ASD is ~4 times more prevalent in boys than in girls (16, 31). The source of this marked difference remains unclear; its basis is likely to be multifactorial but is poorly understood. There is no doubt about the biological etiology of ASD—these disorders are highly heritable. But reasons for the sex ratio in prevalence is a mystery because few of the 100+ thus far identified risk genes reside on sex chromosomes. Baron-Cohen has proposed the “extreme male brain theory” of ASD, attributing sex differences to testosterone levels (32). This is plausible because multiple lines of evidence suggest dysregulated steroidogenesis (33, 34), but testosterone levels are unlikely to fully explain the male/female discrepancy. Rather than searching for the causes of heightened risk in males, one might look, instead, for female protective factors. Girls who are diagnosed with ASD carry significantly more genetic load than boys (35), suggesting adaptive mechanisms such as the ability to effectively mask ASD symptoms (36). If such mechanisms exist, they remain to be discovered. There are also several clinical differences between boys and girls with ASD: more externalizing behavior (e.g., aggressive outbursts, social impairment, hyperactivity, restricted behaviors and interests) in boys and more internalizing symptoms in girls (e.g., depression, anxiety, emotional symptoms) (37). Since diagnostic criteria are based on the much more prevalent male presentation, it is likely that girls, who present somewhat differently, are underdiagnosed (38).

If one omits neurodevelopmental disorders such as ASD, the male/female prevalence ratio is higher in eating disorders than in any other forms of mental illness. Anorexia nervosa affects 0.9–4% of women and roughly 0.3% of men (39). This major difference in prevalence rate is usually attributed not to biology but to contrasts in the gender roles assumed by women and men in adolescence and to the peer pressures to which they are exposed (40). Multiple interacting factors, however, probably play a part. Genetic and epigenetic causes, gonadal hormone levels, and constituents of the immune system have all been considered as potentially contributory to the prevalence ratio. In other words, socialization into gender-specific roles and societal expectations may be exaggerating innate proclivities (41). Because thinness in many cultures equates to female beauty, and beauty serves as a proxy for youth, health and reproductive fitness, eating disorders may, from an evolutionary perspective, reflect an ancient mating strategy—enhancing beauty to attract men who will ensure long term survival of offspring (42). In modern times, this strategy is made increasingly complex because media and the fashion industry define beauty standards.

In most prevalence studies, approximately twice as many women as men are diagnosed with PTSD, and this ratio is consistent across diagnostic systems, measurement methods, ethnicities, cultural backgrounds, and study populations (43). The trauma preceding this disorder is considered to be an exposure to one or more extremely threatening events, such as sexual, physical, or severe emotional abuse, a death in the family, or a serious accident, natural or manmade disaster (44). Symptoms of PTSD are more severe, chronic, and recurrent in women than they are in men (45–47). Associated symptoms/behaviors also differ. Women with PTSD in addition to the core symptoms report dissociative phenomena, somatization, disordered eating, low self-esteem, guilt, and shame. Men, in contrast, present with problems relating to aggression, thrill-seeking, risk-taking, and impulse-control (7). Diagnosable comorbidities induced by trauma also differentiate men and women -e.g., depressive and anxiety disorders accompany PTSD in women while substance use disorders are prevalent in men (48).

We will now explore in greater detail some of the theories that have been proposed to explain sex and gender differences in mental illnesses, using our three selected disorders as examples. The emphasis in each of the theories is different, but they all acknowledge that they are, in themselves, incomplete, and that complementary perspectives are needed for a full explanation of the many factors that result in sex/gender differences.

At conception, most females possess two X chromosomes and most males have one X and one Y. Some male-specific genes are located on the Y chromosome, the most important one being the SRY or testis-determining gene. Once testes develop, the process of subsequent differentiation between the sexes begins (49). Male testes and female ovaries secrete different levels of hormones (estrogens, androgens, progestins, and anti-Müllerian hormone) that subsequently influence most body organs, including the brain. Because females have two X chromosomes, one in each body cell is automatically and randomly inactivated so that males and females end up with an approximately equal total of gene products (50). But ~25% of genes on the inactivated X chromosome escape full inactivation. This process, in addition to epigenetic effects on genes on all chromosomes (50, 51), results in sex/gender differences in the rates of many medical conditions. Proposed mechanisms include the unequal level of gene products expressed by sex biased genes. These are genes that are present in both sexes but that express different quantities of protein products in men and women due to the impact of either epigenetic factors or sex hormone levels (52).

