recent research
has identified autistic traits in some transgender persons. Am J Psychiatry 2017; 174:1155–1162; doi: 10.1176/appi.ajp.2017.17060626

Identifying with the sex opposite to the one assigned at birth terms “trans persons” or “transgender persons” throughout
often evokes serious distress and dysphoria, which are likely to this review. The term “transgender men” is used for persons
be exacerbated by the legal, social, vocational, family, and transitioning from female to male and “transgender women”
relationship consequences of such an identification, including for persons transitioning from male to female. In most cases,
social stigma. Moreover, such distress can lead to a variety of but not all, we refer to persons with gender dysphoria visiting
mental health problems. As a result, many individuals expe- mental health services and seeking gender-affirming surgery.
riencing gender dysphoria search for gender-transition-related Whereas some individuals may not be distressed by expe-
care (1). Our aim in this selective review is to provide an riencing incongruence between their gender identity and
overview of current research in transgender persons, focusing their sex assigned at birth, a formal DSM-5 diagnosis of
on mental health and mental health problems and their neu- gender dysphoria is still necessary in many countries to fa-
robiological and neurocognitive and affective underpinnings. cilitate access to mental health care and other professional
Currently, to facilitate transition, lines of care include cross-sex care that will help in the transition to the experienced gender.
hormone treatment (CSHT) and/or gender-affirming surgery. The term “cisgender persons” is used throughout the review
The beneficial effects of these lines of care on mental health to describe persons who did not transition and who identify
and neurobiology will be scrutinized here. with their sex assigned at birth, for example, a female
Transgender care is complex and highly interdisciplinary. identifying as female and with female sex assigned at birth.
It involves a variety of health professionals, including en-
docrinologists, gynecologists, urologists, surgeons, voice and
EPIDEMIOLOGY
communication specialists and therapists, dermatologists,
and various providers of mental health services (psychia- Epidemiologically, true prevalence rates of transgenderism
trists, clinical psychologists, counselors, etc.) (2). In line with are difficult to establish because of sampling biases (2).
the goals of the World Professional Association for Trans- Concerns in establishing prevalence rates start with the
gender Health (2) to de-psychopathologize being trans- criteria for inclusion, that is, whether the person has already
gender, the terminology has been constantly evolving in undergone gender-affirming surgery, is receiving CSHT, or
recent years. For persons identifying with the sex opposite to has presented with gender dysphoria to a mental health
the one assigned at birth, we use the currently preferred provider. Because of this bias, estimated prevalence rates

See related features: Clinical Guidance (Table of Contents), CME course (p. 1233), and AJP Audio (online)

