Sex and gender in neurodevelopmental conditions

Abstract

Health-related conditions often differ qualitatively or quantitatively between individuals of different birth-assigned sexes and gender identities, and/or with different gendered experiences, requiring tailored care. Studying the moderating and mediating effects of sex-related and gender-related factors on impairment, disability, wellbeing and health is of paramount importance especially for neurodivergent individuals, who are diagnosed with neurodevelopmental conditions with uneven sex/gender distributions. Researchers have become aware of the myriad influences that sex-related and gender-related variables have on the manifestations of neurodevelopmental conditions, and contemporary work has begun to investigate the mechanisms through which these effects are mediated. Here we describe topical concepts of sex and gender science, summarize current knowledge, and discuss research and clinical challenges related to autism, attention-deficit/hyperactivity disorder and other neurodevelopmental conditions. We consider sex and gender in the context of epidemiology, behavioural phenotypes, neurobiology, genetics, endocrinology and neighbouring disciplines. The available evidence supports the view that sex and gender are important contributors to the biological and behavioural variability in neurodevelopmental conditions. Methodological caveats such as frequent conflation of sex and gender constructs, inappropriate measurement of these constructs and under-representation of specific demographic groups (for example, female and gender minority individuals and people with intellectual disabilities) limit the translational potential of research so far. Future research and clinical implementation should integrate sex and gender into next-generation diagnostics, mechanistic investigations and support practices.

Introduction

Sex, gender identity and gendered experiences are known to result in differing qualities and quantities of psychological and biological characteristics. Hypotheses about moderation of individual differences by sex/gender have existed for roughly 100 years¹, and include those regarding the nervous system^2–5. Furthermore, these sex/gender differences can be more or less pronounced among individuals with neurological, psychiatric and neurodevelopmental conditions than among the general population⁶. Understanding the mechanisms whereby sex/gender influence clinical presentation, biology, developmental trajectory and treatment response is of importance for diagnostic assessment, intervention planning, and societal sex and gender equity⁷. Many authorities and organizations now require that research properly accounts for sex and gender⁸.

Considering the effects of sex/gender (Box 1) is particularly relevant for conditions with uneven sex/gender distributions (for example, neurodevelopmental conditions⁹) (Boxes 2,3; Fig. 1). Neurodevelopmental conditions defined by the Diagnostic and Statistical Manual of Mental Disorders, 5th edition, Text Revision (DSM-5-TR)¹⁰ and the International Classification of Diseases, 11th edition (ICD-11)¹¹ differ in sex/gender prevalence across diagnoses, with male children being up to four times more likely to be diagnosed than female children, a difference that, interestingly, seems to attenuate in adulthood^12–15. Gender diversity has also been associated with some neurodevelopmental conditions^16,17 (Box 4). Moreover, behavioural expressions of neurodevelopmental conditions are modulated by sex and gender in nuanced but clinically important ways. Research has generated mixed findings on the moderating effects of sex/gender on genetic processes, and endocrine, neurological and other body systems, in neurodivergent individuals. Many hypotheses have been applied and debated to account for the observed sex/gender effects to achieve better scientific clarity and derive clinically meaningful actions.

Box 1. Sex and gender concepts.

Sex

The umbrella term ‘sex’ includes related but separable biological constructs^7,54,324. Genetic sex components exist at conception, with XY and XX sex chromosomes being the most prevalent forms and, much less frequently, various sex chromosome aneuploidies. The SRY gene on the Y chromosome determines the development of the testes (versus the default development of the ovaries) and the subsequent prenatal testosterone surge, which influences (primarily via the regulation of gene transcription) the development of the genitals and reproductive organs, brain organization and systemic physiology. These influences are further modulated during puberty by sex hormone surges, resulting in physiological effects that last for the rest of life. Other differences in chromosome complement (for example, in genes other than SRY), genomic imprinting, X chromosome inactivation and the escape of such in individuals with XX chromosomes, contribute to additional downstream effects across cell types independent of sex hormone effects. Additionally, mosaicism exists across levels. Therefore, a binary sex assigned at birth label is at best a proxy for the joint or averaged effects of multiple sex-related biological variables. Mechanism-relevant multi-categorical or continuous measures of specific sex-related attributes and mosaicism at each level provide a more accurate approach for studying sex-related mechanisms.

Gender

The umbrella term ‘gender’ includes social–relational constructs that range from individual to societal attributes, and include norms and positional power^7,55,56,324; these constructs change dynamically over the lifespan. Individuals hold a personal sense of their own gender (that is, gender identity) and present themselves in the same or different gender in daily life (that is, lived or expressed gender); both of these gender-related attributes are multi-categorical and can change over time in an individual. An individual’s behaviours are categorized by historically defined, stereotyped characteristics, which can be continuous dimensions that coexist in one person (masculinity and femininity) or qualitatively different social role categories (gender roles). Implicit and explicit stereotypes about gender (gender stereotypes) are prevalent, can be experienced and held by an individual of any gender or sex assigned at birth (enacted gender stigma or discrimination), and internalized as an individual’s belief (internalized gender stigma). Gendered structures and norms (gender ideology) dictate relational patterns such as social support (gender relations) and power structures such as wage gaps, gendered policies and laws (institutionalized gender). These constructs are important moderators or mediators of health-related outcomes⁵⁶. They also afford analytical and interpretational angles for research in which sex-related and gender-related attributes have not been sufficiently differentiated in the first place³²⁴.

Inter-relatedness

Sex-related and gender-related attributes interact to shape human biology, behaviours and health outcomes. It is also important to consider other demographic correlates and social determinants of health — such as race, ethnicity, socioeconomic status, living conditions, food security, age, sexual orientation and language — that interact with sex and gender to influence outcomes (that is, intersectionality³²⁵) and the co-aggregation of clinical conditions (that is, syndemics³²⁶), especially in individuals from multiply-marginalized populations (see Fig. 1).

Box 2. Sex, gender and nervous system development.

Sex and brain development

Sex differences in the genetic architecture of human complex traits and mechanisms contribute to sex-differential traits³²⁷. Sex differences in brain development depend on gonadal hormonal exposure during fetal and postnatal life, and confer different susceptibilities to genetic and environmental exposures^247,268. The sex chromosomes interact synergistically and possibly antagonistically with sex hormones³²⁸. SRY promotes fetal testosterone production in the second trimester, which causes masculinized brain organization in animal models²⁴⁷. Sex-differential gene expression is ubiquitous across human tissues but is largely tissue-specific, reflecting the effects of escape from X chromosome inactivation for X-linked genes, as well as hormone-related transcription factor regulation and sex-differentiated distribution of epigenetic markers for autosomal genes³²⁹. The organization of the human brain does not show dimorphism, but instead a mosaicism of female-typical and male-typical features on multiple continua², which are associated with mental health vulnerability^330,331. Prenatal sex steroid hormonal exposure is longitudinally associated with sex-typical play behaviours, gender identity, sexual orientation and, to a lesser extent, specific cognitive–psychological traits including empathy, aggression, mental rotation and social dominance⁴; nevertheless, very few of these sex-differential traits have been robustly linked to measurable sex-differential brain features⁴. The literature on brain sex/gender differences is inconsistent, although some findings are more established, including on average larger whole-brain, limbic and temporal regional volumes in males, and larger cingulate and prefrontal regional volumes in females^5,332. Some differences have been associated with males performing better on average on perceptual and motor tasks and females on analytical intuitive processing and communication abilities, a difference that has been hypothesized to be mediated by prenatal differences in sex hormones³³². Sex-differential patterns across brain anatomy, function and gene expression are interrelated, and these patterns imply sex-influenced brain development across species³³³, which highlights the importance of measuring brain development against the norm of the same sex³³⁴.

Nature–nurture interplay

Neurodevelopment is not sex-determined but developmentally shaped by the environment (many environmental contexts are gendered) in sex-dependent ways³. Sex-related or gender-related health vulnerability is underpinned by the nature–nurture interplay and the entwined effects of sex-related and gender-related attributes³³⁵. Examples include varied trajectories and increased prevalence of anxiety, mood, stress-related and trauma-related disorders in women and individuals assigned female at birth³³⁶, and differences in the onset and trajectory characteristics of cardiovascular disease attributable to sex-based biology and gender-related factors³³⁷.

Susceptibility to neurodevelopmental conditions

Evidence indicates that sex hormones affect neurotransmitter systems, neural differentiation and synapse formation³³⁸, with critical developmental windows during mid-gestation, shortly after birth and around puberty³³⁹. They also modulate neurotransmitter concentrations and regional glucose metabolism later in life³³⁸. Alterations in intracellular signalling that influence the neurobiology of attention-deficit/hyperactivity disorder are modulated by sex hormones and the menstrual cycle³³⁸. Sex hormones also modulate hypothalamic structure and function, which differ between males and females³⁴⁰. The hypothalamus might be a pivotal structure in Tourette syndrome and other neurodevelopmental conditions³⁴⁰.

Sex/gender differences in dopamine receptor distribution are present in cortical and subcortical regions across humans and animals³⁴¹. Females exhibit higher regional glutamate levels than males in the striatum, anterior cingulate, sensorimotor cortex and cerebellum³⁴². Alterations in dopaminergic and glutamatergic neurotransmission have been implicated in neurodevelopmental conditions^341,342. Studies in rodents have shown alterations in dopamine receptor functioning as a consequence of exposure to stressors and addictive substances, and hyperactivity in association with developmental fluctuation of striatal dopamine receptor densities only in males³⁴¹. Therefore, certain neurotransmitter systems, such as the dopaminergic system, might be more susceptible to interference in males than in females.

Box 3. A note on anti-ableist and neurodiversity-affirmative language.

In recent years, the language used to describe neurodevelopmental conditions and the individuals diagnosed with these conditions has evolved dramatically, specifically with regard to the shift away from overly medicalized and deficit-based language^47,343–345 towards non-pathologizing alternatives, which are rooted in the neurodiversity approach(es) to these conditions^{39,46,346–348} and informed by the preferences of neurodivergent community members^{343,344,349–351}. For instance, although diagnostic manuals such as DSM-5-TR and ICD-11 use the term ‘neurodevelopmental disorders’ to describe this group of conditions, the more value-neutral term ‘neurodevelopmental conditions’ is used in this Review, as it does not imply that neurodivergent individuals are inherently ‘disordered’ or pathological. Although the term ‘neurodevelopmental disabilities’ is a widely accepted alternative that reflects the often-disabling nature of these conditions, it is notable that Tourette syndrome and tic disorders can be diagnosed even when tics do not produce substantial disability or functional impairment¹⁰.