Genes are important even in conditions such as eating disorders, which result from mainly social causes. Twin-based heritability for anorexia nervosa ranges from 48 to 74%, for bulimia nervosa from 55 to 62%, and for binge-eating disorder from 39 to 45% (53). A recent genome-wide association study (GWAS) of ~17,000 anorexia nervosa cases and 55,000 controls yielded 8 significant genetic loci that may contribute to etiology (54). Susceptibility genes, it should be noted, may lie dormant for long periods and only become expressed under specific dietary or perhaps hormonal conditions (55). Sex chromosomes have only recently begun to be included in GWAS screens. In addition, many of the databases used for transcriptome analysis are skewed toward data that has been generated predominantly in males, which means that more genes relevant to sex differences in medical conditions still await discovery (56).

Sex biased genes are present in the central nervous system in great number and may be partially responsible for the neuroanatomical sex differences that have been reported in the human brain. Brain structure develops not only through the action of genes, but also in response to environmental influence so that postnatal experience, which usually differs significantly in girls and boys, affects the size of specialized brain regions. In spite of many reported sex/gender differences in both animal and human brains, these are subject to considerable heterogeneity, even within regions, so that the human brain is not male or female, but best described as a mosaic of relative masculinity and femininity (57). Nonetheless, the question remains as to whether early life programming by testosterone in male fetuses leads to enduring sex differences in brain and behavior (33). Girls who experience elevated androgens in utero tend to display prototypic male behaviors to a variable degree that does not necessarily correlate with structural changes in the genitalia (58). For instance, persons who are born completely insensitive to androgen can be chromosomally XY but phenotypically and psychosexually female (59). This is important because it means that in humans, as in experimental animals, the lack of androgens, rather than the presence of ovarian steroids, is what principally determines femininity.

The developmental trajectory of the brain differs by region and by sex such that the pace of maturation of large areas of the brain, such as the prefrontal cortex and the hippocampus, differs in boys and girls (60, 61). As a result, experiences of trauma or injury, though occurring at comparable ages, may manifest very differently in the two sexes/genders (62). In general, the pace of maturation and development of the nervous system proceeds at a faster pace in girls (63) and, at puberty, comes under the influence of estrogens, which exert protective effects on brain systems via reduction in neuro-inflammation, promotion of synaptic plasticity, and regulation of neurotransmitter signals (64). Growing scientific attention is also being given to sex-based differences in brain circuitry (65), which probably affect response to environmental insult. This means that, although differences in the brain are commonly considered the product of biological sex, gender-based influences may also play a strong, if not equal, role (66).

The central nervous system (CNS) and the hypothalamic-pituitary-adrenal (HPA) axis are both central to the body's response to stress. The role of the CNS is to respond quickly and efficiently to perceived stressors and the role of the HPA axis is to maintain that response for as long as needed, and to return the body to homeostasis once the threat has subsided (67, 68). The HPA axis connects the CNS to the endocrine system while also impacting the sympathetic-adrenal medullary (SAM) axis, the immune system, and the cardiovascular system, as well as directly modulating affective and cognitive processes in the brain (69, 70). A dysregulated HPA axis is central to the trauma and stress response. Dysregulation can result in either an inadequate response to threat and or an excessive stress response that persists long after the threat has passed; both are potentially harmful (71). As a result, the HPA axis is hypothesized to strongly influence the development of stress- and trauma-related disorders such as PTSD. Furthermore, a dysregulated HPA axis can help to explain why people suffering from PTSD respond to traumatic triggers either with hyper-arousal (increased heart rate, respiration, and blood pressure), often resulting in a startle response and a feeling of panic, or with hypo-arousal (decreased heart rate, immobility), a frequent precursor to dissociation. Interestingly, the former response appears more common in men and the latter in women (71), for unknown reasons.

The development and functioning of the HPA axis are influenced by one's genome, the prenatal environment, and by life experiences. HPA functioning is especially sensitive to repeated early life trauma (72). As a result, there are multiple ways in which both sex and gender can affect the HPA axis. Both adrenocorticotropic hormone (ACTH) and free cortisol levels have been reported to increase twice as much in men as in women in response to a psychosocial stress task, though results may depend on the specific stressor studied (69). Some studies have suggested that women's monthly hormonal fluctuations sensitize the HPA-axis (73, 74). This may be one reason why women, more than men, risk developing stress-and-trauma-related disorders. In ASD, on the other hand, multiple lines of evidence suggest that dysregulated steroidogenesis during early development contributes to male preponderance (33, 34). With respect to eating disorders, an activated HPA axis has been closely associated with these conditions although the activation may be secondary, an effect of starvation (75).

