For information about subscribing or purchasing offprints of the published version, with commentaries and author's response, write to: journals_subscriptions@cup.org (North America) or journals_marketing@cup.cam.ac.uk (All other countries).
Activational effects, androgens, corpus callosum, estrogen, female default hypothesis, feminization, masculinization, organizational effects, sensitive periods, sexual differentiation, testosterone
Historically, studies of the role of endogenous hormones in developmental differentiation of the sexes have suggested that mammalian sexual differentiation is primarily mediated by testicular androgens, and that exposure to androgens in early life leads to a male brain as defined by neuroanatomy and behavior. The female brain has been assumed to develop via a hormonal default mechanism, in the absence of androgen or other hormones. The role of ovarian hormones in female sexual differentiation may be complementary to androgen-mediated masculinization because the feminizing effects of ovarian steroids are only found in the absence of perinatal androgen. Ovarian hormones have significant effects on the development of a sexually dimorphic cortical structure, the corpus callosum, which is larger in male than female rats. In the females, removal of the ovaries as late as Day 16 increases the cross-sectional area of the adult corpus callosum. Treatment with low-dose estradiol starting on Day 25 inhibits this effect. Female callosa are also enlarged by a combination of daily postnatal handling and exogenous testosterone administered prior to Day 8. The effects of androgen treatment are expressed early in development, with males and testosterone-treated females having larger callosa than control females as early as Day 30. The effects of ovariectomy do not appear until after Day 55. These findings are consistent with other evidence of a later sensitive period for ovarian feminization than androgenic masculinization.
Reviews on the role of hormones in mammalian sexual differentiation traditionally focus on the effects of neonatal exposure to testicular androgens (e.g., testosterone, the predominant androgen) in males. This emphasis derives from overwhelming data consistently showing demasculinizing effects of postnatal castration in males, and masculinizing effects of perinatal androgens on females. These effects are seen for neuroanatomy, neurochemistry, and behavior, and include reports of cognitive differences among human populations with abnormal androgenic exposure (Collaer & Hines 1995; Masica, Money, Ehrhardt & Lewis 1969; Resnick & Berenbaum 1982) as well as alterations in hypothalamic anatomy following perinatal androgenic manipulations in rats (Gorski 1984).
The role of ovarian hormones has not been as extensively or rigorously investigated. Some researchers, for example, reported that ovarian manipulations had no effect on receptive sexual behavior in rats (Lisk & Suydam, 1967; Whalen & Edwards 1967), but later studies which included evaluation of proceptive behavior did find significant effects (Gerall, Dunlap & Hendricks 1973). Despite the prevailing assumption that the ovaries are not a critical factor in sexual differentiation, a number of reports showing behavioral and neuroanatomical consequences of ovariectomy, as well as low-dose estrogen exposure following removal of the testes, have accumulated over the past 20 years. Based on this accumulating evidence, a handful of researchers have suggested over the past two decades that estrogen plays an active role in differentiation of the female brain (e.g., Dohler, Hancke, Srivastava, Hofmann, Shryne & Gorski 1984b; Dohler 1991; Gerall et al. 1973; Stewart & Cygan 1980; Toran-Allerand 1976, 1992). Yet these hypotheses and findings have not been assimilated into the widely accepted model of androgen-mediated sexual differentiation.
In this target article we review evidence that exposure to ovarian hormones (primarily estrogen) during development is necessary for differentiation of the mammalian female brain. Initially, we discuss mechanisms of hormone action and relevant terminology (Section 2), followed by a discussion of the traditional model of androgen-mediated differentiation, and an abbreviated overview of relevant data (Section 3). We then present a more detailed review of evidence regarding the developmental effects of ovarian hormones and ovarian manipulations (Section 4). Finally, we present the results from a series of studies on the role of perinatal androgens and estrogens in neuroanatomical sexual differentiation of the rat corpus callosum (Section 5).
Androgens comprise a category of chemically related hormones produced primarily in the testes in the male and in the adrenal cortex of both sexes. The ovaries also produce some small amounts. Testicular androgens are the primary catalyst for masculine sexual differentiation in mammals (see Breedlove 1992 or Toran-Allerand 1986 for review). Consequently the term "androgen" is used in this paper to refer to the class of hormones secreted endogenously by the testes, or to exogenous androgen manipulations intended to assess hormonal mechanisms in normal males.
There are a number of different androgens that exert masculinizing effects in different regions and at different times in development, and that may be metabolized from one form to another before acting at the cellular level. For example, testosterone may be converted by the enzyme aromatase into the estrogen, estradiol. Thus in the presence of aromatase, testosterone can ultimately bind to estrogen as well as testosterone receptors. Testosterone may also be converted into dihydrotestosterone, a non-aromatizable form of testosterone that acts at testosterone receptors only. In contrast, the primary steroids produced by the ovaries (estrogen and progesterone) generally act directly at estrogen and progesterone receptors.
The primary mechanism by which steroids appear to influence neuroanatomy, neurophysiology, and behavior is through binding to intraneuronal nuclear receptors in target brain areas and altering neuronal genomic expression (see Toran-Allerand 1986 or McCarthy 1994 for review). Steroid effects are largely modulated by receptor topography, density of receptor populations, and receptor affinity for steroid binding, all of which may differ at any given point in development and as a function of sex. The effects of steroid binding may be expressed as alterations in regional cell growth, proliferation, or death, which may then influence cell number, size, or packing density. Early migrational patterns, dendritic growth, and neuronal myelination may also be altered. Evidence suggests that sexual differentiation is the result of interacting steroid effects on multiple neural parameters. The result of these processes is a complex and incompletely understood pattern of sex differences in neural circuitry and function (for reviews see Breedlove 1992; McCarthy 1994; Tobet & Fox 1992; Toran-Allerand 1986).
Mechanisms of hormone action have traditionally been divided into effects that occur early in development and are permanent (i.e., organizational) versus those that occur later in development and are transitory because they depend upon the presence of circulating hormones (i.e., activational). Within this framework, sex differences in neuroanatomy were assumed to reflect the permanent organizing effects of steroids. Some behavioral effects were interpreted as organizational (e.g., reduced rough and tumble play following early testosterone removal in the male; Meaney 1988; Ward & Stehm 1991), whereas other behaviors, particularly those that could be mimicked by experimental manipulation of circulating steroids (e.g., the "priming" of female rodents for sexual receptivity via exogenous estrogen and progesterone), were considered activational.
Yet accumulated evidence has muddled the organizational/ activational dichotomy. Specifically, the temporal distinction that categorized hormonal effects in development as organizational and hormonal effects in adulthood as activational apparently is not valid. Data which counter the traditional definitions include estrogenic activation of lordosis behavior in female rat pups as young as 6 days (Williams 1986), and changes in hypothalamic anatomy following post-pubertal hormone manipulations (Bloch & Gorski 1988). Some researchers have suggested that the primary organizational/activational distinction now depends upon whether induced changes represent permanent or transient effects, whenever in life they occur (e.g., see Arnold & Breedlove 1985; Stewart 1988; Williams 1986 for discussion). Classification of hormonal effects is also complicated by increasing evidence of ongoing physiological plasticity in the adult brain. Recent evidence shows, for example, that for some neuroanatomical and neurochemical systems the adult female brain is "permanently transient" (e.g., Becker & Cha 1989; Frankfurt, Gould, Woolley & McEwen 1990; Woolley, Gould, Frankfurt & McEwen 1990). In other words, fluctuations in specific neurophysiologic measures occur in response to female hormonal cyclicity. Although these effects could reasonably be excluded from a review on sexual differentiation because they are transient, it nevertheless seems that they constitute part of what makes the female brain distinct (or differentiated) from the male brain. As such, we include these findings in our discussion.
Researchers studying sexual differentiation of complex systems (e.g., behavior) have noted a distinction between the suppression of male or female attributes (demasculinization and defeminization, respectively) versus the enhancement of male and female attributes (masculinization and feminization, respectively; discussed in Toran-Allerand 1986). Certain hormonal manipulations, for example, can suppress female-typical behavior without inducing male behavior, an outcome that would be described as defeminizing but not masculinizing (e.g., see Yahr & Greene, 1992). The application of these terms to sexual differentiation of neuroanatomical structure, where male-female differences tend to be measured on a single axis (e.g., larger/smaller, more cells/fewer cells, etc.), can sometimes complicate use of these definitions. As an example, the sexually dimorphic nucleus of the preoptic area (SDN-POA) is larger in male rats, and increasing the size of the structure in females via early androgen treatment is interpreted as masculinizing. Such enlargement might also be interpreted as defeminizing, because it represents a deviation from the normal female pattern. However, blocking the early action of estrogen in female rats (by an estrogen receptor blocker or estrogen mRNA antisense) apparently decreases the size of this structure (Dohler, Srivastava, Shryne, Jarzab, Sipos & Gorski 1984c; McCarthy, Schlenker & Pfaff 1993). Because these manipulations appear to interfere with the normal process of SDN-POA development in females, their effects are interpreted as defeminizing. Thus for this structure, it appears that increasing size in females reflects masculinization, while decreasing size reflects defeminization.
Testosterone can be intraneuronally converted (aromatized) to estradiol in a variety of species (Beyer, Green, Barker, Huskisson & Hutchison 1994; Hutchison & Beyer 1994a; Hutchison, Beyer, Green & Wozniak 1994b; Hutchison, Beyer, Hutchison & Wozniak 1995; Roselli & Resko 1993; see Toran-Allerand 1986 for review). This locally biosynthesized estrogen may in turn act upon estrogen receptors within neuronal nuclei. It appears that this mechanism plays a critical role in the masculinization process for many mammalian species; androgens such as testosterone are secreted by the testes, but are then converted to estrogen within individual neurons before exerting developmental effects.
Aromatase enzyme has been found subcortically in a variety of species and is known to be involved in sexual differentiation of subcortical structures (see Breedlove 1992 or Toran-Allerand 1986 for review). The role of aromatase at the cortical level has, historically, been less clear. Although aromatase enzyme was localized in fetal and newborn monkey cortex (MacLusky, Naftolin & Goldman-Rakic 1986; Sholl, Goy & Kim 1989), researchers reported difficulty demonstrating estrogen biosynthesis in perinatal rat cortex (MacLusky, Phillip, Hurlburt & Naftolin 1985). More recently, aromatase has been localized in perinatal mouse and rat cingulate cortex (MacLusky, Clark, Naftolin & Goldman-Rakic 1987; MacLusky, Walters, Clark & Toran-Allerand 1994), parietal cortex of fetal guinea pigs (Connolly, Roselli & Resko 1994), and cerebral cortex of young opossums (Fadem, Walters & MacLusky 1993). These findings, coupled with evidence of transient, high-density populations of estrogen receptors in neonatal rat cortex (MacLusky, Chaptal & McEwen 1979a; MacLusky, Lieberburg & McEwen 1979b) and mouse cortex (Shugrue, Stumpf, MacLusky, Zielinski & Hochberg 1990), point to a significant role for biosynthesis of estrogen in sexual differentiation of the cerebral cortex in a variety of species (see MacLusky et al. 1987). These conclusions are in agreement (1) with behavioral research showing that systemic administration of an aromatase-inhibitor to intact male rat pups produced female-like patterns of maze learning, whereas implantation of estradiol into the cortex of neonatally castrated males reinstated male-like maze learning patterns in adulthood (Williams & Meck 1991); and (2) with evidence that biosynthesized estrogen plays a role in sexual differentiation of catecholamine systems in rat frontal cortex (Stewart & Rajabi 1994).
Paradoxically, although female rat pups are exposed to high circulating levels of maternal estrogen in the perinatal period, they are not masculinized. Rather, they are protected from estrogen-based masculinization via a blood-born protein called alpha-fetoprotein (AFP; Raynaud, Mercier-Bodard & Balieu 1971). AFP, which is present in the early perinatal period, binds to circulating estrogen and apparently prevents it from entering the neuron as freely as unbound estrogen (although small amounts of estrogen may pass into the neuron while bound to AFP; see Toran-Allerand 1986). From an evolutionary perspective, AFP appears to protect the female brain from perinatal estrogenic masculinization. Evidence suggests that maximal levels of AFP are seen in rat brain on gestational day 18, and decline to low levels by postnatal day 7, when AFP synthesis is apparently "switched off" (Ali, Kaul & Sahib 1981; Ali & Sahib 1983). This is the same perinatal time frame during which fetal testicular androgens exert masculinizing effects in rats (see discussion below). Unlike the testes, evidence suggests that the ovaries become active in rats and mice around postnatal day 7 (Sokka & Huhtaniemi 1995; Weniger, Zeis & Chouraqui 1993; Mannan & O'Shaughnessy 1991). Thus an active role for the ovaries in development is not at odds with the early presence of AFP.
The earliest reports of an androgenic role in differentiation showed that manipulations of neonatal androgens affected adult sexual behavior. Female guinea pigs exposed to testosterone by various regimes during the prenatal period increased male-typical sexual behavior (mounting). These subjects also decreased female-typical behavior (lordosis) when, as adults, they were gonadectomized, primed with estrogen and progesterone, and tested for sexual receptivity (Phoenix, Goy, Gerall & Young 1959). Similarly, male rats castrated at birth reduced male-typical sexual behaviors and increased feminine behaviors in adulthood (Beach 1974; Whalen & Edwards 1967; Young 1961). These same behavioral patterns were seen in adult male rodents exposed prenatally to stress or alcohol, which disrupts the prenatal testosterone surge in male fetuses (McGivern, Clancy, Hill & Noble 1984; McGivern, Raum, Salido & Redei 1988; Ward 1972, 1983). These effects appear to be mediated by aromatization of testosterone to estrogen, since sexual behavior can be masculinized in females and reinstated in neonatally castrated males with early administration of a synthetic estrogen (e.g., diethylstilbestrol), or high doses of estradiol (Doughty, Booth, McDonald & Parrot 1975; Hendricks & Gerall 1970).
