Migraine is a slow wave across the brain and a messenger released outside it - and the drugs that stop it barely enter the brain

The rest of the day

43 more stories on this beat.

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

1. 02

   The first narcolepsy drug aimed at the cause

   The US Food and Drug Administration has approved oveporexton, sold as Orzeyful, for narcolepsy type 1. It switches on the receptor for orexin, the brain chemical that keeps people awake. [41] Narcolepsy type 1 affects about 1 in 2,000 people; the cells that make orexin are lost. [46] In two 12-week trials of 168 and 105 people, the time patients could stay awake on a 40-minute test rose by 14 to 20 minutes; with placebo it fell slightly. Attacks of sudden muscle weakness fell by 79% to 89%, against 28% to 39% with placebo. [42]

   Why it matters — Every earlier narcolepsy drug pushed the whole brain awake or damped the collapses; this is the first that stands in for the missing signal itself, which is why Nature's briefing compared it to the arrival of the weight-loss drugs. [46][41]

2. 03

   Most people on the new drug got side effects

   In the two trials, 86% to 89% of people on oveporexton had a side effect, against 43% to 54% on placebo. The commonest were urinating more often and short-lived insomnia, and a majority on the drug had them. [42] Its predecessor, TAK-994, was dropped after its 2023 trial because it damaged the liver at higher doses; oveporexton was rebuilt to avoid that, and its liver data so far are clean but short. [43] In rats, a drug of the same class did not directly raise saliva, so the drooling some volunteers reported may have another route. [47]

   Why it matters — Orexin helps the brain run many pathways in the body, not only sleep, so a drug that switches its receptors on reaches the bladder and the night as well as the day. [41][42]

3. 04

   Narcolepsy is being reclassified as an immune disease

   A review of the evidence through December 2025 argues narcolepsy type 1 belongs with the immune-mediated brain diseases. Immune T cells that recognise pieces of orexin have been found in patients, and post-mortem brains show them clustered where the orexin cells lived. [45] Almost everyone with the condition carries one gene variant, HLA-DQB1*06:02, which raises the risk strongly but does not decide it. [45] After the 2009 swine-flu pandemic, cases in children in Finland and Sweden rose more than tenfold, most sharply among the vaccinated. [45] Flu infection or vaccination is listed among the triggers. [46]

   Why it matters — The gene is common and the illness is rare, so what tips a carrier over is a second event, and in 2009 that event was a flu strain that looked, to the immune system, a little like orexin. [45]

4. 05

   The drug also sharpened attention

   A secondary analysis of the 112-person phase 2 trial, published in JAMA Neurology, tested thinking rather than sleepiness. On a ten-minute attention test, people on oveporexton made 9 to 11 fewer lapses than people on placebo, and they made fewer errors on a memory test; the effect sizes were medium to large. [44] Attention is the faculty narcolepsy damages most: an earlier pooled analysis put the deficit at about 0.9 standard deviations. [44] The corresponding author works for Takeda, the drug's maker. [44]

   Why it matters — Cognitive symptoms had never been measured in a trial of these drugs, and attention is where narcolepsy's deficit is largest. [44]

5. 06

   What the old narcolepsy drugs did instead

   Until now every narcolepsy drug treated a symptom. Modafinil raises dopamine and noradrenaline to push the brain awake; methylphenidate does the same more strongly and carries abuse potential; antidepressants damp the collapses. [46] Sodium oxybate, taken at night, is the only one that helps sleepiness, collapses and broken night sleep at once, but it loads the body with salt, and a low-sodium version with 92% less was made for that reason. [46] None of them, the review says, can reverse the disease. [46]

   Why it matters — Each of those drugs works on a different part of the brain from the one that is broken, which is why the review calls them symptomatic relief. [46]

