The gender gap: fact or artifact?

Before examining theories to account for the gender difference in the prevalence of depression, it is important to establish that the female preponderance is not artifactual. Two major reporting artifacts have been cited as potentially contributing to the reported gender gap in depression – a help-seeking artifact, and a recall artifact. The help-seeking artifact refers to a perceived reticence in males to seek treatment or advice for depressive symptoms, which could explain the preponderance of females seeking treatment. However, in a worldwide epidemiological survey for the World Health Organization, Weissman et al. (1996) found that the rates of depression identified in community samples were in accord with those reported from primary care settings. More recently, Kuehner (2003) came to the same conclusion through a systematic review of the literature. The recall artifact postulates that women’s recall is biased in favor of past negative affective states, and thus women report a higher rate of lifetime depression. Kuehner (1999) conducted a controlled test of this by comparing women and men’s reports of depressive symptomatology during a depressive episode, and their recall of these symptoms six months later. He found no recall artifact: there was no sex difference in the differential between the reported severity of symptoms at times one and two, and Kuehner’s more recent systematic review of the literature also concluded that this artifact could not explain gender disparities in mood disorders (Kuehner, 2003).

A third suggested artifact is that the higher rate of sexual abuse and rape experienced by girls and young women accounts for subsequent depression rates, although it is worth noting that this does not constitute an artifact per se, but rather a potential etiological explanation for the gender difference in mood disorders. Kessler (2000) did find that, after rape and sexual trauma were controlled for in a population database, the gender difference in a first episode of depression was halved. However, when traumatic experiences more likely to be experienced by men were also controlled for, the female preponderance was restored. Findings such as these suggest that the increased prevalence of depression in females is not due to artifact (but see Chapter 1).

The gender intensification hypothesis

The gender intensification hypothesis (Hill & Lynch, 1983; Wichstrom, 1999) suggests that gender role orientations become more differentiated between the sexes over the adolescent years, as a result of exacerbated gender socialization pressures during this time. For women, these pressures, both direct and in the form of social learning, are primarily thought to occur through observation of their parents’ marital relations, which emphasize lesser public power and greater responsibility for the domestic sphere and care as part of the female gender role (Obeidallah et al., 1996), and through the socializing effects of parenting behaviors (Sheeber et al., 2002). The hypothesized effect of these socialization experiences is the promotion of assumptions that emphasize collectivity, and a lower sense of self-esteem amongst women. These tendencies, in turn, contribute to the increase in depressive symptoms in women. The consequence of this intensified gender stereotyping, according to these theorists, is that deficits in self-efficacy and instrumentality (i.e., the belief that one can take action to influence personally important outcomes), reflected in low levels of traditionally masculine personality characteristics, may place young adolescent girls at higher risk for depression through greater exposure to experiences that promote learned helplessness (Obeidallah et al., 1996).

One important source of socialization experiences is parenting behaviors throughout childhood and adolescence. Based on reviews of clinical and developmental literature, Hops (Hops et al., 1990; Hops, 1992, 1996) posited two pathways by which parents may inadvertently increase their daughter’s risk for depressive symptomatology and disorder. First, familial socialization processes may serve to normalize and encourage girls’ expression of depressive-like behaviors (e.g., sadness; self-derogation). Indeed, there is some evidence that girls may be differentially socialized to display depressive behaviors during childhood. In a review of parental socialization of emotion, Eisenberg and colleagues (1998) found that, although parents do not typically report reacting differently to girls’ and boys’ emotional displays, observational data suggest that there are indeed differences, “albeit perhaps less than one might expect.” In particular, a series of studies indicated that parents put more pressure on boys to control their emotions and “unnecessary” crying. Block (1983) reported that parents were quicker to respond to crying in girls than in boys. Parents’ meta-messages about the acceptability of emotional expressions are apparently clear to children in that boys expect their parents to disapprove of their expression of sadness more so than girls (Fuchs & Thelen, 1988).