Individuals diagnosed with neurodevelopmental conditions generally are referred to as ‘neurodivergent’ and individuals without these conditions (that is, typically developing individuals) are frequently referred to as ‘neurotypical’. Additional anti-ableist language choices adopted in this Review include referring to core and associated features of neurodevelopmental conditions as traits, characteristics or differences rather than ‘symptoms’ when appropriate, discussing factors that increase the ‘likelihood’ or ‘probability’ of neurodevelopmental conditions rather than ‘risk’, and the use of identity-first language when possible (for example, ‘autistic person’ rather than ‘person with autism’) to describe autism and other conditions where identity-first language is often preferred by stakeholders^{47,343,352,353} (although we recognize this preference might differ by language and culture³⁵⁴). Although a full discussion of language preferences in neurodivergent communities is beyond the scope of this Review, interested readers are directed to several foundational articles in this area for a deeper understanding of these topics^{39,343–345,352,355}.

Fig. 1 |. Why are sex and gender important for neurodevelopmental conditions?

Sex-related biological attributes (blue) and, to a lesser extent, gender-related sociocultural factors (green) serve specific roles in moderating and mediating the sex-differential likelihood of the emergence of neurodevelopmental conditions^{228,229,236,321}, presumably explaining the male preponderance in prevalence¹²³, sex-moderated and sex-specific aetiological factors, and converging mechanisms of neurodivergent development and sex differentiation (such as overlapping downstream molecular pathways²⁷⁷ and neuroendocrine–immune mechanisms²⁶⁸). Sex-related biological and gender socialization mechanisms also converge with mechanisms shaping the lifespan development of neurodivergent individuals^322,323. Sex-related and gender-related attributes also influence how neurodevelopmental conditions are diagnosed and how multilevel phenotypic presentations vary among individuals^106,133. Specifically, these attributes moderate or mediate the behavioural and cognitive manifestations of neurodevelopmental and co-occurring health conditions, account for developmental heterogeneity, shape (and are shaped by) brain development and neurodivergent biology, and impact how and when neurodevelopmental conditions are recognized and diagnosed (especially owing to implicit and explicit gender biases, gender role expectations and gendered stereotypes of neurodivergent presentations)³²³. Findings and reflections then drive the nosological evolution and the operationalization of the diagnostic criteria for neurodevelopmental conditions iteratively¹²³. Ultimately, new knowledge generated by sex-informed and gender-informed research will have a clinical impact⁸ in reducing recognition and diagnostic biases, and in enabling the development and evaluation of sex-specific and tailored biological interventions based on sex-informed neurodivergent biology, and the design and implementation of gender-aware and gender-informed care and support for neurodivergent individuals of all sexes and genders.

Box 4. Gender diversity and neurodevelopmental conditions.

Gender diversity refers to the variations of gender identity and expression that may or may not align with those stereotypically linked to an individual’s sex assigned at birth. When an individual suffers from substantial distress owing to the incongruence between their experienced or expressed gender and their assigned gender (that links to their sex assigned at birth), the diagnostic label of ‘gender dysphoria’ (DSM-5-TR)¹⁰ or ‘gender incongruence’ (ICD-11)¹¹ is used to highlight the clinical support needs. In the context of increasing awareness, recognition and referral for clinical care for gender dysphoria (especially in those assigned female at birth) in the general population^356,357, rates of gender dysphoria and minority gender identities are reported to be particularly heightened in the neurodivergent population^358,359. Studies that used a single item of the Child Behavior Checklist to approximate gender dysphoria or incongruence have shown that 2.5–5.4 of children and adolescents diagnosed with autism, attention-deficit/hyperactivity disorder or intellectual disabilities might have gender-minority identities^17,359,360. Studies using dimensional measures have also found that greater self-reported gender diversity and gender incongruence are more common in autistic than in non-autistic children and adolescents, especially in those assigned female at birth^361,362. In autistic adults, these rates are even higher³⁶³. The presence of gender dysphoria is associated with bullying victimization, poorer mental health and increased suicidal ideation in autistic young people³⁶⁴. Equally, neurodevelopmental conditions are enriched in gender-minority populations across ages^365–367. In particular, the pooled estimate of autism diagnosis prevalence in gender-minority individuals is 11%¹⁶, much higher than in the current general population prevalence. This co-occurrence of gender diversity and neurodevelopmental conditions has been observed across several sociodemographic groups in the past decade; however, hypotheses regarding the developmental mechanisms underlying such co-occurrence are speculative, ranging from shared prenatal sex steroid organizational effects of the brain³⁶⁸ to an impact of neurodivergent information processing on gender cognition and development¹⁶. Regardless, community input to research has resulted in preliminary clinical guidance and service design tailored to the unique needs of neurodivergent gender-minority individuals^369,370, aiming to improve care and support for these historically under-recognized and multiply-marginalized subpopulations.

Neurodevelopmental conditions

Neurodevelopmental conditions arise from divergent architecture, maturation and function of the developing nervous system. They are heterogeneous in presentation, neurobiology and aetiology^9,18, and research has generated various findings indicative of alterations in brain structure and function in the most common neurodevelopmental conditions^19–22. In a US population-based survey performed in 2015–2017, the overall prevalence of neurodevelopmental conditions was estimated at 17.8¹². According to the ICD-11 (ref.¹¹) and DSM-5-TR¹⁰, neurodevelopmental conditions include autism spectrum disorder (hereafter termed autism; see Box 3), attention-deficit/hyperactivity disorder (ADHD), communication (speech and language) disorders, learning disorders (for example, dyslexia), motor disorders (for example, Tourette syndrome) and intellectual disabilities. Although we adopt this definition, broader definitions of neurodevelopmental conditions might also include genetic syndromes (for example, Rett syndrome), congenital brain injury (for example, cerebral palsy), fetal alcohol spectrum disorders, schizophrenia and obsessive–compulsive disorder^23,24.

Neurodevelopmental conditions are usually inborn or develop early in life; some present with unspecific ‘prodromal’ features and are thus detected in later childhood, and more recently the possible existence of later-onset variants, such as in ADHD, has been discussed^25–27. Neurodevelopmental conditions are defined by characteristic differences in social communication, impulsivity, flexibility, attention, speech and language, sensory processing, learning and motor control that diverge substantially from typical development and tend to co-occur^{9–11,18,28,29}. When these traits interfere with subjective wellbeing and social, educational or vocational functioning, a clinical disorder diagnosis is given. Alterations in cognition, such as executive function, social cognition, bottom-up versus top-down processing, phonological and sensorimotor awareness, numerical cognition or general cognitive abilities, are presumed to underlie the defining clinical features³⁰. Neurodevelopmental conditions are highly heritable^31,32 but have multifactorial origins alongside biology–environment interplay^33–35. Psychosocial environments, and society’s awareness, attitudes and support have important effects on neurodivergent people’s health, social participation, experiences of acceptance, wellbeing, quality of life and functional outcomes^36–39.

Many neurodevelopmental conditions are better understood as extreme expressions of traits that are continuously distributed in the population, as opposed to qualitatively distinct entities^40–42, and traditional biomedical and nosological conceptions that exclusively aspire to normalize or cure neurodevelopmental conditions have been increasingly criticized^43–45. The neurodiversity paradigm acknowledges that neurodevelopmental conditions can both result in disability and confer vulnerability to co-occurring health conditions, but this approach conceives neurodevelopmental conditions as parts of variability of the human CNS, not pathology per se^39,43,46,47. Transdiagnostic methods have the potential to promote neurodiversity-affirmative research and practice⁴⁸, owing to their tendency to incorporate and harmonize both medical and social models of disability³⁹. Viable intervention targets within this approach include promotion of optimum developmental and adaptive skills, prevention and intervention for adverse health and social outcomes, quality of life, increasing opportunities for learning and growing to compensate for difficulties, and altering the environment^37,44,49,50. Herein, we favour a strengths-based, non-pathologizing approach to disability to capitalize on an individual’s resources and opportunities, rather than solely focusing on limitations or impairments⁵¹.

Sex and gender constructs

Sex and gender are both multi-component umbrella constructs (Box 1); they are distinct but inter-correlated because of interplaying biological and sociological mechanisms⁵². Sex-related constructs are biological attributes that were initially conceptualized on the basis of the reproduction of organisms and include chromosomal, genetic, gonadal, hormonal, genital and other anatomical and physiological components⁵³. Owing to this conceptual origin, many sex-related attributes are treated as dichotomous (that is, female versus male), but the actual distributions of specific sex-related variables can in some cases be multi-categorical (for example, sex chromosome configurations) or continuous (for example, serum testosterone levels)⁵⁴. Gender refers to constructs of social relations in human societies; these constructs originated in the reproductive arena and extend to psychological, sociocultural and contextual attributes, and power structures^55,56. Ideologies that maintain dominating power structures often closely associate gender with sex and impose a binary view of gender, although empirical data show most gender-related variables to be either multi-categorical (for example, gender identities) or continuous (for example, masculinity and femininity)^54–56. Notably, gender-related attributes can influence biological outcomes that are thought to be primarily driven by sex-related attributes (for example, brain biology, hormonal status and related behaviours)^57,58. Sex-related and gender-related attributes separately and jointly act as moderators of many human health outcomes^7,52.

In this Review, when an empirical finding cannot be clearly delineated as reflecting sex or gender effects specifically, owing to research limitations, we use the term ‘sex/gender’ to highlight this ambiguity in interpretation. Although suboptimal, when studies only report findings by binarized sex/gender, we use the generic terms ‘female(s)’ and ‘male(s)’ to reflect the joint effects of sex and gender attributes, rather than using gender identity-based terms (for example, ‘girls/women’, ‘boys/men’) or claiming that these terms unambiguously represent sex assigned at birth.

Observed differences among individuals of different sexes or gender identities suggest moderating (and in specific cases, mediating) effects of sex and/or gender attributes on typical and atypical development, including neurodevelopmental conditions (Box 2; Fig. 1). Depending on the outcome of interest and the specific sex-related or gender-related variable, such differences vary from dimorphic (that is, existing in two forms) to small differences on a continuum (that is, distributions substantially overlap despite mean differences)⁴. The specific ways in which sex and gender influence the outcome of interest vary according to the persistence of the influence (across development, experience and environment), context-dependence⁵² and the interplay of sex components and gender components across individual development and over generations⁵⁹.

