The Brain

The Three-Pound Universe

Lift a bag of sugar and you are holding about the weight of the thing doing the lifting. The adult human brain weighs around 1.3 to 1.4 kilograms, roughly 2 per cent of body mass, yet it burns about 20 per cent of the energy the body uses at rest. It runs on some 20 watts, less than a dim light bulb, and it has no reserve tank: cut its blood supply and consciousness fades within about ten seconds. Around three quarters of it is water, and most of the dry remainder is fat, much of it wrapped around nerve fibres as insulation.

For most of history nobody was sure the organ mattered. Egyptian embalmers hooked it out through the nose and threw it away while carefully preserving the heart. Aristotle thought the brain was a radiator for cooling the blood. The Hippocratic writers of around 400 BC disagreed, placing joy, grief and madness in the head, and the Roman physician Galen, who had treated gladiators at Pergamon, showed that cutting a nerve in the neck of a squealing pig silenced it: the voice was commanded from the head, not the heart.

Counting its cells took until 2009. In Rio de Janeiro the neuroscientist Suzana Herculano-Houzel dissolved whole brains into a uniform soup of free-floating cell nuclei, stained them and counted samples under the microscope. Her answer was about 86 billion neurons, not the round 100 billion repeated in textbooks, plus a roughly equal number of other cells. The surprise was where they live. Around 69 billion sit in the cerebellum at the back of the head, while the wrinkled cerebral cortex, home of planning, language and memory, holds about 16 billion.

Size alone explains little. An African elephant's brain weighs about 5 kilograms and contains some 257 billion neurons, but nearly all of them are in its cerebellum; its cortex has about a third as many as ours. Albert Einstein's brain, removed by the pathologist Thomas Harvey in Princeton in 1955, weighed 1,230 grams, slightly below the average for a man. What marks out humans is the sheer number of neurons packed into the cortex, and the way they are wired together.

Did you know? The brain has no pain receptors of its own, which is why the surgeon Wilder Penfield in Montreal could stimulate the exposed brains of awake patients, who reported tingles, music and sudden scraps of memory.

The popular idea that we use only 10 per cent of our brains is a myth: scans show activity across the whole organ over the course of a day, and damage to almost any part has consequences. The everyday lesson is about upkeep. An organ this hungry and this delicate depends on steady blood flow, regular meals, decent sleep and a well-run body around it. You carry the most complex object known to science on your shoulders, and it is worth feeding properly.

How Brains Evolved

The first nervous systems had no brain at all. In jellyfish, sea anemones and the freshwater polyp Hydra, neurons form a loose nerve net spread through the body wall, so a touch anywhere ripples outwards like rings on a pond. Animals built this way were drifting through the oceans more than 550 million years ago, and they still do perfectly well: the box jellyfish steers through mangrove swamps with 24 eyes and only a ring of nerves to coordinate them.

The decisive change came with animals that had a front and a back. Once a creature moves head first, it pays to gather sense organs and neurons at the leading end, a trend called cephalisation. Flatworms such as planarians have a pair of head ganglia with nerve cords running down the body; cut one in half and the tail end regrows a new head, brain included. By about 500 million years ago early vertebrates had the three-part plan still visible in every fish, frog and human embryo: forebrain, midbrain and hindbrain.

Mammals added the neocortex, a sheet of six cell layers over the forebrain, more than 200 million years ago. In small mammals such as mice it stays smooth; in larger ones it buckles into folds that pack more surface into the skull, and unfolded the human cortex covers about a quarter of a square metre. Fossil skulls trace the growth in our own line. Australopithecus, like the 3.2-million-year-old skeleton nicknamed Lucy from Hadar in Ethiopia, had a brain of around 400 to 500 cubic centimetres; Homo erectus reached about 900; modern humans average roughly 1,350.

We know this from comparative anatomy, from endocasts (natural or plaster moulds of the inside of fossil skulls) and increasingly from genes. A gene called NOTCH2NL, duplicated only in the human lineage, keeps brain stem cells dividing for longer; its discovery in 2018 by teams in California and Belgium offered one molecular clue to our enlarged cortex.

Evolution also found other routes to cleverness. The octopus has about 500 million neurons, two thirds of them in its arms, in a nervous system that evolved separately from ours. Crows and parrots, with walnut-sized brains, pack their forebrains so densely that some rival monkeys in neuron count. Even the tiny worm C. elegans, with exactly 302 neurons, can learn and remember, and in 1986 its complete wiring diagram became the first ever mapped.