Somatic factors outside the brain, such as immune processes, are capable of influencing the brain and resulting in psychiatric symptoms. In most animals, male sex is associated with a relatively lowered immune response and a relatively higher susceptibility to infection (76). The explanation may lie in the many immune-related genes that are encoded on the X chromosome. Although, as explained earlier, one of the two copies in females is silenced, immune-related genes may escape inactivation in some body cells. As well, immune cells such as T cells and B cells show varying sex-related rates of age-associated decline, with immunosenescence starting earlier in males than in females by ~6 years (76). Furthermore, female hormones play an important protective role in immunity. Estradiol, for instance, has been convincingly shown to reduce the levels of inflammatory cytokines (77). This accords with preclinical animal research, which finds both greater levels of inflammatory signaling molecules and activated immune cells in healthy males than in healthy females (78). From an evolutionary perspective, immune sex differences serve an important purpose because the female immune system needs to suppress an antibody response to male sperm during conception and to male fetuses during pregnancy (79).

Sex differences in the immune system affect the brain, and are seen most notably in the functions of astroglia and microglia, both brain cells that reportedly play a role in psychiatric disease (80). It is currently believed that a combination of genetic sex, plus the organizational and activational effects of gonadal hormones, plus levels of stress hormones contributes significantly to brain immune function (81).

Inflammation experienced either in utero or very early in life is a known risk factor for ASD (81–83), and transcriptome analysis of fetal human brain reveals high levels of gene expression consistent with increased baseline immune reactivity in PTSD (46, 47), which may be transmitted to offspring in a sex-specific manner through varying levels of maternal proinflammatory cytokines over the course of pregnancy (84). Animal research suggests that maternal immune activation or early life inflammation exert negative impact on the brains and subsequent behavior of males, but not females (85).

There are connections between the immune system, the brain, and the gut (86). The composition of intestinal microorganisms appears to affect brain functioning and sex/gender reportedly plays a part in that composition, although study results are inconsistent (87, 88). Men and women tend to eat different foods, take different medications, show different susceptibility to weight gain, and their colonic transit times, on average, differ. These are all factors that affect gut bacteria and potentially influence the expression of many psychiatric disorders (89–91), but this area of research is still in its infancy.

Gonadal hormones play important roles in differentiating the manifestations of mental disorders in men and women (92–94). Besides conferring protection against the effects of injury, inflammation, ischemia and apoptosis, estrogens also promote neuronal growth and enhance neuro-reparative processes such as remyelination. They also serve important regulatory functions in relation to neurotransmitter systems (94). There is evidence that gonadal hormones play a critical role in aiding the epigenetic expression of eating disorder genes (95–97). Gonadal hormones also contribute to our understanding of ASD (98) and PTSD (99).

Though findings are confounded by differences in measurement methods and ascertainment times, time periods associated with female reproduction, such as pregnancy, the postpartum, and menopause are often associated with the emergence or aggravation of psychiatric symptoms (100). For example, women exposed to traumatic events during the early follicular phase of their menstrual cycle show comparatively increased risk of developing PTSD (100). Naturally occurring hormonal fluctuations in women can also affect response to treatment with psychiatric medications (101). The efficacy of certain psychotropic medications appears to decrease during the luteal phase of the menstrual cycle (102) and the risk of drop-outs from trauma-focused cognitive behavior therapy varies according to the time of the month in which it is initiated (7).

An important limitation to note when discussing the influences of gonadal hormones on symptom development is that much of the data is based on animal studies and requires further study in humans (101).

Whereas, sex differentiation is initiated at conception, the influences of gender start to take effect after birth. Parental expectations of offspring, however, precede birth. Parents choose nursery decor, infant clothing, distinctive toys and given names to match cultural gender norms. For example, in many cultures, girls' names allude to beauty and grace while boys' names suggest strength and sturdiness (103). Differentiation along similar lines is widely accepted, as shown by results of a recent study from China. College students, when asked to rate forenames, characterized feminine names as reflecting warmth and masculine names as representing competence (104).