Epidemiology

The estimated prevalence rates and sex/gender ratios of neurodevelopmental conditions vary depending on study design, cohort, age group, informant and country. Meta-analytic or large-scale studies published in the last decade have estimated childhood prevalence as: 5.9% for ADHD⁶⁰, 1.0% for autism⁶¹, 7.7% for speech–language disorders⁶², 9.7% for learning disorders⁶³, 2.8% for tic disorders⁶⁴ and 1% for intellectual disabilities⁶⁵. Far fewer studies have been performed in adult populations. The rates of some neurodevelopmental conditions (for example, ADHD and tic disorders^66,67) are commonly lower in adulthood than in childhood, although associated functional impairment might persist. However, for other neurodevelopmental conditions, such as autism, the number of individuals receiving first-time diagnoses in adulthood has increased substantially in recent years^61,68.

A sex/gender bias towards higher prevalence in males than in females is observed across all neurodevelopmental conditions and ranges from 1.2:1 to 4:1 (Table 1). In a large US parent report survey in 88,530 children aged 3–17 years, the rate of any neurodevelopmental condition, including ADHD, autism, speech fluency disorder (stuttering), learning disorders, intellectual disabilities or any other developmental delay was 12.6% in females and 22.7% in males¹². This sex/gender bias has been found to be more pronounced in childhood than in adulthood in autism, tic disorders and ADHD^{14,19,69–71}, and varies for different subtypes of these conditions^14,72. Level of intellectual functioning modulates the sex/gender ratio in the prevalence of some neurodevelopmental conditions; for example, in autism, more balanced ratios are usually found in individuals with so-called profound autism⁷³ and in autistic individuals with co-occurring intellectual disabilities more generally^14,61.

Table 1 |.

Neurodevelopmental conditions: summary of prevalence, sex/gender ratio, and average differences in behavioural presentation and co-occurring conditions

Neurodevelopmental conditiona	Prevalence (%)	Female to male ratio	Moderation by sex/gender		Refs.
			Behavioural presentation	Co-occurring conditions
Allb	Children: 17.8 F:12.6 M: 22.7	Children: 1:1.7	‐	F: possibly internalizing problems and later detection of some conditions more likely M: possibly co-occurring externalizing behaviours more likely	12,132,133,320
Autism	1.0	1:3	F: less RRB; better social communication, prelinguistic and linguistic functioning, cognitive flexibility and autobiographical memory M: fewer compulsions and less self-injurious behaviour; better impulse inhibition and visuospatial functioning	F: higher rates of epilepsy, endocrine problems, depression, eating problems and internalizing behaviours M: hyperactivity in childhood and externalizing behaviours more likely	15,61,78–80,94,104–106,110–119,122,123,125,126,309
ADHD	Children: 5.9c Adults: 2.8	Children: 1:2.3 Adults: 1:1.6	F: inattentive presentation more likely (childhood); fewer inhibition and cognitive flexibility problems (childhood); fewer complex attention problems (adulthood) M: hyperactive presentation more likely (childhood)	F: possibly higher rates of anxiety, depression and other internalizing behaviours; higher rates of cigarette smoking M: possibly higher rates of delinquent behaviours, ODD, conduct disorder and ASPD (adulthood)	60,72,106,129–133,320
Non-syndromic ID	All: 1.0 Children: 1.8 Adults: 0.5	Mild: 1:1.6 Severe: 1:1.2 Children: 1:1.0–2.2 Adults: 1:1.1–1.4	‐	F: depression (all ages) and dementia (later adulthood) more likely M: personality disorder and psychosis (adulthood) more likely	65,158,160,162
Fragile X syndrome	F: 0.009 M: 0.014	1:1.6	F: milder cognitive difficulties, less evident dysmorphism, and higher adaptive functioning and verbal strengths	F: typically with mild ID or borderline-to-average IQ, co-occurring learning disorder, socioemotional difficulties and psychiatric problems, and 20% have an autism diagnosis M: typically with severe ID and co-occurring ADHD, anxiety and features of autism, and 50–60% have an autism diagnosis	162–164
Communication disorders	Children: 7.7	1:1.7	F: modest advantage in language acquisition in early childhood; grammar problems in childhood and later referral to services more likely	M: higher association with medical conditions in childhood	62,141–143
Learning disorders	Lifetime: 9.7 F: 7.1 M: 12.2	1:1.7	F: literacy advantage, better verbal processing speed, working memory, conceptualization, and orthographic and visuospatial coding M: tendency for mathematical skills to be less challenging, better verbal reasoning	F: higher rates of inattentive and internalizing symptoms M: higher rates of hyperactive and other externalizing behaviours	63,144–148
Motor disorders	All: 5.0–6.0 DCD, severe: 1.5 DCD, probable: 3.0 SMD: 3.0–4.0 TS: 0.77 (adults, 0.05; M, 1.1; F, 0.3) Tic disorders: 3.0 (M, 4.1; F, 3.1)	DCD, severe: 1:2 TS and other tic disorders: 1:1.5	F: onset of tics with OCD; increase of tics during menstrual cycle; higher social impairment; tics more complex and persistent, with a later onset and a later peak of severity; in DCD, gross motor problems and higher social impact more likely M: onset of tics with rage and ADHD; in DCD, fine motor problems more likely	F: possibly higher rates of anxiety and mood disorders with tics M: possibly higher rates of ADHD and disruptive behavioural disorders with tics	64,149,150,154,157

ADHD, attention-deficit/hyperactivity disorder; ASPD, antisocial personality disorder; DCD, developmental coordination disorder; F, females; ID, intellectual disabilities; M, males; OCD, obsessive-compulsive disorder; ODD, oppositional defiant disorder; RRB, restricted and repetitive behaviour; SMD, stereotypic movement disorder; TS, Tourette syndrome.

^aDSM-5-TR or ICD-11 diagnoses.

^bADHD, autism, speech fluency disorder (stuttering), learning disorders, intellectual disabilities or any other developmental delay.

^cBest estimate for children and adolescents.

Co-occurrence of neurodevelopmental conditions with psychiatric and medical conditions is high^74–79. Despite the widespread opinion amongst clinicians that the co-occurrence profiles differ between sexes/genders^80,81 — with males more likely to show externalizing problems such as oppositional defiant disorder and conduct disorder, and females more likely to show internalizing problems such as anxiety and mood disorders — the evidence is inconsistent. Population-based research, as opposed to clinical studies, has often failed to identify these sex/gender differences in psychiatric illnesses^82–84. Alternatively, population-based studies have shown higher rates of suicide in autistic females than in autistic males^85,86, and this finding is particularly pronounced among individuals who are in the average IQ range and have co-occurring ADHD⁸⁷. Sex/gender differences in the causes of premature mortality (which occurs more frequently in autistic and ADHD individuals than in the general population) have also been observed in autism and other developmental disabilities^88,89. For example, autistic males have a higher mortality risk from diseases of the circulatory and nervous systems and cancer, and from external causes, and autistic females have a higher risk from similar causes but also from endocrine diseases and congenital malformations compared with sex-matched non-autistic counterparts^85,87–89. In ADHD, a trend for greater mortality — owing to both natural causes (for example, cardiovascular disease) and unnatural causes (for example, suicide or accidents) — was observed in females compared with males^88,90.

Behavioural phenotypes

Autism

Historically, autistic females have been reported to be more likely to show intellectual disabilities, neurological problems and emotional–behavioural challenges than autistic males^91–96. The consequent presumption that autistic females are more likely to have cognitive and neurological impairments might have contributed to the under-recognition¹⁵ and delayed diagnosis of autism in some females^68,97–102, especially when their autistic features were less obvious¹⁰³. Large-scale studies published in the last 10 years indicate that cognitive and language abilities vary substantially across all sexes and genders in the autistic population, with the most consistent difference being slightly lower restricted and repetitive behaviour scores in females than in males^94,104,105. However, sex-related and gender-related factors are also likely to influence the profiles of some autistic features in subtle ways¹⁰⁶. For instance, autistic females tend to show better social communication skills than autistic males in nuanced ways, which means that autistic features in females might not be sufficiently captured by commonly used instruments^107,108. These social communication skills involve: social attention and motivation; gender stereotypical play; expression of non-verbal communication; impression management skills rehearsed over time^106,109 (Box 5); and (pre)linguistic functioning from communicative gestures, babbling and semantics, to narrative and pragmatic competency^110–114. On average, autistic males and females have similar levels of insistence on sameness; autistic males are more likely to show repetitive motor behaviours and speech, and circumscribed interests, whereas autistic females are more likely to experience compulsions and/or self-injurious behaviours^115–119.

Box 5. Gender, socialization and social coping.

Like neurotypical individuals, neurodivergent people live in societies with varied sociocultural norms, and they grow up navigating and negotiating with the social world with these norms. This socialization experience throughout the course of development shapes individuals’ social coping, including the process of managing their public self-image(s) in social settings (that is, impression management). Although this kind of experience is ubiquitous across people with and without a neurodevelopmental diagnosis, the mismatch between social expectations and neurodivergent behaviours in the neurotypical majority, alongside the unique information processing styles (and constraints) of neurodivergent individuals, render impression management an extremely challenging endeavour for many neurodivergent individuals, that requires intensive efforts and could serve as a major stressor for some¹⁰⁹. Neurodivergent and neurotypical individuals who self-report higher rates of these experiences also typically report worse mental health, although studies have yet to determine the causal direction³¹². The impression management experiences of neurodivergent individuals who strive to cope socially in the world of the neurotypical majority range from the suppression of tics³⁷¹, strategies to compensate and mask attention-deficit/hyperactivity disorder presentations¹³², and the ‘camouflaging’ or ‘passing’ experiences described by some autistic people^{109,312,313,372} and observed in some children with developmental language disorder³⁷³. As gender-related influences such as gender role expectations and gender stereotypes are pervasive and have a profound impact on socialization, female neurodivergent individuals tend to be influenced differently from their male counterparts when it comes to ways and efforts to ‘fit in’ to what the particular society expects of them by suppressing neurodivergent behaviours and forcing the presentation of neurotypical behaviours — all potentially contributing to mental health strain^323,374.

Developmentally, a declining outward autistic presentation is more likely to be observed in autistic females than in autistic males. This sex/gender difference exists even among individuals diagnosed in early childhood^120,121, suggesting the contribution of both sex-related biological and gender socialization influences. Moderating effects of sex/gender on cognitive features associated with autism are mostly observed in executive function and visuospatial domains^122,123: evidence suggests that autistic males are more likely to have challenges with cognitive flexibility and autobiographical memory, whereas autistic females are more likely to have challenges with response inhibition and visuospatial processing¹²⁴. In terms of health conditions, autistic females are more likely to have epilepsy and endocrine–reproductive problems^79,125, depressive disorders⁷⁸, and disordered eating and sleep^126–128, whereas autistic males are more likely to exhibit hyperactivity in childhood¹²⁴.