6. 07

   How narcolepsy is diagnosed, and the trap

   A review for neurologists in Practical Neurology sets out the test. Narcolepsy type 1 is defined by sudden muscle weakness with emotion plus either falling asleep within 8 minutes on a daytime nap test with early dream sleep, or spinal-fluid orexin below 110 pg/mL. [49] In children, dream sleep can begin within 8 minutes of a nap, against 90 minutes normally. [48] The warning is that purely neurological causes of sleepiness are rare next to drugs, poor sleep and other illness, and that ordering tests and starting pills too early are the recurring mistakes. [49]

   Why it matters — A new drug approved only for type 1 makes getting the type right matter more than it did. [41][49]

7. 08

   A birthday cutoff in Wales, and dementia

   When Wales began shingles vaccination, people whose 80th birthday fell just after the start date could have the jab for a year. People born a few days earlier were never eligible. The two groups differ only in a few days of age, which makes them close to a randomised trial. [50] The same team had already reported fewer dementia diagnoses among the eligible. A new analysis finds fewer diagnoses of mild cognitive impairment too, and among people already living with dementia, fewer deaths from it. The effect did not belong to one type of dementia. [50]

   Why it matters — This is the closest thing to a trial that exists, and it says the vaccine helps at every stage, not only before the illness starts. [50]

8. 09

   100 million US records point at the virus itself

   A Nature Medicine analysis of health records from more than 100 million Americans, adjusted for nearly 400 characteristics, found that people with repeated shingles had a higher dementia risk than people with one bout. Shingles vaccines went with lower risk than a pneumonia vaccine used as the comparison. [51] For the older live vaccine, the protection against dementia faded over up to 15 years in step with its fading protection against shingles. Two doses of the newer vaccine protected more than one. [51]

   Why it matters — Protection that rises with dose and fades with the vaccine's own strength is the pattern a real cause leaves, and it points at the chickenpox virus waking up in the nerves. [51]

9. 10

   51% lower with the newer vaccine, at Kaiser

   Kaiser Permanente Southern California followed 65,800 members aged 65 or over who had two doses of the recombinant shingles vaccine between 2018 and 2020, against 263,200 matched unvaccinated members. Dementia risk was 51% lower in the vaccinated. [52] Because people who get vaccines are often healthier, the team also compared them with people who had a tetanus booster; the gap shrank but held, at 27% lower. The effect was stronger in women. [52]

   Why it matters — The tetanus comparison is the honest one, and the effect survived it. [52]

10. 11

    Medicare: 10.45 against 15.73 per 1,000

    In the US Medicare system, 502,845 people who had two doses of the recombinant shingles vaccine were matched to 1,005,690 who had none. New dementia arrived at 10.45 per 1,000 person-years in the vaccinated and 15.73 in the unvaccinated. [53] The hazard was 33% lower in the first three years and 26% lower after, with Alzheimer's disease and vascular dementia both reduced. [53]

    Why it matters — It is the third large dataset pointing the same way, and none of the three is a randomised trial. [53][52][51]

11. 12

    Two explanations, neither proven

    One camp says the virus does it. After chickenpox it sleeps for life in the sensory nerve clusters, and the sleep is not quiet: viral genes keep switching on and the immune system must hold them down. An opinion piece proposes that small, silent reactivations keep priming the brain's immune cells, and that the vaccine shrinks the reservoir. [54] The other camp says the adjuvant does it. BCG, the tuberculosis vaccine, given into the bladder for cancer, has gone with less dementia, and so has an RSV vaccine built on the same AS01 adjuvant as the shingles jab. [55]

    Why it matters — If it is the virus, only shingles vaccines will work; if it is the adjuvant, several vaccines already in use might. [55]

12. 13

    The NIH convened a workshop

    The US National Institutes of Health scheduled a two-day virtual workshop for June 1 and 2, 2026, titled Zoster Vaccine and Dementia Risk Reduction: Evidence and Mechanisms. Three institutes organised it together: ageing, infectious disease, and neurological disorders. [56] Its stated aim was to find the research gaps, and its agenda covered the biology of Alzheimer's disease and how a clinical trial of the idea might be designed. [56]