Estrogen has also been shown to act asymmetrically in the hypothalamus to modify reproductive behavior of the female rat. Estradiol pellets were placed in the left or right ventromedial nucleus during the first two days of life. In adulthood, subjects were ovariectomized and primed with estradiol benzoate and progesterone. Subjects with left-sided implants showed reduced lordosis as compared to right-sided implants and cholesterol controls (Nordeen & Yahr, 1982; Yahr & Greene, 1992). Nordeen and Yahr also found masculinizing effects of estradiol, with local implantation in the right preoptic area leading to increased mounting behavior in adulthood.
Male and female rats differ on a large number of nonreproductive behaviors, including aggressiveness and rough and tumble play, and many of these behaviors are influenced by neonatal exposure to testosterone (for reviews, see Beatty 1979, 1984, 1992). With respect to cognitive behavior, female rats normally learn an active avoidance response more quickly than males, whereas males tend to outperform females on passive avoidance. Moreover, female rats treated neonatally with testosterone and primed with testosterone prior to testing appear to be indistinguishable from males in avoidance learning behavior (see van Haaren, van Hest & Heinsbroek 1990 for review). However, circulating testosterone may not be critical to this sexual dimorphism since others have reported that early neonatal exposure to androgen alone is sufficient to induce a male-like pattern of active avoidance learning in female rats (Denti & Negroni 1975). Avoidance behavior in male rats can also be feminized. For example, prenatal exposure to the androgen receptor blocker cyproterone acetate, followed by postnatal castration, resulted in males with female-like avoidance behavior in adulthood (Scouten, Grotelueschen & Beatty 1975).
With respect to spatial learning, male rats typically do better than females. In general, neonatal castration of males or exposure of females to androgens reverses this adult sexually dimorphic pattern (Dawson, Cheung & Lau 1975; Joseph, Hess & Birecree 1978; Stewart, Skavarenina & Pottier 1975). More recently, Roof (1993a) reported that male rats performed significantly better than females on both the radial arm and Morris water maze. Roof also found that neonatal treatment with testosterone improved spatial ability in female rats to male levels, and this effect was seen as early as 21 days of age. Roof and Havens (1992) reported that neonatal treatment with testosterone led to a male-like pattern of hippocampal anatomy (as measured by size and asymmetry of granule cell layers), and improved maze learning in female rats. Moreover, maze performance correlated significantly with the size of hippocampal granule cell layers.
In a related series of studies, Williams and colleagues manipulated specific components of the extra-maze testing environment and showed that male and female rats use different strategies and rely on different cues in maze learning (Williams & Meck 1987, 1991; Williams, Barnett & Meck 1990a). Williams also showed that estrogen plays a major role in influencing the development of spatial ability in rats by implanting estradiol into the hippocampus or cortex of neonatally castrated males. This reinstated male-like maze learning behavior (Williams, Cohen & Meck 1990b; Williams & Meck 1991). These findings support a developmental role for estrogen biosynthesis (from testosterone) in sexual differentiation of the rat cerebral cortex.
Sex differences in cerebral organization are also seen in non-human primates. Clark and Goldman-Rakic (1989) reported that intact male monkeys made fewer errors than intact females in learning a visual object discrimination reversal task. Lesions to the orbital prefrontal cortex disrupted the ability of males, but not females, to perform the task. Furthermore, females given androgen in early perinatal life performed like normal males, and were similarly disrupted by the lesion.
This male advantage does not generalize across ages and visual learning tasks, since 3-month old male monkeys were slower to learn a set of visual discriminations than were age-matched females (Bachevalier, Hagger, & Bercu 1989). In this study, testosterone levels were obtained from males and estradiol levels from females, and these were correlated against the learning scores. Within the 3-month male monkey group, the rank-order correlation was .95; the higher the testosterone, the slower the learning. There was no significant correlation in any other group. The authors' interpretation was that high testosterone levels temporarily slowed the maturation of the neural systems underlying visual discrimination, because by 6-months of age no sex differences or hormone-behavior correlations were found. A later study showed that ablation of inferior temporal cortex depressed visual discrimination scores in 3-month old female monkeys, but did not affect age-matched males (Bachevalier, Brickson, Hager & Mishkin 1990), an effect that is apparently mediated by testosterone exposure (Hagger & Bachevalier 1991). These results agree with the prior interpretation that testosterone delayed maturation of neural systems underlying the visual discrimination task (see also Bachevalier & Hagger 1991).
Finally, developmental androgen effects have been reported for human cognition, as well as sexual orientation and aggression (see Collaer & Hines 1995; Gouchie & Kimura 1991; Swerdloff, Wang, Hines & Gorski 1992 for reviews). Correlations between androgen level and behavior have been reported with respect to language disabilities (Kirkpatrick, Campbell, Wharry & Robinson 1993), and musical ability (Hassler 1991), and prenatal androgen levels have been correlated with mental rotation skill in girls (Grimshaw, Bryden & Finegan 1995; Grimshaw, Sitarenios & Finegan 1995). Salivary testosterone levels have been positively correlated with spatial ability in women and negatively correlated with spatial ability in men (Gouchie & Kimura 1991).
Neuroanatomical sex differences are present in the rat hypothalamus, most notably in the sexually dimorphic nucleus of the preoptic area (SDN-POA) which is approximately 8 times larger in adult males than females (Gorski, Gordon, Shryne & Southam 1978; Gorski 1984). Castration of males on Day 1 reduces the size of this nucleus in adulthood, whereas the postnatal administration of testosterone to females enlarges the nucleus (Dohler, Coquelin, Davis, Hines, Shryne & Gorski 1982, 1984a; Gorski et al. 1978; Gorski 1984). The SDN-POA can also be enlarged in female rats through the administration of a synthetic estrogen (diethylstilbestrol) which does not bind to AFP (Dohler et al. 1984a; Gorski 1984), indicating that masculinization of this structure is dependent upon the intracellular conversion of testosterone to estrogen. Estrogen may mediate this sexual dimorphism by preventing a developmental loss of neurons within the medial preoptic nucleus (Dodson & Gorski, 1993). Hormonal mediation of neuronal loss also affects sexual dimorphism in the size of the spinal bulbocavernosus nucleus in rats, although differentiation of this nucleus apparently depends on the direct action of androgen, rather than estrogen (see Breedlove 1992 for review). Finally, SDN-POA volume is reduced in male rats exposed prenatally to ethanol, which depresses fetal androgen production (Ahmed, Shryne, Gorski, Branch & Taylor 1991).
Holman and Hutchison (1991) showed that neonatal ovariectomy of female gerbils followed by silastic testosterone implant enlarged the size of the sexually dimorphic preoptic area of the pars compacta (SDApc), as well as the sexually dimorphic suprachiasmatic nucleus, to the size of males. These researchers also found that the volume of the left (but not right) SDApc correlated with ultrasonic courtship vocalizations, a structural-behavioral asymmetry consistent with lateralized hormonal effects on behavior reported by Yahr and Greene (1992). Research has also shown that the volume of the posterodorsal region of the medial nucleus of the amygdala (MApd), and the volume of the encapsulated region of the bed nucleus of the stria terminalis (BNSTenc) are approximately twice as large in male as compared to female rats (Hines, Allen & Gorski 1992), although the hormonal mechanisms underlying these effects were not investigated.
Finally, sex differences are present in the number and pattern of synapses on dendritic spines in the preoptic area of the rat hypothalamus (Raisman & Field 1973), dendritic patterns in the preoptic area of hamsters (Greenough, Carter, Steerman & DeVoogd 1977), and dendritic anatomy in the preoptic area of juvenile macaques (Ayoub, Greenough & Juraska 1983). These differences have been attributed to the effects of neonatal androgen exposure in the male (e.g., Greenough et al. 1977), an assertion consistent with in vitro evidence that testosterone influences neuronal survival and neural outgrowth of cultured rat preoptic cells (Kawashima & Takagi 1994).
Neonatal testosterone also appears to be involved in the sexual differentiation of the cerebral cortex. Diamond (1984) and her colleagues (Diamond, Dowling & Johnson 1981) reported that certain regions of the cortex were significantly thicker in the right hemisphere than in the left in male rats, whereas females showed a non-significant trend toward asymmetry in the opposite direction (see also Kolb, Sutherland, Nonneman & Whishaw 1982; Stewart & Kolb 1988). This effect appears to be mediated at least partly by androgen exposure, since neonatally gonadectomized male rats fail to show the R-L pattern of cortical asymmetry seen in intact males (Diamond 1984; Stewart & Kolb 1988). The male cortical thickness pattern is also reversed by prenatal stress (Fleming, Anderson, Rhees, Kinghorn & Bakaitis 1986; Stewart & Kolb 1988), which depresses and shifts the gestational day 18 testosterone surge in fetal male rats (Ward & Weisz 1980). Finally, the masculinizing effect of androgens on cortical asymmetry in rats appears to be mediated by conversion to estrogen, because perinatal exposure to the aromatase blocker ATD (1,4,6 androstatriene-3,17-dione) reversed the adult cortical thickness pattern in males (Diamond 1991).
More recently, sex differences have been demonstrated in the dendritic branching patterns of prefrontal cortical cells in rats, and these patterns appear to be influenced by gonadal hormonal exposure during development (Kolb & Stewart 1991). In addition, the cortical thickness of the binocular subfield of occipital cortex (Oc1B) is significantly greater in male rats as compared to females (Seymoure & Juraska 1992). There were also sex differences in the apical branching of Oc1B and Oc1M (monocular subfield) neurons, wherein females have longer dendrites, longer terminal branches, and longer bifurcating branches. Reid and Juraska (1992) confirmed the sex difference in binocular cortical thickness, and further reported that this effect reflects higher numbers of neurons and glial cells in males, with no sex differences in soma size or neuronal density. More recently, Reid and Juraska (1995) found sex differences in synaptic junctions in this region, with the higher number of neurons in male cortex relating to higher numbers of synaptic junctions. To our knowledge, the hormonal mechanisms underlying these anatomical sex differences have not yet been determined.
With respect to the hippocampus, Juraska (1991) reported sex differences in hippocampal dendritic anatomy that vary in direction as a function of rearing environment, and are influenced by early androgen exposure. Roof (1993a, 1993b) also reported a sex difference in the granule cell layers of the hippocampus in rats, and found that this effect is modulated by early exposure to testosterone.
Finally, neonatal androgen affects the development of cortical neurotransmitter systems. Monoamine systems innervating anterior cortex in intact female rats develop earlier than in males or androgen treated females (Stewart, Kuhnemann & Rajabi 1991), and evidence suggests that sex differences in development of frontal cortical catecholamine systems may derive from prenatal biosynthesis of estrogen from testosterone (Stewart & Rajabi 1994).
In mammalian sexual differentiation, each critical stage appears to rely on the production of a substance by the male -- Testis Determination Factor, Mullerian Regression Factor, and testosterone. In contrast, the terms "hormonally neutral" and "default" have been used to describe differentiation of the female. Thus, Svare and Kinsley (1987) write "it is important to note that the ovaries and their secretory products do not have a role in the differentiation process. Instead, the critical determinant of internal and external sexual differentiation is the presence or absence of testosterone secreted by the testes" (p. 16). Lisk and Suydam (1967) likewise conclude that "feminization appears to be the neutral condition" (p. 182). Yet researchers have previously suggested that estrogen (presumably of ovarian origin) may play an active role in feminization of the brain (e.g., Dohler et al. 1984b; Dohler 1991; Hendricks 1992; Toran-Allerand 1976, 1992). Toran-Allerand, for example, performed a series of in vitro studies in 1976 and concluded that "these ... experiments suggest ... that no pattern of sexual differentiation need necessarily be intrinsic to nervous tissue but that male and female patterns may both require active induction by steroid" (p. 411). Evidence to support this view has accumulated over the ensuing 20 years, and the following sections will review this evidence. The main findings are summarized in Table 1.
We noted earlier that that AFP synthesis appears to "switch off" around postnatal day 7 in rodents, and AFP levels in both brain and plasma are low by postnatal day 7-10 (Ali et al.1981; Ali & Sahib 1983; Raynaud et al.1971; Raynaud 1973). It is at this time, or later, that estrogen of ovarian origin could freely enter neurons to exert organizational effects. Consistent with this view, Dohler and Wuttke (1975) measured serum levels of follicle stimulating hormone (FSH), lutenizing hormone (LH), and estrogen taken from pups of various ages, and postulated that the feedback regulation of estrogen on phasic gonadotropin release becomes functional between Days 9 and 21 in female rats. More recent serum and tissue analyses from newborn rats and mice have revealed increases in ovarian estrogen in response to hCG (human chorionic gonadotropin), LH, FSH, and cAMP around 1 week after parturition (Sokka & Huhtaniemi 1995; Weniger et al. 1993; Mannan & O'Shaughnessy 1991). Ovarian hypertrophy in response to unilateral gonadectomy has been observed as early as Day 10 in the rat (Gerall & Dunlap 1971), and estrogen synthesis has been observed in neonatal ovarian tissue cultured in vitro (Funkenstein, Nimrod & Lindner 1980; Levina, Gyevai & Horvath 1975). These findings support the view that ovarian estrogen could exert developmental effects in the neonatal period following the decline of AFP.
