We need female mice in neuroscience research | Hacker News Reader

We need female mice in neuroscience research

news.harvard.edu

93 pointsbelter97 comments

I'm reminded of the research from decades ago on the size of infectious particles for airborne disease transmission. Animals were exposed to droplets containing tuberculosis, producing a model for whether other diseases could be transmitted through the air.

That model said COVID-19 could not, which we now know is entirely wrong. As it turns out, tuberculosis must travel deep into the lungs to produce an infection, which is not true for every pathogen.

It's important to check the assumptions underlying new research and make sure they really apply.

We now know what is entirely wrong?

The model predicted COVID-19 could not be transmitted through the air. The model was wrong; COVID-19 is highly transmissible through the air.

The map is never the territory

Good models are useful.

> The model predicted COVID-19 could not be transmitted through the air. The model was wrong; COVID-19 is highly transmissible through the air.

I'm not sure if that's the best way to put it. IIRC, the controversy was if COVID could spread through the air in the form of an aerosol or if it could only spread in the form of droplets. Droplets can't travel as far in the air, and their limitations were the basis of the "6ft social distancing" rules. It turns out COVID can spread as an aerosol, even though that was strongly denied during the early phases of the pandemic.

"… Men and women's brains are connected in different ways which may explain why the sexes excel at certain tasks, say researchers. A US team at the University of Pennsylvania scanned the brains of nearly 1,000 men, women, boys and girls and found striking differences.

Male brains appeared to be wired front to back, with few connections bridging the two hemispheres. In females, the pathways criss-crossed between left and right. ... "

https://www.pnas.org/doi/abs/10.1073/pnas.1316909110

Seems like a pretty fundamental difference

This seems like a wicked study. By starting with 8 year olds, you've selected people who have already developed somewhat. It's quite possible these differences are developed. Partly by reinforcement of norms. We believe boys are more spatially oriented, so we treat them as such, and we implicitly train them to be, this then begins developing those pathways.

But you can't run the study to see if you could develop girls in similar ways. The amount of controls and isolation required to run the study properly is a bit inhumane. You'd need 24 hour control over everything that child learns and encounters.

So while the conclusions may be true, it's possible that it's not an inherent difference due to initial biology.

One thing I wonder about whenever "the brains of men and women have systematically different structures" comes up, is what structures do the brains of trans people demonstrate?

There's not many studies but they generally show that the brain reflects the gender (i.e. a trans woman's brain looks more like a cis woman's than a cis man's), see e.g. https://academic.oup.com/brain/article/131/12/3132/295849?lo...

Thanks :)

Later research found that opposite-sex similarity in some brain features actually correlates with homosexuality.

This wasn't accounted for in the earlier transsexual studies, because they were using subjects who were both homosexual and trans.

Why do you wonder that? Because if brain structure is shaped by environment then it tells us nothing. If it’s not shaped by environment and brain structure is a fundamental biological difference then it provides some signal.

I don't think it matters so much to me why someone has a particular gender identity — if that's physiology or environment — but I think it would still be interesting to be able to determine independently of self-identification.

Here is another study: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8955456/

> For this purpose, we analyzed a sample of 24 cisgender men, 24 cisgender women, and 24 transgender women before gender-affirming hormone therapy. We employed a recently developed multivariate classifier that yields a continuous probabilistic (rather than a binary) estimate for brains to be male or female. The brains of transgender women ranged between cisgender men and cisgender women (albeit still closer to cisgender men), and the differences to both cisgender men and to cisgender women were significant (p = 0.016 and p < 0.001, respectively).

One option to investigate this would be to study children from highly varied cultures around the world. In fact it seems almost obvious that the study of brains is nearly worthless without such diversity. But that would be difficult and expensive.

Also: how much diversity is enough? I strongly suspect that the invention of agriculture had major effects on the social differences between men and women. You would have to include hunter-gatherers in your study to account for this possible effect, and I’m not sure how you would find enough left to do that.

Seems like another deceptive statistical study that probably doesn't even control for brain size. It also seems to have discovered that women probably mix up their logic and "intuitive modes", as we always thought.

And why would they further show statistics by race rather than by brain size? Brain sizes affect the density of connections and proportions of white matter to grey matter, so comparing small-brained women and men to large-brained women and men would seem like an obvious thing to do if you have the data [edit: but this is obviously me just desperately looking for a control.]

Instead, I suspect they were fishing to find out some headline-grabbing result, hoping that (for example) black women were more masculine than white women, or Asian men were more feminine than white men.