    Why it matters — A trial that randomises people to a vaccine and waits for dementia is the missing piece, and the agenda included how such a trial might be designed. [56]

13. 14

    Gout is a crystal, and the crystal needs a threshold

    Gout is what happens when urate, a waste product in the blood, forms needle-shaped crystals in a joint. Urate stays dissolved below about 6.8 mg/dL; above that the blood is oversaturated and crystals can form. [57] Roughly 20% of adult white American men have urate above that line, and the majority of them never get gout, though the risk climbs with the level. [57] When crystals do form, immune cells swallow them and read them as damage, switching on an alarm complex called NLRP3 that releases the signal IL-1 beta. White cells follow, and the joint swells. [58]

    Why it matters — The blood test finds the cause, and most people with the result stay well; the flare needs a second event on top. [57]

14. 15

    Starting the cure brings on the attacks

    Lowering urate is the cure for gout, and its first months are when the flares come. In the 6,183-person CARES trial, flare rates peaked in months 0 to 6, when urate-lowering started, and again in months 6 to 12, after the six months of preventive cover stopped. Only after the first year did flares track the urate level. [60] In a separate 940-man trial, flare risk while starting allopurinol or febuxostat was the same when the dose was raised slowly with cover; younger men, higher starting urate and no visible urate lumps predicted more flares. [59]

    Why it matters — The treatment itself sets off flares at first, which is why the cover exists, and the spike came back when the cover stopped. [60]

15. 16

    An antibody that stopped every flare

    Firsekibart is an antibody that binds IL-1 beta, the alarm signal a urate crystal sets off. In a 162-person phase 2 trial in people just starting urate-lowering treatment, one injection was compared with three months of daily colchicine, the usual cover. Nobody in the 200 mg group had a flare in twelve weeks. The 100 mg group averaged 0.02 flares per person against 0.34 with colchicine. No one stopped for side effects. [62] The trial was open-label, so everyone knew what they were getting. [62]

    Why it matters — It blocks the alarm rather than the crystal, and the comparison was against colchicine, the standard cover, in people starting treatment. [62]

16. 17

    Six months of colchicine cover did not pay

    In a 200-person trial, people starting allopurinol were given either low-dose colchicine or placebo for six months, then followed for six more. Costs were about 1,848 dollars higher per person in the colchicine group over six months and 2,282 dollars over a year. [61] Quality-adjusted life years were a little higher with colchicine at six months and a little lower at one year, and the chance that the cover was good value fell to 1.5% by twelve months. [61] The reason is in the CARES data: the flares the cover held back arrived once the cover stopped. [60]

    Why it matters — A preventive that only moves the flares later buys nothing, which is why the CARES authors suggest the cover may need to run longer, not be dropped. [60]

17. 18

    Reaching the urate target, and the heart

    Among 109,504 British patients newly prescribed a urate-lowering drug, only 27.3% got their urate below 6 mg/dL within a year. [63] Those who did had a 1% higher chance of being alive five years later and 9% fewer major heart events, in a study built to imitate a trial. [63] The team checked its method with control outcomes: reaching target went with fewer gout flares, as it should, and had no effect on bronchitis, cataracts or appendicitis, as it should not. [63]

    Why it matters — Gout was long treated as a sore toe; this says the crystal-forming blood is also a heart risk, and that nearly three quarters of patients never get it down. [63]

18. 19

    Only a third to half get definitive treatment

    A 2025 review of gout care reports that gout is rising worldwide, that only a third to a half of patients ever receive definitive treatment, and that fewer than half of those stay on it. [64] The recommended approach is to treat to a target urate of 5 to 6 mg/dL or lower, and to start the drug alongside three to six months of flare cover. [64] New drugs are coming: a stronger urate blocker called tigulixostat, diabetes drugs of the SGLT2 class, and drugs that make the kidney excrete more urate. [64]