Gender norms are defined as the expectation that boys and girls, men and women, think, feel, talk, and generally behave in gendered ways. The content and rigidity of gender norms and their endorsement by parents and peers vary across cultures and sub-cultures, but the general rule is that boys are expected to be strong and to control their emotions (except anger, which is permitted) whereas girls throughout the world are expected to be amiable, relatively docile, and nurturing (8). Gender traits are often grouped together and referred to as femininity vs. masculinity.

Children learn to act and to think according to gender norms, with parents considered to be the primary channel through which these norms are transmitted (105). Parents interact differently with girl and boy children, encouraging more physical exploration and toughness in boys and more emotional communication, tenderness, and focus on appearance in girls (106). This has been seen as contributing to the high prevalence of eating disorders in girls and women. Parental expectations are reinforced by interpersonal interactions and by the media and can lead to the development, in girls, of vulnerability to body image concerns. Though this is a generality, gender norms frequently encourage girls and women to aim for an idealized body shape and boys and men to nourish a drive for muscularity, accomplishment and power (107, 108).

In general, children are rewarded for what the world around them judges to be desirable behavior and criticized for behaviors that are idiosyncratic (109). In terms of gender socialization, this means that society, including one's parents, teachers, and peers, have low tolerance for personality traits and characteristics traditionally assigned to the opposite gender. Over time, such socialization processes strongly impact the way children see themselves. This happens through mechanisms such as the Rosenthal (Pygmalion) effect (110). Phenomena such as the frequently impaired communicative abilities in boys and underdeveloped mathematical abilities in girls have been linked to this effect.

Stress arises when the environment poses threats to the gender norms that one's culture values. For this reason, men and women are to an extent stressed by different kinds of threats. Women, reportedly, are most threatened by relationship losses whereas threats to status and financial stability are especially stressful for men (111).

Children internalize the gender norms to which they are exposed and, over time, impose these stereotypes on themselves, a phenomenon referred to as self-socialization (112). Children especially take on gender roles modeled by same-sex adults and peers whom they admire and this affects their actions, appearance, hobbies, dreams, educational pursuits, careers, intimate relationships, and eventual parenting styles (113). Structural theories suggest that different gender roles lead to exposure to different kinds of threats and stressors, each calling for different coping strategies, and resulting in gender differences in adaptability and mental health (114). One such theory, role restraint theory, states that, when men and women are allowed to take on similar roles and face similar stressors, differences in their responses disappear (115). Socialization theories address not only exposures but also the development of sex-specific vulnerabilities to stressors and sex-specific coping strategies, e.g., women tending to confide and share, men tending to deny and seek distraction (116).

Though the term “gender identity” is sometimes used to refer to self-perceived gender, it is best defined as a person's deeply felt internal and personal experience of their own gender (117). This includes how one sees oneself, the possibilities and restraints attributable to one's gender, a sense of one's body, societal role, and sexual preferences. It also includes non-binary gender identification (118). Gender identity is most lastingly developed during adolescence, with body image, sexual orientation, and range of potential gender roles becoming central themes (119). A perceived incongruence between society's gender norms and one's gender identity can become a source of emotional problems and felt constraints. For example, masculine gender-role stress, defined as stress resulting from the struggle to live up to masculine ideals, has been associated with PTSD (1, 120). A striking example would be a male veteran who suffers from PTSD after a sexual assault and struggles with the perceived discrepancy between military masculine ideals and the role of victimhood and someone struggling with mental illness (120).

There are multiple sources of gender bias both in mental health research and in the clinical setting. Women have traditionally not been widely recruited for biologically-based mental health research studies because their hormonal fluctuations complicate the task of appropriate matching and because of the concern that an experimental intervention may potentially harm the fetus should an unexpected pregnancy occur. This means that clinical guidance is often still based on research involving mainly men (7). In contrast, many psychotherapeutic trials are conducted primarily on women, simply because more women than men seek help for mental health problems. Furthermore, the vast majority of PTSD research on men is conducted on military personnel or veterans and may not be generalizable to civilian men, let alone to women.