Am J Psychiatry 174:12, December 2017 ajp.psychiatryonline.org 1155

TRANSGENDER RESEARCH IN THE 21ST CENTURY

have traditionally been underestimates, as some gender- Interestingly, in addition to the axis I disorders mentioned
variant individuals (i.e., associating oneself with a sex differ- above, attention in the field has recently shifted to the presence
ent from the one assigned at birth without being dysphoric) of autism spectrum disorders or autistic traits in transgen-
may never present to a mental health care provider. To der persons. While any conclusions would still be premature,
overcome these limitations, recently published prevalence currently available data appear to suggest higher rates of co-
estimates are conducted using meta-analytic methods or occurring autistic traits in trans persons with gender dysphoria
population-based surveys. Using a meta-analysis on reported relative to the general population in Dutch children (N=204)
prevalence rates, Arcelus et al. (3) estimate the prevalence of (15) and British adults (16). In the latter study (16), trans men
persons being transgender at around 4.6 per 100,000, with (N=61) but not trans women (N=198) had higher autism quotient
6.8 for transgender women and 2.6 for transgender men. scores than cisgender men and women but lower scores than
Exploiting nationally available population surveys in the 125 patients with a diagnosis of Asperger’s syndrome. These data
United States, Flores et al. (4) suggest a total prevalence of 0.6% suggest that while autistic traits in trans persons may be higher
for adults in the U.S. population identifying as transgender than in the general population, they may be lower than in the
persons. The prevalence of gender dysphoria has been in- population of people on the autism spectrum and are dependent
creasing worldwide in the past two decades. Potential reasons on sexual orientation or sex.
could be that concerned persons feel freer to disclose their Despite these concerning rates of psychopathology and
dysphoria and to seek help because of more openness in so- axis I and II disorders, emergent longitudinal research has
ciety, changes in legislation, and a trans-affirmative approach begun to document the positive change in mental health
on the part of mental health and medical care providers, the associated with successful transition. This research shows
media, and social media. However, other changes in inter- or significant reductions in psychopathology after transition,
intraindividual factors are also plausible. down to a level resembling normative data (for a compre-
Of note, genetic issues in the synthesis of steroid hormones hensive review, see reference 17). In one longitudinal study of
do not appear to contribute to the prevalence of gender 107 transgender persons (18), symptoms of anxiety, depres-
dysphoria. Individuals with disorders of sexual development, sion, and psychoticism and measures of global severity and
such as congenital adrenal hyperplasia, complete or partial functional impairment were all significantly reduced after
androgen insensitivity syndrome, or other, more rare variants 12 months of CSHT. Similarly, another study (19) also tested the
of disorders of sexual development, do not meet the diagnostic beneficial effects of CSHT on mental health in both a cross-
thresholds for gender dysphoria (5, 6). Rather, it has been sectional and longitudinal design. Cross-sectionally, the study
reported that women with congenital adrenal hyperplasia may found that CSHT alleviated depressive symptoms in trans men
be more likely to have a non-heterosexual orientation (7, 8). but not trans women relative to trans persons not receiving
hormones. With hormone therapy, levels of body uneasiness in
both genders also diminished. Crucially, longitudinally across
MENTAL HEALTH IN TRANSGENDER PERSONS
four time points (at 3, 6, 12, and 24 months of follow-up), CSHT
Much cross-sectional work on mental health in transgender was associated with significant reductions in psychopathology
persons has focused on documenting the psychological or (on the Symptom Checklist–90 global severity index), depressive
psychiatric problems that may affect a transgender person. In symptoms, body uneasiness levels, and gender dysphoria in both
a large initiative across four European countries, Heylens genders (19). In sum, while being trans may be associated with
et al. (9) reported that 38% of transgender persons with increased mental health problems, predominantly affective dis-
gender identity disorder (N=305) currently had axis I di- orders, promising findings are emerging that indicate a reduction
agnoses, the majority of which were affective (27%) and of mental health–related psychopathology, including depression
anxiety disorders (17%). Notably, lifetime prevalences of axis and body dissatisfaction, with time and hormonal treatment.
I diagnosis, affective problems, and anxiety were estimated at Nonetheless, a major concern in trans persons remains the
70%, 60%, and 28%, respectively, while axis II diagnoses were risk of suicidality (suicidal thoughts, suicide attempts, and
estimated at 15%. Substance use disorders, eating disorders, suicide rates) as well as nonsuicidal self-injury, such as
and psychotic disorders were less prevalent, at 16%, 2%, 1%, cutting, hitting, or burning oneself. These behaviors may
respectively. The problematic high rates of psychological reflect a maladaptive way to regulate one’s emotion, or self-
and/or psychiatric problems in trans persons remained even punishment in response to external or internal pressure and
when participants were matched with cisgender compari- social stigma. Because of these high-risk behaviors, the past
son subjects and additional factors such as age, natal sex, and decade has seen a surge in studies assessing the presence of
new assigned sex were taken into account in a retrospective suicidality and nonsuicidal self-injury in trans persons, with
Swedish cohort study (10). While such high rates of psy- steadily increasing sample sizes. Rates of lifetime suicidality
chopathology are consistently reported in the literature, and suicide attempts for trans persons are alarming, ranging
there is disagreement regarding sex specificity. Whereas from 30% to 81% (the prevalence rate of suicide attempts in
some studies document higher prevalences of psychopa- the overall U.S. population is estimated at 4.6%) (9, 20–22).
thology in transgender women relative to transgender men Significant predictors of suicide include a past history of
(11–13), one study reported the reverse (14). maltreatment, gender victimization, depression, substance

1156 ajp.psychiatryonline.org Am J Psychiatry 174:12, December 2017

 MUELLER ET AL.