    Why it matters — The review says the barriers are recognition of the disease and clinicians' adherence to the guidelines, not the drugs. [64]

19. 20

    Febuxostat's worse numbers, explained by its patients

    At Dresden University Hospital, 208 gout patients were followed for an average of 4.8 years. Among those mainly on febuxostat, 18.1% had a heart event, against 7.5% of those mainly on allopurinol; deaths were similar at 6% and 6.9%. [65] But the febuxostat group carried more of almost every risk factor to begin with, and the authors say the result cannot be read as the drug's fault. [65]

    Why it matters — The group that got the second-line drug was sicker to begin with, and then the second-line drug looked dangerous. [65]

20. 21

    Brussels: matched prevention cut kidney colic by three quarters

    A kidney stone is urine chemistry gone over a line, and a Brussels clinic tested treating the chemistry. In 490 stone formers followed for a median of 12.8 months, each patient's own risk factors were corrected. Urine volume rose by 653 ml a day, oxalate and calcium in the urine fell, and citrate, which keeps crystals from forming, rose. [66] Attacks of kidney colic fell from 0.82 to 0.21 per 1,000 days at risk, and procedures to remove stones from 0.31 to 0.21. [66] The more of the four risk factors were controlled, the further colic fell. [66]

    Why it matters — The stone is the event and the oversaturated urine is the cause, so treating the chemistry is treating the odds. [66]

21. 22

    A diabetes drug and stone chemistry

    Diabetes drugs of the SGLT2 class, which make the kidney dump sugar into the urine, cut stone events by 30% to 50% in diabetes trials, for reasons nobody had explained. [67] A Swiss trial called SWEETSTONE was designed to find out. Its 46 non-diabetic stone formers take empagliflozin for two weeks and a placebo for two weeks, in random order, while the lab measures how oversaturated their urine is with three kinds of crystal. [67] A guide for kidney specialists adds that thiazide water pills, the old stone preventive, now mean weighing the patient's diabetes risk. [69]

    Why it matters — If a drug changes the urine's saturation, that is the mechanism; if it does not, the 30% to 50% came from somewhere else. [67]

22. 23

    Calcium pills for thin bones, and stones

    In south-west China, 204 people aged 50 to 89 with osteoporosis had taken 600 mg of calcium carbonate and a vitamin D analogue daily for at least a year. Those who formed kidney stones on the treatment were the ones with a history of recurrent stones, and their urine carried nearly twice as much calcium a day, 1.00 against 0.57. [68] The study is a snapshot, not a trial, and it cannot say how many stones the pills caused. [68]

    Why it matters — The stones arrived in the people whose urine already carried the most calcium, and the authors put the risk down to that. [68]

23. 24

    Stones as the sign of something systemic

    A guide for kidney doctors on stone formers with chronic kidney disease says the two conditions feed each other. Repeated stones raise the risk of kidney failure, and 1.1% of people starting dialysis in the most recent cohort got there through stones. [69] Stones in a young person or with unexplained kidney disease can be the signature of an inherited disorder such as Dent disease or primary hyperoxaluria, and some of those now have gene-silencing drugs. [69] The workup is a 24-hour urine collection, taken while the patient eats and drinks as they normally would. [69]

    Why it matters — A stone is usually treated as a plumbing problem; this guide says it is often the first visible sign of a metabolic one. [69]

24. 25

    A second messenger, a single infusion

    PACAP is a second messenger, related to CGRP, that can start a migraine. Lu AG09222 is an antibody that mops it up. In a phase 2 trial published in the New England Journal of Medicine, 237 people whose two to four earlier preventives had failed got one infusion. Over four weeks, migraine days fell by 6.2 a month on the 750 mg dose against 4.2 on placebo, a difference of two days. [21] In healthy volunteers the same antibody prevented the vessel-widening and the headache that an infusion of PACAP normally causes. [22]

    Why it matters — It is a new target, and the trial chose people whom two to four earlier preventives had already failed. [21]

