Well, well, well. Science decided to play truth or dare again and picked "expose everyone's deepest insecurities." This week: women scientists discover they're terrible at bragging on Twitter, someone finally solved the age-old mystery of "is my baby actually eating or just using me as a pacifier," AI learned to taste your cooking and judge you for it, one brain scan can now predict if you'll forget where you put your keys forever, and mice are making their own Ozempic because apparently even rodents want that summer body.
Pour yourself something caffeinated and pretend you're working. Science is about to make you feel simultaneously amazed and inadequate.
Table of Contents
Women scientists would rather eat glass than tweet about their own papers news
A University of Michigan study just confirmed what every academic conference already knew: women scientists are spectacularly bad at self-promotion online [Nature Communications]. The numbers are brutal: women are 28% less likely than men to tweet about their own papers, even when controlling for everything from field to follower count.
The research team analyzed 23 million tweets about 2.8 million papers from 3.5 million scientists over six years, presumably while procrastinating on their own research. Male self-promotion probability jumped from 2.46% to 10.55% between 2013 and 2018, while women crawled from 1.75% to 7.69%. That's right, ladies, we're losing the Twitter Olympics by a landslide.
Here's the kicker: high-performing women from elite institutions publishing in Nature were the LEAST likely to promote their work. In journals with impact factors above 40, men were 85% more likely to self-promote than women (16.5% vs 8.9%) [Nature Communications]. Apparently, the better your research, the worse you are at telling people about it. It's like inverse networking.
"This isn't just about tweets. It's about who gets seen, cited, and celebrated in science," explained Daniel Romero, the study's co-author. The platform rewards "self-importance, combative behavior, and masculinity," which apparently doesn't align with how women are socialized to behave online. Who could have predicted Twitter would be unwelcoming to women? Besides literally everyone.
Even among Twitter-active scholars, women still self-promoted 9.4% less than men, suggesting this isn't just about platform adoption. It's about deeply ingrained "please don't look at me" syndrome that apparently comes free with every PhD.
Finally, a way to quantify your baby's milk drunk status research
Northwestern University researchers just solved one of parenting's most anxiety-inducing mysteries: exactly how much milk your baby consumed while using you as a 24-hour diner. Their solution? A flexible cord that wraps around your breast like the world's weirdest friendship bracelet and uses electrical impedance to track milk volume in real-time.
The device works by sending tiny electrical currents through breast tissue and measuring voltage differences as milk volume decreases. "Because muscle, fat, bone and tissues conduct electricity differently, bioimpedance can yield an accurate measurement of fat content," explained lead researcher John Rogers. "In a conceptually similar way, we can quantify the change in milk volume within the breast." Yes, they're literally electrocuting your boobs for science.
Clinical testing with 12 mothers over 17 weeks showed strong correlation between device readings and traditional baby weigh-ins. The system requires one "personalization session" during pumping (because, of course, your breasts need their own electrical signature), then provides continuous monitoring through a smartphone app. Finally, data to obsess over at 3 AM!
"Currently, there are no reliable ways to know how much babies are eating when they are breastfeeding," noted pediatrician Jennifer Wicks from Lurie Children's Hospital. The current method involves weighing babies before and after feeding, which is about as accurate as eyeballing how much wine is left in the bottle after book club.
The solution addresses what researchers diplomatically call a "critically important unmet need in maternal and neonatal care," which is science-speak for "parents are losing their minds trying to figure out if their baby is starving." Particularly crucial for premature babies who need precise nutritional monitoring, or just regular parents who need validation that yes, your child is actually eating and not just hanging out.
Check more from the Nature Biomedical Engineering paper or here
Robot tongue achieves what your picky toddler cannot: actually tasting food research
Chinese researchers created an artificial tongue using graphene oxide that can identify flavors with 98.5% accuracy for basic tastes and 75-90% accuracy for unfamiliar samples. Unlike your three-year-old who claims everything tastes "yucky," this device actually provides useful feedback.
The graphene-oxide ionic sensory memristive device (GO-ISMD) measures 4mm × 4mm and uses ion confinement within layered membranes to detect chemical changes. That's smaller than the piece of broccoli your kid has been hiding under their plate for the last hour. The system was trained on 160 different chemicals to build a flavor memory bank, then tested on 40 new samples it had never encountered.
The technology integrates both sensing and computing functions into a single nanofluidic device, using reservoir computing algorithms to process electrical conductivity patterns. Beyond basic tastes (sweet, salty, bitter, sour), the system successfully identified complex combinations like coffee and cola [PNAS]. Whether it can tell Diet Coke from Diet Pepsi remains the true Turing test.