abuse, and young age (20–22). Rates for lifetime nonsuicidal nontrans respondents from the LGB community (27). This
self-injury in the pretransition period are estimated at 38% report also uncovered continuing problems that trans per-
for the total trans population, with 57.7% in trans men and sons face with stigma, harassment, and violence and mal-
26.2% in trans women (11). Although another study reported treatment in their personal and professional lives. Forty-six
much lower rates of nonsuicidal self-injury (19%) (23), both percent of trans persons in the survey reported having felt
studies found that trans men were more at risk of nonsuicidal discriminated against or harassed within a 12-month period
self-injury than trans women. These results are shared by in- because of being perceived as trans (27), including 30% of
vestigations in pediatric populations, which also report elevated trans persons when looking for a job and 23% at the work-
rates of self-harm in children and adolescents with gender place. Such social and environmental factors very likely
dysphoria (24), suggesting that suicidality and self-injurious contribute to mental health problems and suicidality. In a
behavior are prominent across the lifespan in this population. sample of U.S. LGBT youth (N=246, ages 16–20) (28), pro-
Long-term data assessing suicidality and nonsuicidal self- spective LGBT victimization and low social support were
injury after transition are fewer but continue to signal cause associated with suicidal ideation, and self-harm was asso-
for concern. Retrospective data on 1,331 trans persons who ciated with (among other factors) childhood gender non-
visited a university gender clinic (25), with a median follow-up conformity, prospective hopelessness, and victimization.
of 18.5 years, indicated high mortality rates among trans Interindividual factors, including sexual orientation and
women, with a total mortality 51% above the general pop- physical appearance, may also contribute to well-being and
ulation average (rates of total and cause-specific mortality mental health. In an interesting investigation into putative
among trans men did not differ significantly from those of the subtypes within the transgender population, Smith et al. (29)
general population). Causes of death included heart and lung gathered data from 187 trans persons who had completed
problems, neoplasm, HIV/AIDS, and suicide (17 trans women gender-affirming surgery and examined the role of sexual
relative to one trans man). Death by suicide and suicide at- orientation on personality variables, physical appearance,
tempts after transition were equally alarming in the Swedish onset of gender dysphoria (documented gender identity
cohort study by Dhejne et al. (10), who reported an incidence disorder), and psychopathology. Participant’s physical ap-
of 2.7 per 1,000 person-years for death by suicide and 7.9 for pearance (facial hair, height, figure, nose, muscularity,
suicide attempts. Notably, the National Transgender Dis- speech, etc.) was rated in terms of how compatible this ap-
crimination Survey uncovered a high prevalence of 41% for pearance was with the new (preferred) sex. These rat-
suicide attempts in trans persons—double that reported in the ings showed that androphilic trans men compared with
LGB (lesbian, gay, bisexual) community, which is estimated at gynephilic trans women had a higher matching physical
10%220% for the U.S. population (26). Such high rates of appearance. Moreover, androphilic trans women and gyne-
suicide risk in trans persons, even after transition, would philic trans men, compared with gynephilic trans women
mandate that trans persons presenting at the clinic not only be and androphilic trans men, reported more gender identity
assessed for self-injurious behavior but also be monitored disorder symptoms in childhood. Similarly, trans men also
longitudinally. reported more gender identity disorder symptoms at early
Indeed, the recognition of being trans as an identity, rather ages relative to trans women and applied for transition at
than as a mental health disorder (2), is shifting the focus of an earlier age. These results may point to different under-
the clinician’s preassigned role. Given the high rates of mood lying factors and motivations. Smith et al. (29) suggest that
and affective disorders as well as high suicidality in trans whereas androphilic trans women may have a strong aversion
persons, especially before transition, clinicians must be able to their sex organs, motivating earlier transition, gynephilic
to diagnose not only gender dysphoria but also interrelated trans women may have grown up as masculine boys and may
psychological and/or psychiatric problems as well as psy- also have entered marriage and fathered children, thus
chological problems that may masquerade as gender dys- leading to later transition. Similarly, androphilic trans men
phoria. In addition, it is recommended that mental health reported more psychological problems than gynephilic
care providers be familiar with the criteria for diagnosis and trans men, suggesting that a combination of sexual attraction
treatment of gender dysphoria and be able to care for the to a specific gender, physical appearance, and societal expec-
patient during the entire transition process and afterward. tations may exert different effects on individuals with differ-
ent sexual orientations or different age at onset of gender
dysphoria (30).
SOCIAL, ENVIRONMENTAL, AND
The data are currently ambiguous, however, as an Italian
INDIVIDUAL FACTORS
study (19) found that although subjective gender dysphoria,
Social, legal, vocational, and environmental factors contrib- depressive symptoms, and psychopathology decreased with
ute greatly to the well-being of trans persons. Whereas the CSHT, the social and socio-legal factors of gender dysphoria
level of life satisfaction among 6,771 trans persons in a large increased, particularly for trans men. These authors argued
European Union lesbian, gay, bisexual, and transgender for a culturally specific explanation of this effect but re-
(LGBT) survey (N=93,079) was similar to that of the general grettably did not explore this option further. However, the
population, it was nonetheless somewhat lower than that of findings from the European Union survey cited above (27)