Even in a biologically based disorder, such as ASD, there is evidence of diagnostic bias. This is likely due to the fact that the earliest cases of these disorders were observed in boys, so that the diagnostic criteria were established based on characteristic male symptoms. As stated earlier, it is perhaps because of this that ASD remains often undetected in girls, especially when symptoms are mild (37, 121). Sources of data have substantial clinical impact on how diagnostic criteria are formulated, on screening tools, on diagnostic and clinical guidelines, and on the efficacy of the many different ways in which mental illnesses are treated. Most standardized tests, diagnostic questionnaires, and diagnostic criteria themselves are based on the assumption that men and women experience symptoms in the same way and that a specific question is equally well-suited for eliciting relevant symptoms in women and men, whereas this may not be the case (122). In contrast, in the absence of standardized measures, clinicians may preferentially ask questions of one sex/gender and not of the other. For example, women (but not men) are commonly asked during assessments about exposure to sexual trauma and perinatal loss, though men, too, are exposed to such experiences and their sequelae. The same is true for questions about spousal abuse. Women, on the other hand, are unlikely to be asked about ever engaging in acts of aggression. Similar to the Rosenthal effect mentioned previously, when men and women are repeatedly asked questions about specific experiences and/or symptoms, in time, they will become more likely to report such symptoms whereas, when these areas are left unprobed, they stay unreported, especially so if they are perceived as incongruent with gender norms (115). Thus, clinicians may be more likely to overlook the presence of certain disorders in one sex compared to the other.

Furthermore, using dichotomous categories to assess sex and gender forces both male and female study participants into highly heterogenous groups, with no allowance for intra-group variability. It may be time to rethink how we assess sex and gender and introduce continuous measures, allowing for a wider range of gender-normative variability.

Diathesis-stress models have been used across disciplines to integrate multiple baseline vulnerabilities or predispositions, so-called diatheses, with situational stressors (triggers) (131). According to these models, whether or not a specific person develops a disorder depends upon the interaction between pre-existing vulnerabilities and the extent, number, timing, and severity of the stressors they experience.

We propose a diathesis-stress model for sex and gender differences across mental disorders (though we illustrate the model via our three selected disorders—ASD, eating disorders, and PTSD). As indicated in Figure 1 through the use of color gradations, the emphasis on diathesis vs stress varies according to disorder. A diagnostic category thought to originate mostly from diatheses is shown in a darker blue than one mostly caused by exposure to stressors and traumatic events. For example, ASD shows substantial heritability and results mainly from biological causes (132). Moreover, it emerges very early in life before many environmental stresses have had a chance to accumulate. However, in utero stressors, are also part of the genesis of ASD (133), thus a diathesis-stress model is appropriate. Furthermore, disruptions in everyday life and comorbidities elicit symptoms and, thus, are central to the expression and quality of life in ASD (134). In contrast, PTSD requires exposure to a traumatic event and cannot, by definition, develop in the absence of such an event. However, whereas highly resilient individuals may only develop PTSD after a particularly toxic traumatic event, extensive underlying diatheses can cause others to succumb in response to relatively minor stresses. Finally, though most investigators consider social pressures to be the main factors causing eating disorders, there is evidence for substantial individual predispositions (135), thus placing eating disorders in between ASD and PTSD. Please see Table 4.

Few models have addressed the potential effects of multiple diatheses in combination with an accumulation of situational stressors. Such a model (131), however, is truer to life given that the etiology of psychiatric disorders is multifactorial and complex. One complexity is the usual categorization of sex and gender either as a diathesis or as a risk factor (136), but this is an oversimplification because sex/gender exerts its combined effects through a relatively large number of mediating factors (136).

Hyde and colleagues (137) constructed an integrated model of the emergence of sex/gender differences in depression in which they hypothesize that affective, biological, and cognitive factors converge to result in a vulnerability that is only expressed when a severe stressor triggers symptoms, the so-called ABC model (137). Factors a, b, and c interact with each other and with sex/gender to form the diathesis. The presence of a particularly strong risk factor within one of these domains may be sufficient for symptoms to emerge. Equally possible is for relatively weak risk factors across two or three domains to combine and, together, trigger depression. Although Hyde et al.'s model focuses on depression, a similar model can be constructed for other disorders, including those that first emerge in childhood (e.g., ASD) and those that most commonly emerge in adolescence (eating disorders) or adulthood (PTSD).

We suggest that a diathesis-stress model can be created to account for the complex interaction between sex and gender and their combined involvement in the development and expression of multiple mental disorders. Whereas, some aspects of the model will be specific to certain disorders, many elements will be relevant across diagnoses. We will now introduce some of the key elements of such a model.