Did you know? The popular "lizard brain" story, in which a reptilian core is wrapped by an emotional layer and then a rational one, comes from Paul MacLean's 1960s triune-brain theory, which neuroscientists long ago abandoned: evolution remodelled old parts rather than stacking new ones on top.

The lesson is humbling. The human brain is not a finished masterpiece but a renovated old house, sharing its basic plan with a lamprey's and running ancient reflexes underneath. That is why fear can outrun reason, and why sleep, appetite and breathing are handled by parts far older than thought.

Building a Brain

Every brain begins as a flat sheet. About 18 days after conception, a strip of cells along the back of the embryo thickens into the neural plate. Its edges rise, curl towards each other and zip shut into the neural tube, a process finished by about day 28, often before a woman knows she is pregnant. The front of the tube becomes the brain, the rest the spinal cord. If it fails to close, the result is anencephaly at the head end or spina bifida lower down.

That is why folic acid matters. In 1991 the Medical Research Council Vitamin Study, run in 33 centres across seven countries, showed that folic acid taken before and in early pregnancy cut the recurrence of these defects by about 72 per cent. Dozens of countries, including the United States from 1998, now add it to flour, and the United Kingdom has legislated to follow.

By week five the front of the tube has ballooned into five swellings, the forerunners of the cerebral hemispheres, the thalamus, the midbrain, the cerebellum and pons, and the medulla. Then comes a manufacturing frenzy. Over pregnancy, new neurons are made at an average of roughly 250,000 a minute. They are born deep inside, beside the fluid-filled ventricles, and must travel outwards. In the 1970s Pasko Rakic at Yale showed that they climb the long fibres of radial glia like rope ladders, building the cortex inside out: the earliest arrivals form the deepest layer and each later wave passes them to settle above.

The smooth fetal brain begins to fold in the middle of pregnancy, and the main grooves deepen over the last three months. A baby born at term has a brain of about 350 to 400 grams, roughly a quarter of adult weight, which nearly doubles in the first year. The running costs are enormous: Christopher Kuzawa and colleagues estimated in 2014 that at around age five the brain consumes the equivalent of nearly two thirds of the body's resting energy.

After birth, growth is about wiring rather than new neurons. Synapses multiply wildly, then are pruned. Peter Huttenlocher in Chicago counted synapses in post-mortem brains from the late 1970s and found that density in the cortex peaks in early childhood, well above adult levels, before the surplus is trimmed; the visual cortex peaks within the first year, the prefrontal cortex years later. Experience decides what stays. In the 1960s David Hubel and Torsten Wiesel showed that a kitten with one eye kept closed during a critical window never properly wired that eye to its cortex. Insulating myelin reaches the frontal lobes last, into the mid-twenties.

Did you know? In 2015 and 2016 a surge of babies born with unusually small heads in north-east Brazil was traced to the Zika virus, which attacks exactly the stem cells that make cortical neurons.

A brain, then, is built by experience as much as by genes. Folate before conception, no alcohol in pregnancy, and talk, play and sleep afterwards are not optional extras; they are the building materials.

The Neuron

A neuron is a cell built to collect signals, make a decision and pass the verdict on. Its branching dendrites gather inputs from other cells, often thousands of them, through tiny thorn-like spines. The cell body, or soma, holds the nucleus and the protein factories that keep the whole sprawling structure alive. Where the body narrows into the axon sits the trigger zone, the axon initial segment: if the inputs added together push its voltage past a threshold, the neuron fires.

The axon carries that spike away, sometimes a long way. Motor neurons in the lower spinal cord send axons about a metre to the muscles of the foot. Many axons are wrapped in myelin, fatty insulation laid down in segments, with bare gaps called nodes of Ranvier where the signal is renewed. At its far end the axon splits into terminals that form synapses on the next cells. In most neurons information flows one way: in through the dendrites, out along the axon.

Nobody could see this until 1873, when the Italian physician Camillo Golgi, working in a hospital kitchen turned laboratory at Abbiategrasso near Milan, found that silver nitrate stained a few neurons at random, black and complete, against a clear background. The Spaniard Santiago Ramón y Cajal used Golgi's stain to argue that each neuron is a separate cell rather than part of a continuous web. The two shared the Nobel Prize in 1906 while disagreeing in their lectures, and electron microscopes proved Cajal right in the 1950s.