Parents’ reactions to children’s negative emotions may also provide them with gender-differentiated strategies for regulating negative affect. Although the data are limited, some evidence suggests that boys may be encouraged to use distraction and problem-solving more so than are girls (Eisenberg et al., 1998). In fact, one study indicated that school-age children expected fathers to respond to boys’ emotional expressions with problem-solving and mothers to respond to girls’ by focusing on feelings (Dino et al., 1984). Maternal encouragement of emotional expression in response to a stressor was also found to mediate the path from sex to depression by age 15 (Cox et al., 2010). Similarly, in a review of the origins of ruminative coping styles, Nolen-Hoeksema (1998) indicates that failure to teach girls active coping strategies for dealing with negative affect contributes to their greater use of ruminative styles of responding to depressed moods. Further, she suggests that, to the extent girls are told they are naturally emotional, they may have lower expectations that their behavior can influence their affective experiences. A recent study on gender and coping styles supported this, indicating that coping style accounted for 36% of the variance in a sample of 14- to 18-year-olds, and that girls were more likely to use ruminative coping, which in turn significantly predicted depressive symptoms. On the other hand, males were more likely to use problem-solving and distraction coping, which accounted for the protection against depression that has been attributed to masculine traits, as discussed later in this chapter (Li et al., 2006).

Second, differential parental reinforcement of gender-typic behaviors may lead girls to display less instrumental and more relationship-focused behaviors, both of which are related to theoretically derived and empirically supported risk factors for depression. Huston (1983) reported that girls receive more encouragement for dependency and affectionate behavior. They are also reported to receive more support for nurturant play (Ruble et al., 1993) (also see Chapter 1). Interestingly, Block (1983) reported that in Baby X studies, in which infants are “assigned” a gender (i.e., the same baby is labelled a “boy” and a “girl” in interactions with different participants), adults provided more reinforcement for nurturant play when the baby is said to be a girl. Such evidence is compelling in that Baby X studies control for gender differences in children’s actual behavior.

On a related theme, evidence suggests that mothers encourage girls more than boys to have concern for others, to share and to behave prosocially (see Keenan and Shaw, 1997 for a review). However, recent literature has suggested that girls and boys do not actually differ in their levels of prosocial behavior, but instead in how those prosocial behaviors are manifested. A brief longitudinal study indicated that the same parenting behaviors and parental attributions for prosocial behavior led to more feminine (compassionate) prosocial behaviors in preschool girls, and to more masculine (agentic) prosocial behaviors in preschool boys (Hastings et al., 2007). However, the rates of prosocial behavior overall between girls and boys did not differ. Parents may be less attentive to girls’ assertive behavior (Kerig et al., 1993), and may also impede the development of girls’ sense of mastery by limiting their activities and freedom. In a 1983 review of the literature, it was reported that mothers were more likely to give unnecessary assistance to girls than to boys, and were more likely to reward frustration with physical comfort (Block, 1983).

There is also modest evidence that girls have lower self-evaluations of their own efficacy and that such evaluations are related to depressive symptomatology (Avison & McAlpine, 1992; Ohannessian et al., 1999). In an older review, Ruble and colleagues (1993) found that preadolescent girls tended to report lower expectations for success, more maladaptive attributions for success and failure, and poorer self-esteem than age-matched boys. Though gender differences did not emerge in all of these older studies reviewed by Ruble and colleagues, the direction of effects was consistent when gender differences were observed. It is important, however, to remain cognizant of the likelihood that disturbances in perceived self-competence may be sequelae rather than causes of depressive symptomatology. Hence, two longitudinal 17–18-year-old studies by Cole and colleagues (1998, 1999) suggest that children’s underestimates of their own competence emerge as a function of depressive symptomatology and that controlling for depression eliminates the observed gender differences in estimation of competencies.

Thus, although these older data suggest that parents’ early gender-differentiated socialization of children’s emotional and social behaviors may have an effect on children’s ability and motivation to regulate emotion, large gaps remain in the literature. In particular, these largely cross-sectional studies do not provide evidence of current gender differences or that such parental behaviors are predictors rather than consequences of children’s sex-typed behaviors. For example, if girls display more sadness and boys more anger, it would be reasonable to hypothesize that parents’ tendency to discuss sadness with girls and anger with boys emerged consequent to the children’s behavioral propensities. Similarly, it is not clear whether parents’ tendency to be more emotion-focused in response to young girls’ emotions and more problem-focused in response to young boys does not reflect girls’ earlier verbal and emotional development (Keenan & Shaw, 1997). An important caveat is that the research discussed herein focuses on parents’ responses to children’s normative emotional expressions. Though we consider it reasonable to construe depressive symptomatology as being at one end of a continuum that includes normative affective expression, the connection between early socialization of depressive-like behaviors and subsequent depressive functioning remains speculative.