25. 26

    Insurers make patients stop after a year; restarting works

    In Switzerland, the insurer pays for a CGRP antibody for 12 months; then the patient must stop, and may restart if the migraine returns. A Swiss clinic followed 42 patients who did that, after a median pause of 140 days. [23] In the second year the antibody cut migraine days as much as in the first, and the time to halve them was 1.84 months against 2.20 the first time. [23] Guidelines used to recommend a pause after 6 to 12 months; the European Headache Federation now says to continue as long as needed. [23]

    Why it matters — The pause was written for the insurer's year, not the patient's brain, and the headache returns within months of it. [23]

26. 27

    Half responded, or two thirds, depending on the ruler

    In 417 patients on a CGRP antibody at one centre, 50.3% of those with episodic migraine counted as responders by migraine days. By a disability questionnaire it was 69.5%, and by a headache-impact score 67.5%. [24] Counting anyone who improved on any of the three, 84.8% responded. Only 3.3% stopped for side effects. [24] A separate review of 11 chronic-migraine studies flagged the same problem: the studies do not agree on what counts as working. [40]

    Why it matters — A patient whose attacks are as frequent but far lighter is a non-responder on the count that decides whether the insurer keeps paying. [24]

27. 28

    People the trials excluded, treated anyway

    The CGRP trials excluded people with heart risk, cancer, immune suppression and uncontrolled blood pressure. A Spanish study, SAFE-CGRP, looked at exactly those patients across Spanish headache units, including 71 with a cancer history and 23 at moderate-to-high heart risk. [25] Medication overuse fell from 64% of the group to 32% on treatment. Fourteen possible drug-related worsenings were found: four of high blood pressure, seven of Raynaud's, two arthritis flares and one angioedema. All improved when the drug was stopped, and there were no severe events. [25]

    Why it matters — The people most likely to be harmed by blocking a vessel-widener are the ones the trials never tested, and this is the first look at them. [25]

28. 29

    Three in four still on the injection at six months

    At Sutter Health, a Californian health system, 4,683 patients had a self-injected CGRP antibody prescribed, and 484 answered a survey. Three quarters were still on it at six months and more than half at twelve; the commonest reason for stopping was that it did not work well enough. [26] Across long-term studies, 23% stop within a year for any reason and 3% for side effects; the commonest side effect recorded is a cold. [10] Elsewhere, 80 patients a year in rated satisfaction at 77 out of 100, and it tracked the headache-impact score, not the day count. [39]

    Why it matters — Traditional preventives, borrowed from other diseases, lose about half their users; these keep more, and the ones who leave mostly leave because the drug did not work. [26]

29. 30

    The brain's blood flow settles when the drug works

    Researchers scanned blood flow in the brains of 73 migraine patients before and after CGRP antibody treatment, between attacks. Patients whose cortex was over-perfused before treatment and whose flow fell afterwards, and patients without over-perfusion whose flow rose, both moved towards normal. [27] In the second group, 94% halved their headache days. Overall, 74% of the 73 halved their headache days and 32% stopped having them. [27]

    Why it matters — The drug barely enters the brain. [1] The brain's blood flow changed anyway, so whatever the drug does at the covering reaches the cortex. [27]

30. 31

    Blood CGRP: higher in one study, lower in another

    If CGRP drives migraine, blood levels should show it. A registry study measured plasma CGRP in 588 people with migraine and 147 matched controls and found it lower in migraine: 125 against 151 pmol/L, with no difference between subtypes or during attacks. [28] A case-control study of 184 women found it higher, and the high group had four attacks a month against three. [29] Tears may be the better window: in 26 patients CGRP in tear fluid was double that of controls and rose further at the peak of an induced headache. [20]

    Why it matters — Tear fluid sits beside the trigeminal nerve, which is why the authors call it a marker of that nerve's activity; the blood is a long way from it. [20]