Neurons come in striking shapes. Pyramidal cells of the cortex have a triangular body and a long dendrite reaching towards the surface. Purkinje cells in the cerebellum spread a flat fan of dendrites that receives up to about 200,000 inputs. Some neurons are astonishingly specific. In 2005 Rodrigo Quian Quiroga, Itzhak Fried and colleagues, recording from epilepsy patients in Los Angeles who had electrodes implanted for surgery, found single cells in the medial temporal lobe that fired for pictures of Jennifer Aniston, or Halle Berry, or the Sydney Opera House, and hardly at all for anything else.

Did you know? The "Halle Berry neuron" found in 2005 also fired when the patient simply read her name, suggesting it coded the idea of her rather than any one picture.

Most of the neurons you have now were made before birth and are never replaced, which is why diseases that kill them are so hard to reverse; Stephen Hawking lived with motor neurone disease for more than five decades as it destroyed the cells that move muscles. Whether adults grow new neurons in the hippocampus is still debated, with studies in 2018 reaching opposite conclusions. What is certain is that existing neurons keep changing, sprouting and losing spines as you learn and forget. Each one is less a wire than a tiny calculator, and you own about 86 billion of them.

Glia: The Other Half

For a century neuroscience treated most of the brain's non-neuronal cells as packing material. In 1856 the German pathologist Rudolf Virchow named them neuroglia, "nerve glue", imagining a connective putty holding neurons in place. The truth turned out to be far busier. Glia make insulation, feed neurons, mop up chemicals, police infections and help decide which connections survive. The often-repeated claim that glia outnumber neurons ten to one is wrong: careful counts since 2009 find roughly one non-neuronal cell for every neuron across the brain as a whole.

The largest family are the astrocytes, star-shaped cells whose fine processes wrap blood vessels at one end and synapses at the other. A single human astrocyte can touch up to about two million synapses. Astrocytes take up spent glutamate so that signals stay crisp, soak up excess potassium, pass fuel from the blood to neurons and help maintain the blood–brain barrier. They even signal to one another with slow waves of calcium, a quieter conversation running beneath the electrical one.

Oligodendrocytes make myelin in the brain and spinal cord, each one wrapping segments of up to about 50 different axons; in the nerves of the body the job falls to Schwann cells, which wrap one segment each. The Spanish neuroscientist Pío del Río Hortega identified oligodendrocytes and the brain's resident immune cells, the microglia, between 1919 and 1921, using silver stains of his own devising. Microglia never rest: they constantly extend and retract their branches, sampling their surroundings. In 2012 Beth Stevens and colleagues in Boston showed that in young mice microglia swallow synapses that have been tagged by immune proteins, pruning the developing circuit. A fourth group, ependymal cells, line the fluid-filled ventricles, while in the embryo radial glia act as both stem cells and scaffolding.

When glia go wrong the effects are severe. Multiple sclerosis is an immune attack on myelin and the oligodendrocytes that make it. Most malignant brain tumours arise from glial cells, among them the glioblastoma that killed the American senator John McCain in 2018. And when the neuroscientist Marian Diamond in Berkeley examined slices of Einstein's brain in 1985, the feature she highlighted was not his neurons but a higher ratio of glia to neurons in one region, a result critics found hard to interpret.

Did you know? In 2013 researchers in Rochester, New York, transplanted human glial progenitor cells into newborn mice; the animals grew human-type astrocytes and learned several tasks faster than ordinary mice.

The everyday point is that brain health is not only about neurons. Myelin keeps maturing into adulthood and responds to practice; microglia are shaped by sleep, infection and inflammation; astrocytes manage the fuel supply. The other half of the brain does the housekeeping, and without it the house would fall down.

The Action Potential

Every thought you have travels as a crackle of electrical spikes. At rest the inside of a neuron sits at about −70 millivolts compared with the outside, held there by a molecular pump that pushes three sodium ions out for every two potassium ions it brings in. The Danish chemist Jens Skou discovered this sodium–potassium pump in 1957 and shared a Nobel Prize for it forty years later. Simply keeping the battery charged consumes a large share of the brain's energy.

When inputs nudge the axon's trigger zone up to a threshold of roughly −55 millivolts, voltage-gated sodium channels snap open. Sodium floods in, and within about a millisecond the inside swings to around +40 millivolts. The sodium channels then shut themselves, and potassium channels open, letting positive charge out and driving the voltage back down, briefly overshooting below rest. During this refractory period the patch of membrane cannot fire again, so the spike can only travel forwards, regenerating itself along the axon like a burning fuse.

