Mind & Body · Thursday, 1 October 2026

01 Briefing what happened

Scientists replaced half of a mouse's brain with human brain cells, to study illnesses that mice do not get

A Stanford team bred mice with almost no cortex of their own and filled the gap with lab-grown human brain tissue, which wired itself into the mouse's brain and spinal cord.

about 50%

of each mouse's brain, by volume, made of human cells

the mice were bred to lack more than 90% of their own cortex [3][4]

40 days

to grow each clump of human brain tissue in a dish before it went in

four clumps were placed in each mouse [3]

a third

of one side of a rat's cortex: the most an earlier human transplant had filled

this time the human cells had the space to themselves [1]

The lead story — what happened

A Stanford University team led by the neuroscientist Sergiu Pasca bred mice that grow almost none of their own cerebral cortex, the outer layer of the brain used for thinking, memory and the senses. [1][2]
Into newborn mice they placed clumps of human brain-like tissue, called organoids, grown from human stem cells over about 40 days. [3][1]
Within weeks to months the human cells divided, grew blood vessels and filled most of the empty space. [4]
About half of each brain ended up human by volume, though not by number of nerve cells. [5][4]
The human nerve cells connected with the rest of the mouse brain, and some sent fibres down into the spinal cord. [1][5]
The tissue also made a kind of large nerve cell linked to social skills, which mice do not have and which is found in big-brained animals such as whales and primates. [3][5]
Mice born without a cortex did worse than normal mice on a maze memory test, and mice given the human tissue almost fully closed that gap. [3]
When oxygen was cut, the part-human mice developed problems walking, like those in cerebral palsy; ordinary mice cope with low oxygen because they evolved living underground. [3][4]
Pasca said psychiatry has one of the lowest success rates for clinical trials, and that drugs which work well in animals fail badly in patients. [2]
The human tissue is still untidy: most cell types are there, but not sorted into the neat layers of a real cortex. [3][2]
In 2025 the US NIH and FDA pledged to cut animal testing for some medicines, and the NIH announced $87 million for a new organoid centre. [6]
The paper came out in Nature on 16 September, and an editorial in the same journal called it the most extensive merging of human brain cells into an animal so far. [5][7]

Human brain tissue50%
The mouse's own brain50%

About half of each brain ended up human, measured by volume. Counted in nerve cells, the human share is smaller.

Who is involved

Sergiu Pasca

a neuroscientist at Stanford University in California; he led the work and plans to test treatments in the mice
Nita Farahany

a law and philosophy professor at Duke University who sat on the project's volunteer ethics panel; she says she does not know how to regard the animals
Hongkui Zeng

director of brain science at the Allen Institute in Seattle; she says the questions grow if this is tried in pigs or monkeys
Juergen Knoblich

a brain-organoid researcher in Vienna, not involved; he says the data do not show smarter or more aware mice

How it unfolded

2018 Salk Institute scientists report the first human brain organoids implanted into mouse brains [8]
Nov 2025 17 scientists and ethicists call in the journal Science for an international body to oversee brain organoid research [6]
16 Sep 2026 the Stanford mice are published in Nature [5]
Late Sep 2026 a Nature editorial asks who should oversee the field [7]

Where this points

Watch whether an international body for brain organoid research is set up, and whether anyone tries the same thing in bigger, longer-lived animals such as pigs or monkeys. [7][4]

What is pushing on the whole day

The bar and the word are our reading of how hard each one is pushing today. The arrow is where it is heading. The evidence is in the stories below.