Evidence regarding the effects of neonatal ovariectomy (OVX) upon female-typical sexual behavior are mixed. While some researchers have reported no significant effect of removing the ovaries on sexual behavior in female rats (e.g., Lisk & Suydam 1967; Whalen & Edwards 1967), others (e.g., Blizard & Denef 1973) have found that the inhibitory effect of neonatal testosterone exposure on female sexual behavior (lordosis) is suppressed if the ovaries are present during development. Sodersten (1976) similarly reported that intact female rats exposed neonatally to testosterone showed a greater adult lordosis response (when primed with estrogen and progesterone) than females receiving neonatal OVX and testosterone treatment. Furthermore, Sodersten found that postpubertally ovariectomized (OVXd) female rats showed more lordosis after priming than neonatally OVXd females. He concluded that "... although the nature and mechanism of the action of these ovarian secretions remain to be determined we feel that the fact should be recognized that they do exert a modifying influence on psychosexual differentiation" (p. 419).
Other researchers have found that when proceptive or soliciting components of female sexual behavior --- hopping, darting and ear-wiggling --- are examined, ovarian effects become even more evident. Gerall and his colleagues (Gerall et al. 1973) reported higher proceptive behavior in estrogen-primed female rats that received OVX postpubertally as compared to those OVXd neonatally. Similarly, neonatally gonadectomized males that received prepubertal ovarian transplants or low-dose estrogen treatment were more proceptive in adulthood than males that were gonadectomized only. Indeed, the former males exhibited as much proceptive darting behavior as normal females. These findings suggest that lordosis and proceptive behavior are under different hormonal control, and that proceptive behavior may be particularly sensitive to ovarian effects. This conclusion is supported by Ward's (1983) finding that prenatally stressed male rats exhibited lower mounting and increased lordosis in adulthood, but did not exhibit any of the proceptive components of female sexual behavior.
4.3.1. Open-field behavior. Female rats are normally more active in the open field than males. OVX on Day 1 or 8 of life reduced open-field behavior in adult female rats to male levels (Denti & Negroni, 1975; Stewart & Cygan 1980). This agrees with Blizard and Denef (1973), who found that the presence of the ovaries during development suppressed the masculinizing effects of neonatal testosterone treatment on open field behavior in rats. Stewart and Cygan (1980) studied low- and high-dose replacement of estradiol and found that low doses increased (feminized) open-field activity whereas high doses suppressed (masculinized) activity. They concluded that "... estrogens given during the period prior to weaning can have a feminizing effect on adult open-field behavior, and that the sex difference normally observed in adult rats is dependent in part on the presence of the ovaries during a period after birth" (p. 20; see also Stewart, Vallentyne & Meaney 1979).
4.3.2. Plus maze behavior. Zimmerberg and Farley (1993) found that intact adult female rats spend significantly more time in the open arms of a plus maze than males. When females were either exposed neonatally to the estrogen receptor blocker tamoxifen or OVXd at puberty, they spent less time in the open arms as adults. Females that received both treatments spent the least time in the
open. In contrast, neonatal and pubertal manipulations of androgens in males (by administration of the androgen receptor-blocker flutamide and castration) had little effect on plus maze behavior. The authors conclude "... these experiments indicate that female gonadal hormones play an important role both organizationally and activationally in plus maze behavior" (p. 1119). The Zimmerberg and Farley findings were confirmed by Leret, Molina-Holgado, and Gonzalez (1994) who also reported that neonatally OVXd female rats behaved like males when tested in a plus maze paradigm.
4.3.3. Spatial behavior. Krasnoff and Weston (1976) found that sex differences in maze learning emerge around the time of puberty in rats, with females making more errors after puberty whereas male behavior remained essentially unchanged. Intact or neonatally castrated male rats, given low-doses of estrogen neonatally, had lower spatial learning scores in adulthood than untreated males (Dawson et al. 1975). In humans, sex differences in spatial ability (e.g., for mental rotation) have also been reported to appear at puberty (see Halpern 1992 for review), although the hormonal basis for this gender difference is not fully understood. However, fluctuations in spatial ability are seen across the menstrual cycle, and appear to correlate with changes in estrogen level (Hampson 1990; Hampson & Kimura 1988).
4.3.4. Avoidance learning. Sex differences in two-way shuttlebox learning also appear after puberty in rats (Bauer 1978) yet are influenced by ovarian hormones at various times in development. Denti and Negroni (1975) found that neonatal OVX decreased active avoidance performance of adult female rats to male-typical levels. In contrast, postpubertal OVX led to a significant increase in shuttlebox learning in female rats, an effect that was reversed with three days of estrogen replacement (Diaz-Veliz, Soto, Dussaubat & Mora 1989). Diaz-Veliz and colleagues also reported that avoidance learning varied across the estrus cycle, with high estrogen levels associated with increased errors. These latter findings likely reflect activational hormonal effects (given their transient nature), whereas the neonatal ovariectomy appears to have had a permanent (organizational) effect on avoidance learning. It is noteworthy that, in the above example, the activational effects of estrogen on avoidance behavior are opposite to developmental (organizational) estrogen effects.
The mechanisms underlying the emergence of cognitive sex differences at or after puberty are unclear. One possibility is that neural reorganization of some sort occurs at puberty (e.g., see Jernigan, Trauner, Hesselink & Tallal 1991). This would be consistent with evidence that age at puberty is related to spatial abilities in humans (Newcombe & Bandura 1983; Sanders & Soares 1986; Waber 1976; 1977).
4.3.5. Rotation and locomotor behavior. Camp, Robinson and Becker (1984) demonstrated sexual dimorphism of the nigrostriatal system in rats as measured by various tests of amphetamine-induced rotation (with females showing more asymmetry on some tasks and less on others). In addition, Becker and colleagues (Becker & Cha 1989; Castner & Becker 1990) reported that endogenous, or exogenously administered, pulsatile estrogen potentiated the dopaminergic and behavioral locomotor response to amphetamine (AMPH) in female, but not male, rats. This difference appears to underlie higher behavioral responsiviness to AMPH in females as compared to males. OVX depressed striatal dopaminergic release and turnover in females, whereas physiological concentrations of estrogen stimulated dopaminergic release. Castration or estrogen exposure had no similar effect on male striatal tissue. This sex difference appears to emerge at puberty through changes in the response of female striatal tissue to estrogen (Becker & Ramirez 1981). More recently, Forgie and Stewart (1994) reported that in the absence of early testosterone exposure, female rats OVXd on postnatal Days 1 or 25-26 showed less behavioral activation to AMPH than females OVXd in adulthood. This effect was seen for both estradiol-primed and non-primed animals, although activity was highest in primed subjects. These effects were not seen for females exposed neonatally to testosterone.
Dohler and colleagues (Dohler et al. 1984b, 1984c) found that neonatal administration of the estrogen antagonist, tamoxifen, to female rats decreased the size of the SDN-POA in adulthood relative to controls. They proposed that while high levels of estrogen (derived from the intracellular conversion of testosterone) are necessary for the masculinization of this structure, feminization may require some low level of estrogen. This thesis is supported by more recent evidence that neonatal treatment with estrogen mRNA antisense significantly reduces the size of the SDN-POA in intact adult female rats (McCarthy et al. 1993). In a related study, Bloch and Gorski (1987; 1988) reported that the hypothalamic anteroventral preoptic nucleus (AVPv) is significantly larger in female rats than in males, and that postpubertal castration of males followed by treatment with low doses of estrogen and progesterone significantly enlarges this structure. These males also had a smaller SDN-POA compared to control males. Similar effects were not observed in males that were gonadectomized only. More recent work has shown that sex differences in the AVPv emerge at puberty, and are the consequence of increases in AVPv size in females (Davis, Elihu, Shryne & Gorski 1993).
Frankfurt et al. (1990) reported that dendritic spine density of ventromedial hypothalamic neurons varies across the estrus cycle in rats. They found that adult OVX reduced dendritic density whereas estrogen or estrogen plus progesterone replacement increased density. Although such effects are properly described as activational, they demonstrate that circulating hormones can temporarily change neural structure. Furthermore these variations apparently occur in a periodic fashion, as a consequence of hormonal cyclicity, in adult female brains.
4.5.1. Cortical and hippocampal thickness. Ovariectomy on Day 1 was found to increase cortical thickness in 90-day old female rats as compared to sham operated littermates (Diamond, Johnson & Ehlert 1979). In this study, females were also OVXd at 90 or 300 days while littermate controls received sham surgery; in all cases cortical thickness measurements were obtained 90 days after surgery. No significant differences in cortical thickness were found for the 90- or 300-day OVX groups as compared to respective shams. The increase in cortical thickness for the Day-1 OVX group reflected at least in part an increase in neuronal soma size. These findings support the view that ovarian hormones act specifically during pre-pubertal development to affect the thickness of the cerebral cortex.
Pappas, Diamond, and Johnson (1979) sought to determine the relative contributions of estrogen and progesterone. In their first experiment they replicated the neonatal ovariectomy effect on cortical thickness measured at 90 days. In a second experiment, a group of female rats were OVXd on Day 1 while littermates received sham surgery. From 40 to 90 days of age, the OVXd females received daily injections of ethinylestradiol while shams received an equal volume of sesame oil. At 90 days the OVX-plus-estrogen group had significantly thinner cortices than controls, in contrast to the thicker cortices of untreated OVXd groups from prior studies. A third experiment followed the same procedure as experiment 2 except that OVXd females received daily injections of progesterone from Days 40-90 and shams received oil. In this case, the progesterone-treated OVX subjects had significantly thicker cortices than sham controls. These findings suggest that estrogen and progesterone exert different developmental effects on cortical thickness in the female brain.
In later studies Diamond and colleagues found the right cerebral cortex of the male rat to be significantly thicker than the left throughout life, whereas the cerebral cortex of females showed no significant asymmetry (but a trend to L>R; Diamond, Johnson, Young & Singh 1983). To investigate the role of ovarian hormones in this sexual dimorphism, females received OVX on Day 1 and female littermates received sham surgery. Cortical thickness measurements at 90 days showed that shams exhibited no significant asymmetry, whereas the OVXd females exhibited a male pattern of right-significantly-greater-than-left in the visual cortex (Diamond, Dowling & Johnson 1981). Stewart and Kolb (1988) later replicated the described sex difference, but did not find that ovariectomy reversed the cortical asymmetry pattern in females.
Finally, despite evidence of ovarian effects on cortical anatomy as reviewed above, Diamond, Murphy, Akiyama & Johnson (1982) failed to find any difference in hippocampal thickness between OVX and sham control female rats.
4.5.2. Dendritic spine density. Stewart and Kolb (1994) reported that OVX of intact adult female rats on Day 150 increased the dendritic arbor of pyramidal neurons of the parietal cortex as compared to controls. Similar effects were seen on dendritic arbor of pyramidal neurons of females treated neonatally with testosterone, although adult OVX of these subjects exerted no further effects. These findings may reflect a transient response to hormonal change, similar to changes observed in hypothalamic (Frankfurt et al. 1990) and hippocampal dendritic anatomy, following adult OVX of untreated females. Gould, Woolley, Frankfurt & McEwen (1990), for example, found that OVX in adulthood decreased dendritic spine density on CA1 pyramidal cells in the hippocampus of female rats, and that this effect was blocked by the concurrent administration of estrogen and progesterone. They suggested that variations in the density of these spines may accompany the estrus cycle, which was confirmed by Woolley, Gould, Frankfurt and McEwen (1990).
Evidence also supports developmental ovarian effects on cortical dendritic anatomy. Munoz-Cueto, Garcia-Segura & Ruiz-Marcos (1990) found that the development of dendritic spines in visual cortex occurs later in intact male as compared to intact female rats. After Day 20, dendritic spine numbers continue to increase for males, whereas females show a significant decrease. The female-typical loss of dendritic spines was prevented by OVX on Day 30, leading to a higher number of cortical dendritic spines among ovariectomized females as compared to intact females by Day 60. Munoz-Cueto et al. postulated that estrogen exerted inhibitory effects on cortical dendritic spine growth in both sexes, with aromatized testosterone delaying development in males during the early period, and ovarian estrogen promoting a loss of spines during the later time period (Days 20 to 60).
Given the long-standing observation that masculinization of many systems is dependent on the biosynthesis of estrogen, one might question how estrogen could exert masculinizing effects in males and concomitant feminizing effects in females. We suggest three inter-related mechanisms that could account for these dimorphic effects: (1) sex differences in estrogen levels (high for aromatized estrogen, relatively low for estrogen of ovarian origin); (2) sex differences in critical periods of estrogen action; and (3) variation in topographic distribution and density of target estrogen receptor populations as a function of sex (Brown, MacLusky, Shanabrough & Naftolin 1990; DonCarlos & Handa, 1994; Kuhnemann, Brown, Hochberg & MacLusky 1994) and postnatal age (MacLusky et al. 1979a and b; Miranda & Toran-Allerand, 1992; O'Keefe & Handa, 1990; Shugrue et al. 1990). Evidence for a temporal distinction in sensitive periods for testicular and ovarian effects is discussed below.