This is the action potential, and it is all-or-nothing: a neuron does not fire a bigger spike for a stronger stimulus, it fires more often, up to several hundred times a second. Alan Hodgkin and Andrew Huxley worked out the mechanism at the Marine Biological Association laboratory in Plymouth using the giant axon of the squid, up to about a millimetre thick, wide enough to slide a wire electrode inside. They recorded a spike from within the axon in 1939, were interrupted by the Second World War, and published their mathematical model in 1952; it won them a Nobel Prize in 1963 and is still used today.

Speed depends on insulation. In thin, bare axons spikes crawl at around one metre per second; in thick, myelinated ones they leap from node to node at up to about 120 metres per second, a trick called saltatory conduction. That is why pulling your hand from a hot pan is quick but the slow, burning ache arrives a moment later along thinner fibres. In 1976 Erwin Neher and Bert Sakmann in Göttingen went further, recording the tiny current through a single ion channel with a glass pipette sealed onto a cell membrane.

Did you know? Tetrodotoxin, the poison in Japanese fugu pufferfish, plugs sodium channels so that nerves cannot fire, which is why chefs in Japan need a special licence to prepare the fish.

The idea of animal electricity is older still: in Bologna in the 1780s Luigi Galvani made dissected frog legs twitch with metal and sparks. Today the same channels explain everyday medicine. Lidocaine at the dentist blocks sodium channels in pain fibres, and many epilepsy drugs calm overexcitable ones. Your sensations, decisions and movements are all written in one code of near-identical spikes; what changes is only where and how often they fire.

The Synapse

Neurons do not touch. Between the end of one and the surface of the next lies a gap about 20 nanometres wide, roughly one four-thousandth of the width of a human hair. The British physiologist Charles Sherrington named this junction the synapse in 1897, from the Greek for "clasp", long before anyone could see it. For decades scientists argued over how signals crossed: the "sparks" camp said electrically, the "soups" camp said chemically.

The soups won, helped by a dream. On the night before Easter Sunday in 1921 the pharmacologist Otto Loewi, working in Graz in Austria, woke with an idea, scribbled it down and in the morning could not read his own writing. The dream returned the next night, and this time he went straight to the laboratory. He stimulated the vagus nerve of one frog heart, slowing it, then transferred the fluid bathing it to a second heart, which slowed too. A chemical, later identified as acetylcholine, had carried the message.

Here is what happens, all within about a millisecond. A spike arrives at the axon terminal and opens calcium channels. Calcium triggers tiny bubbles called vesicles, each holding a few thousand molecules of neurotransmitter, to fuse with the membrane using a molecular zipper of SNARE proteins, a discovery that earned Thomas Südhof, James Rothman and Randy Schekman a Nobel Prize in 2013. The transmitter spills into the cleft, binds receptors on the far side and opens or closes channels there. Then it is cleared away, sucked back by transporters or broken down by enzymes, so the line is ready for the next message. In the 1950s Bernard Katz in London showed that transmitter is released in these fixed packets, or quanta. A minority of synapses skip the chemistry altogether and pass current directly through channels called gap junctions.

Your brain has something like 100 trillion synapses, and many poisons work by sabotaging them. Curare, the South American arrow poison studied by Claude Bernard in Paris in the 1850s, blocks acetylcholine receptors on muscles. Botulinum toxin, the Botox of cosmetic clinics, cuts the SNARE proteins so no transmitter is released. Nerve agents such as the Novichok used in Salisbury in 2018 do the opposite, jamming the enzyme that clears acetylcholine so muscles cannot relax.

Did you know? The slogan "neurons that fire together wire together" is not Donald Hebb's own phrase from 1949; it is a later paraphrase popularised by the neuroscientist Carla Shatz in 1992.

Synapses are where learning lives. Each one can be strengthened or weakened by use, which is how practice turns into skill and how a single striking event can leave a lasting memory. Most psychiatric drugs act here too. Every thought you have is a cascade of tiny chemical splashes across gaps you will never see.

Chemical Messengers

The brain speaks with more than a hundred chemical words, but two do most of the talking. Glutamate, the same amino acid that gives parmesan and soy sauce their savoury taste, is the main "go" signal, released at most excitatory synapses. GABA is the main "stop", making the next neuron less likely to fire. Thought depends on the balance between them. Too much glutamate and neurons fire uncontrollably, as in a seizure; boost GABA with alcohol, benzodiazepines or many anaesthetics and the brain slows, sedates and finally switches off.

