A Half-Truth Is a Whole Lie: On the Necessity of Investigating Sex Influences on the Brain

Abstract

Sex influences are proving to be extremely widespread on brain function, including the human brain. Ample evidence now proves that the sex of subjects can influence, ever reverse, findings, hence conclusions, at all levels of brain science, down to the molecular level, often in completely unanticipated ways. Thus the still-prominent assumption that sex influences may be safely ignored by neurobiologists is invalid and must be abandoned. The failure to properly consider the issue fills the literature with conclusions tenuous at best, false at worst. The continuing, widespread resistance to investigating sex influences among brain scientists, a resistance largely rooted in deeply entrenched biases against the topic, is becoming increasingly scientifically indefensible and strongly retards progress in our field.

I started to understand the truth of the Yiddish proverb in this article's title as it relates to sex influences on the brain about 10 yr ago, shortly after my own research into the neurobiology of emotional memory drew me into the issue of sex influences on brain function. I attended a seminar about stress effects on hippocampal function on the molecular level. This work, the speaker argued, would be very important for understanding two disorders in particular: posttraumatic stress disorder (PTSD) and clinical depression. In response to my question about whether she was using male or females rats, she replied, “Male rats. I don't go near female rats!” When I asked why, she explained that, in essence, the nice story she was developing in male rats did not hold up well, if at all, in female rats. Although she acknowledged that her work did not generalize from male rats to female rats, she freely generalized her work to the understanding of humans, and worse, to the understanding of disorders that mostly affect women!

In fairness to the speaker, she was operating under one of neuroscience's still dominant, albeit false, assumptions, namely, that biological sex does not matter all that much, hence may be safely ignored. Many neuroscientists cling to outdated assumptions about the general importance of the sex influence issue. For example, a recent article on how to make better animal models of human disorders failed to even note in passing the pervasive problem of investigators exclusively using male animals to model disorders that affect women, sometimes even while knowing that their findings generalize poorly from the male to female animal (1).

Neuroscience today relies overwhelmingly on male animals (2), ironically due in part to a widespread belief that “sex matters” at least in that circulating sex hormones in females may “complicate” results. However, investigators invariably draw their conclusions about both sexes, implying that their findings will apply equally well to both males and females. Perhaps as an initial simplifying assumption, this extreme reliance on male subjects as a proxy for both sexes was defensible. For decades, sex influences on the brain appeared to most neuroscientists as restricted to a tiny fraction of the brain (mostly the hypothalamus) and a single class of behaviors (sex behaviors). Sex influences found elsewhere were generally viewed as of marginal interest and not worth sidetracking the search for more “fundamental” mechanisms that studying only males supposedly allowed.

Even if this view was once defensible, it is no longer, for the simple reason that abundant evidence now shows that sex influences on brain function are ubiquitous, found at every level of neuroscience from the behaving human to the ion channel (3, 4). These sex influences can modify, negate, even reverse findings, hence conclusions. For example, and not uncommonly, clear phenotypic effects of a gene “knockout” in one sex may not exist at all, or even be reversed, in the other sex given the same knockout (4). Sex differences are often unanticipated, which may tempt investigators to not believe them. In my own experience with the domain of emotional memory, for example, evidence for sex influences on amygdala function, although initially greeted with skepticism, now appears increasingly accepted, as indicated for example by recent a metaanalysis (5). And it is often not clear why they should exist, even though they do. But neither the fact that we did not anticipate a result, nor that we do not immediately know what it means, grants us license to dismiss the result. Rather, both are a call for action, for further research, especially when the key variable involved is as evolutionarily fundamental as is sex, one of the two driving forces of evolution according to Darwin (6).

Many examples illustrate why “sex matters,” and conversely, why the continued, widespread avoidance/trivialization of the issue by neuroscientists must no longer be considered acceptable. Here, I highlight just a few, focusing on examples from human subject research.

