Sex bias and omission in neuroscience research is influenced by research model and journal, but not reported NIH funding

Abstract

Neuroscience research has historically demonstrated sex bias that favors male over female research subjects, as well as sex omission, which is the lack of reporting sex. Here we analyzed the status of sex bias and omission in neuroscience research published across six different journals in 2017. Regarding sex omission, 16% of articles did not report sex. Regarding sex bias, 52% of neuroscience articles reported using both males and females, albeit only 15% of articles using both males and females reported assessing sex as an experimental variable. Overrepresentation of the sole use of males compared to females persisted (26% versus 5%, respectively). Sex bias and omission differed across research models, but not by reported NIH funding status. Sex omission differed across journals. These findings represent the latest information regarding the complex status of sex in neuroscience research and illustrate the continued need for thoughtful and informed action to enhance scientific discovery.

Keywords: Female; Male; Neuroscience; Sex bias; Sex difference; Sex omission; Sex reporting; Women’s health.

Figure 1.

Research model, NIH funding status, and journal of neuroscience articles published in 2017. A) Reported research models. Mice were the most commonly employed species, followed by humans, rats, non-human primates, and cell lines. B) Reported NIH funding status. A majority of articles reported NIH funding. C) Journal. The majority of articles were published by the Journal of Neuroscience, followed by the Journal of Neurophysiology, Neuron, Nature Neuroscience, Science and Nature.

Figure 2.

Distribution of sex bias and omission in neuroscience articles published in 2017. A) The majority of studies employed both males and females. Sex bias remains present; studies that employed only males made up the second largest proportion of the dataset. Sex omission also persists; articles not reporting sex comprised the third largest proportion of the dataset. B) The majority of articles using both male and female animals do not report analyzing sex as an experimental variable. Of articles employing both males and females, only ~15% of articles incorporated sex as a biological variable. Articles were characterized as reporting sex as a variable if any statement or statistical test indicated that data from males and females were compared, regardless of outcome and whether or not data were reported.

Figure 3.

Sex bias and omission vary considerably by research model. A) All research models. Articles were coded into twenty six different research model categories. Some research model categories exhibited considerable sex bias, such as ferrets and oscine birds, while others did not, such as gerbils. Other research model categories generally neglected to report sex, most notably immortalized cell lines, while some research model categories widely report sex, such as rabbits. B) Mice. C) Humans. D) Rats.

Figure 4.

Sex bias and omission do not vary by reported NIH funding. Levels of sex bias seem similar between articles reporting and not reporting NIH funding. Likewise, the proportion of articles reporting sex does not vary by reported NIH funding.

Figure 5.

Sex omission but not sex bias varies by journal. The proportion of articles reporting sex varies by journal, as does the proportion of articles that employ male and females without considering sex as a biological variable. Levels of sex bias seem similar between journals. In three out of the six journals analyzed, the majority of articles reported employing both males and females independent of whether sex was considered an experimental variable. In two journals, articles reporting the use of both males and females represent a plurality. Across all journals, the proportion of articles solely employing males is larger than the proportion of articles solely employing females.

Figure 6.

Changes in sex bias and omission across time in rats and mice. Data obtained from rats and mice from both the current study and a previous study that analyzed the same journals (Will et al. 2017) allows for assessment of sex omission and bias between 2010 and 2017. Sex omission dramatically decreased from 2010 to 2013, and then continued to decrease from 2013 to 2017. Sex bias presents a complex presentation. More articles increasingly report the use of males and females between 2010 and 2017, reaching a plurality by 2017. Only a relatively small but growing proportion of articles include evaluate sex as an experimental variable. However, the proportion of articles solely employing males remains much higher than articles solely employing females.