Am J Psychiatry 174:12, December 2017 ajp.psychiatryonline.org 1157

TRANSGENDER RESEARCH IN THE 21ST CENTURY

support such conjectures, given substantial differences phenotype for trans persons in many sexually dimorphic
among member states in perception of discrimination, social fascicles yet disagree on the parameters tested and the
stigma, and hate-related crime. However, more converging precise order of group findings. In 3-T diffusion tensor im-
evidence and research endeavor in this domain are needed aging studies, Kranz et al. (38) documented widespread
before more definitive conclusions can be drawn. In addi- group differences in mean diffusivity between four groups
tion to research into these social and environmental factors, (hormonally untreated transgender men and women and
much recent work has examined the underlying neurobiol- cisgender men and women). Cisgender women had the
ogy in transgender persons. highest mean diffusivity values, followed by transgender
men, transgender women, and cisgender men. Moreover, no
differences were found in fractional anisotropy. Directly
STRUCTURAL NEUROANATOMY AND
contrasting these findings, Rametti et al. documented, in
RESTING-STATE ACTIVITY
separate reports on trans men (39) and trans women (40),
Despite intensive searching, no clear neurobiological marker differences in fractional anisotropy, with larger values for
or “cause” of being transgender has been identified. Yet cisgender men relative to cisgender women, values for trans
functional MRI (fMRI) and structural MRI studies of the men being more similar to cisgender men, and values for trans
brain in trans persons have been surging over the past decade women falling between those of the cisgender groups.
to chart the complex contributions of underlying neurobi- Although markedly fewer longitudinal studies or studies
ology. Hunting for a neurobiological etiology in trans persons, in trans persons after gender-affirming surgery and/or CSHT
the majority of studies in this domain have examined whether have been published, the few that have are promising. Con-
the brains of trans persons resemble those of their sex firming histological findings of a female-sized hypothalamus
assigned at birth, resemble those of their gender identity, or in trans women (37), Hulshoff Pol et al. (41) documented a
are intermediate to either sex. These structural MRI studies decrease in hypothalamus size in this cohort with CSHT.
have predominantly focused on persons pretreatment and Conversely, Kim et al. (42) found an increase in hypothalamic
have reported mixed findings. Various factors contribute to volume in trans men, suggesting plasticity and sensitivity of
these mixed results, including the use of different regions of the hypothalamus to CSHT to shift toward the identified
interest, as well as statistical power issues, with sample sizes sex. A treatment study examining cortical thickness after
ranging from single case studies to around 24 participants per 6 months of CSHT (43) indicated a variety of cortical
group, with slightly higher samples for cisgender control thickness increases with androgen treatment in trans men
groups (31). In some of the largest studies available in hor- and decreases with estrogens and antiandrogens in trans
monally untreated transgender persons, gray matter volume women. Again providing complementary evidence, while
(32, 33) and total brain volume (33) were similar for trans that study reported a volume increase in the global ven-
persons relative to their sex assigned at birth. Findings in tricular system in trans women, other groups have docu-
other global measures, such as in the corpus callosum, are mented specific reductions in third ventricle size in trans men
mixed and do not show any difference in size in trans persons with hormonal treatment (35, 41). Some, however, have
before (34) or after CSHT (35), although they indicate a suggested that this finding is related to volumetric changes in
corpus callosum shape consistent with their gender identity adjacent gray matter structures, including the hypothalamus
(36). As for subcortical structures, putamen volume is (41), a conjecture requiring further confirmation. However,
either larger (32) or smaller (33) in trans women relative to serious caveats of currently available structural MRI studies
cisgender men, with other studies suggesting a larger volume include small numbers of transgender persons and cisgender
in trans women relative to cisgender women (35). Coming comparison subjects (41), an absence of a cisgender com-
close to providing a neurostructural correlate of being a trans parison group (to establish baseline values for cisgender men
person, a valuable histological study in 42 postmortem brains and women) (43), an absence of pretreatment data (35), or
(37) reported that the size of the hypothalamic uncinate separate publication of trans men and trans women, thus not
nucleus (INAH-3) in 10 trans women resembled that in allowing direct comparison (39, 40). Moreover, given the
cisgender women, that is, it was consistent with their gender clinical implications of early versus late age at onset of gender
identity rather than their sex assigned at birth. Even though dysphoria and sexual orientation (30), such variables deserve
the trans persons in this postmortem study had been hor- future scrutiny.
monally treated, the lack of testosterone did not appear to be In addition to characterizing gray and white matter dif-
the primary causative factor, given the absence of such an ferences in transgender persons, structural and functional
anatomical effect in nontrans castrated men, five of whom connectivity studies have also begun to emerge (44, 45). In a
were also examined in the study. These data highlight the structural connectivity study of transgender persons before
importance of paying close attention to regionally specific, CSHT, Hahn et al. (44) documented decreased structural
minute structures when assessing which brain areas conform hemispheric connectivity ratios for transgender persons
to gender identity rather than sex assigned at birth. relative to cisgender persons in subcortical brain areas. In an
In comparison to these studies on gray matter volume, the MRI study examining local resting-state activity, Mueller
few available white matter studies agree on an intermediate et al. (45) recently observed that circulating androgen levels

1158 ajp.psychiatryonline.org Am J Psychiatry 174:12, December 2017

 MUELLER ET AL.