The first transmitter identified was acetylcholine, which carries commands from nerves to every skeletal muscle and helps the brain stay alert and attentive; Henry Dale in London and Otto Loewi in Graz shared a Nobel Prize for its discovery in 1936. Other messengers work less like telephone lines and more like radio broadcasts. Small clusters of cells in the brainstem and midbrain spray dopamine, noradrenaline and serotonin over wide areas, tuning motivation, alertness and mood rather than carrying specific messages.

In 1957 and 1958 the Swedish pharmacologist Arvid Carlsson drained rabbits of dopamine with the drug reserpine. They slumped motionless until he injected L-DOPA, which the brain converts into dopamine, and they were up and moving within minutes. Carlsson realised dopamine was a transmitter in its own right, and the finding led directly to L-DOPA treatment for Parkinson's disease. In 1969 Oliver Sacks gave the drug to survivors of the 1920s sleeping-sickness epidemic in a Bronx hospital, people frozen for decades, a story he told in Awakenings. Carlsson shared a Nobel Prize in 2000.

The body also makes its own opiates. In 1973 Candace Pert and Solomon Snyder at Johns Hopkins found receptors in the brain to which morphine binds, and in 1975 John Hughes and Hans Kosterlitz in Aberdeen isolated natural chemicals that fit them, the enkephalins, members of the family now called endorphins. Peptides such as oxytocin act as messengers too, and even a gas, nitric oxide, carries signals between cells.

Did you know? About 90 per cent of the body's serotonin is made not in the brain but in the gut, where it helps to control the movement of food.

Popular shorthand often gets these chemicals wrong. Dopamine is less a "pleasure molecule" than a signal of wanting, and of rewards that turn out better than expected. The claim that depression is simply a serotonin shortage also has weak support, as a 2022 review of the evidence concluded, even though drugs that act on serotonin, such as fluoxetine (approved in the United States as Prozac in 1987), help many people. Every pill, drink and cup of coffee that changes how you feel works by borrowing, blocking or imitating one of the brain's own words.

Fuel and the Blood–Brain Barrier

The brain is the body's most demanding customer. Around 750 millilitres of blood pass through it every minute, about 15 per cent of the heart's output at rest, delivering roughly a fifth of the body's oxygen and some 120 grams of glucose a day. Unlike muscle or liver, the brain stores almost no fuel. Interrupt the supply, as in a faint or a cardiac arrest, and consciousness goes within about ten seconds; after a few minutes without oxygen, neurons begin to die.

Glucose is the usual fuel, but not the only one. In studies in the 1960s George Cahill and colleagues in Boston followed people through long fasts and showed that after several days the brain switches much of its demand to ketones, made by the liver from fat. That flexibility is why humans survive weeks without food, and why the ketogenic diet, devised at the Mayo Clinic in 1921, still controls seizures in some children with epilepsy.

The supply line is also a fortress wall. In 1885 the German scientist Paul Ehrlich injected dyes into the blood of animals and found that every organ took up the colour except the brain. In 1913 his student Edwin Goldmann injected dye into the fluid around the brain instead, and only the brain stained. Something separated brain from blood. Electron microscopes revealed it in 1967: the cells lining brain capillaries are stitched together by tight junctions, leaving no gaps for substances to slip through. Together with wrapping cells called pericytes and the endfeet of astrocytes, they form the blood–brain barrier.

Oxygen, carbon dioxide and small fat-soluble molecules pass through it easily; most other things need a dedicated transporter. Glucose rides a carrier called GLUT1. Dopamine itself cannot get in, which is why people with Parkinson's disease take L-DOPA, which uses an amino acid carrier and is converted once inside. The barrier also keeps out the vast majority of drug molecules, one reason brain diseases are so hard to treat. A few spots lack it on purpose, such as the area postrema in the brainstem, which samples the blood for toxins and triggers vomiting.

Did you know? Heroin is morphine with two small chemical groups added; they make it more fat-soluble, so it crosses the blood–brain barrier faster and hits harder.

The blood also carries away heat, since a busy brain warms itself. And a common myth needs correcting: hard thinking barely raises the brain's total energy use. Most of its fuel goes on constant background activity, as Marcus Raichle's scanning work in St Louis showed. The practical lessons are simple. Eat regularly, treat confusion in a person with diabetes as an emergency, and look after the heart and blood vessels, because the brain lives entirely on what they deliver.