ADHD

In children diagnosed with ADHD, males are more likely to express hyperactivity than females, and females are more likely to show inattention as the predominant presentation^129,130; similar patterns have also been observed in adults¹³¹. Evidence also indicates that females are less likely than males to show ‘classic’ ADHD (for example, displaying disruptive and impulsive behaviours)^132,133. This observation should however be considered alongside recognition biases¹⁰⁶; for example, ADHD symptom ratings by teachers and parents tend to be higher for male than for female children, despite no differences in directly observed ADHD behaviours¹³⁴. According to a meta-analysis published in 2021, among children diagnosed with ADHD, males experience more inhibition and cognitive flexibility difficulties than females¹²⁹; less is known about the differences between adult males and adult females, although the results of another meta-analysis suggest that males are more likely to experience impaired complex attention than females¹³⁵. Some studies indicate that males diagnosed with ADHD are more likely to show antisocial behaviours and to receive oppositional defiant disorder or conduct disorder diagnoses, whereas females are more likely to experience internalizing and emotion regulation problems^106,130. ADHD has been causally associated with a greater risk of cigarette smoking in female than male adolescents¹³⁶. In a Danish population-based registry study, compared with male individuals diagnosed with ADHD, female individuals diagnosed with ADHD had greater rates of co-occurring autism, intellectual disabilities, oppositional defiant disorder and conduct disorder, personality disorders, schizophrenia, substance use disorders and suicidal behaviour⁸². In adults specifically, the results of registry-based studies indicate that ADHD is more strongly associated with anxiety, depression, bipolar disorder and personality disorders in females than in males, and more strongly associated with schizophrenia, substance use disorders and hypertension in males than in females^137,138.

Communication disorders

Limited research exists on sex/gender influences on language disorders and hardly any is available on stuttering, speech sound disorder or social (pragmatic) communication disorder¹³⁹. In the general population, recent research unexpectedly revealed that male children produce more ‘protophones’ (that is, speech-like vocalizations) than female children in the first year of life¹⁴⁰. On the other hand, there is a modest female advantage in communicative gestures and receptive and expressive vocabulary, especially in the early stages of language acquisition¹⁴¹. However, the existence of sex/gender differences in language is controversial, with evidence published over the last 10 years indicating that these effects are smaller than previously assumed¹⁴². In a longitudinal population study of primary school children aged 4–6 years, the association between speech–language dysfunctions and medical diagnoses was higher in males than females, and females diagnosed with language disorders were more likely to have morphosyntactic problems than males¹⁴³. Importantly, in this study, screening for language disorders identified more males than females; indeed, females are often older than males when they are referred to clinical services¹⁴³. Overall, there seem to be sex/gender differences in language disorders, which currently appear subtle, but influence the timing and rates of detection and service access.

Learning disorders

The literature regarding clinically relevant effects of sex and gender on learning disorders is scarce. In a population-based study, female children showed a literacy advantage over males, with better spelling and reading fluency¹⁴⁴. Regarding mathematical skills, findings are mixed, with a tendency for more female than male children to have challenges in this area^144–146. In a clinically enriched community study of children, males showed lower reading performance than females, but this pattern was mostly accounted for by an over-representation of males in the low end of test score distributions; further, this sex/gender difference in reading is fully mediated by differences in processing speed, inhibition and verbal reasoning¹⁴⁶. The findings of other cognitive studies in dyslexia also suggest that females have better verbal working memory and conceptualization, and orthographic and visuospatial coding than males¹⁴⁷. The co-occurrence of behavioural challenges with dyslexia might differ among sexes/genders. For example, in a study published in 2000, dyslexia was associated with inattentive patterns in both female and male children but was associated with hyperactive and impulsive symptoms only in males; the latter symptoms might be more likely to trigger externalizing behaviours and subsequent clinical referrals in males¹⁴⁸.

Motor disorders

The nature of motor, vocal and combined tics seems to be similar across sexes/genders; however, evidence suggests that compared with males, females are less likely to experience simple tics and more often have complex tics¹⁴⁹. Tics in females also seem to start later than in males, with a later peak of symptom severity, and are more likely to be persistent and/or worsen with age than in males¹⁴⁹. Evidence suggests that complex tics in females can also surge during the oestrogenic phase of the menstrual cycle^150,151. Furthermore, some findings suggest that on the simple-to-complex sensory phenomena and motor phenomena continua — from premonitory urges to obsessive urges or thoughts, and from simple tics to compulsive tics — males tend to more commonly manifest simple urge and tic behaviours¹⁵², whereas females more commonly manifest obsessive–compulsive-like tic behaviours¹⁵³ and show an increase in complexity of tics with age¹⁵⁴. However, these findings might have been confounded by different age distributions among male and female samples, with females being older¹⁵³.

The onset of tics in concert with co-occurring conditions might also differ according to sex/gender. Evidence suggests that females more commonly have an onset combined with compulsion, whereas males frequently have tics that co-occur with rage and ADHD^150,153,154. Among individuals with Tourette syndrome, females are less likely than males to have co-occurring ADHD but are more likely to have anxiety and mood problems^{150,154–156}. Overall, females seem to experience greater tic-related impairment in social participation and quality of life than males^150,156.

In a survey-based study in adults with a high likelihood of developmental coordination disorder, females were more likely to report gross motor and non-motor difficulties and a negative impact on activities and participation, whereas males were more likely to report fine motor difficulties¹⁵⁷.

Intellectual disabilities

Little research exists on sex/gender differences in non-syndromic intellectual disabilities. In adolescents and adults with intellectual disabilities, depression seems to be more common in females, and psychotic and personality disorders more common in males^158,159. Evidence from a registry study suggests that, in later adulthood, females with intellectual disabilities are more likely to have dementia than males with intellectual disabilities¹⁶⁰. The literature on behavioural phenotypes in syndromic intellectual disabilities is comprehensive¹⁶¹ but few studies exist on sex/gender differences. This scarcity is probably because some of these conditions are rare, which results in small sample sizes, or sex-specific (for example, Rett syndrome, which almost exclusively affects females).

In fragile X syndrome (FXS), a relatively common syndromic form of intellectual disability caused by a mutation (CGG trinucleotide repeat expansion) in the fragile X messenger ribonucleoprotein 1 (FMR1) gene, the female-to-male ratio is about 1:1.6 (ref.¹⁶²). Males with FXS often have severe intellectual disability and co-occurring ADHD, anxiety and features of autism. By contrast, females with FXS typically have mild intellectual disability or borderline-to-average IQ, with co-occurring learning disorders, socioemotional difficulties and psychiatric problems¹⁶³. These differences are likely to be the result of males being hemizygous for the mutation whereas females carry heterozygous mutations and have one functioning copy of the FMR1 gene¹⁶⁴. Despite females with FXS having milder cognitive difficulties and dysmorphism, higher adaptive functioning and stronger verbal communication skills, they do not necessarily have better outcomes than males with FXS¹⁶³.

The effects of sex/gender on behavioural phenotypes of the main neurodevelopmental conditions are summarized in Table 1.

Access and response to intervention

Neurodivergent females and their parents have clearly voiced the need for sex and gender to be considered in relation to therapeutic intervention and support^165,166; however, few studies have examined whether access and response to intervention differs between neurodivergent females and males. In ADHD, evidence suggests that female children (but not adults) are less likely than male children to receive medication if no co-occurring externalizing challenges are present^167,168. Regarding pharmacodynamic profiles for extended-release formulations of methylphenidate for ADHD, compared with males, females show a better response at 1.5 h after administration and an inferior response at 12 h¹⁶⁹. In Tourette syndrome, a study published in 1989 found that haloperidol yields a better treatment response in females than in males¹⁷⁰. In autism, trials of social skills group interventions showed better effects on social responsiveness in females than in males^171,172, despite many female individuals struggling to fit in in these predominantly male groups. In short, sex-related and gender-related attributes can be important moderators of access and response to therapies and support across biological and social levels, but their exact roles (both statistically and mechanistically) remain to be clarified in well-powered trials.

Brain biology

Autism

Structure.

With regard to autism-associated changes in brain morphology, more evidence supports an influence of sex/gender on the spatial pattern of these changes than simply on their effect size^173,174. According to a systematic review, sex/gender-by-diagnosis interaction effects are predominantly observed in brain regions that also tend to vary by sex/gender in typical development; for example, visual, limbic, cerebellar, sensorimotor, ventral attention and default mode network regions¹⁷⁵. In these regions, autistic males usually exhibit higher and autistic females lower values in cortical thickness, volume and surface area than their neurotypical counterparts¹⁷⁵. According to a large-scale analysis published in 2020, both autistic males and females have greater cortical thickness than neurotypical individuals, although these differences are concentrated in different brain regions in the two sexes¹⁷⁶. Another large-scale study published in 2022 used a deep-learning framework that had been trained on structural images of neurotypical brains to predict biological sex. By applying this prediction model to two large samples of structural brain images from autistic people, the researchers identified a replicable shift towards more male-typical brain anatomy in both autistic males and females¹⁷⁷. In females this shift was driven mainly by features in regions associated with visual, face and speech processing, whereas in males it was driven mainly by regions associated with reward and motor processing¹⁷⁷.

Studies have found that, compared with sex-matched neurotypical children, autistic females but not males have a smaller head circumference in the first year of life¹⁷⁸. Brain overgrowth between 1 and 5 years of age has been observed in both autistic males and autistic females, and notably involves partly different brain regions or shows different developmental trajectories in different sexes/genders^179–182. In a longitudinal study that spanned early to middle childhood, brain overgrowth of autistic males was predominantly driven by increased grey matter growth, whereas autistic females showed slower grey and white matter growth trajectories compared with neurotypical peers¹⁸². In the same cohort, developing brain regions that have monosynaptic connections with the amygdala showed qualitatively different volumetric alterations in autistic females and autistic males¹⁸³, and amygdala volume correlated with internalizing behaviours only in females¹⁸⁴. Studies in adolescents and adults have identified region-specific volumetric sex/gender differences, especially in limbic regions; in addition, greater white matter reductions were observed in autistic females in youth and in autistic males in adulthood¹⁷⁵. Alterations in grey matter density covariance (a measure of structural connectivity) in salience, executive control and default mode networks have been found to differ between preschool-aged autistic females and males¹⁸⁵. In another study, both autistic males and autistic females showed alterations compared with neurotypical peers in grey-to-white matter boundary sharpness; however, these alterations occurred in different brain regions in females and males and with larger effects in females¹⁸⁶.