In a meta-analysis of research dating from the 1950s to 1990, Lytton and Romney (1991) concluded that, although there is modest evidence of parental encouragement for sex-typed activities, overall the evidence does not support differences in parental restrictiveness or encouragement of either achievement or dependency as a function of child gender. Hence, it appears that the evidence for the shaping of differential activities for girls and boys is stronger than that for other areas of gender-socialization. Additionally, a more recent longitudinal study that examined intensification of gender roles at the beginning of adolescence found that, although masculinity did predict lower levels of depression among both girls and boys, no difference in masculine traits was found between the sexes. Thus, the finding did not account for the higher rates of depression in girls (Priess et al., 2009).

Limitations of the gender intensification hypothesis

Aside from the limited number and lack of contemporary studies that would allow strong causal inferences to be drawn regarding socialization and the emergence of the gender gap in mood disorders, there is also some literature that is inconsistent with this hypothesis. Studies documenting the trajectories of children’s own gender role concepts have shown that the rigidity of children’s gender stereotypes tends to lessen as adolescence approaches. These studies have also found that boys are likely to labor under more rigid self-imposed sex-types than girls (Banerjee & Lintern, 2000). Furthermore, gender intensification theorists cite the emergence of the gender gap at adolescence as support for their argument; however, as will be discussed below, the onset of the gender difference in depression is not predicted by age as such, but rather by the pubertal status of the individual (Angold et al., 1999). If female depression is a consequence of a broad societal pressure applied to girls when they reach a certain age or stage of schooling, this should not be the case.

Another problematic finding with regards to gender intensification theories of adolescent depression is that the rates of female depression have not markedly lessened over the last 50 years (Weissman et al., 1996). While the feminist movement has yet to yield true gender equality, it is clear that the status and opportunities that adolescent girls may expect at present far outstrip those available 50 years ago. Gender intensification theories would logically predict that a rise in the financial and social power of women would be accompanied by a commensurate fall in depression onset at that developmental stage where the assumption of adult female roles is hypothesized to be paramount to adolescent girls’ sense of self-worth and efficacy. This matter, however, may not be as straightforward as it first appears. Thus, if the socialization of female roles in terms of reduced instrumentality and increased sensitivity to social relationships has not changed fundamentally over time, then the increased opportunity for women and girls may actually increase the gap between their behavioral repertoire and the demands placed on them. Nevertheless, the lack of change in the gender gap in rates of depression over a period of dramatic historical change in the type and range of roles socialized in young women does seem puzzling if gender socialization is the key process driving this phenomenon.

Another contradiction within these theories is their heavy emphasis on the timing of gender intensification as a corollary to the timing of the emergence of the gender gap in affective disorders. As noted earlier, social learning and gender typing have been shown to begin in the first few years of life. If identification with a feminine stereotype is as strongly linked to depression as some of these theorists suggest, the finding that the rates of depression in prepubertal children are equal between genders is counterintuitive (Gelman et al., 2004). A more sophisticated version of the gender intensification hypothesis posits that what is socialized in girls during childhood are reductions in instrumentality and increased experiences of helplessness and dependency (see earlier; Sheeber et al., 2002). These socialization experiences then constitute a diathesis that interacts with the developmental demands of early adolescence to create greater risk for depression in females. However, even this very plausible view of the role of socialization pressures during adolescence needs to explain what it is about the developmental demands of early adolescence that is specifically associated with the emergence of depressive disorders in vulnerable females (as opposed to, say, anxiety where gender differences are seen much earlier in life; Lewinsohn et al., 1998).

