When Jeffrey Mogil, now a Professor of Pain Studies at McGill University, first entered the field, he expected it to be less complex than psychiatric disease research. Pain was evolutionarily ancient, he reasoned, and probably tightly controlled. “But it turns out that is wrong,” he said. “Pain is just as complex as anything else.”
One challenge is that pain is difficult to measure directly. In animal experiments, researchers infer an internal experience from behavior, while human pain ratings can obscure substantial differences in what people are actually experiencing. Responses to the same stimulus, injury, or treatment can vary widely between individuals.
“We know less than we think we know when we ask people normal questions about their pain,” Mogil said. Ask someone to rate their pain from zero to ten, and a score of six says remarkably little about what that six consists of — spontaneous pain, hypersensitivity, numbness, dysesthesia, or some combination.
This variation matters because it reflects the underlying biological differences that shape how pain develops and how individuals respond to treatment. Mogil's research has shown that genetic background can change the apparent importance of a pain mechanism. For example, the role of calcitonin gene-related peptide (CGRP) in heat-induced pain depended on which mouse strains were studied. In some strains, the evidence pointed strongly to CGRP; in others, it appeared to play no role at all.
However, researchers cannot account for every source of biological variation without making experiments increasingly complex and costly. So which variables are worth prioritizing?
For Mogil, sex is an obvious place to start. By studying male and female animals, researchers can investigate whether the same pain mechanisms operate in both sexes — or whether similar responses arise through different biological pathways.
What sex differences reveal about pain
You don’t need to give anything up to study sex.
—Jeffrey Mogil, McGill University
It is relatively straightforward to incorporate sex into experimental design. By including equal numbers of male and female animals, researchers can investigate potential sex differences without necessarily increasing the total number of animals used.
“You don’t need to give anything up to study sex,” Mogil said. Equal representation will not give every study enough statistical power to detect subtle differences, but it can allow researchers to identify larger effects without increasing the overall sample size.
For Mogil, this has never required making sex differences the primary object of an experiment. “I’ve never set out to study or find sex differences,” he said. “I simply put myself in a position to see them if they’re there, and they happen to be there a lot.”
One of the clearest examples of a qualitative sex difference came from Mogil’s own work. In 2015, he and colleagues reported that male and female mice developed similar levels of mechanical pain hypersensitivity after nerve injury but relied on different immune cells. In males, the response depended on microglia, part of the innate immune system; in females, it appeared to involve cells of the adaptive immune system, particularly T cells.
Eleven years later, the broad finding has held up, although the biology has become more complicated. “There have been about two dozen studies that found some version of the sex difference,” Mogil said. Other studies have challenged parts of the original picture, and the pathway may depend on the particular pain state.
More recent work has begun to map these differences in greater detail. In a 2025 study led by Tuan Trang, researchers identified pannexin-1 channels as a shared component of neuropathic pain pathways in both sexes, but in different immune cells. In males, pannexin-1 acts in microglia and promotes the release of vascular endothelial growth factor (VEGF). In females, it acts in T cells and drives the release of leptin, a hormone involved in energy balance and immune signaling.
“So there is a shared component, but the cellular biology around it is entirely sex-dependent,” Mogil said.
The work challenges the assumption that similar pain behaviors necessarily reflect the same underlying mechanisms. “Behind the scenes, more or less the same phenotype can be produced by completely different pathways. That’s ultimately very surprising.”
This distinction could matter when researchers develop drugs to target those pathways. A treatment aimed at a shared component such as pannexin-1 might affect both sexes, whereas drugs targeting downstream components could have very different effects.
The clinical consequences of overlooking sex
Mogil suspects that failing to recognize sex-dependent mechanisms may already have affected pain drug development.
Migraine provides a useful example of why those differences can be difficult to evaluate clinically. The disorder affects women more than twice as often as men, while men account for only around 10–18 percent of participants in many migraine trials. As a result, it’s much harder to assess treatment efficacy in male patients with confidence.
Mogil argued that animal data can help fill this gap in the clinical evidence. Animal studies have raised questions about whether the biology underlying migraine pain is the same in males and females. One molecule of particular interest is CGRP, which plays an important role in migraines and is targeted by several migraine treatments. In a study in rodents, CGRP triggered migraine-like pain behaviors in females but not males under the same experimental conditions.