Replicable sex/gender-by-diagnosis interactions have been identified in fibre tracts (primarily in white matter microstructural properties, such as fractional anisotropy) that commonly show sex/gender differences in neurotypical individuals (for example, the corpus callosum, corona radiata, cingulum and inferior fronto-occipital fasciculus)^175,187. White matter tract development deviations from sex/gender-specific age norms have been reported in individuals diagnosed with autism, ADHD and early-stage schizophrenia¹⁸⁸. Interestingly, such deviations, especially in tracts connecting prefrontal cortex regions, are shared between autistic individuals and their undiagnosed female siblings¹⁸⁹. In early childhood, there is evidence for sex/gender-by-diagnosis interaction effects on callosal structural connectivity but not tract development trajectories^190,191. Overall, the evidence indicates that structural brain organization differs between autistic females and males when compared with neurotypical counterparts, especially with regard to the spatial pattern of alterations and their developmental trajectories. Many of these sex/gender differential alterations associated with autism overlap with brain structures that also show sex/gender differences in neurotypical populations.

Functional organization.

Functionally, regions-of-interest studies have found sex/gender-by-diagnosis interaction effects on resting-state functional connectivity of the default mode, salience and ventral attention networks, as well as limbic regions and the cerebellum, although the direction of the effects varies between studies and conditions^175,187. Sex/gender differences in amygdala resting-state connectivity were attenuated in autistic compared with neurotypical children¹⁹². Voxel-mirrored homotopic connectivity, a proxy for cross-hemispheric processing, showed replicable occipital sex/gender-by-diagnosis interaction effects¹⁹³. Task-based functional MRI (fMRI) studies have used predominantly social processing tasks and have identified sex/gender-by-diagnosis interaction effects predominantly in limbic and default mode network regions. Where sex/gender-by-diagnosis interactions were found, the most common findings were hypoactivation in autistic males compared with autistic females^175,187.

An imbalance between excitatory and inhibitory neural activity has been proposed as a core mechanism of autism^194,195. Interestingly, one study that estimated excitation–inhibition balance using resting-state fMRI data found that autistic male adults, but not autistic female adults, showed an excitation–inhibition imbalance towards higher excitation compared with their neurotypical counterparts in default mode network and other areas¹⁹⁶. In autistic female adults, excitation–inhibition balance was associated with greater estimated camouflaging¹⁹⁶ (Box 5). The results of studies in animal models and gene expression patterns in humans have also led to the speculation that the emergence of autism is driven by brain plastic reactions following genetic and environmental events; such plastic reactions occur more readily in males than in females and involve perceptual or language-related brain regions, leading to the male-preponderant prevalence of autism and autistic cognitive strengths¹⁹⁷. Taken together, functional brain alterations in autism seem to differ between females and males, but owing to the heterogeneity of study designs and findings, drawing firm conclusions about how they differ would be premature.

Neurochemistry.

In the light of the excitation–inhibition imbalance hypothesis of autism^194,195, studies have investigated the neurochemical correlates of autism, focusing on glutamate and &-aminobutyric acid (GABA), the predominant excitatory and inhibitory neurotransmitters in the CNS. In a study that used magnetic resonance spectroscopy (MRS), neurotypical sex/gender differences in glutamate–glutamine and N-acetyl compounds in bilateral posterior thalamic radiations and the left centrum semiovale were partly attenuated or reversed in autistic individuals¹⁹⁸. N-Acetyl compounds are thought to have an important role in longer-term energy expenditure and myelin synthesis¹⁹⁸. Another MRS study found a sex/gender-by-diagnosis interaction effect in GABA concentration in the left dorsolateral prefrontal cortex; thalamic GABA concentration was positively correlated with autistic traits in autistic females and negatively correlated with these traits in autistic males¹⁹⁹. A study using PET and simultaneous MRI found that, compared with neurotypical individuals, the association between cortical thickness and GABA_A receptor binding potential in the left postcentral gyrus was different in autistic male adults but not in autistic female adults²⁰⁰. In summary, studies on sex/gender-specific alterations in autism-related neurochemistry are scarce to date, with some limited evidence for differences in excitatory and inhibitory neurotransmission in particular regions.

ADHD

Structure.

Both males and females diagnosed with ADHD have reductions in the volume of some brain regions compared with their neurotypical peers. Volume reductions in basal ganglia, medial prefrontal and premotor cortex regions have been found predominantly in males, and reductions in lateral premotor and prefrontal regions seem to be more common in females²⁰¹. In children, reduced total brain volume has been associated with both ADHD and higher polygenic scores for ADHD, with both these effects stronger in males than in females²⁰². Reduced caudate volume in male adults diagnosed with ADHD compared with neurotypical male adults correlated with hyperactive–impulsive symptoms, whereas no such effect was observed in female adults²⁰³. In contrast, volume reductions in the putamen and thalamus were associated with ADHD symptoms in female but not in male children diagnosed with ADHD²⁰⁴.

A study in school-aged children identified sex/gender-by-diagnosis interaction effects on measures of white matter microstructure, which also correlated with inhibitory control performance²⁰⁵. More specifically, compared with neurotypical participants, males diagnosed with ADHD had alterations in fractional anisotropy in primary motor regions, whereas ADHD females had alterations in medial orbitofrontal regions. Another study, which assessed callosal fibre tracts, found a sex/gender-by-diagnosis interaction effect on radial diffusivity in the posterior parietal tracts; diffusivity was altered only in females diagnosed with ADHD and correlated with attention control in both sexes/genders²⁰⁶. Taken together, the brain structural correlates of both ADHD diagnosis and traits differ partly between males and females, potentially underlying sex/gender-related differences in ADHD behavioural characteristics.

Functional organization.

Evidence from one study suggests that, compared with neurotypical peers, school-aged ADHD females display more alterations in resting-state functional connectivity between striatal and prefrontal regions than males²⁰⁷. Functional connectivity between these regions also correlated with increased delay discounting of monetary rewards, a proxy for altered reward functioning and impulsivity, that was elevated in ADHD females only. A study that explored lateralization found, in addition to sex/gender-independent effects of ADHD, more leftward lateralization in male individuals diagnosed with ADHD and more rightward lateralization in female individuals diagnosed with ADHD²⁰⁸. In neurotypical adults, inattention and hyperactivity traits were associated with striatal resting-state connectivity to different regions in males and females²⁰⁹. Emerging evidence also indicates sex/gender-specific alterations in functional connectivity between the limbic striatum and frontal default mode network regions in children diagnosed with ADHD, with increased connectivity in ADHD compared with neurotypical children only in males²¹⁰. Overall, brain functional connectivity alterations associated with a diagnosis of ADHD, specific ADHD traits and reward functioning seem to differ between female and male individuals, especially in the limbic and prefrontal regions.

Neurochemistry.

One small-sample study in children using single-voxel proton MRS found a sex/gender-by-diagnosis interaction effect for N-acetylaspartate concentration in the right dorsolateral frontal white matter, with especially low concentrations in ADHD females, potentially indicating elevated levels of neuronal death or injury²¹¹. Another study found sex/gender differences in cerebellar choline and glutamate–glutamine concentrations with lower levels in females and interaction effects between age and sex/gender for choline, creatine and glutamate–glutamine concentrations in the pregenual anterior cingulate cortex in adults diagnosed with ADHD²¹². Choline is associated with energy metabolism and creatine with myelination. These findings might be potential evidence for sex/gender differences in brain maturation delays linked to ADHD, although the study assessed adults. In sum, there is preliminary evidence for sex/gender-specific neurochemical mechanisms underlying ADHD, affecting not only neurotransmitters but also other metabolites linked to brain maturation or injury.

Dyslexia

The results of some studies indicate that dyslexic female individuals have more frequent or pronounced changes in brain volume at the gross level than dyslexic male individuals¹⁴⁷. These changes include reduced whole-brain volume, reduced right hemispheric grey matter and bilateral white matter volume, and an increased left-hemispheric grey matter to white matter ratio, as compared with neurotypical individuals. A study that assessed whole and lobar brain volumes across participants from three countries did not find sex/gender-by-diagnosis interactions²¹³. The results of some studies indicate that some dyslexia-related alterations are specific to female individuals; for example, reductions in right hemispheric grey matter, especially sensorimotor regions (precuneus, cuneus, precentral and postcentral gyrus) and the left visual word form area^147,214. In contrast, altered left–right asymmetry of the planum temporale²¹⁴, and reductions in left-hemispheric grey matter cortical thickness in the middle and superior temporal gyri, orbitofrontal cortex, inferior frontal gyrus, temporoparietal cortex and fusiform gyrus have been observed in dyslexic male individuals compared with neurotypical male individuals¹⁴⁷. Preliminary evidence suggests that structural brain alterations are more anatomically widespread in dyslexic females than in dyslexic males, who have a clearer involvement of the classic reading network regions¹⁴⁷. One ¹H-MRS study assessed neurotransmitter metabolite concentrations in the anterior cingulate cortex and identified higher levels of choline and myoinositol in association with reduced processing speed across dyslexic and neurotypical female children; this association was not observed in male children²¹⁵. Taken together, there is preliminary evidence that dyslexic females have more widespread brain alterations than males.

Tourette syndrome

Preliminary findings indicate that, compared with neurotypical children, male children with Tourette syndrome have alterations in cortical and callosal thickness and basal ganglia volume; these alterations were not observed in female children with Tourette syndrome^149,216. One study found reduced cortical thickness in prefrontal, orbitofrontal and parietal regions in males with Tourette syndrome compared with females with Tourette syndrome, as well as a negative correlation between tic severity and post-central and pre-central regions that was significantly stronger in males than in females²¹⁷. Another study in children with Tourette syndrome identified a sex/gender-by-diagnosis interaction effect on grey matter volume of the bilateral hypothalamus²¹⁸. In a study from 2001 that included participants aged 6–63 years, sex/gender differences in parieto-occipital volume (including both grey and white matter) observed in the neurotypical group were diminished in the Tourette syndrome group, especially in adults²¹⁹. Overall, current evidence suggests that males with Tourette syndrome might have more widespread brain alterations than their neurotypical counterparts than do females with Tourette syndrome, and that typical brain sex/gender differences are reduced in individuals diagnosed with Tourette syndrome.

