I can usually predict how my students will react to claims about brains and development that are based on rodent research – and no, it has nothing to do with a squeamishness about mice. If they think the ultimate claim being made is cool and interesting, like the way that mice seem to genetically inherit sensitivity to a smell their grandfather was trained to fear, or that an insufficiently nurturing mother can make them anxious, they have no problems with mice research. In fact, they immediately start looking for more practical applications to human; when I recently shared the story of how injecting a promiscuous vole species with vasopressin could make them more monogamous, some students immediately envisioned a future where women at bars would have little vials of vasopressin (or its fellow player in love, oxytocin) to help attract a mate. If they don’t like the final conclusion, though – because they don’t want to stop drinking coffee when they get pregnant, perhaps – suddenly they become thorough skeptics in how much rodents can really tell us about human behavior.
A healthy dose of skepticism about research claims is one of my goals for my students, but I prefer it to be based on the methods used in a study, not whether they like or dislike the results. And as we are now discovering, there may be even more cause for concern about rodent research than we had thought.
Warnings about the limits of rodent research are not exactly new, just in need of better PR. A few years ago I was startled by the very thorough expose about mice research that Daniel Engber ran on Slate. I am not directly tied to any rodent research – I am a psychologists and not a neuroscientist specifically because I was not interested in the required lab that involved studying and then “sacrificing” rats – and I knew far more about how mice are usually tested (Skinner boxes, enriched environments, water mazes, and so on) then where the mice themselves come from. Turns out that they are purebred mice, so that any researcher ordering a dozen Black 6 or Sprague-Dawley critters can be confident about exactly what she is getting. Not only do we have the mouse genome mapped, we can “knockout” virtually any gene we choose to determine exactly what impact that gene has on behavior. For scientists who like to control absolutely everything about the study, including all the quirks and foibles of the participants, such mice are a dream.
But just a purebred dogs suffer from the health consequences of too much inbreeding and not enough genetic variety, our purebred mice strains may have breed-specific characteristics that make them poor matches to humans. Engber spoke of research that revealed how supposedly “healthy” mice used as controls in diet studies may have really been sedentary and obese, not healthy at all; the second part of his expose emphasized how the popular Black 6 mouse is in fact highly sensitive to pain, resistant to painkillers, and prone to addiction, even compared to other mice. I have no doubt there are some human beings with a similar profile, but which, and how many, and do we really want all of our understanding and drug development to be based on that profile?
And now, it’s not just the strain of rodent used in a study we need to be cautious about. It’s the strain of experimenter – or rather, the sex of the experimenter. Currently going viral in the science news is a study – by Jeffrey Mogil, the same researcher who identified the Black 6’s super sensitivity to pain – showing that mice respond to pain differently when a male researcher is in the room. Turns out, males of any species, including humans, stress out mice; thank you, pheromones. That stress produces stress hormones, which either dampen pain or at least convince the mice not to show it – something similar to how at least one runner managed to finish a race on a broken leg. Stress does far more than just mask our pain, though. It can disrupt our ability to remember, it makes rats more aggressive, and too much stress can even lead to “learned helplessness” – I can’t do anything about it, so I’ll just give up on trying.
Suddenly an awful lot of research can be called into question: Is this discovery about pain, or learning, or any other behavior, only true when stress is involved? How many studies couldn’t find what they were looking for, because sometimes there was a stressful male experimenter collecting data, and sometimes an unstressful female experimenter, and the stress hid a more interesting effect?
Mogil suggests that the mice would be less stressed out if there were also a female human experimenter present, or if they had time to get used to the male…but most studies don’t even report the sex of the experimenter, let alone how long they were in the room. And what if it was a female experimenter collecting data, but a male research assistant has recently been in the room cleaning cages? I’m not even sure how many scientists would have kept track in their own records, or have kept those records.
This new caution does not invalidate all previous mouse research, of course; it should in fact inspire a new renaissance of mouse research to tease apart what studies might have involved more mouse stress and which involved less, and what that means for any conclusions we have drawn. It will make me feel better about my own studies, however. Humans, and children in particular, may be messy and idiosyncratic with a host of individual differences that can make it a real struggle to find a clear pattern in their behavior. But I think we may have a better handle on understanding each other – and therefore figuring out what kinds of things might be influencing our results – than we have on understanding mice.