Tests that start in mice High↑

Stanford put human brain cells into mice so that treatments for brain illnesses mice do not get can be tested in them. [2] A Louisville team showed in mice that a sugary diet helps a gut germ called C. difficile settle in. [9] Light therapy protected brain cells in mice with a Parkinson's-like illness, and early human trials are under way. [10]

Links that may not be causes High→

People who climbed stairs had a 39% lower risk of dying of heart disease, but the studies only watched them. [11] A 2026 paper says fitter people may simply be the ones who take the stairs. [12] Older people eating the most ultra-processed food had a 58% higher dementia risk, and that study cannot prove the food caused it. [13]

Cancer survival rising, unevenly Building↑

Seven in ten US cancer patients now live at least five years, a record. [14] Five-year breast cancer survival runs from 39.1% in the WHO's African region to 88.5% in the Americas. [15] Inside the US, richer city counties saw cancer deaths fall first and fastest. [16]

US vaccine advice in dispute Building↑

The surgeon general nominee told senators she has seen no good evidence that vaccines cause autism. [17] The US CDC says legal uncertainties left its 2025 flu vaccine advice in place for this season. [18] COVID shots are expected to be covered by insurance despite a legal fight over the panel that guides insurers. [19]

The rest of the day

22 more stories on this beat.

Each with its own sources. None of these is a link to the story above.

02
Surgeon general pick clears a Senate committee

On 30 September a US Senate health committee voted along party lines to send Nicole Saphier's nomination as surgeon general to the full Senate. [20] The surgeon general is the US government's top doctor and public health educator. [20] Saphier, a radiologist who has appeared on Fox News, is President Trump's third pick, after two earlier nominees failed to win enough support. [20] At her hearing on 16 September she said she had seen no reputable evidence that vaccines cause autism. [17]

Why it matters — No date is set for the full Senate vote. [20] Almost 3,300 US measles cases had been confirmed this year by 11 September, more than in all of 2025, which is why her answers on vaccines drew attention. [17]
03
Sugar during antibiotics, and fewer gut bacteria

Researchers at Memorial Sloan Kettering Cancer Center in New York recorded more than 40,000 food items eaten by 158 people with blood cancers during weeks in hospital on antibiotics. [21] Antibiotics kill many gut bacteria, and the variety of those bacteria matters for digestion and health. [21] For every extra 100 grams of high-sugar food, about one large milkshake, the drop in variety was on average 24% bigger. [21] The sugar also helped Enterococcus faecium, a germ behind hospital infections, to grow. [21] New Scientist reported the study on 30 September. [21]

100 gExtra sugary food
24%Bigger loss of gut variety
In patients taking antibiotics, each extra 100 grams of sugary food went with a 24% bigger drop in gut bacteria variety.

Why it matters — It is a link seen in what patients chose to eat, not a trial that cut their sugar. [21] It matches a mouse study from the University of Louisville in which a high-sugar diet let a gut germ called C. difficile settle in and made its illness worse. [9]
04
WHO meets to pick the 2027 flu strains

WHO experts met in the week of 21 September to choose which flu strains go into vaccines for the southern hemisphere's 2027 season. [22] Flu viruses keep changing, so the vaccine is updated twice a year, once for each half of the world, using samples from WHO's worldwide network of labs. [22] New variants can appear after the choice is made, leaving the vaccine a poorer match. [23] A study of 63 past seasons found a repeatable method could have improved the match in 51 of them. [23]

Why it matters — Choosing three months later would have improved the match further in 14 of the 63 seasons, if faster vaccine methods allowed it. [23] In the US, the CDC says legal uncertainties mean its July 2025 flu vaccine advice still stands for the 2026-27 season. [18]
05
Stair climbers had fewer heart deaths

A review of nine studies covering 480,479 people aged 35 to 84 found that regular stair climbers had a 39% lower risk of dying from heart and blood-vessel disease. [11] Their risk of dying from any cause was 24% lower. [11] It was published on 13 August 2026 in the American Journal of Cardiovascular Drugs. [11] The studies watched what people did and assigned nobody to climb stairs, so they cannot show the stairs caused the benefit. [11]

Heart and vessel deaths39%
Deaths from any cause24%
How much lower the risk was for stair climbers, in each kind of death. The studies show a link, not a cause.