The concept of critical, or sensitive, periods in development has played a central role in theories of sexual differentiation. Thus, it has been generally agreed that testosterone exerts masculinizing effects on the CNS of male rats during the period between about gestational day 17 and postnatal days 8-10, depending on the system being studied (Rhees, Shryne & Gorski 1990a; Rhees, Shryne & Gorski 1990b; but see Bloch & Mills 1995). This perinatal period of sensitivity to the masculinizing effects of testosterone appears to be similar in mice (e.g., see Wagner & Clemens 1989). However, a different set of temporal parameters appears to apply to female brain development. The sensitive period for permanent structural and behavioral ovarian effects does not end by day 10 in rodents, as generally appears to be true in males, but (depending on the system under study) appears to extend quite late in life. Support for a later sensitive period in females includes evidence that: (1) the critical period for feminization of sexual behavior apparently extends up to puberty in female rats (Gerall et al. 1973); (2) exposure to low doses of estrogen as late as Day 30 to 40 results in feminized open-field behavior in OVXd rats (Stewart & Cygan 1980); (3) OVX on Days 25-26 decreases the locomotor response to AMPH in female rats (Forgie & Stewart 1994); (4) ethinylestradiol exposure from Days 40 to 90 leads to a thinner cortex in OVXd female rats (Pappas et al. 1979); (5) post-pubertal castration of male rats followed by low-dose estrogen and progesterone treatment increases the size of the AVPv, and decreases the size of other sexually dimorphic nuclei (Bloch & Gorski 1988); and (6) OVX of female rats on Day 30 prevents the female-typical decrease in cortical pyramidal dendritic spines (Munoz-Cueto et al.1990). These findings all suggest a sensitive period for ovarian feminization that extends up to or around puberty in rodents.
Thus it may be that early (< Day 10) high levels of intra-cellular estrogen interact with sex and age specific estrogen receptor populations to "masculinize," whereas later (> Day 10) and lower levels of estrogen interact with age and sex specific estrogen receptor populations to "feminize." Work by Stewart and Cygan (1980; see also Stewart et al. 1979) nicely illustrates this distinction by showing masculinizing effects on OVXd female rats with early high dose estrogen treatment (25 ug estradiol benzoate on P2 and 3) and feminizing effects on OVXd female rats with later low dose estrogen replacement (silastic implants of estradiol 17B on P30-40, delivering physiological levels of about 108 pg/ml serum) on the same variable -- open field behavior.
The literature reviewed here provides strong evidence that ovarian hormones influence the development of the female brain. The findings do not refute or contradict the profound evidence of androgen-mediated masculinization, but suggest that ovarian hormones may exert parallel influences on the development of brain and behavior in the female, in the absence of early androgens. This in turn compels us to broaden the concept of sexual differentiation, by recognizing that both testicular and ovarian hormones are active participants. This idea was anticipated more than 15 years ago by Stewart and Cygan who wrote in 1980 that "while both testicular and ovarian hormones contribute to normal male and female behavioral development, their actions are not merely reciprocal and probably occur at different times in development" (p. 24).
In the next section, we review data showing that ovarian and testicular hormones each play a critical role in neuroanatomical differentiation of the corpus callosum (CC).
We have systematically investigated the role of neonatal gonadal hormones on callosal development in the rat, prompted by the finding that the corpus callosum is significantly larger in adult male than female Purdue-Wistar rats (Berrebi, Fitch, Ralphe, Denenberg, Friedrich & Denenberg 1988). In our initial study, entire litters of male and female pups received handling stimulation between birth and weaning, or were nonhandled controls. Handling was included because of prior data showing that this procedure affects the development of cerebral laterality and may influence callosal size as well (Denenberg 1981). Handling consisted of removing the newborn pups from the maternity cage, leaving the mother in the cage, placing each pup into a 1 gallon can containing wood shavings, leaving them for 3 minutes, and returning them to the home cage (Denenberg 1977). This was done daily from Day 1 through 20, with weaning on Day 21. Subjects were then group housed with same-sexed littermates. At 110 days they were perfused, the brains were removed, and a mid-sagittal section of the callosum was obtained. Using a projection microscope the callosum was magnified and drawn.
Males were found to have a larger absolute cross-sectional callosal area than females. Further, there was a Sex x Handling interaction, with handled (H) males having the largest callosum, followed by nonhandled (NH) males, then NH females, with H females having the smallest callosa. H males differed significantly from H females, and NH males differed significantly from NH females. However, the magnitude of the sex difference was greater within the H animals than within the NH rats, and this was the cause of the significant interaction (see Table 2).
Since males and females also differed on brain weight, the data were re-calculated as relative values (Berrebi et al. 1988). Callosal area for each subject was divided by that animal's brain weight taken to the two-thirds power. [The two-thirds correction was used to convert the index of brain size from a 3-dimensional to a 2-dimensional measure.] However, the same significant effects were obtained as for the absolute CC values.
We then developed a software program, Stereology, to expand and simplify our data analyses (Denenberg, Cowell, Fitch, Kertesz & Kenner 1991a). The outline of the callosum was traced onto a digitizing tablet, and the computer calculated the following callosal parameters: area, perimeter, length, and 99 widths measured along the longitudinal axis of the callosum. The Berrebi et al. data were re-analyzed using the Stereology program (Denenberg, Berrebi & Fitch 1989). A comparison of computer-generated callosal values with those from Berrebi et al found a mean error of 0.27%. [The Berrebi et al values in Table 2 are from the computer program to maintain comparability with the other data in that table.]
Next, the measures for callosal area, perimeter, length, the 99 widths, and brain weight were entered into a factor analysis (Denenberg et al. 1989). The 99 widths fell into seven oblique factors. An eighth factor had significant loadings on brain weight, callosal length, and callosal perimeter. Callosal area did not load on any factor. Interestingly, a highly similar pattern of 7 oblique width factors was obtained from independent analyses of human CC width measures (Denenberg et al. 1991a). This parallel was somewhat startling, given the profound shape differences between rat and human CC, and could possibly suggest that common structural factors in the CC (e.g., discrete fiber bundles) underlie the observation of statistical loading on discrete factors. The 7 width factors obtained in the Denenberg et al. (1989) analysis of the rat CC have subsequently been used in all rat CC analyses.
In the Denenberg et al. (1989) analysis, the variables of brain weight, callosal length, and callosal perimeter were in a factor by themselves, not associated with any of the seven width factors nor with callosal area. Statistically, this means that brain weight is independent of callosal area and width. To verify this, we correlated each animal's brain weight against its seven factor width scores and against its callosal area. Out of 64 correlations, only six were significant at the .05 level, and one of those had a negative sign (Denenberg et al. 1989). This is essentially a chance distribution (p>.10). In subsequent studies we have confirmed that brain weight is uncorrelated with corpus callosum (CC) area or any of the width factors (Fitch, Cowell, Schrott & Denenberg 1991; Mack, Fitch, Cowell, Schrott & Denenberg 1993).
The correlation pattern is directly relevant to the question of whether one should use absolute or relative CC values in assessing experimental treatments. Simply finding that there are significant differences in group means on two or more variables is not sufficient evidence to cause one to use relative measures (e.g., males have a larger CC area than females and also have brains that weigh more). It is also necessary that there be a significant association between two variables within a group before one needs to make an adjustment (the same is true for a covariance analysis). As an example, women weigh on average less than men, and women score lower on average than men on certain tests of spatial ability (Halpern 1992). One cannot draw any conclusion concerning an association between these two variables from such data. That can only be done if a significant correlation exists between weight and spatial scores within each gender.
In three independent studies we have failed to find any evidence that CC area and brain weight are related. In addition, when we adjusted for brain weight in the Berrebi et al. study, we still obtained the same significant effects as we had obtained using absolute CC values. These several findings allow us to conclude that cross-sectional callosal area and the seven callosal width factors can be evaluated in absolute terms and do not need to be corrected for brain weight. This issue also has relevance for the field of human callosal research, in which some studies use absolute measures and others use corrected ones. We discuss this issue further in the context of reviewing the human literature on callosal sex differences.
The Berrebi et al. findings demonstrated a clear sexual dimorphism in callosal size in the Wistar rat, whether handled in infancy or not. This result has since been independently replicated in Long-Evans rats (Nunez, Kim & Juraska 1995; Zimmerberg & Scalzi 1989; Zimmerberg & Mickus 1990), and in Sprague-Dawley rats (Mack, McGivern, Hyde & Denenberg 1996b).
Because many sexual dimorphisms in the brain are influenced by hormones, we set out to study the role of hormonal exposure in development of the sexually dimorphic rat CC. The cumulative results from this series of studies are displayed in Table 2, and Figures 1 and 2. Table 2 presents the CC area means listed by reference and treatment. Figures 1 and 2 summarize the effects of various hormonal manipulations on callosal size in males and females as compared to control female values. The means and SE's in these figures were taken from the data in Table 2, and weighted as a function of N before pooling. Unless otherwise noted, all experiments described below were conducted with animals handled in infancy, because we wished to maximize the baseline sex differences. In all our CC analyses we have looked at callosal area, perimeter, length, and the seven regional width factors (that were derived from the factor analysis procedure described above). Typically when a manipulation alters callosal size there are significant effects on callosal area, one or both of the two anterior width factors, and one or both of the two posterior width factors. For purposes of this review only callosal area data are reported because, with one exception to be discussed below, these accurately reflect the findings from the complete data set.
[Insert Figures 1 and 2 about here]
5.4.1. Testosterone administration to females. In our first study we found that a single sc injection of 1 mg. testosterone propionate (TP) administered to handled 4-day old female pups was sufficient to significantly increase their adult CC area as compared to oil-treated female littermates (Fitch et al. 1990a) (Figure 2). Indeed, the increase was so large that the TP female CC values did not differ significantly from those of male littermates. We then repeated the Fitch et al. experiment with nonhandled rats, fully expecting to find the same effect (Denenberg, Fitch, Schrott, Cowell & Waters 1991b). To our surprise, TP did not increase callosal size in NH females. We then did a second experiment with both H and NH animals. We replicated the finding that H females given TP had significantly larger callosa, and we also replicated the finding that NH females given TP were unaffected with respect to callosal size. We interpreted these data as suggesting a synergy between the presence of testosterone and the effects of handling on adrenal corticosteroids (Denenberg, Brumaghim, Haltmeyer & Zarrow 1967; Meaney, Aitken, Bhatnagar, Van Berkel & Sapolsky 1988). This hypothesis is supported by evidence that handling alters glucocorticoid receptor levels in the cortex and hippocampus (Meaney et al. 1988), and that adrenalectomy alters myelination of the cerebral cortex (Meyer & Fairman 1985). These findings support the notion that handling (and subsequent adrenal changes) might interact with gonadal steroids to influence differentiation of the cerebral cortex, particularly the CC which is a myelinated structure.
The finding that TP treatment must be associated with handling in order to significantly enlarge the female's callosum suggests that the mechanism underlying this effect is more complex than simple exposure of the female to androgen. Because evidence supports the relative importance of estrogen biosynthesis to sexual differentiation of the rodent cerebral cortex (reviewed above), we examined the relative influence of testosterone's aromatized metabolite, estrogen, on the development of callosal size (Fitch et al. 1990a). We did not see any significant effects on adult callosal size when handled female pups were exposed to a synthetic estrogen, diethylstilbestrol (10 ug DES on Day 4; dose derived from Dohler et al. 1984a), although DES females did have significantly higher body weights than oil controls (thus establishing its effectiveness; Fitch et al. 1990a). This failure to find an effect of DES on callosal size suggests that masculinization of this structure may not depend exclusively on aromatization, although clearly this issue will require further study.
5.4.2. Developmental effects of TP treatment. Handled females were given TP or oil on Day 4, and littermates from each condition (as well as male controls) were sacrificed at 30, 55, or 90 days (Fitch, Cowell, Schrott & Denenberg 1990b). Callosal size for both TP treated females and males was significantly greater than for intact (control) females by 30 days of age, and there were no significant differences between TP females and males. This pattern was also seen at 55 and 90 days of age (Figure 3).
[Insert Figure 3 about here]
5.4.3. Castration of males. To investigate the role of endogenous testicular androgens in callosal development, we castrated handled male pups on Day 1 of life (Fitch et al.1990a) (Figure 1). Contrary to our expectations, the callosal size of these males was not affected in adulthood. Denenberg et al. (1991b) also failed, in two experiments, to see an effect of Day 1 castration on CC size.
These findings suggested that testosterone exposure in the prenatal and early (<24 hour) postnatal period had already exerted organizing effects in males. Therefore, we conducted another experiment in which an androgen receptor blocker, flutamide, was administered to pregnant dams (25 mg/kg; Fitch et al. 1991a). Based on Ward and Weisz's (1980) report that the prenatal surge of testosterone occurs on prenatal Day 18, and work by Neri et al. (Neri, Florance, Koziol & Van Cleave 1972) showing demasculinizing effects of flutamide on male reproductive structure and behavior, we chose gestational Day 17 to start the flutamide treatment. Control dams received the vehicle (polyethylene glycol) only. Prenatal flutamide exposure was followed by neonatal castration of male pups. Surgery consisted of placing a small horizontal incision in abdominal skin and muscle, and visualizing and removing the testes. All other pups (flutamide treated females and controls of both sexes) received sham surgery, which consisted of a skin and muscle incision only. All surgeries were performed under cryogenic anaesthesia. In this study, unlike those described above, non-handled animals were used. Results showed that callosal area was significantly smaller among treated males compared to control males, and did not differ from that of female littermates (Fitch, Cowell, Schrott & Denenberg 1991a) (see Figure 1). These results have since been replicated with handled animals (Fitch 1990; Fitch, Cowell, Schrott & Denenberg, in prep) (Figure 1).