On the mis-presentation and misinterpretation of gender-related data: The case of Ingalhalikar’s human connectome study

https://www.pnas.org/doi/full/10.1073/pnas.1323319111

> Instead, I suspect they were fishing to find out some headline-grabbing result, hoping that (for example) black women were more masculine than white women, or Asian men were more feminine than white men.

That’s quite a bomb to throw with zero evidence. Like, why does your mind even jump to that?

> “People have been making this assumption that we can use male mice to reliably make comparisons within and across experiments, but our data suggest that female mice are more stable in terms of behavior despite the fact that they have the estrous cycle,” Datta said.

Isn't this quote mixing up two uses of "stable"? They stuck with male mice to be able to compare to other studies, not to reduce variability within the study (which should not be a goal in itself).

If this shows anything, it is that we know way too little about biological factors underlying behavior, and that variation in other factors (age is an obvious one, but there are also genes that affect this behavior) should be investigated too.

He was actually referring to both aspects. It's all explained further down in the article:

>Males also exhibited individuality of behavior, but they had more behavioral variation within a single mouse and between mice than females.

Hmm, what is "behavioral variation within a single mouse and between mice"?

I get variation between mice, but is variation within a single mouse the range of behaviour?

Yes. Say today the mouse engages in 3 behaviors (groom, walk around edge of enclosure, then dart across it) but tomorrow spends the whole day huddled in the corner (1 behavior, not exhibited yesterday). That’s sort of what they mean by within mouse variability.

Expanding the article's scope from "we should study mice of both sexes" to boiling the ocean with "we should IDIC mice studies" doesn't help here. Invoking age makes researchers more resistant to accounting for female biology in their research, because it gives them a ready excuse to say "we can't afford the combinatorial matrix of sex and age, so we're just going to study healthy adult males". In a world of theoretically infinite resources, it absolutely makes sense to widen the scope of mice studies to include all possible variations of mice, by sex and by age and by genetics. In today's actual world, it does not.

But then I don't see why people are willing to draw conclusions from mouse studies in the first place, certainly not medical. They don't even generalize over all mice.

If anyone still takes them more seriously than a mere hint, I can't see why they would criticize them for excluding females. This study at least doesn't show much difference, and many other factors are excluded anyway.

My wife works in the neuroscience field and has done research in this very topic.

A few general thoughts she's expressed in the past:

(1) Research often dismisses the sexual cycles as insignificant or play it down significantly, then generalize behavior they observe during experiments... without controlling for cycles.

(2) Because of this I know she (and many others) will dismiss observational studies without replication.

(3) Studying females requires extra care -- experiments involving females, should measure the cycle of each mouse to keep a record

The takeaway ought not to rely on a comparison of behavior stability but rather the fact that have studies on females is just intrinsically useful.

The issue is rather complicated. Yes, you want to have a diverse sample, no questions here. However, there might be issues with the statistical design. Mice are territorial animals, and male mice get especially aggressive in a presence of a female. Put a couple of male mice in a cage, they'd be fine; add a girl in there, and there will be casualties.

That creates extreme stress in both male and female mice, and stress is an important confounding factor in neuroscience and beyond. Grouping mice by sex in different cages is obviously incorrect - cage effect is a major factor in animal research.

Not sure how these problems are solved now - maybe there are statisticians in the audience who can shed some light on that.

> Put a couple of male mice in a cage, they'd be fine;

> Grouping mice by sex in different cages is obviously incorrect - cage effect is a major factor in animal research.

What does this mean?

To simplify, this means: if you have a following setup

Cage 1: male1-control, male2-control, male3-treatment, male4-treatment Cage 2: female1-control, female2-control, female3-treatment, female4-treatment

and you see females responding to treatment better than males, what does that mean? It can be any of the following:

0: females respond differently to the treatment due to their biochemistry (say, some Y-chromosome gene is interfering with the treatment) 1: females respond differently to the treatment due to social effects (a male would respond to treatment but not in an environmnet where he has to compete with other males) 2: there is a cage effect (say, it is slighly warmer, or more light, or fresh breeze, or something else mice care about) that causes the animals in one cage to respond differently.

Only with the above setup, you cannot know which answer is correct.

Statisticians have a bunch of tricks to tackle some of the issues above. For example, you can crossover (swap cages after some time) or increase the number of cages and randomise across cage conditions (say, you have 20 cages, 10 all-male and 10 all-female, spread around the facility to cover the variability of the cage conditions). Still, you won't be able to distinguish between #0 and #1.