Why it matters — A 2026 opinion paper argues that people who are already fitter are the ones who take the stairs, which would make stairs a sign of health rather than its cause. [12] It also warns that hard climbing can strain the hearts of some heart patients. [12]
06
Plant fats and longer lives after prostate cancer

Researchers followed 4,884 US men whose prostate cancer had not spread, from a long-running study of male health workers, for a median of almost 13 years. [24] Men who ate the most saturated fat after diagnosis were 24% more likely to die than those who ate the least. [25] The extra deaths were mostly from heart disease and other cancers, not from prostate cancer. [25] It appeared in JAMA Network Open on 2 September 2026. [25]

Why it matters — About 3.5 million US men live with prostate cancer, often for many years, so what kills them is often something else. [25] Swapping a tenth of calories from animal fat to plant fat went with a 16% lower risk of death, though the men chose their own food. [25]
07
COVID rising in US wastewater

COVID virus levels in US wastewater have risen steadily since July, most in the western and southern states, Scientific American reported on 21 September. [19] Positive tests are rising too, but hospital admissions have not followed yet. [19] Nearly half of current samples are descendants of a strain called XFG, which has led infections for more than a year, so most immune systems have met something close to it. [19]

Why it matters — This year's updated vaccines are aimed at XFG. [19] Wastewater, then positive tests, then hospital beds is the usual order, so doctors are watching for the third step. [19] Infection still causes long COVID, though nobody can yet say how often. [19]
08
Seven in ten US cancer patients pass five years

In January 2026 the American Cancer Society, a US charity that counts the country's cancers, reported that 70% of people diagnosed between 2015 and 2021 were alive five years later, compared with people without cancer. [14] That is the highest it has recorded, up from about half in the mid-1970s. [14] Survival with myeloma, a bone-marrow cancer, rose from 32% to 62% since the 1990s, and with liver cancer from 7% to 22%. [14]

Mid-1970s50%
Diagnosed 2015-2170%
Share of US cancer patients alive five years after diagnosis. The newest figure describes people diagnosed up to 2021.

Why it matters — The society's chief scientist named less smoking first among the reasons: the share of active smokers fell from 44% to 11% since the 1960s. [26] The same report projected more than 2 million new US cases and over 625,000 deaths in 2026. [14]
09
First world map of breast cancer survival

In July the WHO published the first five-year breast cancer survival estimates for all 194 of its member countries, for women diagnosed from 2017 to 2021. [15] The global middle figure was 77.8%. [15] It ranged from 39.1% in the WHO's African region to 88.5% in the Americas, and from 41.9% in low-income countries to 87.3% in high-income ones. [15][27] Only 67 countries had cancer registries with real survival data, so the rest were estimated. [15]

WHO African region39.1%
Europe84%
The Americas88.5%
Share of women alive five years after a breast cancer diagnosis, 2017 to 2021, by WHO region.

Why it matters — Survival was lowest where breast cancer is found at a later stage. [15] The figures are meant as a starting point for the WHO's goal of cutting early breast cancer deaths by 2.5% a year. [15]
10
Richer US counties gained most against cancer

US cancer death rates fell 32% from 1991 to 2019. [16] A 2026 study of 2,954 counties in the British Journal of Cancer found that coastal, city and higher-income counties improved first and fastest. [16] The counties with the highest death rates to begin with also improved a lot. [16] The data came from more than 21 million cancer deaths on death certificates. [16]

Why it matters — The national fall is real, and where a person lives changed how much of it reached them. [16] The authors call for programmes aimed at the places left behind. [16]
11
Medicaid expansion and five-year cancer survival

Medicaid is US public health insurance for people on low incomes, and from 2014 most states widened who could get it. [28] A 2025 study compared 1,423,983 cancer cases in adults aged 18 to 59 in 26 states that widened it and 12 that did not. [28] Five-year survival rose more in the states that widened it, by about 2.6 percentage points more in rural areas and about 1.5 more in high-poverty areas. [28]

Why it matters — The gains were biggest for pancreatic, lung and bowel cancer. [28] As of June 2025, ten states, mostly in the South and Midwest, had still not widened it. [28]
12
Lung cancer survival tracks county smoking

A 2026 study linked US cancer registry data covering 85% of the population to county smoking rates, air pollution and social conditions. [29] People with lung cancer in counties where more than 20% of adults smoke had a 14% higher risk of dying than those in counties with the lowest smoking. [29] Living with fine air pollution above 12 micrograms per cubic metre also went with lower survival. [29]