Finally, although the degree of gender gap in depressive disorders is unevenly distributed, such that women in disadvantaged sectors of society (e.g., women of color, women living in poverty, single mothers and those with less than a high school education) are disproportionately affected (Everson et al., 2002), the fact that the gender gap is also reliably observed in more privileged social groups casts doubt on the validity of a theory which posits membership of a devalued social group as the primary causative influence on depression.

Gender roles and individual differences in personality and emotion

An alternative possibility posits that there are more fundamental gender differences in affective functioning that may place females at greater risk for affective disorders, and that these differences are reflected in gender divergences in personality and emotional functioning. Indeed, investigations into developmental changes in gender identity often use self-report measures of traits traditionally thought to be more characteristic of males than females or vice versa. In the Personal Attributes Questionnaire, for example, the “masculine” item endorsements include: independent, active, competitive, making decisions easily, self-confident, not giving up easily, and standing up well under pressure (Spence et al. 1974). The finding that high scores on these traits are protective against depression certainly seems to be robust (e.g., Hoffman et al., 2004), although given that a number of these items overlap with depressive symptoms (e.g., difficulty making decisions, low self-confidence, hopelessness, and poor resilience) these associations may be, in part, tautological.

Indeed, the assumption that these are inherently masculine traits may need reevaluation. When these traits are considered in their own right, decoupled from their description as “masculine,” the argument becomes circular, in claiming that, for example, low self-confidence leads to depression, which is in part indexed by low self-confidence (Barrett & White, 2002). In addition, several recent studies suggest that “feminine” traits do not predict depression at all, but “masculine” traits are predictive of lower levels of depression, specifically in adolescence (Priess et al., 2009; Li et al., 2006). For example, Priess and colleagues suggest that, based on these and other findings (e.g., Wichstrom, 1999) perhaps the framing of masculinity and femininity should be rethought to be more synonymous with female versus male traits, which show secular changes over time with the changing social environment/landscape.

One way to reframe the relationship between gender roles and depressive phenomena is by taking into account the correlation between sex role inventory scores and broader personality traits. For example, Francis and Wilcox (1998) found that high scores on the masculinity scale of the Bem Sex Role Inventory (Bem, 1974) were associated with low neuroticism and high extraversion, whereas high scores on femininity were associated with high neuroticism. Neuroticism is a temperament trait that may place individuals at higher risk for depression and anxiety, and is also reliably found to be higher in females (Fanous et al., 2002). O’Shea (2002, cited in Parker & Brotchie, 2004) found that, although adolescent girls did record higher neuroticism scores than boys, high neuroticism was a strong predictor of first onset depression, regardless of gender. In fact, adjustment for neuroticism greatly attenuated the gender differential in first onset depression in this sample. This suggests that it may be neuroticism, rather than gender role per se, that confers risk for depression, although this still begs the questions as to why female gender and typical female gender role descriptions are associated with higher levels of neuroticism.

A second way to conceptualize traits that place one at risk for depression is to examine the relationships between negative emotionality (related to neuroticism), approach (related to positive emotionality) and effortful control. As early as childhood, effortful control is more characteristic of females, whilst approach is more characteristic of males. Negative emotionality tends to be higher in females, especially after the onset of adolescence (Martel, 2013). Of these broad characteristics, there is evidence that negative emotionality is a general marker of risk for psychopathology and that low approach is related to internalizing disorders such as depression. These sex differences may contribute to the larger subset of females at risk for depression (Martel, 2013). The author emphasizes, however, that these are broad group differences and that there are females and males who exhibit the opposite genders’ average pattern of traits and are thus at risk for a different psychopathology presentation than the larger subset of their gender (Martel, 2013).