This is not really my area though, so I want a knowledgeble reader to forgive my ignorance. I am a simple bioinformatician who was never too deeply involved into this type of pre-clinical biostatistics. I'm sure someone on HN has an answer how to properly design experiments like that. My message was that it's a bit more complicated than just 'we need female mice in research' - and I saw way too many irresposible biologists sacrificing 100s of mice in experiments where a correct inference couldn't be made.

Findings reveal that despite hormonal fluctuations, female mice exhibit more stable exploratory behavior than their male peers

Completely agree. Another issue plaguing some animal based neuroscience research is shoddy sample size. Understandable, given that some of the experiments are so damn challenging, but in the long run this does no favors to reproducibility.

That being said, for both of these issues, grants should award more money to fund the expansion in research protocol. Expecting any change without additional resources doesn't help.

We only need identical clones . Everything else is mediocrity

Female research is quite harder than males using animals. The sex hormone could affect brain significantly which then ask for much big sample size. Besides due to the lack of previous studies, it might be hard to give a plausible explanation of the findings without test the sex cycle, but even with the knowledge of the cycle, which hormone?

Imagine it being 2023 and you are asking for more animals to torture. We should be demanding advancements in computational chemistry and biology, in vitro methods, and synthetic biology. Minor scientific advancements should not come at the expense of sentient creatures.

Imagine it being 2023 and you are asking to torture specifically female animals more.

What does being male or female matter? It is about adding to suffering in the world instead of removing suffering.

Why is the comment downvoted? Animal suffering is a legitimate moral concern, and I don't think the comment is inappropriate. Instead of downvoting, a better approach would be to point out to the author that Harvard has apparently a policy towards animal studies [1] that is worth exploring further.

[1] https://research.harvard.edu/2021/02/04/animal-studies/

You're right, but downvoted because it is much, much harder to do what you're asking. But it's just bad science to ignore or not want to do something because it's harder. Modeling biological or physiological systems is something we do via animal sacrifice as well. My own research uses mice. I think current versions of LLMs could be used for synthesizing the framework of such models, at the very least. It would certainly make for an interesting open source framework, and would be pretty valuable if done right. There are some options already, but they are too specialized and not geared towards systems biology or physiology. Mainly the way I think about it is that a mouse can't reason with you about why you shouldn't do what you do to it. So you, as the more analytically mature being have to make up your own reasons. Is it worth it? That would be a digression for this post.

But more importantly, a lot of bad science is done while wasting away animal life. Regulatory bodies can at least do something to make it less cheap or harder to justify the use of animal models. Or at least ensure a higher standard of expectations for having used animal sacrifice.

Does this mean that until now, they did experiments only on male mice?

I wonder what the variation looks like in social behaviors across hormonal states.

Since mice are not humans the gender of the mice might be irrelevant.

So much nutritional studies have been done on animals and yet we know that different animals react totaly different to eg saturated fats and atherosclerosis.

Animals are a garbage proxy for anything human related.

Maybe we should be doing more invasive neuroscience research on humans.

Unfortunately, with political correctness these days, people start whining about "ethics" and "human rights" when you experiment on human brains (fnord).

Depending on what you define as 'experiment', there's quite some recent neuroscience research using invasive techniques in humans. Which as far as I'm aware is approved by an ethical committee, but that's not quite the same as 'whining' :)

The gender of researchers also effects outcomes: https://www.science.org/content/article/male-scent-may-compr...

... technically, the "sex" of the researchers.

Well there's a potential paper right there. Take male/female researcher and pump them full of hormones and see if it also affects the mice behaviour too.

Don’t know why you’re being downvoted. Gender isn’t a real thing biologically, it is a societal construct. It describes human behavior. So the gender role of the researcher can’t possibly have an effect on the experiment as it is imaginary. Only the sex of the researcher is what can trigger a response, as there is a biological basis.

I’ve never heard human behavior called imaginary before.

> it is a societal construct

That has a two-way relationship with biological factors, statistically at least, meaning it's not entirely social either.

It is 100% social.

Link to the study?

According to who? [1]
https://www.psychologytoday.com/ie/blog/old-school-parenting...

[1] The WHO?: https://www.who.int/health-topics/gender

> Since mice are not humans the gender of the mice might be irrelevant.

Actually, it's not. It's well known that medications can have different effects or different dosage requirements depending on menstrual cycle.

And if you drop out a drug candidate in animal trials because you don't account for genetic and hormonal gender as well as menstrual cycle, you can end up dropping prematurely.

Plot twist: the male mice were asked to state their gender identity and as a result the studies are still conducted using exclusively male mice. /s

Please don't post ideological flamebait to HN. We're trying for something else here.

https://news.ycombinator.com/newsguidelines.html

Reply on news.ycombinator.com