5.4.4. The sensitive period for testosterone effects.
Because the results described above were seen in the absence of both the pre- and postnatal testosterone surge, it was impossible to delineate the exact sensitive period for CC masculinization in males. To further assess this window, nonhandled Sprague-Dawley pups were delivered by Cesarean section on Day 22 of gestation and, within 20 minutes thereafter, males were castrated or received sham surgery (Mack et al. 1996b). Female pups received no treatment. This procedure eliminated the postnatal androgen surge in castrated males, while leaving the prenatal surge intact. At 110 days the CCs of castrated and sham males did not differ, but both groups had larger CCs than their female littermates. This finding indicates that the sensitive period for testosterone-mediated sexual differentiation of the male CC begins prenatally, presumably in conjunction with the prenatal testosterone surge. This assertion is supported by the fact that a sexual dimorphism in callosal size can be seen as early as postnatal day 3 in rats (Zimmerberg & Scalzi 1989), as well as evidence that this callosal sex difference is eliminated by prenatal exposure to alcohol (which is thought to suppress the prenatal surge of testosterone in the male fetus; McGivern et al. 1988).
Our results showed that the prenatal period marks the beginning of CC sensitivity to testosterone, but it was not clear when the window "closed." That is, in the absence of prenatal testosterone, would the window of CC sensitivity to testosterone extend postnatally? Prior results showed that Day 4 TP administration to handled female rats increased their callosal size to that of males (Fitch et al. 1990a), suggesting that the sensitive window extends to at least postnatal Day 4. Moreover, additional studies showed that the end of the sensitive period apparently falls somewhere between postnatal Days 4 and 8 because TP administered to females on Day 4 significantly increased callosal size whereas TP administered on Days 8, 12, or 16 did not (Fitch, Cowell, Schrott & Denenberg 1991b) (Figure 2). Thus the window of CC sensitivity to testosterone appears to begin prenatally, and end between postnatal Days 4 and 8. This sensitive period is consistent with the reports of others. Breedlove and Arnold (1983) found that castration on Day 1 did not demasculinize the size of the bulbocavernosus spinal nucleus in male rats, whereas prenatal treatment with flutamide, combined with castration, did. Wagner and Clemens (1989) found that TP treatment to female mice on postnatal Days 1, 3, and 5, but not 7, 9, and 11, significantly increased the size of this nucleus to that of males.
As an aside, the argument can be made that the developmental parameters we have derived from the effects of testosterone on intact female rats cannot be generalized to mechanisms of endogenous androgen exposure in males. To fully address the validity of this assumption it will be necessary to repeat the above experiments by administering TP to prenatally flutamide-treated, postnatally castrated males rather than intact females.
5.5.1. Tamoxifen effects. The first hint that callosal size is affected by ovarian hormones derived from a study where tamoxifen (10 ug TX; dose based on Dohler et al. 1984c) was given to 4-day old male and female pups. This compound is an estrogen antagonist, and our primary reason for using it was to block the action of aromatized estrogen in males (as a test of the aromatization hypothesis for CC differentiation). TX was given to female littermates primarily to provide a matched control. To our surprise, TX resulted in a near-significant (p<.06) increase in callosal area in females (Figure 2), but did not affect callosal size in males (Fitch et al. 1990a). In addition, TX significantly increased two of the females' callosal width regions at a significance level of p<.01. In the same experiment, the synthetic estrogen diethylstilbestrol (DES) was administered to males and females on Day 4. However, DES did not significantly affect callosal area in either sex (Figure 2), suggesting that the mechanism of tamoxifen action was not through incidental activation of estrogen receptors (which can occur when receptors are bound by tamoxifen), but rather, by blocking the binding of endogenous estrogen to receptors. Another possibility is that tamoxifen interfered with ovarian development and consequent function, but this interpretation also points to a critical role for the ovaries in CC development.
5.5.2. Ovariectomy effects. The findings described above suggested (but did not prove) that the increase in the size of the female callosum was a consequence of temporary estrogen "removal" during development. One way to test this hypothesis directly was to remove estrogen via ovariectomy, and measure callosal size in adulthood. We did this experiment -- handled females received ovariectomy on postnatal Days 8, 12, or 16. Surgery was performed under inhalation anaesthesia, and consisted of a dorsal horizontal incision in skin and muscle, followed by visualization and removal of the ovaries, along with the tips of the uterine horn (this facilitated complete removal of ovarian tissue). Sham surgery consisted of skin and muscle incision only. Results from this study showed that all three OVXd groups had significantly larger callosa than sham-operated female controls in adulthood (Fitch et al. 1991b) (Figure 2). Further, the three groups did not differ among themselves for callosal width or area. These results suggest that ovarian hormones are acting at and beyond Day 16, much later than for TP which was effective in increasing callosal size when given to females on Day 4 but not 8.
The above results have been replicated twice, for both handled and non-handled females (Mack, Fitch, Cowell, Schrott & Denenberg 1993; Mack, Cowell & Denenberg 1992) (Figure 2). Thus, ovariectomy altered callosal size regardless of the presence or absence of handling, whereas TP effects on the female callosum were observed only in combination with handling. This distinction may relate to the later sensitive period for ovarian as compared to TP manipulations, but the mechanisms underlying this interaction will clearly require further study.
5.5.3. Effects of ovariectomy and estrogen replacement. Since OVX removes the primary source of both estrogen and progesterone, it was impossible to determine which hormone (or both) influences callosal size in females based on the results above. The finding that TX treatment increased callosal size in female rats pointed to estrogen as the causal agent. To test this idea directly we did an experiment in which the ovaries were removed on Day 12 (as above), and females were provided with estrogen replacement via a low-dose silastic estradiol (E) implant inserted under the skin on Day 25 (Mack et al. 1993). A sub-group of OVXd females received blank controls. In adulthood, those OVXd females with the E implant had significantly smaller callosa than littermates receiving ovariectomy only. In fact they were smaller than intact female littermates as well (see Figure 2). Our results agree with those of Pappas et al. (1979), who found that ethinylestradiol given on Days 40 to 90 decreased cortical thickness in OVXd female rats. These findings emphasize that the sensitive period for ovarian effects extends much later than that of testosterone, in this case as late as Day 25.
5.5.4. Developmental OVX effects. In a related study, handled females were given TP or oil on Day 4, and littermates from each condition (as well as male controls) were sacrificed at 30, 55, or 90 days (described above; Fitch, Cowell, Schrott & Denenberg 1990b). In addition, other female littermates received sham or OVX surgery on Day 12 (as above), and were also sacrificed at 30, 55, or 90 days. The effects of TP on callosal size were significant by 30 days of age (paralleling early sex effects), whereas the effects of OVX were not evident until 90 days of age, well after puberty in female rats (Fitch et al. 1990b) (Figure 3).
5.5.5. What about the adrenals?. To further support the assertion that ovarian hormones have direct effects on cortical differentiation, it was necessary to demonstrate that the removal of the ovaries did not lead to an increase in adrenal androgen output and hence indirect "masculinization." In fact, the opposite effect was found. Female rats were OVXd on day 12 or received sham surgery (as above), and were sacrificed at 70 days of age by rapid decapitation (trunk blood samples obtained between 8:15 and 10:15 AM; see Fitch, McGivern, Redei, Schrott, Cowell & Denenberg 1992 for further details). OVXd females were found to secrete half as much androstenedione, the primary adrenal androgen, as sham-operated female littermates. Moreover, OVXd females exposed to a novel environment prior to sacrifice did not show the stress-mediated rise in androstenedione observed for intact females.
5.5.6. Evidence that estrogen exerts organizing effects. A key question is whether anatomical changes in the callosum reflect permanent organizational steroid effects, or activational effects as a function of estrogen levels at the time of sacrifice. Three sets of data support the organizational hypothesis. First, the effects of Day 12 OVX were observed at 90, but not 55 or 30, days of age (Figure 3). This argues against an activational mechanism because the ovaries are certainly active in intact females by Day 30. Second, we examined intact adult female rats for phase of estrus at the time of sacrifice, and then measured ovarian weight and uterine weight in addition to our standard measures of callosal size. Although a significant effect of estrus on uterine weight was observed, no relationship between callosal size and estrus was found (Mack, Fitch, Hyde, Seaman, Bimonte, Wei & Denenberg 1996a). Third, we OVXd a group of females at 78 days, well after puberty, and did sham surgery on female littermate controls. At 110 days we found no significant difference between the CC of the two groups (F<1.0) (Mack et al. 1996a). These three sets of findings strongly favor the hypothesis that ovarian hormones, primarily estrogen, exert permanent anatomical (organizational) effects upon the callosum.
5.5.7 Summary. These combined findings establish that (1) the removal of ovarian hormones in early life leads to callosal enlargement; (2) these effects can be countered by the administration of estrogen; (3) the sensitive period for this phenomenon extends at least through Day 25 of life, considerably later than for testosterone effects; (4) these findings do not reflect secondary effects on adrenal androgen output; (5) ovarian effects on the callosum do not interact with handling in the same manner as androgenic manipulations; (6) developmentally, the expression of this effect begins considerably later than that of testosterone; and (7) these are permanent organizational effects, not transitory (activational) ones.
The accumulated findings from multiple laboratories lead to the conclusion that ovarian hormones act during a sensitive period that extends at least through puberty, and perhaps beyond, to organize the brain of the female. Therefore, the actions of testicular and ovarian hormones contribute to the existence of anatomical differences that characterize the male and female brain. Many issues are raised by this conclusion, some of which are discussed below.
The results reviewed here prompt the consideration of feminization as a process that occurs in parallel with masculinization. The two processes are qualitatively different and operate during different developmental periods. In order for the brain to become sexually differentiated, males need exposure to testicular androgens during the perinatal period (roughly from embryonic day 17 through postnatal days 8-10 in rodents), and females need exposure to ovarian secretions including, but not necessarily limited to, estrogen, during a later period that may extend to or even beyond puberty. (For a discussion of ovarian factors other than estrogen that may also be found to influence female development, see McCarthy 1995.)
Given the presence of two processes, one must consider the extent to which they interact in vivo, particularly in an experimental condition where an intact female is treated with androgen. For example, testosterone exposure in infancy combined with handling stimulation is sufficient to enlarge the female callosum to the size of a male, yet to what extent TP affects or redirects the feminization process that normally occurs in the female is unclear. Some findings suggest that the presence of the ovaries may modify the developmental actions of androgens (e.g., Blizard & Denef 1973). Alternatively, exposure to androgens may alter the activity of the ovaries themselves, greatly confounding the interpretation of endogenous processes (e.g., Barraclough 1961). Consequently, one must consider the implications of these findings for the common research practice of using intact females as "controls" in hormonal manipulation experiments. Although we originally based our conclusions about the actions of androgen on the sexual differentiation of the callosum on comparisons between TP-treated intact and oil-treated intact females, it is possible that this comparison has limited validity. It is unclear to what extent endogenous ovarian hormones interact with exogenous hormonal manipulations. Certainly in future research ovariectomized females should be used as a base-line control to assess developmental hormonal effects, as gonadectomized males are the common control for hormonal manipulations in the male.
This leads to a consideration of the comparison between androgen-deprived males (prenatal flutamide followed by gonadectomy) and neonatally ovariectomized females. These two experimental conditions represent gonadally "ahormonal" states. If hormonal status were the only determinant of callosal size, they should be identical. However, this is not the case. Androgen-deprived males have callosa the size of intact females, whereas ovariectomized females have callosa the size of males. This implies that there are nonhormonal factors, possibly genetic ones, that are also involved in callosal development and sexual differentiation in general. In support of this position, anatomical sexual dimorphisms have been found in Marsupial embryos before the onset of gonadal hormone production in either sex (Wai-Sum, Short, Renfree & Shaw 1988; see also discussion by Pilgrim & Hutchison 1994).
It will be of considerable interest to determine the cellular dimensions of the gross anatomical effects reviewed here. We have examined the distribution of axon types in the genu of the rat and found major sex differences (Mack, Boehm, Berrebi and Denenberg 1995). Females had a higher proportion of unmyelinated axons than males in terms of number of axons and area taken up by axonal fibers. However, the total area occupied by neuronal material did not differ between the sexes, indicating that the sex difference is due to the partitioning of axon types and not to differences in the amount of nonneuronal constitutents.
Juraska and Kopcik (1988) also reported sex differences in the ultrastructure of the splenial (bulbous anterior) portion of the rat's corpus callosum and showed that these were influenced by exposure to an enriched environment. Unmyelinated axons were found to outnumber myelinated axons in the splenium by a factor of 10:1, and females from both conditions had significantly more unmyelinated axons than males. Enrichment increased the number of myelinated axons in females, and increased the diameter of myelinated axons in males (see also Juraska 1991).
Consistent with the above findings on control animals, Kim and Juraska (1990) found that female rats had more unmyelinated axons in the splenium than males at 25, but not 15, days of age. Because loss of unmyelinated axons occurs relatively early in development while axons myelinate and increase in diameter later in life (Berbel & Innocenti 1988), it is possible that early androgen versus estrogen exposure differentially affects these two processes. It is also worth noting that ovarian hormones promote a female-typical loss in cortical dendritic spines in rats between Days 20 and 60 (Munoz-Cueto et al. 1990), suggesting that ovarian estrogen may exert inhibitory effects on callosal connectivity later in life, perhaps by influencing axonal withdrawal. In a recent study, Nunez, Kim and Juraska (1995) attempted to delineate early androgenic effects on CC ultrastructure in the rat, but obtained no significant results. This is likely due to the fact that TP was given to non-handled females, and males were castrated on Day 1 -- both manipulations we have previously shown to be without effect on rat CC anatomy (Denenberg et al 1991b; Fitch et al 1990a). A more effective procedure would be to use a prenatal testosterone blockade such as described in Fitch et al. (1991a). Although the hormonal mechanisms underlying observed sex differences in callosal ultrastructure are as yet unknown, it is important to emphasize that significant cellular differences do characterize the male and female callosum.