Skull, Meninges and Fluid

The living brain is soft, about the consistency of firm jelly, and it is protected accordingly. Outermost comes the scalp, then the skull: eight cranial bones, including the frontal bone, two parietals, two temporals and the occipital, locked together by wavy joints called sutures. In a newborn they are still separate, with soft gaps called fontanelles that let the head squeeze through the birth canal and leave room for the brain to double in size; the largest closes at around 18 months to two years.

Inside the bone lie three membranes, the meninges, whose names come from medieval Latin translations of Arabic medical texts. The dura mater, "tough mother", is a leathery sheet lining the skull, folding inwards to separate the hemispheres and carrying large veins called sinuses. Beneath it is the delicate arachnoid, named for its spider-web strands, and then the pia mater, "tender mother", clinging to every fold of the brain's surface. Between arachnoid and pia lies a space filled with cerebrospinal fluid.

That fluid, about 150 millilitres at any moment, is made by the choroid plexus in the brain's ventricles at roughly 500 millilitres a day, so it is replaced several times daily. It cushions the brain and, by buoyancy, cuts its effective weight from around 1,400 grams to about 50. In 1891 Heinrich Quincke in Kiel introduced the lumbar puncture, drawing fluid from the lower back to diagnose infection. When fluid cannot drain, it builds up as hydrocephalus. In 1955 John Holter, an engineer in Philadelphia whose baby son Casey had the condition, built a silicone valve shunt that became the model for treatment worldwide.

The armour has weak points. A blow to the thin bone of the temple can tear the middle meningeal artery and cause an epidural haematoma: the person may seem fine, then decline over hours, as happened to the actress Natasha Richardson after a fall on a beginners' ski slope in Quebec in 2009. Torn veins under the dura cause subdural bleeds, common in older people after minor falls. Bacterial meningitis inflames the membranes and can kill within a day; in 2015 the United Kingdom became the first country to vaccinate all babies against meningococcal group B.

Did you know? Holes drilled in living skulls, called trepanation, are among the oldest surgery known; many Neolithic and Inca skulls show bone that healed around the hole, so the patients survived.

Even this old anatomy still surprises. In 2015 teams in Virginia and Helsinki described lymphatic vessels in the meninges, drainage channels that textbooks had missed, and studies in mice suggest that fluid flushes through the brain more briskly during sleep. Wear the helmet, take any blow to the head seriously, and remember that a good night's sleep may be the brain's way of doing the washing-up.

The Brainstem

The brainstem is the stalk that joins the brain to the spinal cord, a column of tissue barely thicker than a thumb. It has three stacked segments: the midbrain at the top, the bulging pons in the middle and the medulla tapering into the spinal cord below. Ten of the twelve pairs of cranial nerves leave from here, carrying the movements and sensations of the eyes, face, tongue and throat. Every message travelling between body and brain passes through it, which makes it the busiest and least forgiving corridor in the nervous system.

The medulla runs the machinery you never think about. In 1991 Jeffrey Smith, Jack Feldman and colleagues at UCLA found a small cluster of cells in the rat medulla, the pre-Bötzinger complex, that keeps producing a breathing rhythm even in a thin slice of tissue kept alive in a dish. Nearby sensors track carbon dioxide in the blood and set the pace; other nuclei adjust heart rate and blood pressure. When polio attacked the medulla in the Copenhagen epidemic of 1952, patients could no longer breathe for themselves, and medical students squeezed air into their lungs by hand, in shifts, for weeks. That crisis gave birth to the modern intensive care unit.

The brainstem is also the brain's on switch. In 1949 Giuseppe Moruzzi and Horace Magoun, working in Chicago, stimulated a web of neurons running through its core, the reticular formation, in anaesthetised cats. The brain's electrical rhythm jumped instantly from the slow waves of sleep to the fast, flat pattern of alertness. Clusters within this core, such as the locus coeruleus with its noradrenaline and the raphe nuclei with their serotonin, spray their chemicals over the entire forebrain. This is why a lesion the size of a pea in the upper brainstem can cause coma, while far larger injuries to the cortex may not.

The pons carries traffic from the cortex to the cerebellum and helps paralyse the body during dreaming sleep. A stroke there can produce locked-in syndrome. In December 1995 Jean-Dominique Bauby, editor of French Elle, woke from such a stroke fully aware but able to move only his left eyelid. Blinking as an assistant recited the alphabet, he dictated The Diving-Bell and the Butterfly, published in 1997 days before his death.