In accordance with a model presented by Parker and Brotchie (138), designed to account for sex/gender differences in depression, Christiansen (1) first suggested that the diathesis component of the model represents an inherent preparedness toward responding to a stressor with excessive fear and anxiety, which, when a person is exposed to trauma, significantly increases the risk of developing symptoms (1, 138). This inherent preparedness is, on average, more marked in women than in men and, thus, represents a biological sex difference in PTSD vulnerability (139). There is evidence that women, as a group, have a greater likelihood than men to be genetically predisposed toward fear (140). In line with other models (130, 131), we acknowledge that diatheses may vary in their degree of specificity, with some being disorder-specific (e.g., early childhood mealtime conflicts as a specific risk factor for eating disorder in adolescence), while others are shared among multiple disorders (e.g., negative affectivity or childhood trauma) (141, 142). Elwood and colleagues (136) suggest a possible hierarchy of diatheses that are more or less essential for the development of a specific disorder. Their model supports our earlier statement, that a single high-level diathesis may be sufficient for a person to develop a disorder when exposed to stress, whereas a lower-level (less specific) diathesis can only lead to symptoms when combined with other pre-existing vulnerabilities.

McKeever and Huff (131), in their diathesis-stress model of PTSD, suggest two types of diatheses: ecological and biological, which influence each other. Moreover, either one can elicit specific stressors from the environment. Ecological diatheses consist of risk factors linked to an individual's self (e.g., developmental history, coping mechanisms, social support). The distinction between ecological and biological diatheses is particularly relevant in a model of sex/gender differences, as gender would probably be considered an ecological diathesis and sex a biological diathesis, depending on the disorder. The distinction is further relevant to a model spanning multiple disorders, with some being biologically influenced (e.g., ASD) and others (e.g., PTSD) being more ecologically based. Our model distinguishes among different ecological diatheses by separating out intrapersonal factors, interpersonal/relational factors, and societal factors. Our model also differentiates between sex and gender because we believe that biological diatheses will mostly (but not exclusively) rely on sex, whereas societal diatheses will mostly rely on gender, and intra- and interpersonal factors will likely rely equally on both. A non-conclusive list of diatheses related to sex/gender differences in mental disorders can be found in Table 5.

Finally, whereas diathesis-stress models generally focus specifically on personal vulnerabilities to disorders, we think of the presence or absence of external resources also being diatheses. An example would be the presence of social capital serving as protective factor against symptom development (143).

Biological vulnerabilities in any one disorder are often specific to that disorder, but some are more general. One example is the involvement of the same neurotransmitter in multiple disorders, which is why selective serotonin re-uptake inhibitors (SSRIs), for instance, are effective in several different categories of mental disorder. Biological vulnerabilities relevant to sex/gender differences are likely to include genes on X and Y chromosomes (but also on autosomes), impairments of brain structure and function, insufficient or excessive levels of gonadal and other hormones, disturbances of HPA axis reactivity, immune system, and/or microbiome irregularities, as addressed earlier. Biological diatheses are often present at birth and strongly impacted by sex. However, gender also affects biological factors through experience, particularly through the influence of socialization and gender-specific stressors that lead, for instance, to the activation of the HPA axis or to epigenetic arousal of dormant genes.

Intrapersonal diatheses include affective and cognitive vulnerabilities that, along with pre-existing symptoms and disorders, serve as predispositions to new disorders. Affective factors consist of temperament, negative affectivity, sometimes referred to as neuroticism, and mood. Most research agrees that women score higher than men on measures such as sadness, grief, negative affectivity, and anxiety proneness, whereas men tend to score higher on measures of aggression, anger, and impulsivity. Common cognitive vulnerabilities include rumination, maladaptive coping styles, poor self-perceived coping efficacy, negative appraisals, and negative schemas. Cognitive schemas are best understood as internal working models of how individuals view themselves, other people, and the world. Understanding gender norms, gender roles and gender-related schemas, and one's expectations of oneself and others fall into this category.