Cellular difference have also been examined in the human CC. Aboitiz (1992), for example, found regional differences in postmortem human callosal fiber composition, whereby thin fibers were most dense in the anterior callosum, decreased gradually to the posterior mid-body, and increased again in the posterior region. Large-diameter fibers showed a complementary pattern, with a peak density in the posterior mid-body, and decreasing density in the anterior and posterior poles. In the very posterior-most region, however, this pattern reversed, with a local increase in large-diameter fibers and a decrease in thin fibers. Based on these regional patterns, Aboitiz concluded that callosal regions which connect primary and secondary sensory and motor areas are characterized by more fast-conducting, large-diameter fibers, whereas callosal regions inter-connecting "association" and prefrontal areas contained more small-diameter, slow-conducting, lightly-myelinated fibers. Aboitiz also reported that the vast majority of fibers in the human CC were small-diameter (or thin) fibers, and that callosal area measures were correlated with total number of thin fibers. Aboitiz (1992) suggested that a larger callosa may reflect greater inter-hemispheric connectivity of association regions. Finally, he found no sex differences in callosal fiber composition or fiber patterns in the human CC.
6.3.1. Are there sex differences? In the human literature there have been many reports of sex differences in cerebral lateralization or functional organization (e.g., Kimura 1987; Kimura & Harshman 1984; McGlone 1980; Shaywitz et al. 1995), but the primary finding of structural sexual dimorphism in human cerebral cortex has been for the corpus callosum. The initial paper by deLacoste-Utamsing and Holloway (1982), which stated that the splenium of the callosum is larger in women than men, stirred enormous public and scientific interest. Subsequently, this finding has been contradicted by many reports of no human CC sex differences (e.g., Aboitiz 1992; Bell & Variend 1985; Bleier, Houston & Byne 1986; Byne, Bleier & Houston 1986; Demeter, Ringo & Doty 1985; Kertesz, Polk, Howell & Black 1987; Nasrallah, Andreasen, Coffman, Olson, Dunn, Ehrhardt & Chapman 1986; Oppenheim, Lee, Nass & Gazzaniga 1987; Parashos, Wilkinson & Coffey 1995; Pozzilli, Bastianello, Bozzao, Pierallini, Giubilei, Argentino & Bozzao 1994). On the other hand, one report (Allen, Richey, Chai & Gorski 1991) found sex differences in the shape and maximum width of the splenium favoring females, but no sex differences in overall size, or size of subdivisions.
A recent meta-analysis of 49 studies published between 1982 and 1994 found no evidence of a significant overall sex difference favoring females in total callosal size, or the size or shape of the splenium, whether or not an appropriate adjustment was made for brain size using analysis of covariance or linear regression (Bishop & Wahlsten 1996). In fact, a small, but significant, overall effect on CC size favoring males was observed. The authors conclude that "The widespread belief that women have a larger splenium than men and consequently think differently is untenable" (Bishop & Wahlsten 1996, page number to be added when paper published).
At least two major reasons for the confusion characterizing the literature on sex differences in the human CC have been identified: (1) use of incorrect statistics in testing for sex differences, and (2) failure to account for age and handedness effects.
6.3.1.1 Pseudostatistics. Several reports have shown the CC to be larger in women than men, but in each case these measures were "corrected" by dividing the CC by a measure of overall brain size (e.g., Johnson, Farnworth, Pinkston, Bigler & Blatter 1994; Holloway 1990; Holloway, Anderson, Defendini, & Harper 1993; Steinmetz, Staiger, Schlaug, Huang & Jancke 1995). This raises the issue of absolute versus relative measurement. We addressed this issue in Section 5.2 with respect to our rat research, and pointed out that it is necessary to have significant correlations between callosal size and some index of brain size within each group before one is justified in using relative values. A simple example can illustrate this principle. On average, there is no sex difference between men and women on IQ tests. However, female brains are smaller than male brains, and weigh less. One could obtain an estimate of brain size from cranial measurements or neuroimaging, divide this number into the person's IQ score, and obtain a score that measures "IQ per unit brain tissue." On such a measure females would be significantly superior to males. The reason we do not use such a statistic is that research has established that there is no within-group correlation between IQ and brain size.
Bishop and Wahlsten (1996) were only able to find four studies that reported separate correlations between brain weight and CC area for males and females. The average was .29 (significantly different from zero). However, other researchers have failed to find significant correlations (e.g., Clarke and Zaidel 1994; Kertesz et al. 1987; Parashos, Wilkinson & Coffey 1995). As discussed earlier, we also failed to find any correlation between brain weight and CC size in our rat studies.
Even if one finds a significant correlation in a particular sample, the practice of dividing one number by another is not appropriate unless the correlation is high. The correct statistical procedure is to use a regression analysis or analysis of covariance to remove the linear effects of the second variable. Bishop and Wahlsten (1996) found seven studies that used analysis of covariance to remove the effects of brain or cortex size. In six of these no sex difference was found in CC size. In one study (Holloway, 1990) females were found to have a significantly larger CC. However, in three other studies Holloway did not find a significant sex difference in CC size when brain weight was used as a covariate (Holloway et al. 1993).
Even though one can rationalize the procedure of adjusting CC size for brain size if there is a significant within-group correlation, there is a compelling biological argument for only assessing the absolute measure: in so far as the size of the callosum is related to its cellular constituents, the absolute number of axons and whether they are myelinated or unmyelinated, is of critical importance in determining functional activity (see Section 6.2 above). In this sense there is no logical need to correct for what the rest of the brain is doing.
In conclusion, the procedure of dividing brain size into CC area as a "correction factor" is incorrect, and, because the female brain is typically smaller, can lead to false results suggesting a larger "relative" CC in females as compared to males.
6.3.1.2. Age effects. Another source of confusion involves failure to control for age of subjects. Cowell, Allen, Zalatimo, and Denenberg (1992), for example, re-analyzed the Allen et al. (1991) data and examined how the size of seven regional CC widths varied as a function of age. Trend analyses showed higher order Sex x Age interactions in three of the regions, with males reaching a peak callosal size in their twenties whereas females did not attain their maximum widths until age 41-50. Cowell et al. addressed the issue of failure to find replicable sex differences and stated that " ... the practice of pooling over age in adults has been responsible, in part, for inconsistent [sex difference] reports" (p. 191).
6.3.1.3. Handedness. The critical nature of handedness is nicely illustrated by Witelson's (1989) study of the isthmus, the narrow area just before the splenium. Her studies show that sex effects on the human CC are complicated when consistency of hand usage is taken into consideration. Witelson determined hand consistency by administering behavioral tests for such acts as writing, drawing, toothbrushing, throwing a ball, and using scissors. Subjects were scored as right, either, or left for each item. Consistent right-handers (CRH) were those subjects who used their right hand for all activities or had only right and either scores. The remainder, called non-consistent right handers (NCRH), used their left hand for at least one activity, including writing. Witelson found the isthmus portion of the callosum to be significantly larger in NCRH men. This finding has recently been confirmed by Denenberg, Kertesz, and Cowell (1991c) who used right-handed writing subjects and obtained the same results as Witelson. In a later study Cowell, Kertesz, and Denenberg (1993) examined consistent and non-consistent left handers and contrasted them with consistent and non-consistent right-handers (using writing hand to define handedness group). Among the left-handers, those who were consistent in hand usage had the largest isthmus area regardless of sex; they ranked just below the non-consistent right-handed males. In a related study, Habib, Gayraud, Olivia, Regis, Salamon, and Khalil (1991) used both right- and left-handed subjects, and found the anterior body of the CC to be larger in NCRH males, compared to CRH males, with no difference between the two female groups.
6.3.2. Yes, there are sex differences. The human data indicate that callosal sex differences are influenced by several variables including age, handedness, callosal region, and probably other measures as yet unknown. Although many human CC studies examined one or two of these variables simultaneously, few have conducted a careful analysis of specific callosal regions, and accounted for critical variables such as sex, age, and hand-preference (using a more sensitive measure than simply "writing hand"). Yet this fine-grained analysis is apparently required in order to observe consistent group difference in anatomy of the human CC.
Such a conclusion is not entirely at odds with the animal literature since evidence has shown interactions between sex and environmental exposure when callosal ultrastructure is examined in the rat (e.g., Juraska 1991), as well as interactive age and hormonal effects on callosal size (e.g., TP effects are seen early and OVX effects are seen later; Figure 3), and finally, interactions between sex, handling experience, and hormonal manipulations, on mean width of specific callosal regions (data not presented here but see Berrebi et al. 1988; Fitch et al. 1990a). We did not present our regional callosal data here because, in the rat model, both sex and hormone manipulations exerted such pervasive effects that anterior and posterior callosal regions, and overall callosal area were all affected. Because overall callosal area effectively characterizes sex and hormonal effects on the rat CC, we employed this simplification to document the role of ovarian hormones upon brain development. Nevertheless, it should be recognized that many other variables do exert significant influence on regional callosal anatomy in animal models. As both human and animal research progresses in teasing apart the relative importance of various developmental and structural factors on CC measurement, it is highly likely that more parallels between human and animal data will become evident.
In summary, according to the traditional model of sexual differentiation, mammalian sexual differentiation is primarily mediated by androgens of testicular origin and the presence of these androgens in early life produces a "male" brain as defined by neuroanatomy and behavior. In contrast, the female brain has been assumed to develop via a hormonal default mechanism, in the absence of androgen or other hormones. In the first part of this target article we reviewed literature supporting an active role for ovarian hormones in sexual differentiation. We then presented data demonstrating significant effects of ovarian hormones on a sexually dimorphic cortical structure, the corpus callosum, that is larger in male as compared to female rats.
In the female rat, removal of the ovaries as late as Day 16 increased the area of the corpus callosum in adulthood. However, treatment with low-dose estradiol starting on Day 25 prevented this increase. Callosal size was also increased by a combination of handling female rats in infancy and administering testosterone prior to Day 8. The sensitive period for TP effects upon CC size starts around prenatal day 17 and is over by postnatal day 8. In contrast, the sensitive period for estrogen action extends at least through postnatal day 25. Further, the effects of androgen treatment were expressed early in development, with males and testosterone-treated females having larger callosa as early as Day 30, whereas the effects of ovariectomy did not appear until after Day 55.
These data support the view that ovarian hormones play an important role in the development of the female brain and that the temporal parameters and mechanisms of "ovarian feminization" are markedly different from those of androgenic masculinization. Such findings speak to the need to complement our current model of androgen-mediated sexual differentiation of the brain with what is now known about the parallel role of the ovaries.
Aboitiz, F. (1992) Brain connections: interhemispheric fiber systems and anatomic brain asymmetries in humans. Biological Research 25: 51-61.
Ahmed, I.I., Shryne, J.E., Gorski, R.A., Branch, B.J. & Taylor, A.N. (1991) Prenatal ethanol and the prepubertal sexually dimorphic nucleus of the preoptic area. Physiology and Behavior 49: 427-32.
Ali, M. & Sahib, M.K (1983) Changes in alpha-fetoprotein and albumin synthesis rates and their levels during fetal and neonatal development of the rat brain. Brain Research 282: 314-7.
Ali, M., Kaul, H.K., & Sahib, M.K. (1981) Ontogeny and distribution of alpha-fetoprotein in feto-neonatal rat brain. Brain Research 227: 618-21.
Allen, L., Richey, M., Chai, Y. & Gorski, R. (1991) Sex differences in the cerebral cortex and the callosum of the living human being. Journal of Neuroscience 11; 933-942.
Arnold, A.P. & Breedlove, S.M. (1985) Organizational and activational effects of sex steroids on brain and behavior: a reanalysis. Hormones and Behavior 19; 469-498.
Ayoub, D., Greenough, W. & Juraska, J. (1983) Sex differences in dendritic structure in the preoptic area of the juvenile Macaque monkey brain. Science 219; 197-198.
Bachevalier, J. & Hagger, C. (1991) Sex differences in the development of learning abilities in primates. Psychoneuroendocrinology 16: 177-88.
Bachevalier, J., Brickson, M., Hagger, C. & Mishkin, M. (1990) Age and sex differences in the effects of selective temporal lobe lesions on the formation of visual discrimination habits in rhesus monkeys (Macaca mulatta). Behavioral Neuroscience 6; 885-899.
Bachevalier, J., Hagger, C. & Bercu, B. B. Gender differences in visual habit formation in 30 month-old rhesus monkeys. Developmental Psychobiology 22; 585-599.
Barraclough, C.A. (1961) The production of annovulatory, sterile rats by single injections of testosterone propionate. Endocrinology 68: 62-67.
Bauer, R. (1978) Ontogeny of two-way avoidance in male and female rats. Developmental Psychobiology 11: 103-116.
Beach, F.A. (1974) Behavioral endocrinology and the study of reproduction. Biology of Reproduction 10: 2-18.
Beatty, W. (1979) Gonadal hormones and sex differences in non- reproductive behaviors in rodents: organizational and activational influences. Hormones and Behavior 12: 112-163.
Beatty, W. (1984) Hormonal organization of sex differences in play fighting and spatial behavior. In Progress in Brain Research, Vol. 61, eds. G.J. De Vries, J.P.C. De Bruin, H.B.M. Uylings, and M.A. Corner. Elsevier Science Publishers. p. 315-329.