Frank Porreca, David Dodick, and colleagues reanalyzed publicly available FDA trial data for three gepants, drugs commonly used for acute migraine treatment. Across the pooled data, treatment effects were statistically significant in women, while estimates in men were smaller and did not reach statistical significance. “If those studies were adequately powered,” Mogil said, “I think they would show that the drugs don’t work in men.”
Nevertheless, a recent European Headache Federation review concluded that the available evidence does not currently support changing CGRP-targeted migraine treatment according to sex. “This interpretation ignores the animal data,” Mogil said, adding that “clinicians often don’t believe in animal data, and often don’t even know it exists.”
From representation to sex-specific evidence
While disparities in sex representation need to be addressed, representation alone does not produce sex-specific evidence.
Either they never ran the analysis, or they didn’t think the analysis was important enough to report.
—Jeffrey Mogil, McGill University
Mogil examined studies published in Pain between 2015 and 2019. He found that although men and women were included, analyses assessing the role of sex were often not reported. Researchers routinely list the sex composition of a study in the demographic table and then never return to it. According to Mogil, around 40 percent of studies involving human participants reported whether their main findings were similar in men and women, while roughly 60 percent did not address the question.
“Either they never ran the analysis, or they didn’t think the analysis was important enough to report,” he said.
Without testing for sex-dependent treatment effects, researchers may never know if a failed trial reflects a treatment that does not work, or one whose benefits were obscured by the way the results were analyzed. One example that Mogil kept returning to was MorphiDex.
MorphiDex is a combination of morphine and dextromethorphan that was developed to treat chronic pain in the 2000s. However, it failed to show an advantage over morphine alone in clinical trials. The result surprised Mogil because his earlier animal research had suggested that the mechanism would work in males but not females. He contacted the first author of the published report to ask if the clinical results showed the same pattern.
The response, Mogil recalled, was simple: “I don’t know. We never looked.”
According to Mogil, the researcher later tried to revisit the data but was unable to access them because the company’s legal department had sealed the file. Whether formal comparison of treatment effects in men and women would have changed the conclusion of the trial is therefore unknown.
“I think this may have happened a bunch of times,” Mogil said, “but I can’t prove it.”
For Mogil, the lesson is that researchers should design studies from the outset to test whether treatments work differently in men and women, and report principal findings clearly for each sex.
When better tools cannot fix the wrong question
Questions of experimental design become ever more important as neuroscience grows increasingly sophisticated. Researchers can now manipulate and measure biological systems with extraordinary precision. “What we can do with optogenetics, DREADDs, tissue clearing, and viral vectors is mind-boggling,” said Mogil. But he worries that the field has become much better at interrogating mechanisms than at questioning the models used to study them.
“Very few scientists pay attention to model development and validity,” he said. “I don’t think physics envy has ever left biology. Most biologists are very interested in molecules and techniques that help them understand molecules, but not in constructs. They’re very reductionistic.”
Pain research offers a simple example of this reductionism. For years, pain studies routinely relied on responses to von Frey fibres or heat stimuli to measure evoked hypersensitivity, treating these responses as proxies for pain itself. But sensitivity to an external stimulus does not necessarily capture the ongoing, spontaneous pain that patients experience.
It matters what questions we ask.
—Jeffrey Mogil, McGill University
Mogil believes the field is beginning to move beyond these limited measures. “The old guard continues to ignore me completely,” said Mogil, “but young scientists are doing experiments in better ways and measuring more important things.” He sees the growing effort to measure spontaneous pain as one sign that the field is beginning to ask better questions.
Yet research that questions the validity of experimental models can be difficult to fund. Mogil has spent five funding cycles seeking support for a project centered on construct validity, only to repeatedly encounter the same criticism from reviewers: that the work looks backwards rather than forward.
“They literally keep saying, ‘Your grant doesn’t look forward; it looks backward. We want to look forward,’” he said. “And I say, ‘If we don’t solve the problem of what we’re looking at, looking forward isn’t going to help. Looking backward might.’”
Progress therefore depends on more than acquiring better tools or identifying the next molecular target. “It matters what questions we ask. It matters what symptoms we’re modeling. It matters both the internal and external validity of our studies, and people should pay more attention to this than they do.”
Acknowledgements
This interview was conducted in collaboration with COST Action CA24168, a European Initiative to Enhance the Current SABV Policy in Preclinical Biomedical Research (EU-SABV), supported by COST (European Cooperation in Science and Technology) and funded by the European Union. We thank Natasha A. Karp, Esther Pearl, Ivana Jaric, and Nikolaos Kostomitsopoulos for their valuable input.