Why it matters — The study used county figures, not each patient's own smoking, so it shows where the risk gathers rather than why. [29] The authors point to helping people quit and cutting pollution. [29]
13
Red-light therapy: some real uses, much hype

Helmets, masks and beds that shine red and near-infrared light are sold for everything from ageing skin to ADHD, and one projection puts the market above $1 billion by 2030. [10] A 2025 review by more than 20 specialists found it worked for some ulcers, nerve damage, skin burns from radiation and pattern hair loss. [10] The US FDA has approved one device, for dry age-related macular degeneration, an eye disease. [10] Scientific American reviewed the science in March 2026. [10]

Why it matters — The leading idea is that an enzyme in mitochondria, the cell's power plants, absorbs the light and makes more energy. [10] Brain benefits have been seen in mice with a Parkinson's-like illness, but getting enough light through a human skull is still unsolved. [10]
14
High blood pressure in your thirties

Researchers used national health insurance records to follow 291,887 Koreans from age 30 to 40, then for about ten years more. [30] Those whose top blood-pressure number ran about 10 points above their peers for a decade had a 27% higher risk of heart disease after 40. [30] Their risk of kidney disease was 22% higher. [30] The work was presented at an American Heart Association meeting and reported on 4 April 2026, before peer review. [30]

Why it matters — Doctors quoted said raised readings in people under 40 are often watched rather than treated. [30] The study added up how high and how long, rather than using one reading. [30]
15
Five minutes more a day

An analysis of data from 150,000 adults in the UK, the US and Scandinavia was reported by BBC Future in May 2026. [31] It estimated that five extra minutes a day of moderate activity could prevent about one in ten early deaths. [31] Moderate means brisk walking, cycling or climbing stairs. [31] The gain is largest for people who do almost nothing, and smaller for people who are already active. [31]

Why it matters — It is an estimate from data on how active people were, not a trial. [31] It points the same way as step-count studies, which find most of the benefit at the low end. [32]
16
7,000 steps did most of the work

A review of 57 studies with more than 160,000 people who wore step counters for several years found 7,000 steps a day brought most of the benefit linked to 10,000. [32] Compared with 2,000 steps, 7,000 went with a 25% lower risk of heart disease and a 47% lower risk of dying during the studies. [32] It was published in Lancet Public Health in August 2025. [32]

Why it matters — Going past 7,000 added only a modest extra benefit. [32] Like the stairs review, it shows a link, not proof that walking caused it. [32]
17
Very hard exercise mostly did not harm hearts

A review from Queensland University of Technology in Australia, accepted for publication in December 2025, read 14 studies on whether years of very long, hard exercise harm the heart. [33] Most found it was not harmful. Some found the benefit stops growing beyond a certain amount. [33] Only two suggested extra exercise could become harmful, and both had serious problems in how they were done. [33]

Why it matters — Every one of the 14 studies had possible sources of bias, which the authors say weakens them. [33] WHO guidelines advise at least 150 to 300 minutes a week of moderate activity. [33]
18
A migraine pill tried on endometriosis tissue

Endometriosis is tissue like the womb lining growing outside the womb, and it has few treatments that do not work through hormones. [34] Researchers looked for existing drugs that might act on a protein called ROR1, which is raised in endometriosis lesions. [34] Rimegepant, an approved migraine pill, slowed endometriosis cells in a dish. [34] In tissue grown from three patients' lesions, two responded and one did not. [34] The study was published in December 2025, after earlier work showed the drug shrank lesions in mice. [34]

Why it matters — Re-using a drug already approved for something else can be faster than inventing one. [34] Different answers from three patients' tissue show why a lab result is only a first step. [34]
19
Isolation and memory in 24 countries

A 2025 study pooled five long-running ageing studies covering 101,581 older people in 24 countries. [35] More social isolation went with lower scores on memory, orientation and planning tests. [35] The link was weaker in countries with stronger welfare systems and richer economies. [35] It was stronger among the oldest people, women and people with less money. [35]