in trans men were associated with local resting-state activity transition. In transgender men, testosterone treatment may
in the frontal cortex and the cerebellum, an effect that was deepen the voice to approach male vocal pitch. However, in
moreover linked to CSHT duration, at least in the cerebellum. transgender women, estrogens do not have such an effect and
No such effects of estrogen were found in transgender vocal surgery and/or speech therapy may be needed (2).
women. Taken together, structural MRI studies appear to Indeed, speech therapists, speech-language pathologists, or
show, depending on the brain region queried, an intermediate speech-voice clinicians may all contribute to helping a
phenotype or a phenotype consistent with gender identity in transgender person accomplish the pitch, intonation, speech
trans persons that may shift further toward the experienced rate, or phrasing patterns as well as the nonverbal commu-
gender with CSHT (with other brain areas being consistent nication of their desired sex (2). Related fMRI research has
with sex assigned at birth). Yet, a unique neurobiological examined how a male or female voice is perceived in trans
locus of “being transgender” remains to be identified, and women. Junger et al. (49) examined voice gender perception
more corroborative evidence utilizing larger and better- of male and female voices in a German sample of hormonally
powered controlled studies is needed before any definitive treated (N=16) and untreated (N=17) trans women in com-
conclusions can be drawn. Joint imaging initiatives to parison to cisgender men (N=21) and women (N=20). These
compensate for these shortcomings are lacking at this stage. male or female voices were presented “pure” or were mod-
Nonetheless, some of these structural discoveries have led to ified and presented in semitone steps toward one or the other
important work on the functional implications of these gender. Whereas cisgender men differed from cisgender
changes. women in their response times when identifying voices of the
opposite sex, reaction times were in between for transgender
women. Consistent with such a response pattern, when
FUNCTIONAL NEUROANATOMY
voices were morphed toward the other sex, neural activation
Although currently even more limited than structural im- in the superior frontal gyrus in trans women was more similar
aging studies, fMRI studies on neurocognitive or affective to that of cisgender women and less similar when compared
processes have emerged, creating an intriguing line of work to cisgender men.
based on earlier anatomical and histological findings. Indeed, Other neurocognitive work in transgender persons has
exploiting the early discovery of a female-like hypothalamus focused on functions commonly hypothesized to show
(37) and bed nuclei of the stria terminalis (BNST) (46) in “typical” sex differences in performance and to assess the
transgender women, researchers recently aimed to probe the extent to which CSHT would shift performance toward the
functional implications of this finding. The hypothalamus and gender identity in trans persons. Two such sex-stereotypical
BNST are centrally connected with other limbic structures tasks are mental rotation (a spatial cognition task, said to favor
such as the amygdala and the hippocampus and occupy a males) and language and verbal fluency tasks (said to favor
central role in hormonal signaling in sex-differentiated be- females). Starting with the first of these, in the largest cross-
havior. Studies set out to assess this impact in olfactory sectional study to date on mental rotation abilities in trans
processing. Males and females are sensitive to estratetraenol persons, hormonally treated trans women (N=18) exhibited
and androstadienone, respectively, both steroid compounds reduced parietal lobe activation, a region commonly active
with pheromone capability in humans and known to activate during mental rotation, relative to cisgender men but in-
the hypothalamus. Berglund et al. (47) discovered that ac- creased activation in the right orbital and dorsolateral pre-
tivation of the hypothalamic network to olfactory encounter frontal cortex relative to cisgender women (50). By contrast,
of androstadienone in trans women was similar to that of no group effects were observed for trans men. Given the
cisgender women and different from that of cisgender men. absence of pretreatment data in that study, it is difficult to
Burke et al. (48) replicated this finding with 3-T fMRI in a assess whether the observed effects resulted from the CSHT
larger sample of either prepubertal or pubertal children and or existed prior to it. Interestingly, a partial answer to this
adolescents with gender dysphoria (36 children and 38 ado- question is provided by Schöning et al. (51), who documented
lescents) and without gender dysphoria (39 children and increased parietal cortex activation during mental rotation in
41 adolescents). Interestingly, these additional developmental cisgender men relative to both hormonally treated and un-
findings revealed that the sexually differentiated response of treated trans women. This would suggest that transgender
the hypothalamus was already present when comparing women are more similar to their gender identity in their
prepubertal cisgender boys with cisgender girls but that the neural response during mental rotation, an effect that may
difference in youths with gender dysphoria only began to occur prior to hormonal treatment.
emerge in adolescence and was consistent prepubertally with In tasks hypothesized to favor females, findings have been
sex assigned at birth. mixed. In a pretreatment sample of adolescents with gender
Whereas olfaction is a largely subliminal process, voice dysphoria, Soleman et al. (52) reported better behavioral
and speech perception are both processed sub- and supra- performance in trans girls in a verbal fluency task (i.e., word
liminally and contribute much to the perception of gender. production). Curiously, these trans girls resembled neither
Therefore, attention in transgender care is also directed their sex assigned at birth nor their gender identity but
toward the perceived voice and speech before and after produced more words (phonetic fluency) than any of the