Primary applications target medical restoration of taste perception for stroke victims and people with neurodegenerative conditions. "It will be a huge success if we can give an individual back some taste experiences," commented neuroethicist Andrea Lavazza. Though honestly, some of us might prefer not tasting airline food.
The current system remains "relatively bulky" with "large energy demands," requiring significant miniaturization before clinical deployment. But imagine the possibilities: DoorDash listings with actual taste samples, or wine tastings where you don't have to pretend you know what "hints of elderberry" means.
Your brain scan just called you old (and it's probably right) news & research
Duke University researchers developed a tool called DunedinPACNI that sounds like a Harry Potter spell but actually predicts dementia risk using just one six-minute MRI scan. The system analyzes 315 structural brain features to calculate biological aging pace, basically telling you if your brain is aging like fine wine or gas station sushi.
The validation study of over 50,000 brain scans found that individuals identified as fastest aging were 60% more likely to develop dementia in subsequent years [Duke Today]. The top 10% of fast agers showed 61% higher risk of progressing to mild cognitive impairment or dementia compared to average agers. It's like getting a speeding ticket for aging.
"Our jaws just dropped to the floor" when seeing initial results, recalled lead researcher Ahmad Hariri, which is exactly what you want to hear from the scientist analyzing your brain. The system works by tracking how the same individuals age over time, avoiding the "kids these days" bias that affects comparisons between different generations.
The algorithm uses standard T1-weighted MRI scans available worldwide and requires no specialized equipment beyond what's already gathering dust in most hospitals. Researchers trained the model on 860 participants from a New Zealand birth cohort followed since 1972, because apparently New Zealanders age in scientifically useful ways.
Clinical implementation could begin within a few years, though researchers emphasize the need for reference norms across diverse populations. Because finding out your brain is aging faster than expected is apparently something we need to democratize globally.
Check the article in Nature Aging or see more here
Mice achieve what millions of humans cannot: making their own weight loss drugs news & research
Japanese researchers at the University of Osaka used CRISPR to transform mice into biological Ozempic factories, achieving 34% weight reduction that lasted over 24 weeks from a single treatment. That's right, these mice are literally their own pharmacy, and they didn't even need insurance pre-approval.
The team used lipid nanoparticles to deliver CRISPR components targeting liver cells, specifically the albumin gene locus. Modified hepatocytes continuously produced and secreted exenatide into the bloodstream, with detectable drug levels lasting up to 28 weeks. It's like having a tiny pharmaceutical company in your liver, minus the price gouging.
Results matched mice receiving continuous synthetic drug via osmotic pumps, with treated animals showing 29% reduction in food intake and metabolic improvements equivalent to normal diet-fed controls [Communications Medicine]. These mice basically got the benefits of willpower without actually having any.
"We hope that our design of a one-time genetic treatment can be applied to many conditions that do not have exact genetic causes," explained lead author Keiichiro Suzuki. Because why take medication every week when you can just reprogram your DNA once and call it a day?
Of course, permanent genetic modification carries unknown long-term risks, like possibly turning into a mouse or developing an uncontrollable urge to run on wheels. Competing approaches from companies like Fractyl Health target similar outcomes through pancreatic gene therapy, with human trials potentially beginning in 2026. Until then, we're stuck with our regular human livers that stubbornly refuse to make their own pharmaceuticals.
Check the Nature Communications Medicine article for more, or look more here
And there you have it, another week where science looked at our problems and said, "hold my pipette." We've got women scientists who'd rather publish in obscurity than tweet, breast pumps that finally joined the 21st century, robot tongues judging your cooking, brain scans predicting your future forgetfulness, and mice living their best genetically modified lives.
What made you question reality this week? Are you ready to strap electrodes to your chest for data-driven parenting? Planning to train an AI tongue to taste-test your questionable leftovers? Already worried about what that brain scan might reveal?
Hit reply with your thoughts, gender-based Twitter anxieties, or applications to become a self-medicating mouse. We read every email, even the ones clearly written by someone procrastinating on self-promotion.
If this newsletter found its way to you like a woman scientist's unannounced publication, subscribe at cerealbio.tech for your weekly dose of "science is getting weird again." Our readership keeps growing despite our complete inability to promote ourselves on social media.
Stay hydrated, check your brain age, and maybe just let the robots taste your food first, Prateek & Jere
P.S. - To our women scientist readers: We see you not tweeting about your groundbreaking research. We get it. But also, please tell us about your cool science because we're nosy and Twitter won't do it for you.
And if you’re still not convinced, no worries! Just unsubscribe here. You can always check back on us later.