Did you know? The urge that ends a breath-hold comes mainly from rising carbon dioxide, not falling oxygen, and the brainstem will not let you ignore it for long.

One myth deserves retiring: the brainstem is not a leftover "reptilian brain" bolted beneath a human one. Reptiles have forebrains too, and our brainstem evolved alongside the rest. Doctors still test its reflexes (pupils, gag, cough) to judge coma and brain death, and every night it decides when you fall asleep. Thought lives higher up, but life itself is negotiated here.

The Cerebellum

Tucked beneath the back of the cerebrum sits the cerebellum, Latin for "little brain". It is about a tenth of the brain's volume, yet in 2009 the Brazilian neuroscientist Suzana Herculano-Houzel counted roughly 69 billion of our 86 billion neurons inside it, close to four out of five. Most are tiny granule cells, packed into a sheet so tightly folded that it would stretch over a metre if pulled flat.

The first clues came from damage. In 1824 the French physiologist Pierre Flourens removed the cerebellum from pigeons: they could still fly and walk, but clumsily, overshooting and staggering. During the First World War the neurologist Gordon Holmes examined soldiers in London with bullet wounds to the cerebellum and catalogued what went wrong: hands that trembled more as they neared a target, movements broken into jerky pieces, speech slurred into separate syllables. The cerebellum does not start movements. It makes them smooth, accurate and well timed.

Its circuit is strikingly regular, the same wiring repeated across thousands of narrow leaves called folia. Mossy fibres bring in news from the body and the cortex to the granule cells, whose axons run along each folium as parallel fibres. These cross the flat, fan-shaped dendrites of the Purkinje cells, first drawn by the Czech anatomist Jan Evangelista Purkyně in 1837. Each Purkinje cell also receives a single, powerful climbing fibre from the inferior olive in the brainstem. In 1969 and 1971 David Marr and James Albus proposed that this climbing fibre is a teaching signal that fires when a movement goes wrong, and in the 1980s Masao Ito in Tokyo showed that it weakens the parallel-fibre synapses active at the same moment. The Purkinje cell, which can only inhibit, then sends its corrected output to the deep cerebellar nuclei.

The cerebellum works by prediction: it forecasts the sensory result of each command and corrects the error. In 1998 Sarah-Jayne Blakemore and colleagues at University College London showed this is why your own touch feels muted. In 2014 doctors in Jinan, China, described a 24-year-old woman born without a cerebellum; she had learned to walk at about seven and still swayed and spoke unsteadily, proof of how much the rest of the brain can compensate, and how much it cannot.

Did you know? You cannot tickle yourself because your cerebellum predicts exactly what your fingers will do and turns the sensation down.

The old idea that the cerebellum is only for movement is fading; Jeremy Schmahmann in Boston described in 1998 how damage to it also blunts planning, language and mood. Alcohol hits it early, which is why roadside sobriety tests ask drivers to walk a straight line. Every skill you polish, from typing to a tennis serve, is tuned in this small, crowded structure.

The Thalamus

At the centre of the brain, above the brainstem, sit two egg-shaped masses about the size of walnuts, one in each hemisphere. Together they form the thalamus, from the Greek for "inner chamber". Almost everything you see, hear, touch and taste passes through it on the way to the cortex. Only smell takes a different route, running straight from the nose to the olfactory cortex.

The thalamus is divided into relay nuclei, each dedicated to one stream. Signals from the eyes go to the lateral geniculate nucleus, then on to the visual cortex at the back of the head; sound goes through the medial geniculate to the auditory cortex; touch, pain and taste pass through the ventral posterior nuclei. Motor loops from the cerebellum and basal ganglia pass through other nuclei on their way to the frontal lobe, and the large pulvinar helps steer attention.

Calling it a relay undersells it. Only a small minority of the synapses on the relay cells come from the senses; the rest arrive from the cortex and the brainstem, letting the brain adjust what reaches it. Wrapped around the thalamus is the thalamic reticular nucleus, a thin shell of inhibitory neurons that acts as a gate. The Romanian-born neuroscientist Mircea Steriade, working in Quebec, showed in the 1980s and 1990s how this shell and the cortex trade bursts of firing to produce the sleep spindles of light sleep, when the gate swings shut.