Why it matters — The researchers used a method meant to reduce the chance that the link runs the other way, with failing memory making people withdraw, and the link held. [35] The WHO says isolation and loneliness affect about a quarter of older people. [36]
20
Nearly half of older people have two illnesses

A 2025 review of 49 studies covering 729,043 people aged 60 or over found that 46% had two or more long-term illnesses at the same time. [37] The share was 48% in women and 41% in men, and lower in studies from China than elsewhere. [37] The WHO says there were 1.1 billion people aged 60 or over in 2023, and the number will nearly double by 2050. [36]

Why it matters — The WHO estimates that about 15% of adults aged 70 and over live with a mental disorder, most often anxiety or depression. [36] Being 70 or older raised the odds of having several illnesses at once. [37]
21
Watching human brain cells pass messages

Johns Hopkins researchers hit living brain tissue with an electric pulse and froze it a fraction of a second later, to photograph brain cells passing messages. [38] They did it first in mouse tissue, then in tissue removed from six people having surgery for epilepsy. [38] The same recycling of tiny message-carrying bubbles seen in the mice was running in the human cells. [38] It was published on 24 November 2025 in the journal Neuron. [38]

Why it matters — The junctions where brain cells pass messages go wrong in Parkinson's disease. [38] The team hopes to compare inherited and non-inherited forms of the illness this way. [38]
22
Ultra-processed food and dementia risk

Harvard researchers followed 5,370 US adults with an average age of 64 for nearly nine years, testing memory and attention every two years. [13] Those who reported eating the most ultra-processed food, such as sugary drinks, deli meats and fast food, had a 58% higher risk of dementia than those who ate the least. [13] Those eating the most whole, little-processed food had a 41% lower risk. [13] It was published on 3 June 2026. [13]

Why it matters — Diet was asked about once, at the start. [13] The study cannot show that the food caused the extra dementia cases. [13] Ultra-processed food is also linked to diabetes, high blood pressure and obesity, which are known risks for memory loss. [13]
23
More US men diagnosed with prostate cancer

An American Cancer Society study published on 2 September 2025 found more US men were diagnosed with prostate cancer from 2014 to 2021, reversing the fall from 2007 to 2014. [39] Advanced cases rose by 6% a year in men aged 55 to 69, and by 6.2% a year in men aged 70 and over. [39] Black men were twice as likely to die of it as white men. [39]

Why it matters — The researchers said screening rates, access to care, environment and genes could all help explain the gaps between groups. [39]

02 Lesson why it matters

Research in mice carries over only where mice work like us

Scientists test ideas in mice because they cannot test them safely in people, so a result carries over only where a mouse's body works the way ours does.

The twist

Mice cope with low oxygen because their ancestors lived underground, so an ordinary mouse cannot show the brain damage that low oxygen causes in a newborn baby. The Stanford mice could show it, because the part being damaged was human.

The picture

Ordinary mousecopes

Part-human mousewalks badly

Both were short of oxygen. Only the mouse with human brain cells developed walking problems like a child with cerebral palsy, which is the gap that makes it useful.

How it works

Some experiments would harm a person, so they are tried in animals such as mice first
A mouse shares much of its body's machinery with ours
Where its body works differently, the result does not carry over to people
So researchers put human parts into the mouse, or check the result in human tissue
Only a test in people finally settles whether it works in people

The same force, elsewhere today

Where this chain is also running, in today's other stories.

Sugar during antibiotics

A mouse study found a sugary diet helped a gut germ take hold, and the New York team then checked the link in the real meals of 158 patients.
Red-light therapy

Light protected brain cells in mice with a Parkinson's-like illness, and the open step is whether enough light can get through a human skull.
Zap-and-freeze at Johns Hopkins

The method was run on mouse tissue first, then on tissue from six epilepsy operations, to check the same process runs in human cells.
A migraine pill for endometriosis

Instead of an animal, the team grew lesion tissue from three patients, and the drug worked on two of them and not on the third.