With respect to measures of emotion, there is some evidence for gender differences across a range of emotional processes (i.e., emotion perception, reactivity, regulation and experience), although negative findings are also reported (Barrett et al., 1998). Most of these studies have used self-report measures, which are open to language, retrospective and stereotype biases, as noted earlier (Fischer, 2000; Fugate et al., 2009). A recent review by Whittle and colleagues (2011) found that these sex differences in emotionality are not limited to self-report methodologies, but are also observed in brain measures derived from neuroimaging studies of emotional processes. Specifically, they found evidence that women recruit different neurocircuitry to men during perceptual emotion processing, which may, in some cases, lead to more accurate or faster processing, although in others may result in overreactivity. Also, while it appears that women show greater neural activation during reactivity to a range of negative emotional stimuli (particularly involving the amygdala), men show greater activation during reactivity to negative emotional stimuli that signal cues of dominance. A growing number of imaging studies also suggest that women and men use different strategies to downregulate negative emotions, and that these strategies might be mediated by different neural circuitry. Further, some research suggests that men may engage in more efficient automatic or unconscious emotion regulation when exposed to emotional stimuli, which may result from greater integration of cognitive and emotional neural circuits. Finally, sex differences in neural activity associated with reactivity to positive emotional stimuli appears to depend on the type of positive emotion and the stimulus modality. Overall, this review suggests that sex differences in emotional processes that are likely to confer greater risk for affective disorders are also evident in methodologies that directly measure neural processes that are outside of volitional control (hence ruling out a role for reporting and other self-report biases).

Pubertal development and the diathesis for affective disorder

The relevance of pubertal changes to the developmental examination of the gender gap lies primarily in the potential role of hormonal changes at puberty as catalytic agents for the development of depression in those women placed at risk by temperamental predisposition. Angold and colleagues (1999) reported that the transition to Tanner stage III (an index of body shape change and pubertal stage) of puberty predicted the increase in rates of DSM-IV unipolar depression in girls, exerting a much-larger effect than chronological age. This supports the view that changes in rates of depression at adolescence are specifically related to the physical and physiological changes of puberty, rather than to broad psychosocial factors common to girls at a particular stage of adolescence. However, it does not rule out the possibility that societal pressures are the primary precipitators of depression, but are prompted not by age but by the visible manifestations of puberty. This problem was later clarified by the addition of hormonal variables into the model (Angold et al., 1999), which eliminated the effect of morphological status, strongly implicating the effects of estrogen in the development of depression in adolescent girls.

Importantly, there are considerable individual differences in the timing and rapidity of pubertal development (Ellis, 2004; Mendle et al., 2010), with the timing of puberty (i.e., the timing of changes in pubertal status relative to peers) being found to be the most consistent pubertal predictor of vulnerability to adolescent onset depression and other poor outcomes during adolescence (Mendle et al., 2010). Specifically, early onset of puberty has consistently been associated with depressive symptoms in girls (Graber et al., 1997; Kaltiala-Heino et al., 2003). The literature for boys is more mixed, with different studies finding that earlier or later onset of puberty (and some finding that both early and late onset) is associated with depressive symptoms and disorders (Graber, 2009).

At puberty, hormonal levels begin to fluctuate cyclically over a broader spectrum in girls than in boys. Estrogen in particular is recognized as playing an important role in mediating female sensitivity to stress. Estrogen apparently acts as an anxiolytic, and thus the cyclical withdrawal of estrogen that occurs shortly prior to menstruation may be analogous to the physiologic effects of anxiolytic withdrawal, creating a greater sensitivity in menstruating women to the anxiogenic and depressogenic effects of negative life events. At the onset of puberty for men, on the other hand, the increase of testosterone has been suggested to have a protective effect against depression and anxiety, although it tends to increase aggressive and risk-taking behaviors (Seeman, 1997).

It is important to note that the cyclical release and withdrawal of estrogen alone is not suggested to be the causative factor for the sudden increase in the incidence of depressive episodes in adolescent girls. Rather, this cycle is thought to result in a biological “kindling” that increases the stress reactivity of at-risk individuals to negative life events. In this respect, negative perceptions of gender role expectations or a socialized limitation in the behavioral repertoire of women may indeed come into play as proximal risk factors, but are regarded as precipitants of the development of mood disorder in girls who are, by virtue of temperament and biology, already at risk. In other words, the interaction of temperament and hormonal change may create a vulnerability diathesis upon which social factors may act as a catalyst to generate gender difference in rates of depression.

In support of the putative interaction between pubertal estrogen changes and negative life events, Brooks-Gunn and Warren (1989) found that, while the increase in pubertal estrogen levels accounted for 4% of variance in increased negative affect reported by 100 adolescent white girls, the joint contributions of estrogen rise and negative life events accounted for only 17% of the variance. It would appear that the onset of puberty sensitizes young women to the stress of negative life events, and possibly desensitizes young men, via the protective influence of testosterone. Prepubertal girls and boys tend to show similar correlations between number of life stressors and depressive affect. However, after the onset of puberty, the relationship between stress and negative affect strengthens for young women, and declines almost to elimination in young men (Angold et al., 1999).