In our model, coping strategies and appraisals are considered post-stressor factors in that they are often situationally specific and associated with a particular stressor and a person's response to that stressor, as well as with subsequent symptoms. However, there are overall general coping styles and attributional styles that remain relatively stable over a person's life, and, as such, should be considered diatheses. For instance, women have repeatedly been found to ruminate more than men, to appraise situations as more threatening/severe than men do, and to score higher on maladaptive aspects of emotion-focused coping (144). These are enduring traits that serve as diatheses for all internalizing disorders. Furthermore, they exacerbate existing disorders, including neurodevelopmental disorders such as ASD, by undermining levels of functioning and quality of life, making it more difficult for already vulnerable individuals to cope with their symptoms. By contrast, enduring tendencies toward anger, aggression, and lack of impulse control serve as diatheses for externalizing disorders. Finally, our model also places internalized gender roles, masculine and feminine traits, gender identification, and sexual preferences into the category of intrapersonal vulnerabilities.

Interpersonal aspects of our model include sex/gender differences related to insecure attachment, poor or negative social support (145), and victim blaming (146). Attachment is usually considered an intrapersonal factor, but we have included it in this category, as the quality of early attachments exerts a strong influence on subsequent relationships. Though women, as a group, are not less securely attached than men, they do score higher on the attachment anxiety dimension of attachment, which has been linked to later symptom development (147). With respect to social support, women have consistently been found to be recipients of both positive and negative social support, more so than men. Though positive support serves to buffer the impact of stressors, the effects of negative support (hurtful social interactions) may carry greater weight and undermine mental health.

Recurrent patterns of intimate interpersonal interactions with friends, family members, and lovers (e.g., abusive relationships, response to sexual intimacy, enjoyment of sex, relationship gender roles) fall into the category of interpersonal diatheses, whereas we categorize one off stressful interactions (e.g., fights, assaults) as stressors.

The last group of diatheses relating to sex/gender differences in mental disorders are social and environmental: socio-economic status, caste, residential area, educational level, work status, degree of independence, marginalization, politics and other societal factors that set individuals apart. Social hierarchies affect men's and women's access to financial resources, support, protection, financial aid, medical and psychological treatment, police assistance, and legal justice and matter greatly in the aftermath of a traumatic exposure. They are strongly influenced by gender. Factors such as financial dependence, poor access to healthcare, and exposure to domestic abuse put women at a much higher risk than men for developing trauma-related disorders, even though men, as a group, are more exposed to industrial accidents, criminal victimization, and the trauma of war (7).

Gender-based socialization, gendered violence, misogyny (and, to a much lesser extent, misandry), homophobia, and transphobia, also belong in the social diathesis category, whereas the internalized version and enactment of such phenomena fit best under intrapersonal diatheses. Though we designated gender roles as intrapersonal (because they arguably result from internalized gender norms), we have called gender-role stress a social diathesis as it results from a mismatch between internalized gender identity and society's gender-based expectations.

We note in our model that sex and gender has the power to modulate societal diatheses, for example through the ability of secondary sexual characteristics, dress, hair styles, the application of cosmetics or scents, and plastic surgery to elicit and change the behavior of others, leading to changes in one's social status. Traits associated with masculinity, such as assertiveness and aggressiveness, are also able to alter established social hierarchies.

The stress component of our model is composed of a range of events, from severe to minor. A non-exclusive list of such events shown previously to contribute to sex/gender differences in mental disorders is shown in Table 6. Aversive childhood events such as childhood sexual abuse are associated with multiple mental disorders later in life. But even a relatively minor break in routine upsets a child with ASD and can trigger a cascade of symptoms. As proposed elsewhere (1), the stress component of this model is much more likely to be affected by gender than by sex.

Men, on average, experience a higher number of traumatic stressors than women. These is particularly the case for non-sexual torture, physical violence, severe accidents, gang violence, and witnessing severe injuries and death of others during both war and peacetime (46, 48, 144). Women, on the other hand, experience obstetrical trauma, sexual violence, and betrayal trauma, including intimate partner violence. With respect to everyday stressors, female gender role norms result in women being more embedded than men in the social networks of family and friends. As a result, women are exposed to more non-romantic interpersonal conflicts, losses, and indirect stressors [i.e., negative events that happen to others, such as the divorce or illness and death of close friends (148)]. Though slightly dated, one study of married heterosexual American couples found that wives reported more everyday stressors than husbands, especially in terms of household stressors, family demands, demands from others, and arguments with children (148). In turn, husbands reported more work stressors, stressors related to transportation, and arguments with non-family others. These differences may be accurate but may also result from reporting bias, as men and women differ, both in term of what stresses them, and how willingly they report such stressors. According to a classical paper, unmarried women live longer than married women, whereas married men live longer than single men (149). This may no longer apply because the division of labor between wife and husband has changed markedly over the years (at least in the West), but married women still take on additional caretaker tasks in families, as well as being more affected than men by partnership and family conflict, whereas men mostly benefit from the positive social support they receive through marriage (150).