Beatty, W. W. (1992) Gonadal hormones and sex differences in nonreproductive behaviors. In Handbook of Behavioral Neurobiology, Vol. 11, eds. A. A. Gerall, H. Moltz, and I. L. Ward. Plenum Press, New York. p. 85-128.
Becker, J.B. & Cha, J. (1989) Estrous-cycle variation in amphetamine induced behaviors and striatal dopamine release assessed with microdialysis. Behavioral Brain Research 35: 117-125.
Becker, J.B. (1990) Direct effect of 17B-estradiol on striatum: sex differences in dopamine release. Synapse 5: 157-164.
Becker, JB, & Ramirez, VD. (1981) Experimental studies on the development of sex differences in the release of dopamine from striatal tissue fragments in vitro. Neuroendocrinology 32: 168-173.
Bell, A.D. & Variend, S. (1985) Failure to demonstrate sexual dimorphism of the corpus callosum in childhood. Journal of Anatomy 143: 143-147.
Berbel, P. & Innocenti, G.M. (1988) The development of the corpus callosum in cats: a light- and electron-microscopic study. The Journal of Comparative Neurology 276: 132-156.
Berrebi, A.S., Fitch, R.H., Ralphe, D.L., Denenberg, J.O., Friedrich Jr., V.L. & Denenberg, V.H. (1988) Corpus callosum: region-specific effects of sex, early experience, and age. Brain Research 438: 216-224.
Beyer, C., Green, SJ., Barker, PJ., Huskisson, NS. & Hutchison, JB. (1994) Aromatase-immunoreactivity is localized specifically in neurons in the developing mouse hypothalamus and cortex. Brain Research 638: 203-10.
Bishop, KM & Wahlsten, D (1996) Sex differences in the human corpus callosum: Myth or reality? Neuroscience and Biobehavioral Reviews, in press.
Bleier, R., Houston, L. & Byne, W. (1986) Can the corpus callosum predict gender, age, handedness, or cognitive differences? Trends in Neuroscience 9: 391-394.
Blizard, D. & Denef, C. (1973) Neonatal androgen effects on open-field activity and sexual behavior in the female rat: the modifying influence of ovarian secretions during development. Physiology and Behavior 11; 65-69.
Bloch, G.J. & Gorski, R.A. (1987) Estrogen treatment in adulthood affects the size of medial preoptic structures in the male rat. Society for Neuroscience Abstracts 13: 1165.
Bloch, G.J. & Gorski, R.A. (1988) Estrogen/progesterone treatment in adulthood affects the size of several components of the medial preoptic area in the male rat. Journal of Comparative Neurology 275: 613-622.
Bloch, GJ. & Mills, R. (1995) Prepubertal testosterone treatment of neonatally gonadectomized male rats: defeminization and masculinization of behavioral and endocrine function in adulthood. Neuroscience and Biobehavioral Reviews 19: 187-200.
Breedlove, A.S. & Arnold, A.P. (1983) Hormonal control of a developing neuromuscular system II. Sensitive periods for the androgen-induced masculinization of the rat spinal nucleus of the bulbocavernosus. Journal of Neuroscience 3: 424-432.
Breedlove, S.M. (1992) Sexual differentiation of the brain and behavior. In Behavioral Endocrinology, eds. J.B. Becker, S.M. Breedlove, and D. Crews. MIT Press, MA. p. 39-68.
Brown, T.J., MacLusky, N.J., Shanabrough, M., & Naftolin, F. (1990) Comparison of age- and sex-related changes in cell nuclear estrogen- binding capacity and progestin receptor induction in the rat brain. Endocrinology 126: 2965-2972.
Byne, W., Bleier, R. & Houston, L. (1986) Individual, age and sex related differences in human corpus callosum: a study using magnetic resonance imaging. Society for Neuroscience Abstracts 12: 719.
Camp, D., Robinson, T. & Becker, J. (1984) Sex differences in the effect of early experience on the development of behavioral and brain asymmetries in rats. Physiology and Behavior 33: 433-439.
Castner, S. A. & Becker, J. B. (1990) Estrogen and striatal dopamine release: a microanalysis study. Society for Neuroscience Abstracts 16: 130.
Clark, A.S. & Goldman-Rakic, P.S. (1989) Gonadal hormones influence the emergence of cortical function in nonhuman primates. Behavioral Neuroscience 103: 1287-1295.
Clarke, J.M. & Zaidel, E. 1994. Anatomical-behavioral relationships: corpus callosum morphometry and hemispheric specialization. Behavioral Brain Research 64: 185-202.
Clarke, S., Kraftsik, R., Van der Loos, H. & Innocenti, G.M. (1989) Forms and measures of adult and developing human corpus callosum: is there sexual dimorphism? Journal of Comparative Neurology 280: 213-230.
Collaer, M.L., & Hines, M. (1995) Human behavioral sex differences: A role for gonadal hormones during early development? Psychological Bulletin 118: 55-107.
Connolly, PB., Roselli, CE., & Resko JA. (1994) Aromatase activity in developing guinea pig brain: ontogeny and effects of exogenous androgens. Biology of Reproduction 50: 436-41.
Cowell, P. E., Allen, L. S., Zalatimo, N. S. & Denenberg, V. H. (1992) A developmental study of sex and age interactions in the human corpus callosum. Developmental Brain Research 66: 187-192.
Cowell, P. E., Kertesz, A. & Denenberg, V. H. (1993) Multiple dimensions of handedness and the human corpus callosum. Neurology 43: 2353-2357.
Davis, E.C., Elihu, N., Shryne, J.E. & Gorski, R.A. (1993) Evidence for post-pubertal onset of the volumetric sexual dimorphism and postpubertal growth of the anteroventral periventricular nucleus of the rat hypothalamus. Society for Neuroscience Abstracts 2: 1312.
Dawson, J., Cheung, Y.M. & Lau, R.T.S. (1975) Developmental effects of neonatal sex hormones on spatial and activity skills in the white rat. Biological Psychology 3: 213-229.
deLacoste-Utamsing, C.M. & Holloway, R.L. (1982) Sexual dimorphism in the human corpus callosum. Science 216: 1431-1432.
Demeter, S., Ringo, J. & Doty, R.W. (1985) Sexual dimorphism in the human corpus callosum? Society for Neuroscience Abstracts 11: 868.
Denenberg, V. H. (1977) Assessing the effects of early experience. In Methods in Psychobiology, ed. R.D. Myers. Academic Press. p. 127-147.
Denenberg, V. H. (1981) Hemispheric laterality in animals and the effects of early experience. Behavioral and Brain Sciences 4: 1-49.
Denenberg, V. H., Berrebi, A. S. & Fitch, R. H. (1989) A factor analysis of the rat's corpus callosum. Brain Research 497: 271-279.
Denenberg, V. H., Brumaghim, J. T., Haltmeyer, G. C. & Zarrow, M. X. (1967) Increased adrenocortical activity in the neonatal rat following handling. Endocrinology 81: 1047-1052.
Denenberg, V. H., Kertesz, A. & Cowell, P. E. (1991c) A factor analysis of the human's corpus callosum. Brain Research 548: 126-132.
Denenberg, V.H., Cowell, P.E., Fitch, R.H., Kertesz, A. & Kenner, G.H. (1991a) Corpus callosum: multiple parameter measurements in rodents and humans. Physiology and Behavior 49: 433-437.
Denenberg, V.H., Fitch, R.H., Schrott, L.M., Cowell, P.E. & Waters, N.S. (1991b) Corpus callosum: interactive effects of testosterone and handling in the rat. Behavioral Neuroscience 105: 562-566.
Denti, A. & Negroni, J. (1975) Activity and learning in neonatally hormone treated rats. Acta Physiologica Latinoamerica 25: 99-106.
Diamond, M., Dowling, G. & Johnson, R. (1981) Morphologic cerebral cortical asymmetry in male and female rats. Experimental Neurology 71: 261-268.
Diamond, M., Johnson, R., Young, D. & Singh, S. (1983) Morphological hemispheric differences in the rat brain: male-female; right-left. Experimental Neurology 81: 1-13.
Diamond, M., Johnson, R.E. & Ehlert, J. (1979) A comparison of cortical thickness in male and female rats - normal and gonadectomized, young and adult. Behavioral Neurology 26: 485-491.
Diamond, M., Murphy, G., Akiyama, K. & Johnson, R. (1982) Morphologic hippocampal asymmetry in male and female rats. Experimental Neurology 76: 553-566.
Diamond, M.C. (1984) Age, sex and environmental influences. In Cerebral Dominance - The Biological Foundations, eds. N. Geschwind and A.M. Galaburda. Harvard University Press. p. 134-146.
Diamond, M.C. (1991) Hormonal effects on the development of cerebral lateralization. Psychoneuroendocrinology 16: 121-129.
Diaz-Veliz, G., Soto, V., Dussaubat, N. & Mora, S. (1989) Influence of the estrous cycle, ovariectomy and estradiol replacement upon the acquisition of conditioned avoidance responses in rats. Physiology and Behavior 46: 397-401.
Dodson, R.E., & Gorski, R.A. (1993) Testosterone propionate administration prevents the loss of neurons within the central part of the medial preoptic nucleus. Journal of Neurobiology 24: 80-8.
Dohler, K. (1991) The pre- and postnatal influence of hormones and neurotransmitters on sexual differentiation of the mammalian hypothalamus. International Reviews in Cytology 131: 1-57.
Dohler, K. D., Coquelin, A., Davis, F., Hines, M., Shryne, J. E. & Gorski, R.A. (1984a) Pre- and postnatal influence of testosterone propionate and diethylstilbestrol on differentiation of the sexually dimorphic nucleus of the preoptic area in male and female rats. Brain Research 302: 291-295.
Dohler, K. D., Hancke, J. L., Srivastava, S. S., Hofmann, C., Shryne, J. E. & Gorski, R.A. (1984b) Participation of estrogens in female sexual differentiation of the brain: neuroanatomical, neuroendocrine and behavioral evidence. In Progress in Brain Research, Vol. 61, eds. G.J. De Vries, J.P.C. De Bruin, H.B.M. Uylings, and M.A. Corner. Elsevier Science Publishers. p. 99-117.
Dohler, K., Coquelin, A., Davis, F., Hines, M., Shryne, J. & Gorski, R. (1982) Differentiation of the sexually dimorphic nucleus in the preoptic area of the rat brain is determined by the perinatal hormonal environment. Neuroscience Letters 33: 295-298.
Dohler, K.D. & Wuttke, W. (1975) Changes with age in levels of serum gonadotropins, prolactin, and gonadal steroids in prepubertal male and female rats. Endocrinology 97: 898-908.
Dohler, K.D., Srivastava, S., Shryne, J., Jarzab, B., Sipos, A. & Gorski, R.A. (1984c) Differentiation of the sexually dimorphic nucleus in the preoptic area of the rat brain is inhibited by postnatal treatment with an estrogen antagonist. Neuroendocrinology 38: 297-301.
DonCarlos, L.L., & Handa, R.J. (1994) Developmental profile of estrogen receptor mRNA in the preoptic area of male and female neonatal rats. Developmental Brain Research 79: 283-289.
Doughty, C., Booth, J.E., McDonald, P.G. & Parrot, R.F. 1975. Effects of oestradiol benzoate and synthetic estrogen RU-2858 on sexual differentiation in the neonatal rat. Journal of Endocrinology 67: 419-424.
Fadem, BH., Walters, M., & MacLusky, NJ. (1993) Neural aromatase activity in a marsupial, the gray short-tailed opossum (Monodelphis domestica): ontogeny during postnatal development and androgen regulation in adulthood. Developmental Brain Research 74:199-205.
Fitch, R. H. Neonatal hormones and sexual dimorphism in the rat corpus callosum. 1990. Ph.D. dissertation, University of Connecticut.
Fitch, R. H., Cowell, P. E., Schrott, L. M. & Denenberg, V. H. (1991a) Corpus callosum: perinatal anti-androgen and callosal demasculinization. International Journal of Developmental Neuroscience 9: 35-38.
Fitch, R. H., Cowell, P. E., Schrott, L. M. & Denenberg, V. H. (1991b) Corpus callosum: ovarian hormones and feminization. Brain Research 542: 313-317.
Fitch, R.H., Berrebi, A.S., Cowell, P.E., Schrott, L.M. & Denenberg, V.H. (1990a) Corpus callosum: effects of neonatal hormones on sexual dimorphism in the rat. Brain Research 515: 111-116.
Fitch, R.H., Cowell, P.E., Schrott, L.S., & Denenberg, V.H. (1990b) Corpus callosum: neonatal hormones and development. Paper presented at the International Society for Developmental Psychobiology meeting, Cambridge, England.
Fitch, R.H., McGivern, R.F., Redei, E., Schrott, L.M., Cowell, P.E. & Denenberg, V.H. (1992) Neonatal ovariectomy and adrenal-pituitary responsiveness in the adult rat. Acta Endocrinologica 126: 44-48.
Fleming, D., Anderson, R.H., Rhees, R., Kinghorn, E. & Bakaitis, J. (1986) Effects of prenatal stress on sexually dimorphic asymmetries in the cerebral cortex of the male rat. Brain Research Bulletin 16: 395-398.
Forgie, ML., & Stewart, J. (1994) Effect of prepubertal ovariectomy on amphetamine-induced locomotion in adult female rats. Hormones and Behavior 28: 241-260.