Where you've seen this

Crash-test dummies

a dummy shows what a crash does to a person only where it is built like a person

Model planes in wind tunnels

a small model predicts a real plane only if air flows over it the same way

Flight simulators

they train pilots well for the problems they were built to copy, and not for others

The catch

A part-human brain still sits in a mouse body with a short mouse life, so it shows how young human brain cells grow, not how a grown human brain works.

And the whole of it

A person who takes a new medicine for a brain illness meets the last step of a chain that usually began in an animal. Each person along that chain, from the lab bench to the clinic, sees only their own step of it.

03 Truth what's really going on

What is really going on

Stanford built mice whose brains are about half human by volume because drugs for brain illnesses that work in ordinary animals keep failing in patients. [2][4] The rules for how to treat such animals are still being written, and no international body yet oversees this work. [7][6]

Why it works on us — The words part-human mouse make a reader picture an animal that thinks like a person, and the scientists quoted say there is no sign of that. [2][3]

Who gains

Pasca's lab at Stanford — It now has a living animal in which to test treatments for illnesses like cerebral palsy, which it says it plans to do. [4]
Organoid researchers in the US — US health agencies pledged to rely less on animal tests, and the NIH announced $87 million for a new organoid centre. [6]
Nicole Saphier — The committee vote leaves only the full Senate between her and the surgeon general's job, after two earlier nominees failed. [20]
People with cancer in rural and high-poverty areas of states that widened Medicaid — Their five-year survival rose more than in states that did not widen it. [28]
Makers of red-light devices — Proven uses for a few conditions lend weight to a market projected above $1 billion by 2030, while experts warn of considerable hype around other uses. [10]

Who pays

The mice — They are bred without most of their own cortex, and pups born with a whole cortex are mostly culled so the others survive. [1]
Women with breast cancer in the WHO's African region — Five-year survival there was 39.1%, against 88.5% in the Americas, and survival was lowest where cancers are found late. [15]
Black men in the US — They were twice as likely as white men to die of prostate cancer. [39]
Cancer patients given sugary drinks during antibiotics — Doctors often suggest milkshakes or smoothies to boost nutrition, and sugary food went with a bigger loss of gut bacteria. [21]
Older people who are isolated in poorer countries — The link between isolation and weaker memory was stronger where welfare systems and economies were weaker. [35]

What nobody knows yet

Open questions from across today’s stories — ours included.

01

Whether the part-human mice feel or experience anything an ordinary mouse does not.

They behave much like ordinary mice, but there is no agreed way to measure this, and an ethicist on the project's own panel says she does not know how to regard them. [4][7]
02

Who will oversee research like this.

An international body was proposed in 2025, and its form, its rules and who should run it are still being worked out. [7][6]
03

Whether drugs tested in these mice will do better in patients.

The study did not test a drug in them, and the mice are expected to be used in a small number of labs for a few specific illnesses. [1][2]
04

Whether climbing stairs protects the heart or just marks people who are already fit.

Every study in the review watched people rather than assigning them, and a 2026 paper argues the link may run the other way. [11][12]
05

Whether eating less sugar during antibiotics would protect gut bacteria.

The New York study recorded what patients chose to eat, and nobody has yet given one group less sugar and compared the two. [21]
06

Which strains the WHO chose for the 2027 southern hemisphere flu shot, and how well they will match.

We could not read the WHO's published report on the choice, and a match can only be judged once that season's viruses circulate. [22][23]
07

Whether this autumn's COVID rise will reach US hospitals.

Virus in wastewater and positive tests are up, and hospital admissions have not risen yet. [19]
08

How many women actually survive breast cancer in most countries.

Only 67 of 194 countries had registry data, so the rest of the WHO's figures come from a statistical model. [15]

04 Hope carry this

Seven in ten people diagnosed with cancer in the US between 2015 and 2021 were alive five years later, the highest share the American Cancer Society has recorded. In the mid-1970s it was about half.

Also true today

Survival with myeloma, a bone-marrow cancer, rose from 32% to 62% since the 1990s, and survival with liver cancer from 7% to 22%.
Mice born without most of their cortex did almost as well as ordinary mice on a maze memory test once human brain tissue had grown into the gap.
A review of 14 studies found that years of very long, hard exercise mostly did not harm the heart.