31. 32

    The aura wave can set off a seizure-like burst, in rats

    The slow wave behind aura was triggered in the hippocampus of awake rats, the memory region that also regulates mood. When it reached the lower hippocampus it set off a seizure-like discharge on both sides of the brain, including the side the wave never touched, and the rats shook like wet dogs. [30] It is the first experimental evidence that the wave can trigger seizure-like activity in a brain without epilepsy. [30] In hemiplegic migraine, a rare form with one-sided weakness, EEG during the aura shows flattened activity on the affected side in 71% of recorded cases. [31]

    Why it matters — The authors suggest it may explain the seizures that sometimes follow a migraine aura. [30]

32. 33

    When a migraine is something else

    A systematic review of case reports from 1977 to 2024 asked what else produces migraine-like headaches. Stroke, epilepsy and head injury were the commonest causes found. [32] Only 11 cases met the strict definition of a symptomatic migraine, and all of them had aura; five were tangled blood vessels in the brain, two were narrowed neck arteries and one a torn one. [32] Most patients in the reports had an MRI, and the drugs used were the ordinary migraine ones: anti-inflammatories in 30.5% and triptans in 18.2%. [32]

    Why it matters — All eleven definite cases had aura, so a migraine-like headache with aura is where the rare causes in this review sat. [32]

33. 34

    18.8% of students with PMS had migraine

    A multi-centre study screened female college students with premenstrual syndrome and found migraine in 18.8% of them. The students with migraine slept worse and scored higher for depression, anxiety and stress. [33] Premenstrual syndrome itself is reported in anywhere from 12% of women in France to 98% in Iran, a spread the authors put down to how it is measured; the pooled estimate is 47.8%. [33]

    Why it matters — Migraine is three times commoner in women and new cases peak at 18 to 24. [16] This is what that looks like in a lecture hall, with the menstrual cycle on top. [33]

34. 35

    Face flushing and migraine share a messenger

    Rosacea is a chronic skin condition of the face that flushes and stings. A review argues it and migraine are one kind of disorder in two places. Both run on the same messengers, CGRP and PACAP, released from sensory nerve endings, and both involve the same heat-and-irritant sensors, the TRP channels, and the same nerve-driven inflammation of small vessels. [34] The face is served by the trigeminal nerve, the same nerve that serves the brain's covering. [1]

    Why it matters — If the two conditions are one mechanism, the CGRP drugs made for one may be tested on the other. [34]

35. 36

    Migraine and the body's energy supply

    Migraine affects an estimated 12% of Americans, and each year about 2.5% of people with episodic migraine slide into the chronic form, 15 or more headache days a month. [13] One line of research treats migraine as a brain energy shortfall: attacks follow fasting and low blood sugar, and people with insulin resistance report worse attacks. [13] A review of the gut's role puts the prevalence at 18.9% of women and 11.4% of men and the annual cost to the United States at 20 billion dollars, and lists diet, gut bacteria and inflammation among the influences. [35]

    Why it matters — The trigeminal story explains the pain; it does not yet explain why one brain has a lower threshold on a given day, and energy is one candidate. [13]

36. 37

    In Israel, the new drugs did not change who gets prevention

    A national study in an Israeli health fund compared 356,441 migraine patients with 4,257,890 controls. Three quarters were women, the average age was 31, and new cases peaked in women aged 18 to 24 at 7.5 per 1,000 a year. [16] Patients carried more other illnesses than controls, with a higher stroke risk and less diabetes. [16] Only 9.6% used any preventive drug in 2018 and 8.8% in 2022, after the CGRP drugs had arrived. [16]

    Why it matters — A drug with a number-needed-to-treat of three to six changes nothing if fewer than one patient in ten is offered prevention at all. [9]

37. 38

    Four pills against the same messenger

    The antibodies are injections; the gepants are pills and a nasal spray that block the CGRP receptor. Four are approved: atogepant for prevention, ubrogepant and zavegepant for attacks, and rimegepant for both. [12] In trials they cut monthly migraine days and gave pain freedom at two hours more often than placebo, with mostly mild side effects. [12] Among the antibodies, a network analysis of 16 trials and 9,123 people ranked galcanezumab highest for effect and fremanezumab best for the balance of effect and safety. [15]