Intellectual disabilities

Sex/gender differences in the brain correlates of FXS have been observed. Caudate enlargement, a striking feature in FXS, is more pronounced in males than in females¹⁶⁴, whereas a decrease in the size of the cerebellar vermis, which correlates with IQ, seems to be similar in females and in males with FXS²²⁰. During eye-gaze processing, females with FXS show reduced insula and fusiform gyrus activation, whereas males with FXS show increased amygdala and insula activation, as compared with neurotypical individuals¹⁶⁴.

Sex/gender moderation effects on brain structure and function in non-syndromic intellectual disabilities remain largely unexplored. Furthermore, the effects of IQ are usually insufficiently captured in neuroimaging studies of sex/gender in neurodevelopmental conditions^188,221, with the majority of studies controlling for IQ as a nuisance covariate or simply excluding participants with intellectual disabilities¹⁷⁴. The results of a meta-analysis indicate that whole-brain volume positively correlates with IQ across sexes/genders in both neurotypical and neurodivergent cohorts²²². However, in an fMRI study, the resting-state brain connectivity features that predicted higher IQ differed between neurotypical males and females²²³.

Table 2 summarizes the average sex/gender differences and sex/gender-moderating effects that have been reported for brain structure, function and neurochemistry in neurodevelopmental conditions.

Table 2 |.

Summary of average sex and gender differences in brain structure, function, and neurochemistry in neurodevelopmental conditions

Category	Autism	ADHD	ID	LDa	MD
Brain structure	Sex/gender-specific alterations in GM and WM volume, cortical thickness, and white matter tracts notably in regions and tracts with sex/gender differences in neurotypical individuals173,175,187 Sex/gender-specific alterations in callosal structural connectivity191 ↓ WM integrity in autistic female youth and autistic male adults compared with neurotypical peers175	F: ADHD associated with ↓ GM volume in lateral premotor and prefrontal regions201, and altered WM microstructure in medial orbitofrontal regions205 and posterior parietal tracts of corpus callosum206; ADHD symptoms associated with ↓ putamen and thalamus volumes204 M: ADHD associated with ↓ GM volume in basal ganglia, medial prefrontal and premotor cortex regions201; stronger total ↓ brain volume in childhood202; reduced caudate volume that correlated with hyperactive-impulsive symptoms203; and altered WM microstructure in primary motor cortex205	In FXS: caudate enlargement more pronounced in males than in females164,220	F: dyslexia associated with more widespread and gross-level volume reductions, ↓ GM in right hemisphere, and ↓ WM in left hemisphere147,214 M: dyslexia associated with altered left-right asymmetry of planum temporale surface area, and ↓ left hemisphere GM and WM volume in reading network regions147,214	Normative sex/gender differences in parietooccipital volume diminished in TS, especially in adults219 ↓ cortical thickness in prefrontal, orbitofrontal and parietal regions in males compared with females with TS217 M: TS associated with altered cortical and callosal thickness and basal ganglia volume in childhood149,216
Brain function	Different alterations in functional connectivity among sexes/genders, e.g. in default mode and salience/ventral attention networks, limbic regions and cerebellum175 F: autism associated with ↓ voxel-mirrored homotopic connectivity of dorsolateral occipital cortex193	In ADHD: more right-ward lateralization of resting-state signal variability in females and more left-ward lateralization in males208 F: ADHD associated with altered resting-state functional connectivity between striatal and prefrontal regions, related to reward processing207 M: ADHD associated with ↑ functional connectivity between limbic striatum and frontal default mode network regions210	Neurotypical males and females differ in which resting-state brain connectivity features predict higher IQ scores223 In FXS: During eye-gaze processing, females show reduced insula and fusiform gyrus activation, and males show increased amygdala and insula activation164	Insufficient information	Insufficient information
Brain neurochemistry	Compared with neurotypical individuals: reversed or reduced sex/gender differences in the concentrations of glutamate-glutamine and N-acetyl compounds in bilateral posterior thalamic radiations and the left centrum semiovale198; females and males show different alterations in GABA concentration in the left dorsolateral prefrontal cortex and opposite direction of correlation between thalamic GABA and autistic traits199 M: association between cortical thickness and GABAA receptor binding potential in the left postcentral gyrus different in autistic compared with typically developing males200	Cerebellum choline and glutamate-glutamine concentrations greater in males than in females diagnosed with ADHD212 F: ADHD associated with reduced N-acetylaspartate concentration in right dorsolateral frontal WM in childhood211	Insufficient information	F: ↑ choline and myoinositol in the anterior cingulate associated with reduced processing speed in dyslexic and neurotypical children215	Insufficient information

ADHD, attention-deficit/hyperactivity disorder; F, females; FXS, fragile X syndrome; GABA, (γ-aminobutyric acid; GM, grey matter; ID, intellectual disabilities; LD, learning disorders; M, males; MD, motor disorders; TS, Tourette syndrome; WM, white matter.

^aHere dyslexia.

Genomics

Most heritable traits are subject to different liabilities by sex²²⁴. In relation to autism, the so-called female protective effect model proposes that autistic females inherit more genetic variants that increase the likelihood of autism compared with autistic males^225,226. Higher autism prevalence (compared with that in the general population) occurs among individuals who have a close autistic female relative and the likelihood of autism increases in later-born offspring siblings of autistic females, compared with autistic males (known as the Carter effect)^227–229. Similarly, in the context of ADHD, mothers diagnosed with ADHD transmit a greater likelihood of ADHD to offspring than fathers²³⁰, although this has not been found in all studies²³¹. Genetic propensity for various neurodevelopmental conditions can be associated with sex-differentiated phenotypic heterogeneity; for example, male siblings of females with anxiety disorders have elevated relative likelihoods of ADHD or any neurodevelopmental condition, although this effect could also represent a diagnostic sex/gender bias²³².

Molecular genetic studies indicate that autistic females carry more common genetic variants associated with autism than autistic males. Polygenic scores for autism, based on the weighted sum of susceptibility alleles present, are increased in autistic females without intellectual disabilities²²⁹, in non-autistic mothers and female siblings²²⁸, and in mothers with broader subdiagnostic autism phenotypes²³³. In a separate study, no consistent within-trait sex differences in heritability, referring to the proportion of the phenotypic variance explained by all measured single nucleotide polymorphisms (SNP), were found in an analysis of data from 20 psychiatric and neurodevelopmental conditions, including autism and ADHD²³⁴. This is despite the existence of several shared sex-specific SNP statistical associations between autism and ADHD. Autistic females are also more likely than autistic males to carry a higher burden of de novo rare variants^235,236, which is inversely related to polygenic score load²³⁶. In addition to dominant mutations, biallelic recessive mutations are descriptively more prevalent in autistic females than in autistic males²³⁷. In relation to structural variants, some neurodevelopmental condition-related copy number variations display phenotypic sex differences; autism is more common in males with 22q11.2 duplication and 16p11.2 deletion syndromes than in females with these syndromes²³⁸. These findings suggest that the sex-differential probability of developing autism is relevant to sex-related factors that modify the effects of common and rare variants associated with autism.

Sex chromosome aneuploidies (SCA) are associated with increased prevalence of autism, ADHD, and language and cognitive deficits, particularly in association with male phenotypes²³⁹. For example, autism diagnosis prevalence is almost double in individuals with XYY syndrome (on average 30%) than in individuals with XXY syndrome (18%) and XXX syndrome (15%), although it is still higher in individuals with the latter two syndromes than in the general population²³⁹. ADHD symptoms and diagnoses are also markedly increased across all these three types of sex chromosome trisomy²³⁹. Chromosomal dosage might influence SCA phenotypes owing to excess or reduced sex chromosome number; for example, a supernumerary Y chromosome is particularly associated with reduced social and attentional functioning²⁴⁰.

X chromosome inactivation (XCI) could also account for phenotypic heterogeneity²⁴¹. X-linked neurodevelopmental conditions, FXS and Rett syndrome, are associated with sex-differentiated behavioural and developmental outcomes in human and animal studies²⁴² although this is not universal for X-linked genes²⁴³. Evidence suggests that XCI skewing can cause phenotypic variability among female family members with FXS, but evidence regarding the role of XCI in the Rett syndrome phenotype is inconclusive²⁴⁴. Some genes escape XCI (that is, they continue to be expressed from both the active and the inactive chromosome) and contribute to complex trait variation and SCA phenotypes via dosage effects associated with increased expression of genes that have not been silenced²⁴⁵. Y chromosome genes that could contribute to sex bias in neurodevelopmental conditions include NLGN4Y, an autism candidate gene²⁴⁶, and SRY²⁴⁷, a brain-expressed gene that stimulates male gonadal development and fetal testosterone production. Some autosomal genes show sexually differentiated expression in the context of neurodevelopmental conditions; for example, genes involved in the Wnt signalling pathway in ADHD²⁴⁸ and SHANK1 in autism²⁴⁹. In studies in animal models, the candidate autism genes Chd8 (ref.²⁵⁰) and Cntnap2 (ref.²⁵¹) show sex-differentiated expression.

In summary, genetic contributions to sex differences in neurodevelopmental conditions are small and polygenic, and large sample sizes are typically required to detect the effects²⁵². Studies investigating the so-called female protective effect in autism have used partly overlapping samples, and independent replication is therefore required^{225,227–229,236}. A publication from 2019 describes a method for the estimation of X-linked gene dosage from the summary statistics of genome-wide association studies, which will aid in the investigation of X-linked variation in complex traits²⁵³. Studying XCI and escape genes in humans is methodologically challenging²⁵⁴ and difficult to relate to human brain tissues²⁴⁴, although catalogues of XCI across human tissues have helped identify genes contributing to sex-specific differences²⁵⁵. See Table 3 for a summary of the findings presented in this section.

Table 3 |.