Finally, sex/gender differences in the types of traumatic and non-traumatic events that men and women are exposed to may influence the type of symptoms they develop. For example, perhaps as a result of being more commonly exposed to betrayal trauma, women may be more likely develop self-doubts that lowers their self-esteem (151). Another example referred to earlier: whereas sexual assault often causes multiple problems with intimacy, sexuality, and self-worth in both men and women, male victims may, in addition, subsequently struggle with what they perceive to be a threat to their masculinity (120).

In accordance with other diathesis-stress models (e.g., 120), only immediate stressors are represented in the stressor component of our model. Once the stress has either subsided or become chronic, it may become part of the diathesis going forward, influencing the response to subsequent stressors. For example, a specific incident of racism, sexism, or homophobia may initially serve as a stressor and subsequently contribute to societal (e.g., perceived systemic racism or institutional sexism), intrapersonal (e.g., rumination), interpersonal (e.g., problems with intimacy), and biological diatheses (e.g., a more reactive HPA axis). The reverse is also true. Diatheses contribute not only to symptomatology but also to the types of stressors to which men and women are differentially exposed. Furthermore, an acute episode of severe mental illness may itself serve as a stressor, thus perpetuating the cycle of stress and distress. In this way, the diathesis and the stressor components of our model are impacted by each another.

Though peri- and post-stressors do not fit neatly into either the diathesis or the stressor part of the model, they play a very important role in the development of many psychiatric disorders. In the case of PTSD, two meta-analyses have found that pre-traumatic risk factors account for only a small percentage of variance in PTSD development compared to peritraumatic and posttraumatic risk factors (152). Thus, a full diathesis-stress model can fully explain sex/gender differences in PTSD only to the extent that it can find a way to include effects of risk factors occurring at the time of and following a trauma. Similarly, for other disorders, the manner in which an individual appraises, responds to, and copes with a given stressor is often central to severity and chronicity. Although very relevant to chronicity, few diathesis-stress models have explicitly included peri and post-stressor factors (153).

It could be argued that peri-stressor factors should be part of the stress component of the model, as it is not the event in itself, but rather the full experience of it, e.g., perception, appraisal, and response, that is central to symptom development. However, this solution does not accommodate post-stressor factors that are vital to the development, and especially the maintenance, of symptoms.

It is important to note that peri-and post-stressor risk factors are strongly linked to pre-stressor risk factors. How a person responds to an event is, in large part, dependent on personality, affectivity, neurobiology, and past experience. Thus, we argue here that, since peri- and post-stressor factors are simultaneously influenced by diatheses and by the type and severity of the stressor, while also contributing to the overall experience and effect of the stressor, they deserve separate representation in the model. We have linked them to the diathesis-stressor interaction by referring to them as “stress reactions,” which includes both acute and long-term responses. Acute responses are appraisals of the stressor plus the biological and psychological responses that occur during and shortly after stress exposure. More long-term responses are continued appraisal, situational coping, social support, and the reactions of others. While these stress reactions contribute to both the type and the severity of symptoms, they are also, in turn, affected by the symptoms that emerge. For example, an individual may appraise early symptom development as either a natural reaction to a traumatic event or, at the opposite extreme, as a sign of losing one's mind. This may, in turn, either decrease or increase the number and severity of symptoms, as well as affect re-appraisal of the traumatic event. Though this interaction is bidirectional during the early aftermath of the stressor, we have chosen in our illustration of the model to include only arrows going from stress reactions to disorders/symptoms in Figure 1, in order to acknowledge that, over time, this association goes mostly in one direction.

Stress reactions are particularly relevant to consider in a diathesis-stress model of sex/gender differences as men and women tend to differ widely in both their acute and their long-term reactions to stress. As previously discussed, sex/gender differences have been demonstrated in the immediate physiological and psychological stress response, e.g., in hypo- vs. hyperarousal, perception of social support, coping ability, and treatment-seeking. It is our belief that sex/gender differences in these early responses are central to the emergence of clinically relevant symptoms and to their maintenance.