Am J Psychiatry 174:12, December 2017 ajp.psychiatryonline.org 1159

TRANSGENDER RESEARCH IN THE 21ST CENTURY

FIGURE 1. Overview of Key Issues and Findings in Current LIMITATIONS AND FUTURE PERSPECTIVES
Transgender Research
Many studies in transgender persons suffer from methodo-
logical shortcomings, including small or very small sample
Key findings
sizes, ranging from 10 to 25 trans persons per group (with
• Prevalence of number of people presenting to gender
clinics is increasing some exceptions, e.g., 48). Low statistical power is particu-
• Evidence of biological etiology is still limited larly problematic when the aim is to show the presence of an
• Trans persons experience a high rate of affective disor- intermediate phenotype, which may make it difficult to show
ders, especially depression and anxiety, as well as a high significant differences when results for the group of interest
risk of suicidality and nonsuicidal self-injury (trans persons) fall in between those of the control groups
• Recent research trends include preliminary findings of (cisgender persons). Thus, better-powered and multicenter
elevated autistic traits in trans persons studies are needed. Some initiatives in that direction have
• Gender-affirming surgery and/or cross-sex hormone already formed, such as the European Network for the In-
treatment decrease mental health problems, body uneas- vestigation of Gender Incongruence (ENIGI initiative),
iness, and gender dysphoria but not necessarily suicidality
which aims to assess differences in gender dysphoria among
• A high rate of trans persons find discrimination both in
personal and professional lives
different European countries. Indeed, the ENIGI initiative
• Search for neurobiological correlates of transgenderism
revealed that some sociodemographic characteristics, in-
is still ongoing cluding education level, age, and number of persons
• There have been promising functional MRI findings in (trans men or trans women) presenting to the clinic, may
voice perception and olfaction already differ among European countries (Norway, Germany,
Belgium, and the Netherlands) (54), suggesting the utility of
Key issues
including such factors in future work. Indeed, such country-
• Structural MRI research is often discrepant but consis-
specific phenomena may account for some of the differences
tently shows some effects
reported in the literature, for example, whether psychopa-
• Sample sizes are increasing in behavioral research but
remain small in neurobiological studies thology is higher in trans men or trans women (9, 12, 14). Even
• Currently there is no neuro-affective research in trans though such initiatives may need to overcome country-
persons specific limitations, including type of CSHT, legal struggles
• There have been no randomized controlled trials yet on with acknowledgment of transition status, and recruitment
cross-sex hormone treatment pool (i.e., only through referral, mental health centers, or
• There is a paucity of data in trans adolescents focus and activist groups), they allow sample sizes large
enough to assess critical factors related to mental health and
care, including quality of life and satisfaction before and after
transition. Specifically, longitudinal research is urgently
other three groups. Yet, at the brain level, group differences needed, given that treatment protocols and available care in
revealed a larger activation in cisgender boys relative to specific countries have changed dramatically over the past
cisgender girls in the Rolandic operculum, and brain activity few years. During such longitudinal work, the effects of
in the transgender groups was at levels in between these CSHT on psychology during the transition process also re-
groups. In a small study of CSHT (six trans men and eight quire more attention. Likewise, neurobiological work would
trans women), Sommer et al. (53) examined two language- see significant improvements in explanatory power during
based tasks (verb generation and categorical decision) and joint initiatives or pooling of samples.
reported an increase in language activation in both trans- More comparative work between countries is also needed
gender groups after treatment, an effect that was associated to establish optimal procedures, and the effects of CSHT on
with posttreatment estradiol levels for both sexes. Taken the brain and behavior require further confirmation and
together, unfortunately, findings from this line of work ex- validation. Fortunately, the World Professional Association
amining sex-typical and sex-atypical neurocognitive differ- for Transgender Health has set out clearly defined treatment
ences are inconclusive because of very small samples (51, 53), standards providing clinical guidance to meet the health care
effects that failed to reach statistical significance (52), lack of needs of trans persons that may be adjusted depending on the
pretreatment data (50), or non-sex-specific hormonal effects cultural context or the individual and that consequently
(affecting both sexes) (53). While presenting an encouraging provide a standardized yet sufficiently flexible framework
first set of preliminary studies to build upon, no conclusions (Standards of Care for the Health of Transsexual, Trans-
can yet be drawn as to the effect of CSHT on sex-typical gender, and Gender-Nonconforming People, Version 7) (2).
cognitive performance or the presence of pretreatment
neurocognitive differences. Surprisingly, despite the high
CONCLUSIONS
rates of affective disorders in trans persons, no published
research to date has examined the neural correlates of Since the beginning of the century, much progress has been
emotional or affective responding in this population. made in charting the mental health needs of transgender

1160 ajp.psychiatryonline.org Am J Psychiatry 174:12, December 2017

 MUELLER ET AL.