Consider first a recent, high-profile paper in Nature entitled “Post-traumatic stress disorder is associated with PACAP and the PAC1 receptor,” which examined relationships between a pituitary peptide and stress responses (7). An equally valid title for the article could have been “Post-traumatic stress disorder is not associated with PACAP and the PAC1 receptor,” or perhaps, “Post-traumatic stress disorder is associated with PACAP and the PAC1 receptor in women only.” The reason is that the data indicated that the pituitary peptide is related to PTSD in women and not in men. The former result is of course no more important, theoretically or practically, than the latter. To be sure, starting with the abstract, the authors are clear about the sex difference in their data. Indeed, had they ignored potential sex influences, they may well have come to a false negative conclusion about the potential role of the peptide in PTSD. But emphasizing that the peptide appears to be involved in PTSD in women deemphasizes the equally important point that it appears not to be in men.

Consider next a recent investigation of the effects of a stressor on memory consolidation. Yonelinas et al. (8) wanted to determine whether a postlearning stressor could enhance memory consolidation as assessed with a particular tool commonly used by cognitive scientists to study memory strength (called “ROC” curves). Subjects in this study viewed a series of pictures and shortly thereafter jumped out of an airplane. That activated a stress response and enhanced memory for some of the pictures seen shortly before but only in men. The stressor did not enhance memory in women. Thus the answer from this study to the question of whether postlearning stress could enhance memory was yes (in men) and no (in women). Crucially, failure to consider sex would have led the authors to a false negative conclusion about the effects of stress on memory strength (personal communication).

Sex influences are equally apparent in the human brain, as evidenced by a growing number of brain imaging studies. Importantly, sex influences are becoming increasingly apparent even in the “resting” human brain. The amygdala, a critical structure for an array of emotion-connected function, exhibits a striking sex by hemisphere difference at rest, such that the left hemisphere amygdala is far more interconnected with the rest of the brain at rest in women than in men, whereas the right hemisphere amygdala is more connected with the rest of the brain in men than in women (9, 10). Focusing on the brain more generally, Tomasi and Volkow (11) also found widespread patterns of functional connectivity among human brain regions at rest that differed substantially depending on the sex of the subjects and the hemisphere at issue. A similar conclusion was reached in another, very large study of resting human brain function (12). Of course, the greater the sex influences on resting brain function, the more all studies of human brain function will have to be considered studies of sex influences, because no matter the subject under investigation, the sexes will start from different baselines.

Circulating sex hormones, although certainly not the full explanation of sex influences on brain, are critical to consider, despite the increased complexity they may imply. Monthly variation in sex hormones, for example, influences patterns of brain connectivity both within and between the hemispheres (13). And recently, we presented evidence suggesting that hormonal contraception, which powerfully alters the natural sex hormone milieu, alters what women retain in memory from an emotional event (14). With each report such as these, the assumption that one may safely ignore sex hormone levels (in both men and women) in studies of brain and cognition weakens.

Finally, sex influences are clearly so pervasive, so nontrivial, so potentially important in any given situation that one may now question whether the continued insistence among neuroscientists to use almost exclusively male animals is not only questionable as a research strategy but is becoming scientifically indefensible. Again, a specific example makes the point, this time from animal research. Mogil and Chanda (15) are at the forefront of understanding sex influences on brain mechanisms of pain. Despite the general reluctance to study sex influences in that domain (as in most of neuroscience), the accumulated evidence clearly demonstrates that neural pain mechanisms are both similar and different in males and females, with the “different” part in no danger of being overinvestigated at present. The data already allow the conclusion that studying only males will both lead, and mislead, the search for effective pain treatments in women. After surveying the field, Mogil and Chanda (15) concluded that “Given these facts, one might argue that basic scientists are shirking their responsibility to half of the human population by avoiding direct animal models of them. We simply do not see any valid excuses remaining for the continued exclusion of female rodents from basic science studies of pain.”

The facts support this view. Surveying all the evidence, one sees it becoming increasingly scientifically indefensible for neuroscientists to continue their excessive reliance on studying and reporting findings from only males while stating or implying that their findings will help us equally understand both men and women. No valid excuses for this as a default assumption among brain scientists in 2012 remain.