Summary of sex and gender influences in genetic factors and sex hormones associated with neurodevelopmental conditions

Category	Autism	ADHD	ID	MD
Clinical genetics	↑ Autism prevalence in close relatives of autistic females compared with close relatives of autistic males227–229	↑ Likelihood of ADHD in offspring of mothers with ADHD compared with offspring of fathers with ADHD in some studies23 231 ↑ ADHD prevalence, ↑ any neurodevelopmental diagnosis in siblings of females with anxiety disorders compared with siblings of males with anxiety disorders232	Insufficient information	Insufficient information
Common variants	↑ PGS for autism in autistic females without ID compared with males229, non-autistic mothers compared with fathers, female siblings of autistic individuals compared with male siblings228, and in mothers with broader autism phenotype compared with fathers233	No specific within-trait sex-specific SNP heritability234 Shared sex-specific SNP associations between ADHD and autism234	Insufficient information	Insufficient information
Rare variants	↑ Prevalence of rare de novo variants and biallelic recessive mutations in autistic females compared with autistic males235–237 Autism is a common outcome in males with 22q11.2 duplication and 16p11.2 deletion CNV syndromes238	Insufficient information	No associated sex differences in X-linked rare variants243	Insufficient information
Sex chromosomes and autosomes	↑ Autism prevalence in people with SCA (XXX, XXY, XYY, XXYY)239 Y-linked genes NLGN4Y246 and SRY247 might increase male prevalence Sex differences in SHANK1 gene expression249 Sex differences in Chd8 (ref.250) and Cntnap2 (ref.251) animal models	↑ ADHD prevalence in people with SCA (XXX, XXY, XYY, XXYY)239. Sex differences in gene expression in Wnt pathway248	↑ Language and cognitive deficits in people with SCA (XXX, XXY, XYY, XXYY)239 XCI thought to ↑ phenotype heterogeneity in FXS242	Insufficient information
Sex hormones	↓ 2D to 4D digit ratio, ↑ facial landmark masculinity260,261; ↑ androgens and oestradiol in amniotic fluid (prenatally)263,264; low maternal serum oestradiol265, compared with sex-matched neurotypical groups ↑ Autism in offspring of mothers with PCOS, regardless of offspring sex266	↓ 2D to 4D digit ratio compared with neurotypical individuals260 ↑ ADHD in offspring of mothers with PCOS, regardless of offspring sex266	Insufficient information	↑ Chronic tic disorder in offspring of mothers with PCOS, regardless of offspring sex266
Immune response	Autism or related features ↑ with in utero exposure to sodium valproate in humans and murine models33 and with sex hormone differences in rodents271 ↑ Autism in offspring of mothers with obesity272 Weak-to-moderate association between autism and the effect of EDCs on immune cell function275	Insufficient information	Insufficient information	Insufficient information

No sex/gender differences in these factors reported for learning disorders. ADHD, attention-deficit/hyperactivity disorder; CNV, copy number variation; EDCs, endocrine disrupting chemicals; FXS, fragile X syndrome; ID, intellectual disabilities; MD, motor disorders; PCOS, polycystic ovary syndrome; PGS, polygenic score; SCA, sex chromosome aneuploidies; SNP, single nucleotide polymorphisms; XCI, X chromosome inactivation; XXX, triple X syndrome; XXY, Klinefelter syndrome; XYY, XYY syndrome.

Sex hormones

Steroid hormones, particularly androgens, might be biological mediators (and potential moderators for genetic aetiological factors²⁵⁶) of demonstrated sex differences in autism prevalence. The ‘extreme male brain’ and fetal androgen theories^257,258 posit that in utero exposure to androgens increases stereotypically male-prevalent cognitive traits (‘systemizing’), which are on-average higher in autistic than in non-autistic individuals, proportionately more than stereotypically female-prevalent cognitive traits (‘empathizing’)²⁵⁹. Supporting these theories, physical markers of in utero androgen exposure such as lower 2D to 4D digit ratio²⁶⁰ or increased facial landmark masculinity²⁶¹ are associated with increased prevalence of autism and ADHD diagnoses. In utero androgen exposure is associated with neurodevelopmental outcomes in rodent models (via testosterone aromatization to oestradiol)²⁶². However, in primates the mechanism is more complex and yet to be fully clarified²⁴². A small number of studies have found associations between higher concentrations of androgens²⁶³ and oestrogens (particularly oestradiol) in amniotic fluid²⁶⁴ and an increased likelihood of an autism diagnosis in male children. In contrast, low maternal serum oestradiol was associated with offspring autism diagnoses²⁶⁵. Maternal polycystic ovary syndrome, a condition associated with elevated testosterone levels, is associated with an increased likelihood of autism, ADHD and chronic tic disorder diagnoses in offspring regardless of sex²⁶⁶. However, there is no consistent association between in utero androgen exposure and autistic traits^266,267.

The role of steroid hormones is intricately linked to the placenta, the in utero environment and inflammation²⁶⁸. Indeed, the immune system seems to act bidirectionally with the endocrine system. In animal models, immune-related signalling and cellular actions are implicated in brain masculinization, sex-specific synaptic patterning and social behaviours of the animals^269,270. For example, disrupted prostaglandin signalling during a critical period of cerebellar development is associated with deficits in play behaviour in male rodents²⁶⁹. Sodium valproate is an important teratogen associated with cognitive deficits and autism in humans³³. Mice exposed to sodium valproate in utero showed sex-specific differences in steroid metabolism in the brain²⁷¹. Evidence suggests that maternal obesity results in early maturation of the fetal hypothalamic–pituitary–adrenal axis, which might alter fetal neural programming and promote sex-specific neurodevelopmental outcomes^272,273. In addition to metabolic and neuroendocrine effects, obesity is associated with a pro-inflammatory state, suggesting a role for maternal immune activation in the development of neurodevelopmental conditions²⁷².

Evidence from animal and human studies also suggests that endocrine-disrupting compounds such as bisphenol A affect neurodevelopmental outcomes via immune cell functioning²⁷⁴, although most human studies are limited by imprecise measures of exposure and more rigorous epidemiological studies are warranted²⁷⁵. Sexually dimorphic outcomes, potentially mediated by disrupted oestrogen receptor-’ signalling, have been associated with altered expression of autism-related genes in the transcriptomes of exposed rat pups²⁷⁶. Converging findings from transcriptomic studies indicate that autistic individuals have increased cerebral cortical expression of genes associated with glial and immune functions, and decreased expression of neuronal and synaptic genes, as compared with neurotypical individuals²⁷⁷. Interestingly, these findings mirror sex-differentiated gene expression in neurotypical individuals: compared with neurotypical females, neurotypical males have increased expression of glial and immune-related genes and decreased expression of neuronal and synaptic genes. This observation provides insight into sex-differentiated neurobiology, but there is a need to overcome limitations such as under-representation of females in these cohorts in order for the findings to be truly informative.

Immunology

Sex/gender differences in immunological responses are present across the lifespan and contribute to differential susceptibility to infections and autoimmune diseases in the general population²⁷⁸. Specific immune alterations such as maternal immune activation and infection, as well as differential cytokine and chemokine profiles, have been linked to neurodevelopmental conditions, particularly autism and ADHD²⁷⁹. Maternal prenatal infections increase the likelihood of autism (though potentially not causally²⁸⁰), intellectual disabilities and ADHD in offspring, and the presence of specific inflammatory disease states during pregnancy (for example, maternal asthma, diabetes or autoimmune diseases) has been associated with autism, ADHD and Tourette syndrome^272,281. Evidence of neuroinflammation has been observed in post-mortem studies in individuals who had neurodevelopmental conditions, and gene expression studies indicate microglial over-activation^277,282,283. Moreover, parental immunological disorders (for example, ankylosing spondylitis, ulcerative colitis, autoimmune thyroid disease and atopic/allergic diseases) are associated with autism and ADHD in offspring^284,285. Shared genetic factors or familial causes of neurodevelopmental conditions and autoimmune disorders have also been identified²⁸⁶.

Nevertheless, research on the intersections of sex/gender and immunology in neurodevelopmental conditions remains scarce²⁷⁹. A Norwegian registry study found higher rates of psoriasis, ulcerative colitis and Crohn’s disease in females than in males diagnosed with ADHD²⁸⁷. Animal studies have generated some evidence that the effects of immune activation are moderated by offspring sex²⁷⁰. For example, male, but not female, offspring of exposed mice exhibited elevated IL-6 cytokine levels and repetitive-restricted behaviours compared with sex-matched offspring of control mice^288–290. However, other studies found more pronounced effects of maternal immune activation on female offspring^291,292. In short, although solid evidence exists that some immunological mechanisms differ for females and males in the general population, and it is likely that immune processes are altered in neurodevelopmental conditions, the study of sex/gender differences in immune responses and mechanisms in neurodivergent populations is still in its infancy.

Gut–brain axis and microbiome

Evidence suggests that the gut–brain axis and microbiome can modulate brain development and behaviour²⁹³, impacting mental health and neurodevelopmental outcomes²⁹⁴ while also interacting with sex-related attributes²⁹⁵. In the general population, some gastrointestinal disorders are more common in males (for example, diarrhoea) and some are more common in females (for example, functional dyspepsia and irritable bowel syndrome); they are also more often associated with anxiety and depression in females²⁹⁶. Associations between neurodevelopmental conditions, particularly autism and ADHD, and gastrointestinal problems are established but specific microbiome signatures have yet to be identified^77,297,298. Nevertheless, specific antigenic determinants (epitopes) found on microorganism surfaces have been reported to correlate with sex/gender in autism²⁹⁹. In addition, in animal studies, sex-specific gut microbiota profiles were detected in the BTBR mouse model of autism, and prenatal exposure to elevated levels of microbiome products generated subtle sex differences in social behaviour and activity levels at different points during development^291,300. To summarize, although research shows that the gut–brain axis and microbiome influences on mental health issues and neurodevelopmental conditions might affect females and males differently, exact findings on mechanisms are lacking.

Conclusions and future directions

The evidence discussed in this Review demonstrates that multiple sex-related and gender-related factors drive the biological and behavioural variability in neurodevelopmental conditions (see Fig. 2 for summary). The strength of this evidence varies — from established to emerging — and is drawn from epidemiological, behavioural, cognitive, neurobiological, genetic, endocrinological and immunological research. Notably, the mechanisms by which sex-related and gender-related attributes exert their effects on neurodevelopmental outcomes are likely to include a wide range of interacting biological, psychological and sociocultural processes. As neurodevelopmental conditions are operationalized diagnostically via sets of prototypical traits and behaviours, it is important that clinical assessments and subsequent recommendations for services are well informed by the nuanced sex and gender differences in the clinical gestalt of neurodivergent individuals (Table 1). Research has been instrumental in raising clinical awareness of such nuances, especially in autism and ADHD^106,133, and improving the timely and effective identification and support of neurodivergent females. Clarifying the biological mechanisms and specific processes that mediate sex/gender influences on neurodevelopmental phenotypes will bring new opportunities for sex-tailored and gender-tailored biological interventions (Fig. 2; Tables 2,3). In autism, there is more support for qualitative sex/gender brain differences, with more widespread and qualitatively different network alterations in females compared with males. Similarly, increasing evidence suggests that the neural correlates of ADHD, dyslexia, Tourette syndrome and FXS might be different in females than in males.