persons alongside essential discoveries of underlying neu- 12. Gómez-Gil E, Trilla A, Salamero M, et al: Sociodemographic, clin-
robiology (33, 37, 38) and associated function (47, 48) ical, and psychiatric characteristics of transsexuals from Spain.
Arch Sex Behav 2009; 38:378–392
(Figure 1). Available data suggest high rates of affective
13. Hoshiai M, Matsumoto Y, Sato T, et al: Psychiatric comorbidity
disorders (9, 17–19) that may decrease with treatment (18, 19), among patients with gender identity disorder. Psychiatry Clin
although suicide risk remains a cause for concern (9, 20, 22, Neurosci 2010; 64:514–519
26). The currently increasing efforts to characterize the 14. Haraldsen IR, Dahl AA: Symptom profiles of gender dysphoric
changes associated with transition to the preferred sex are patients of transsexual type compared to patients with personality
disorders and healthy adults. Acta Psychiatr Scand 2000; 102:
promising (19, 41, 45) but require larger-scaled collaborations
276–281
for future validation and confirmation (54). We hope that 15. de Vries AL, Noens IL, Cohen-Kettenis PT, et al: Autism spectrum
such a research agenda will also contribute to a reduction in disorders in gender dysphoric children and adolescents. J Autism
social stigma and an environment in which all persons receive Dev Disord 2010; 40:930–936
optimal care regardless of sex or sexual orientation. 16. Jones RM, Wheelwright S, Farrell K, et al: Brief report: female-to-
male transsexual people and autistic traits. J Autism Dev Disord
2012; 42:301–306
AUTHOR AND ARTICLE INFORMATION 17. Dhejne C, Van Vlerken R, Heylens G, et al: Mental health and gender
dysphoria: a review of the literature. Int Rev Psychiatry 2016; 28:
From the Department of Experimental Clinical and Health Psychology,
44–57
Ghent University; and the Department of Endocrinology and the Center
18. Colizzi M, Costa R, Todarello O: Transsexual patients’ psychiatric
for Sexology and Gender, Ghent University Hospital, Ghent, Belgium.
comorbidity and positive effect of cross-sex hormonal treatment on men-
Address correspondence to Dr. Mueller ([email protected]). tal health: results from a longitudinal study. Psychoneuroendocrinology
Dr. T’Sjoen has received grants (as principal investigator) from AstraZeneca, 2014; 39:65–73
Bayer Schering, Ipsen, and Sandoz; consulting fees (as advisory board 19. Fisher AD, Castellini G, Ristori J, et al: Cross-sex hormone treatment
member) from Ipsen and Novartis; and speaking fees from Ferring and and psychobiological changes in transsexual persons: two-year
Novartis. The other authors report no financial relationships with com- follow-up data. J Clin Endocrinol Metab 2016; 101:4260–4269
mercial interests. 20. Clements-Nolle K, Marx R, Katz M: Attempted suicide among
Received June 8, 2017; revision received Aug. 30, 2017; accepted Sept. 1,
transgender persons: the influence of gender-based discrimination
2017; published online Oct. 20, 2017.
and victimization. J Homosex 2006; 51:53–69
21. Haas AP, Rodgers P, Herman JL: Suicide attempts among trans-
gender and gender non-conforming adults: findings of the National
REFERENCES Transgender Discrimination Survey. Los Angeles, UCLA, Williams
1. American Psychiatric Association: Diagnostic and Statistical Manual Institute, Jan 2014
of Mental Disorders, 5th ed: DSM-5. Washington, DC, American 22. Reisner SL, White JM, Bradford JB, et al: Transgender health
Psychiatric Association, 2013 disparities: comparing full cohort and nested matched-pair study
2. Coleman E, Bockting W, Botzer M, et al: Standards of Care for the designs in a community health center. LGBT Health 2014; 1:177–184
Health of Transsexual, Transgender, and Gender-Nonconforming 23. Davey A, Arcelus J, Meyer C, et al: Self-injury among trans indi-
People, Version 7. Int J Transgend 2012; 13:165–232 viduals and matched controls: prevalence and associated factors.
3. Arcelus J, Bouman WP, Van Den Noortgate W, et al: Systematic Health Soc Care Community 2016; 24:485–494
review and meta-analysis of prevalence studies in transsexualism. 24. Holt V, Skagerberg E, Dunsford M: Young people with features of
Eur Psychiatry 2015;30:807–815 gender dysphoria: demographics and associated difficulties. Clin
4. Flores AR, Herman JL, Gates GJ, et al: How many adults identify as Child Psychol Psychiatry 2016; 21:108–118
transgender in the United States? Los Angeles, UCLA, Williams 25. Asscheman H, Giltay EJ, Megens JA, et al: A long-term follow-up
Institute, June 2016 study of mortality in transsexuals receiving treatment with cross-sex
5. Jürgensen M, Kleinemeier E, Lux A, et al: Psychosexual develop- hormones. Eur J Endocrinol 2011; 164:635–642
ment in adolescents and adults with disorders of sex development: 26. Haas AP, Eliason M, Mays VM, et al: Suicide and suicide risk in
results from the German Clinical Evaluation Study. J Sex Med 2013; lesbian, gay, bisexual, and transgender populations: review and
10:2703–2714 recommendations. J Homosex 2011; 58:10–51
6. Mattila AK, Fagerholm R, Santtila P, et al: Gender identity and gender 27. European Union Agency for Fundamental Rights: EU LGBT Survey:
role orientation in female assigned patients with disorders of sex European Union Lesbian, Gay, Bisexual And Transgender Survey.
development. J Urol 2012; 188:1930–1934 Vienna, European Union Agency for Fundamental Rights, 2014
7. Berenbaum SA, Beltz AM: Sexual differentiation of human behavior: 28. Liu RT, Mustanski B: Suicidal ideation and self-harm in lesbian,
effects of prenatal and pubertal organizational hormones. Front gay, bisexual, and transgender youth. Am J Prev Med 2012; 42:
Neuroendocrinol 2011; 32:183–200 221–228
8. Zucker KJ, Bradley SJ, Oliver G, et al: Psychosexual development of 29. Smith YLS, van Goozen SHM, Kuiper AJ, et al: Transsexual sub-
women with congenital adrenal hyperplasia. Horm Behav 1996; 30: types: clinical and theoretical significance. Psychiatry Res 2005; 137:
300–318 151–160
9. Heylens G, Elaut E, Kreukels BPC, et al: Psychiatric characteristics in 30. Nieder TO, Herff M, Cerwenka S, et al: Age of onset and sexual
transsexual individuals: multicentre study in four European coun- orientation in transsexual males and females. J Sex Med 2011; 8:
tries. Br J Psychiatry 2014; 204:151–156 783–791
10. Dhejne C, Lichtenstein P, Boman M, et al: Long-term follow-up of 31. Smith ES, Junger J, Derntl B, et al: The transsexual brain: a review of
transsexual persons undergoing sex reassignment surgery: cohort findings on the neural basis of transsexualism. Neurosci Biobehav
study in Sweden. PLoS One 2011; 6:e16885 Rev 2015; 59:251–266
11. Claes L, Bouman WP, Witcomb G, et al: Non-suicidal self-injury in 32. Luders E, Sánchez FJ, Gaser C, et al: Regional gray matter variation
trans people: associations with psychological symptoms, victimi- in male-to-female transsexualism. Neuroimage 2009; 46:904–907
zation, interpersonal functioning, and perceived social support. J 33. Savic I, Arver S: Sex dimorphism of the brain in male-to-female
Sex Med 2015; 12:168–179 transsexuals. Cereb Cortex 2011; 21:2525–2533