Frankfurt, M., Gould, E., Woolley, C. S. & McEwen, B. S. (1990) Gonadal steroids modify dendritic spine density in ventromedial hypothalamic neurons: a golgi study in the adult rat. Neuroendocrinology 51: 530-535.
Funkenstein, B., Nimrod, A. & Lindner, H.R. (1980) The development of steroidogenic capability and responsiveness to gonadotropins in cultured neonatal rat ovaries. Endocrinology 106: 98-106.
Galaburda, A.M., Corsiglia, J., Rosen, G.D. & Sherman, G.F. (1987) Planum temporale asymmetry, reappraisal since Geschwind and Levitsky. Neuropsychologia 25: 853-868.
Gerall, A. & Dunlap, J.L. (1971) Evidence that the ovaries of the neonatal rat secrete active substances. Journal of Endocrinology 50: 529-530.
Gerall, A., Dunlap, J. & Hendricks, S. (1973) Effects of ovarian secretions on female behavioral potentiality in the rat. Journal of Comparative and Physiological Psychology 82: 449-465.
Geschwind, N. & Behan, P.O. (1982) Left-handedness: association with immune disease, migraine, and developmental learning disorders. Proceedings of the National Academy of Science 79: 5097-5100.
Geschwind, N. & Galaburda, A.M. (1985) Cerebral lateralization: biological mechanisms, associations, and pathology, Pts. I,II,and III. Archives of Neurology 42: 428-654.
Gorski, R. (1984) Critical role for the medial preoptic area in the sexual differentiation of the brain. In Progress in Brain Research, Vol. 61, eds. G.J. De Vries, J.P.C. De Bruin, H.B.M. Uylings, and M.A. Corner. Elsevier Science Publishers. p. 129-145.
Gorski, R.A., Gordon, J.H., Shryne, J.E. & Southam, A.M. (1978) Evidence for a morphological sex difference within the medial preoptic area of the brain. Brain Research 148: 333-346.
Gouchie, C. & Kimura, D. (1991) The relationship between testosterone and cognitive ability patterns. Psychoneuroendocrinology 16: 323-34.
Gould, E., Woolley, C.S., Frankfurt, M. & McEwen, B.S. (1990) Gonadal steroids regulate dendritic spine density in hippocampal pyramidal cells in adulthood. Journal of Neuroscience 10: 1286-1291.
Greenough, W.T., Carter, C.S. & Steerman, C. (1977) Sex differences in dendritic patterns in hamster preoptic area. Brain Research 126: 63-72.
Grimshaw, G.M., Bryden, M.P. & Finegan, J.K. (1995) Neuropsychology 9: 68-79.
Grimshaw, G.M., Sitarenios, G., & Finegan, J.K. (1995) Mental rotation at 7 years: Relations with prenatal testosterone levels and spatial play experiences. Brain and Cognition 29: 85-100.
Habib, M., Gayraud, D., Olivia, A., Regis, J., Salamon, G., & Khalil, R. (1991) Effects of handedness and sex on the morphology of the corpus callosum: A study with brain magnetic resonance imaging. Brain and Cognition 16: 41-61.
Hagger, C. & Bachevalier, J. (1991) Visual habit formation in 3-month-old monkeys (Macaca mulatta): reversal of sex difference following neonatal manipulations of androgens. Behavioral Brain Research 45: 57-63.
Halpern, D.F. (1992) Sex Differences in Cognitive Abilities. Lawrence Erlbaum Associates, London.
Hampson, E. & Kimura, D. (1988) Reciprocal effects of hormonal fluctuations on human motor and perceptual-spatial skills. Behavioral Neuroscience 102: 456-459.
Hampson, E. (1990) Variations in sex-related cognitive abilities across the menstrual cycle. Brain and Cognition 14: 26-43.
Hassler, M. (1991) Testosterone and musical talent. Experimental & Clinical Endocrinology 98: 89-98.
Hendricks, S.E. & Gerall, A.A. (1970) Effect of neonatally administered estrogen on development of male and female rats. Endocrinology 87: 435-459.
Hendricks, S.E. (1992) Role of estrogens and progestins in the development of female sexual behavior potential. In Handbook of Behavioral Neurobiology, vol. 11: Sexual Differentiation, eds. A.A. Gerall, H. Moltz, and I.L. Ward. Plenum Press, New York. p. 129-155.
Hines, M., Allen, L.S. & Gorski, R.A. (1992) Sex differences in subregions of the medial nucleus of the amygdala and the bed nucleus of the stria terminalis of the rat. Brain Research 579: 321-326.
Holloway, RL (1990) Sexual dimorphism inthe corpus callosum: Its evolutionary and clinical implications. In ed.G. Sperber Apes and Angels: Essays in Anthropology in Honor of Phillip V. Tobias. New York: Wiley-Liss, pp 221-228.
Holloway, RL, Anderson, PJ, Defendini, R. & Harper, C (1993) Sexual dimorphism in the human corpus callosum from three independent samples: relative size of the corpus callosum. American Journal of Physical Anthropology 92: 481-498.
Holloway, R.L. & deLacoste, M.C. (1986) Sexual dimorphism in the human corpus callosum: an extension and replication study. Human Neurobiology 5: 87-91.
Holman, S.D., & Hutchison, JB. (1991) Lateralized action of androgen on development of behavior and brain sex differences. Brain Research Bulletin 27: 261-5.
Hutchison, JB. & Beyer, C. (1994a). Gender-specific brain formation of oestrogen in behavioral development. Psychoneuroendocrinology 19: 529-41.
Hutchison, JB., Beyer, C., Green, S., & Wozniak, A. (1994b). Brain formation of oestrogen in the mouse: sex dimorphism in aromatase development. Journal of Steroid Biochemistry and Molecular Biology 49: 407-15.
Hutchison, JB., Beyer, C., Hutchison, RE., & Wozniak, A. (1995) Sexual dimorphism in the developmental regulation of brain aromatase. Journal of Steroid Biochemistry and Molecular Biology 53: 307-13.
Jernigan, T.L., Trauner, D.A., Hesselink, J.R. & Tallal, P.T. (1991) Maturation of human cerebrum observed in vivo during adolescence. Brain 114: 2037-2049.
Johnson, S.C., Farnworth, T., Pinkston, J.B., Bigler, E.D. & Blatter, D.D. 1994. Corpus callosum surface area across the human adult life span: effect of age and gender. Brain Research Bulletin 35: 373-7.
Joseph, R., Hess, S. & Birecree, E. (1978) Effects of hormone manipulations and exploration on sex differences in maze learning. Behavioral Biology 24: 364-377.
Juraska, J (1991) Sex differences in "cognitive" regions of the rat brain. Psychoneuroendocrinology 16: 105-119.
Juraska, J. (1984) Sex differences in dendritic response to differential experience in rat visual cortex. Brain Research 295: 27-34.
Juraska, J. M. & Kopcik, J. R. (1988) Sex and environmental influences on the size and ultrastructure of the rat corpus callosum. Brain Research 450: 1-8.
Kawashima, S. & Takagi, K. (1994) Role of sex steroids on the survival, neuritic outgrowth of neurons, and dopamine neurons in cultured preoptic area and hypothalamus. Hormones & Behavior 28: 305-12.
Kertesz, A., Polk, M., Howell, J. & Black, S. (1987) Cerebral dominance, sex, and callosal size in MRI. Neurology 37: 1385-1388.
Kim, J.H.Y. & Juraska, J.M. (1990) Sex differences in the ultrastructural development of the rat corpus callosum. Society for Neuroscience Abstracts 16: 321.
Kimura, D. & Harshman, R. (1984) Sex differences in brain organization for verbal and non-verbal functions. In Progress in Brain Research, Vol. 61, eds. G.J. De Vries, J.P.C. De Bruin, H.B.M. Uylings, and M.A. Corner. Elsevier Science Publishers. p. 423-439.
Kimura, D. (1987) Sex differences, human brain organization. In Encyclopedia of Neuroscience. Birkhauser Press, Boston. p. 1084-1085.
Kirkpatrick, SW., Campbell, PS., Wharry, RE., & Robinson SL. (1993) Salivary testosterone in children with and without learning disabilities. Physiology & Behavior 53: 583-6.
Kolb, B. & Stewart, J. (1991) Sex differences in dendritic branching of cells in the prefrontal cortex of rats. Journal of Neuroendocrinology 3: 95-99.
Kolb, B., Sutherland, R. J., Nonneman, A. J. & Whishaw, I. Q. (1982) Asymmetry in the cerebral hemispheres of the rat, mouse, rabbit and cat: the right hemisphere is larger. Experimental Neurology 78: 348-359.
Kopcik, J.R., Seymoure, P., Schneider, S.K., Kim-Hong, J. & Juraska, J.M. (1992) Do callosal projection neurons reflect sex differences in axon number? Brain Research Bulletin 29: 493-7.
Krasnoff, A. & Weston, L. (1976) Puberal status and sex differences: activity and maze behavior in rats. Developmental Psychobiology 9: 261-269.
Kuhnemann, S., Brown, TJ., Hochberg, RB., & MacLusky, NJ. (1994) Sex differences in the development of estrogen receptors in the rat brain. Hormones & Behavior 28: 483-91.
Leret, M.L., Molina-Holgado, F. & Gonzalez, M.I. (1994) The effect of perinatal exposure to estrogens on the sexually dimorphic response to novelty. Physiology and Behavior 55: 371-373.
Levina, S.E., Gyevai, A. & Horvath, E. (1975) Responsiveness of the ovary to gonadotropins in pre- and perinatal life: oestrogen secretion in tissue and organ cultures. Journal of Endocrinology 65: 219-223.
Lisk, R.D. & Suydam, A.J. (1967) Sexual behavior patterns in the prepubertally castrate rat. Anatomical Records 157: 181-189.
Mack, C. M., Boehm G. W., Berrebi, A. S. & Denenberg, V. H. (1995) Sex differences in the distribution of axon types within the genu of the rat corpus callosum. Brain Research 697: 152-156.
Mack, C.M., Cowell, P.E. & Denenberg, V.H. (1992) Corpus callosum: interactive effects of handling and ovariectomy in the rat. Society for Neuroscience Abstracts 18: 213.
Mack, C.M., Fitch, R.H., Cowell, P.E., Schrott, L.M. & Denenberg, V.H. (1993) Ovarian estrogen acts to feminize the rat's corpus callosum. Developmental Brain Research 71: 115-119.
Mack, C.M., Fitch, R.H., Hyde, L.A., Seaman, A.J., Bimonte, H.E., Wei, W. & Denenberg, V. (1996a) Lack of activational influence of ovarian hormones on the size of the female rat's corpus callosum. Physiology and Behavior, 60: 431-434.
Mack, C. M., McGivern, R. F., Hyde, L. A. & Denenberg, V. H. (1996b) Absence of postnatal testosterone fails to demasculinize the male rat's corpus callosum. Brain Research, in press.
MacLusky, N.J., Chaptal, C. & McEwen, B.S. (1979a) The development of estrogen receptor systems in the rat brain and pituitary: postnatal development. Brain Research 178: 143-160.
MacLusky, N.J., Clark, A.S., Naftolin, F. & Goldman-Rakic, P.S. (1987) Estrogen formation in the mammalian brain: possible role of aromatase in sexual differentiation of the hippocampus and neocortex. Steroids 50: 459-474.
MacLusky, N.J., Lieberburg, I. & McEwen, B.S. (1979b) The development of estrogen receptors in the rat brain and pituitary: perinatal development. Brain Research 178: 129-142.
MacLusky, N.J., Naftolin, F. & Goldman-Rakic, P.S. (1986) Estrogen formation and binding in the cerebral cortex of the rhesus monkey. Proceedings of the National Academy of Science 83: 513-516.
MacLusky, N.J., Phillip, A., Hurlburt, C. & Naftolin, T. (1985) Estrogen formation in the developing rat brain: sex differences in aromatase activity during early postnatal life. Psychoneuroendocrinology 10: 355-361.
MacLusky, NJ, Walters, MJ, Clark, AS., & Toran-Allerand, CD. (1994) Aromatase in the cerebral cortex, hippocampus, and mid-brain: ontogeny and developmental implications. Molecular and Cellular Neurosciences 5: 691-8.
Mannan, M.A., & O'Shaughnessy, P.J. (1991) Steroidogenesis during postnatal development in the mouse ovary. Journal of Endocrinology 130: 101-6.
Masica, D.N., Money, J., Ehrhardt, A.A. & Lewis, V.G. (1969) IQ, fetal sex hormones, and cognitive patterns: studies of the testicular feminizing syndrome of androgen insensitivity. Johns Hopkins Medical Journal 124: 34-43.
McCarthy, M.M. (1994) Molecular aspects of sexual differentiation in the rodent brain. Psychoneuroendocrinology 19: 415-27.
McCarthy, M.M. (1995) How about sexually differentiating factors other than estrogen? Psycoloquy 95.6.32.sex-brain.7.mccarthy.
McCarthy, N.M., Schlenker, E.H. & Pfaff, D.W. (1993) Enduring consequences of neonatal treatment with antisense oligodeoxynucleotides to estrogen receptor messenger ribonucleic acid on sexual differentiation of rat brain. Endocrinology 133: 433-439.
McCormick, C.M., Witelson, S.F. & Kingstone, E. (1990) Left-handedness in homosexual men and women: neuroendocrine implications. Psychoneuroendocrinology 15: 69-76.
McGivern, R., Raum, W., Salido, E. & Redei, E. (1988) Lack of prenatal testosterone surge in fetal rats: alterations in testicular morphology and phys