Pubertal brain development, emotional processes and risk for mood disorders

The body of research reviewed earlier suggests that both the stage and timing of puberty affect emotional and behavioral dysregulation, and consequently vulnerability to mental health problems. However, the mechanisms underpinning these effects are still poorly understood (Ge & Natsuaki, 2009). Hormonal changes of puberty may be associated with changes in dysregulated affect and (perhaps) vulnerability to affective psychopathology, but evidence of the direct hormonal effects is mixed (Graber, 2009). Rather, it is the action of these hormones on the development of social, cognitive and affective systems in the brain (and the resultant psychological changes) that may be most likely to explain puberty-related trends in the emergence of symptoms. Recent reviews have made it clear that we still lack a full understanding of the relationship between pubertal processes and brain development (Blakemore et al., 2010).

The principle focus of studies of adolescent brain development has been on age-related changes (e.g., Giedd et al., 1999), although based on findings such as peaks in gray matter density that seem to coincide with timing of puberty onset (Blakemore et al., 2010) some researchers have suggested that brain developmental patterns align better with pubertal development rather than age per se (Peper et al., 2011). Brain regions associated with reward processing (i.e., mesolimbic circuitry mediating reward-related behavior) are densely innervated by gonadal steroid receptors. Several pubertal hormones are considered neurosteroids, as they are centrally converted or synthesized de novo by the brain. These and other neurosteroids are thought to play a role in regulating neurotransmitter systems associated with affective and social responsiveness, including dopamine, serotonin, endogenous opioids, oxytocin and vasopressin (Epperson et al., 1999). Furthermore, gonadal steroids have their own direct effects on affective processes (McEwen, 2001). Thus, although the main function of pubertal changes may be to increase sensitivity to sexual stimuli and increase sexual motivation (Sisk & Foster, 2004), there is also evidence that increases in affective reactivity reflect a more general reorientation to social and emotional aspects of the environment, and this may in turn result in heightened vulnerability to emotional and behavioral disorders (Nelson, et al., 2005).

Importantly, although there is informative animal work regarding the links between puberty and brain structure and function, as noted earlier relatively little is known about this relationship in humans (Blakemore et al., 2010). Recently, a small number of adolescent MRI studies have investigated, in more detail, the relationships between brain structure and function, gender and hormones that change at puberty. For example, a structural MRI study by Peper and colleagues (2009a) reported an association between testosterone levels and global gray matter density in males (but not in females): females showed a negative association between estradiol levels and both global and regional gray matter density. Neufang and colleagues (2009) found a positive relationship between pubertal stage and gray matter volume in the amygdala, and a negative relationship between these measures and hippocampal volume, regardless of gender. In addition, females showed a positive relationship between estrogen levels and limbic gray matter, whilst males showed a negative relationship between testosterone and parietal cortex gray matter volume. Finally, Goddings et al. (2014) recently reported that pubertal development is significantly related to structural volume in a range of brain regions (i.e., amygdala, hippocampus and corpus striatum including the nucleus accumbens, caudate, putamen and globus pallidus) in both sexes.

Trajectories of white matter development also differ as a function of pubertal hormones (Peper et al., 2009 a,b; Perrin et al., 2008, 2009). For example, Herting and colleagues (2012) found pubertal development to be associated with patterns of white matter structure in the insula, and, in boys, testosterone predicted white matter integrity in sexually dimorphic regions as well as whole-brain white matter; in girls, estradiol showed a negative relationship with white matter. These findings all suggest a significant role for puberty, distinct to that of chronological age, in sex differences in structural brain development during early adolescence. The implications of these puberty-related sex differences in brain development for behavior and mental health are yet to be determined.