Am J Psychiatry 174:12, December 2017 ajp.psychiatryonline.org 1161

TRANSGENDER RESEARCH IN THE 21ST CENTURY

34. Emory LE, Williams DH, Cole CM, et al: Anatomic variation of the 45. Mueller SC, Wierckx K, Jackson K, et al: Circulating androgens cor-
corpus callosum in persons with gender dysphoria. Arch Sex Behav relate with resting-state MRI in transgender men. Psychoneuro-
1991; 20:409–417 endocrinology 2016; 73:91–98
35. Mueller SC, Landre L, Wierckx K, et al: A structural MRI study in 46. Zhou J-N, Hofman MA, Gooren LJ, et al: A sex difference in the
transgender persons on cross-sex hormone therapy. Neuroendo- human brain and its relation to transsexuality. Nature 1995; 378:
crinology 2017; 105:123–130 68–70
36. Yokota Y, Kawamura Y, Kameya Y: Callosal shapes at the midsagittal 47. Berglund H, Lindström P, Dhejne-Helmy C, et al: Male-to-female
plane: MRI differences of normal males, normal females, and GID. transsexuals show sex-atypical hypothalamus activation when
Conf Proc IEEE Eng Med Biol Sci 2005; 3:3055–3058 smelling odorous steroids. Cereb Cortex 2008; 18:1900–1908
37. Garcia-Falgueras A, Swaab DF: A sex difference in the hypothalamic 48. Burke SM, Cohen-Kettenis PT, Veltman DJ, et al: Hypothalamic
uncinate nucleus: relationship to gender identity. Brain 2008; 131: response to the chemo-signal androstadienone in gender dys-
3132–3146 phoric children and adolescents. Front Endocrinol (Lausanne)
38. Kranz GS, Hahn A, Kaufmann U, et al: White matter microstructure 2014; 5:60
in transsexuals and controls investigated by diffusion tensor imaging. 49. Junger J, Habel U, Bröhr S, et al: More than just two sexes: the neural
J Neurosci 2014; 34:15466–15475 correlates of voice gender perception in gender dysphoria. PLoS One
39. Rametti G, Carrillo B, Gómez-Gil E, et al: White matter micro- 2014; 9:e111672
structure in female to male transsexuals before cross-sex hormonal 50. Carrillo B, Gómez-Gil E, Rametti G, et al: Cortical activation during
treatment: a diffusion tensor imaging study. J Psychiatr Res 2011; 45: mental rotation in male-to-female and female-to-male transsexuals
199–204 under hormonal treatment. Psychoneuroendocrinology 2010; 35:
40. Rametti G, Carrillo B, Gómez-Gil E, et al: The microstructure of 1213–1222
white matter in male to female transsexuals before cross-sex hor- 51. Schöning S, Engelien A, Bauer C, et al: Neuroimaging differences in
monal treatment: a DTI study. J Psychiatr Res 2011; 45:949–954 spatial cognition between men and male-to-female transsexuals
41. Hulshoff Pol HE, Cohen-Kettenis PT, Van Haren NEM, et al: before and during hormone therapy. J Sex Med 2010; 7:1858–1867
Changing your sex changes your brain: influences of testosterone 52. Soleman RS, Schagen SE, Veltman DJ, et al: Sex differences in verbal
and estrogen on adult human brain structure. Eur J Endocrinol 2006; fluency during adolescence: a functional magnetic resonance im-
155(suppl):S107–S114 aging study in gender dysphoric and control boys and girls. J Sex Med
42. Kim TH, Kim SK, Jeong GW: Cerebral gray matter volume variation 2013; 10:1969–1977
in female-to-male transsexuals: a voxel-based morphometric study. 53. Sommer IE, Cohen-Kettenis PT, van Raalten T, et al: Effects of cross-
Neuroreport 2015; 26:1119–1125 sex hormones on cerebral activation during language and mental
43. Zubiaurre-Elorza L, Junque C, Gómez-Gil E, et al: Effects of cross- rotation: an fMRI study in transsexuals. Eur Neuropsychopharmacol
sex hormone treatment on cortical thickness in transsexual indi- 2008; 18:215–221
viduals. J Sex Med 2014; 11:1248–1261 54. Kreukels BP, Haraldsen IR, De Cuypere G, et al: A European Net-
44. Hahn A, Kranz GS, Küblböck M, et al: Structural connectivity work for the Investigation of Gender Incongruence: the ENIGI
networks of transgender people. Cereb Cortex 2015; 25:3527–3534 initiative. Eur Psychiatry 2012; 27:445–450

1162 ajp.psychiatryonline.org Am J Psychiatry 174:12, December 2017