    Why it matters — Migraine is the leading cause of disability in women under 50, and the treatments for it now come in three forms instead of one. [14]

38. 39

    What the wave leaves behind in neurons

    When the slow wave crosses a mouse cortex, neurons shed tiny membrane packets called vesicles, and a proteomic study read what was in them. The cargo showed neurons rewiring their gene expression, their scaffolding, their stress response and their metabolism, not only inflammation, which the authors call an adaptive response. [36] The same wave runs in injured brains: it complicates more than half of severe acute brain injuries and adds to the damage after a stroke by overloading the tissue's energy supply. [17][37]

    Why it matters — A migraine aura is a healthy brain doing, briefly, what a stroke-damaged brain does repeatedly, and a vesicle in the blood might one day show it happened. [36]

39. 40

    The 200 papers migraine research stands on

    A bibliometric study pulled the 200 most-cited migraine papers and mapped them. They came from 45 journals and 45 countries, were written by 4,409 researchers at 1,592 institutions, and the most influential authors were Richard Lipton and Peter Goadsby. The United States led in volume. [38] The hot topics are treatment, the CGRP system, genetics, the trigeminal pathway and, lately, the gut. [38]

    Why it matters — The map says where the attention goes: treatment and CGRP first, and lately the gut. [38]

40. 41

    Over 50 genes linked to stuttering

    A review of the genetics of stuttering, a speech disorder that affects 1% to 3% of people, counts more than 50 associated genes found so far, many near genes already tied to neurological traits. [70] The data came from four sources: a stuttering research project, a US adolescent survey, a hospital biobank and the customer database of 23andMe. [70] Hospital billing codes captured a prevalence of 0.01%, a hundredth of the real figure, which is why the studies had to recruit directly. [70]

    Why it matters — A condition that hospital records count at one hundredth of its real rate cannot be studied from hospital records, which is why the genes came from volunteers and a consumer DNA company. [70]

41. 42

    Speech rhythm, musical rhythm and songbirds

    The ability to hear the rhythm of speech, called prosody, was tested in 1,501 people and their genomes scanned. Fourteen suggestive signals turned up. The genes involved were enriched in the brain regions songbirds use to learn their songs, and shared inheritance with word reading and with keeping a musical beat. [71] It is the first genome-wide study of the trait, and small, so the authors call it initial. [71]

    Why it matters — The ear for the rise and fall of a sentence may run on the same machinery as a bird learning a tune, and it predicts how a child reads. [71]

42. 43

    Teaching babies to talk before the delay shows

    Some genetic conditions predictably delay speech, and a paper on precision medicine in speech therapy describes intervening before the delay appears. Babble Boot Camp coached parents of infants with classic galactosemia, a metabolic disorder with a known speech risk, from the first months of life, and a clinical trial found it worked; pilots are running in Down syndrome. [72] The paper also describes a shared trait in some dyslexia, a weaker filtering of sensory input, as a possible target. [72]

    Why it matters — For most children the delay is the first sign; a genetic diagnosis moves the start of help to before there is anything to see. [72]

43. 44

    Your DNA repeats lengthen as you age

    Stretches of repeated DNA letters can grow, and some, when they grow far enough, cause inherited diseases. A study of more than 900,000 people in the UK Biobank and the US All of Us programme measured them. Common repeats in two genes lengthen in blood cells as people age, so most genomes carry repeats expanding across a lifetime. [73] Twenty-nine inherited variants sped up or slowed that growth, and an expanded repeat in a gene called GLS went with a 14-fold higher chance of end-stage kidney disease. [73]

    Why it matters — A genome is usually described as fixed at birth; in these stretches it is still being written, at a pace set by other genes. [73]