Fig. 2 |. Simplified schema of the influences of sex-related and gender-related attributes on brain development.

Visualization of different levels at which sex-related and gender-related attributes influence brain development, interacting with environmental factors. Sex-related attributes can influence brain development before or shortly after birth and throughout life, whereas the influence of gender-related attributes occurs after birth and throughout the lifespan. SRY encodes sex-determining region Y protein, which stimulates male gonadal development and fetal testosterone production. NLGN4Y encodes neuroligin-4 and is a candidate autism gene. F, females; M, males; SCA, sex chromosome aneuploidies; XCI, X chromosome inactivation; ↑, factor increases the likelihood of developing and/or presenting neurodevelopmental conditions; ↓, factor decreases the likelihood of developing and/or presenting neurodevelopmental conditions. ^aSee Fig. 1 for more examples of sex-related and gender-related attributes.

Limitations of existing evidence

Much of the evidence implicating sex hormones in the mechanisms of sex differences in brain development and of neurodevelopmental conditions relies on studies in animals. Research in humans is limited by methodological challenges and an incomplete understanding of the complex interactions between genetic, endocrine, metabolic and immune factors. Mechanistic research is required to characterize the human sex steroid metabolome, the actions of sex steroids across development and their interaction with genes and downstream epigenetic factors, alongside their interplay with gender-related experiential factors in human societies.

There remain major caveats in current research that limit our understanding and the translatability of knowledge. Substantial methodological problems arise from the inappropriate conflation of sex and gender attributes and the systemic under-representation of specific subpopulations. Foremost, although gender terminology is used increasingly in research, most studies have investigated the moderating effect(s) of sex-labels assigned at birth, which only serve as a proxy of the joint effects of sex-related biological factors and gender-related attributes beyond gender identity. It is rather the exception that biomedical research properly assesses gender-related attributes³⁰¹ (Box 1), and there are very few countries that register sex beyond the male/female binary and gender identity. The convenient but problematic use of one single, binary variable to represent sex or gender (even conflating sex and gender) without delineating specific sex-related and gender-related components has to change³⁰². Further, many studies include small numbers of females and gender-minority individuals and therefore frequently lack the statistical power to conduct sex-focused and gender-focused analyses. Some research has excluded females or treated sex/gender as confounds to adjust for, rather than examining sex and gender attributes as substantive predictors, moderators or mediators in their own right, which is suboptimal^187,303.

Other caveats and controversies are also prominent. First, most research focuses on autism, and to a lesser extent on ADHD and tic disorders, whereas research for other diagnoses is selective or absent. Beyond specific genetic syndromes (for example, FXS), there is little research on sex/gender effects in intellectual disabilities, which is also reflected in the autism literature insofar as autistic people with intellectual disabilities are disproportionately under-represented^304,305. This is especially the case for those with high support needs, whose pronounced cognitive and/or language impairments necessitate round-the-clock care, that is, so-called profound autism⁷³. Second, sex/gender ratios, behavioural and biological phenotypes, and co-occurring conditions change over time, and there are cohort effects³⁰⁶. In ADHD, hyperactivity, impulsivity and inattentive patterns change over the course of development, and research published over the last two decades has proposed the existence of a potential ‘late-onset’ form, although this remains controversial^{25,27,307,308}. The limited number of cross-sectional and especially longitudinal studies on the moderating and mediating effects of specific sex and gender components prevents us from drawing firm conclusions regarding their roles in neurodivergent developmental trajectories¹⁷⁵. Third, some widely cited constructs and models require reconsideration and refinement. Discussions about the validity and conceptual clarity of the so-called female (autism) phenotype^106,309,310, camouflaging^{109,311–315} and female protective effect^226,316 are ongoing. Fourth, it is important to determine the extent to which the observed sex/gender effects in neurodevelopmental conditions diverge from established sex/gender effects in the general population^123,317. Typically, larger sex/gender differences are reported in clinical studies, but infrequently the same patterns emerge in population-based studies^78,82–84. Fifth, the intersectionality of ethnocultural and sex/gender influences in neurodevelopmental conditions requires far more attention. For instance, in contrast to popular belief, the largest sex/gender differences in psychological constructs in the general population, and possibly neurodevelopmental conditions, are reported in egalitarian, high-income countries³¹⁸. Sixth, polarization exists in public and scientific discourses regarding sex/gender issues, including risks of both over-emphasizing and under-emphasizing sex/gender differences and influences for ideological reasons over scientific reasons. Research should seek to achieve a balanced and sober approach, unaffected by populism. Finally, although a pivotal goal of sex/gender-related research is to generate evidence that enables the best possible support for individuals of any sex and gender, studies on sex/gender influences of intervention effects are in their infancy. Inclusive and equitable intervention research to address the question ‘what works for whom’ must generate evidence that is sex/gender-informed, sex/gender-aware and sex/gender-tailored.

Future directions

We should aim to develop a better understanding of sex-informed and gender-informed measurements and outcomes via collaborative and participatory approaches with neurodivergent people and their families. We should recruit large-enough samples with stratification for specific sex-related and gender-related factors to provide the necessary information for sex-tailored and gender-tailored understanding and individualized prediction, and to quantify behaviours and biology in relation to general population sex and gender norms. We should account for specific sex-related and gender-related factors in proposed mechanisms rather than lumping these constructs together. Applying this research strategy is especially important, as hitherto, evidence does not allow robust separation of shared, non-shared and quantitatively different neurodevelopmental features by sex and gender, which muddles mechanistic understanding. To make real progress, future research should avoid using a single, conflated sex/gender label, but rather measure specific sex-related and gender-related attributes more precisely to test theory-driven hypotheses and allow post hoc and data-driven explorations. Study samples need to be diversified to ensure representation not only across sexes and genders, but also across ethnic, socioeconomic and cultural groups to address intersectionality. Inclusion of people with intellectual disabilities across research areas is crucial to ensuring that findings can be generalized to the neurodivergent population at large. Accordingly, new longitudinal studies will be able to capture the heterogeneity of developmental trajectories of the neurodivergent populations and clarify the developmental contributions of sex-related and gender-related factors to different life outcomes. With these improvements, we are optimistic that research will be able to integrate sex and gender into next-generation diagnostic, intervention and support practices favouring personalized profiles^29,319.

Key points.

• Sex-related and gender-related factors moderate or mediate biological and behavioural variability in neurodevelopmental conditions; they exert influence through diverse mechanisms that act at multiple levels.

• Moderating effects of sex and/or gender are supported by epidemiological, behavioural, cognitive, neurobiological, genetic, endocrinological and immunological evidence, especially in autism, but also in attention-deficit/hyperactivity disorder (ADHD), Tourette syndrome, dyslexia and fragile X syndrome.

• In autism and ADHD, research has started to raise awareness of clinical nuances to enable better and earlier identification of neurodivergent females.

• In neuroimaging studies, the neurobiological correlates of autism and ADHD show some overlap between males and females, but also some qualitative differences; preliminary evidence indicates that there are differences in the presentation of fragile X syndrome, dyslexia and Tourette syndrome in the brain between female and male individuals.

• Sex hormones are implicated in the mechanisms underlying sex differences in brain development and in susceptibility to neurodevelopmental conditions, but much of the evidence relies on preclinical studies.

• Future research needs to address under-representation of females, gender minorities and specific diagnoses; challenges of interpretation; insufficient understanding of specific sex and gender attributes; developmental variabilities; intersectionality of ethnocultural and sex/gender factors; and sex/gender influences of intervention.

Glossary

Gender diversity

The variations of gender identity and expression in the human populations that may or may not be aligned with those stereotypically linked to a person’s sex assigned at birth.

Gender-minority individual

A person whose gender identity is different from their gender and sex assigned at birth; for instance, someone who identifies as transgender or non-binary.

Genomic imprinting

Parent-of-origin-specific allelic expression of genes whereby only one copy of a gene is expressed while the other copy is suppressed; both maternally and paternally inherited genes can be subject to genomic imprinting.

Impression management

The ubiquitous human tendency to present favourable impressions of oneself in front of others during social interactions, such that one can achieve interpersonal or pragmatic rewards.

Lateralization

The tendency for selective neural or cognitive processes to be more strongly represented in or specialized to one brain hemisphere.

Liabilities

The sum of genetic and environmental factors that contribute to the development of a multifactorial phenotype; the genetic liability threshold model proposes that there is a continuous liability distribution within a population for a binary trait outcome and the distribution has a threshold that divides the population into two (that is, those with and those without the trait).

Morphosyntactic

An alternative term for ‘grammar’; grammar includes morphology and syntax, where morphology is about words and their formation, and syntax is about sentences and their formation.

Mosaicism

In genetics, mosaicism refers to the situation where an individual or organism has more than one genetic lineage within their cells, which can arise secondary to genetic mutation (including sex chromosomes) in the zygote; regarding neurophenotypes, mosaicism refers to the idea that the degree of typical ‘maleness’/’femaleness’ of specific features in the brain of an individual is not internally consistent.

Profound autism

A new administrative term proposed by The Lancet Commission on the future of care and clinical research in autism to describe autistic individuals with high support needs (that is, requiring 24-h access to an adult who can care for them if concerns arise, being unable to be left completely alone in a residence, and not being able to take care of basic daily adaptive needs), often owing to substantial intellectual disability, very limited language, or both⁷³.

Sex assigned at birth

The sex label (also termed ‘sex at birth’ or ‘sex designated at birth’) recorded at a person’s birth (for example, on a person’s birth certificate), typically given based on a child’s reproductive system and related physical characteristics (for example, genital anatomy).

Sex chromosome aneuploidies

The presence of additional sex chromosome(s) (X or Y) beyond the normal complement, associated with neurogenetic conditions; for example, X0 (Turner syndrome), XXY (Klinefelter syndrome), XXX (trisomy X, triple X syndrome), XYY (XYY syndrome, Jacobs syndrome).

Transdiagnostic methods

Approaches that reduce adherence to the conventional categorical diagnostic nosology or replace it with new frameworks that characterize medical, psychiatric and neurodevelopmental conditions in terms of relevant trait dimensions instead of discrete categories²¹.

X chromosome inactivation

The random silencing of one copy of the X chromosome per cell in individuals with XX chromosomes to compensate for dosage effects.