Forbes and colleagues (2010) found that individuals with more advanced pubertal maturation exhibited less striatal and more medial prefrontal reactivity to reward; further, testosterone was positively correlated with striatal reactivity in boys during reward anticipation and negatively correlated with striatal reactivity in girls and boys during reward outcome. Goddings and colleagues (2012) examined both social and basic emotions in early adolescent girls. They found that increased hormone levels (independent of age) were associated with higher left anterior temporal cortex (ATC) activity during social emotion processing. This latter finding is particularly important as it suggests that the effects of pubertal development may be somewhat specific to social cognitive brain function, which may be relevant given the strong link between social processes and vulnerability to depression.

Few published studies have directly investigated whether the association between mental health symptoms and pubertal processes might be mediated by neurobiological variables. Whittle and colleagues (2012) examined the relationship between pubertal timing, brain structure and depressive symptoms during early adolescence. This study revealed that a larger volume of the pituitary gland, a key component of the hypothalamic-pituitary-gonadal (HPG) and hypothalamic-pituitary-adrenal (HPA) axes, mediated the relationship between early pubertal timing and depressive symptoms in female and male adolescents. These findings are consistent with neurobiological mechanisms being responsible for the link between exposure to pubertal hormones and depressive symptoms in adolescents.

In conclusion, there are good reasons to suspect that pubertal hormonal changes affect social, cognitive and affective systems in the brain, and that the resultant psychological changes might be critical mediators of the relationship between pubertal development and the emergence of symptoms of mental disorder during early adolescence. However, the extant human literature on the relationship between pubertal processes and brain development remains preliminary and far more work is needed to describe the biological and psychosocial mechanisms that underlie these developmental effects fully.

Pubertal change and interpersonal stressors

An alternative account of the relationship between pubertal development and affective disorders is suggested by examining whether the relation between pubertal hormone changes and depression in women is mediated by negative life events. In this respect, it is notable that the majority of stressors preceding the onset of adolescent depression are interpersonal (Cyranowski et al., 2000). Consistent with this observation, it has been proposed that the relationship between the quality of family relationships and depressive symptoms may be stronger for females than males (Kavanagh & Hops, 1994). During adolescence, females tend to gain independence more slowly from their families (Huston & Alvarez, 1990) and have relationships that are both more disclosing and more conflictual with their parents than do males (Noller, 1994; Montemayor, 1983). Many older studies using self-report measures found a stronger negative correlation between cohesive and supportive family relationships and depressive symptoms amongst girls than boys (Avison & McAlpine, 1992; Rubin et al., 1992; Slavin & Rainer, 1990; Windle, 1992), although a study by Sheeber and colleagues (1997), which included observational as well as self-report measures of family functioning, reported that the relationship between family functioning and depression was equivalent for females and males. This suggests that the stronger link between family processes and depression amongst females may be primarily determined by the way depressed states affect female perceptions of the family environment rather than objectively observable features.

The correspondence between the findings of a specific association between interpersonal stress and risk for depression on the one hand, and the evidence of increased affiliative proclivities amongst women (both as compared to men and after menarche) on the other, led Cyranowski and colleagues (2000) to examine potential biological substrates for female affiliative behavior. Nonhuman mammal research has strongly implicated the hypothalamic neuropeptide oxytocin in affiliative and caregiving behaviors (Depue & Lenzenweger, 2001; see also Chapter 8). Oxytocin transmission is thought to be regulated by estrogen and progesterone levels, giving rise to the idea of a hormonally driven pubertal increase in affiliative proclivity in females (Cyranowski et al., 2000). The review by Gordon and colleagues (2011) suggests that reproductive hormones, including testosterone and estradiol, potentiate the release of oxytocin. This potential connection between pubertal development and an increase in biologically controlled sensitivity to social stressors allows for a synthesis of psychosocial, biological and stress-response precipitants to adolescent onset depression, with the important caveat that the role of oxytocin in human female affiliative behavior is yet to be fully understood. Oxytocin plays a role in many complicated reproductive behaviors, including the formation of social (not just sexual) bonds (Lee et al., 2009). Although the formation of affiliative networks is in itself a protective factor against depression, the increased desire and need for these relationships, as well as the burden of increased care-taking within relationships, may mean that affiliative failures form potent stressors for girls who are temperamentally at risk, as has been emphasized by recent evolutionary models of depression.