Damage reveals how central it is. In 1906 the Paris neurologists Joseph Dejerine and Gustave Roussy described patients who, after a small thalamic stroke, felt burning pain on one side of the body from the lightest touch. Karen Ann Quinlan, whose coma prompted a landmark New Jersey right-to-die case in 1976, was found at autopsy to have the worst damage in her thalamus, while her cortex was comparatively spared. And in 1986 Elio Lugaresi in Bologna described fatal familial insomnia, an inherited prion disease that destroys parts of the thalamus: sufferers lose the ability to sleep and die within months to a few years.

Did you know? Fewer than one in ten of the synapses on the visual thalamus's relay cells come from the eye; most of the rest come from the cortex itself.

The gate can sometimes be reopened. In 2007 Nicholas Schiff and colleagues in New York implanted electrodes in the central thalamus of a man who had been minimally conscious for six years after an assault; with stimulation he began to name objects and eat by mouth. Anaesthetics, sleep and attention all work partly by tuning thalamic traffic. Whenever you suddenly notice your own name across a noisy room, the gate has just opened.

Hypothalamus and Pituitary

The hypothalamus weighs about four grams, less than one per cent of the brain, and sits just above the roof of the mouth, behind the eyes. Yet it holds the set points of the body: temperature, hunger, thirst, sleep timing and the hormones of stress, growth and reproduction. Hanging beneath it on a thin stalk is the pea-sized pituitary gland, cradled in a saddle of bone called the sella turcica, the "Turkish saddle".

For decades the pituitary was called the master gland. The real master turned out to be the brain above it. In the late 1940s and 1950s the British physiologist Geoffrey Harris argued that the hypothalamus controls the front of the pituitary through chemicals carried in a tiny private system of blood vessels, the portal veins. Proving it took a race. Roger Guillemin and Andrew Schally, rival teams in the United States, ground up millions of sheep and pig brains from slaughterhouses to isolate the first releasing hormones, TRH in 1969 and GnRH in 1971, and shared a Nobel Prize in 1977.

The system runs on loops. The hypothalamus releases a hormone such as CRH; the anterior pituitary answers with ACTH; the adrenal glands produce cortisol; and cortisol returns to the brain and switches the first step down. This negative feedback works like a thermostat, keeping levels in range. The back of the pituitary works differently: it is made of nerve endings from hypothalamic neurons that release oxytocin, for labour and breastfeeding, and vasopressin, which tells the kidneys to save water.

Experiments and patients mapped its other jobs. In 1940 Albert Hetherington and Stephen Ranson, in Chicago, made small lesions in the rat hypothalamus and produced animals that ate until they became grossly obese; damage to a neighbouring zone later produced rats that would starve beside full bowls. Robert Wadlow of Alton, Illinois, whose pituitary poured out growth hormone, reached 2.72 metres, the tallest person reliably measured. People who lack vasopressin can pass well over ten litres of urine a day.

Did you know? The back half of the pituitary is not really a gland at all, but the ends of nerve cells whose bodies sit in the hypothalamus above.

Fever shows the thermostat at work. Chemicals released during infection reset the hypothalamic set point upwards, so you shiver under a blanket on a warm day; aspirin and ibuprofen work partly by blocking that signal. Long-term stress, shift work and crash dieting all disturb these loops. A region the size of an almond decides when you feel cold, hungry or thirsty, and the rest of the body obeys.

The Limbic System

In 1878 the French surgeon Paul Broca noticed a ring of cortex curling around the inner edge of each hemisphere and named it le grand lobe limbique, from the Latin limbus, a border. He thought it was mainly for smell. In 1937 the American anatomist James Papez proposed instead that a loop running from the hippocampus through the mammillary bodies, the thalamus and the cingulate cortex was the brain's circuit for emotion. In 1952 Paul MacLean gave the whole collection a name: the limbic system.

Its members are a mixed cast. The hippocampus, curled inside the temporal lobe, files new memories; its output cable, the fornix, arches forward to the mammillary bodies. The almond-shaped amygdala flags threats and anything that matters, and drives the hypothalamus to raise heart rate and release stress hormones. The nucleus accumbens and the dopamine cells of the ventral tegmental area generate wanting. The cingulate cortex, the orbitofrontal cortex and the insula weigh up conflict, value and the felt state of the body.

The evidence came largely from damage. In 1937 Heinrich Klüver and Paul Bucy in Chicago removed the temporal lobes of rhesus monkeys, including both amygdalae; the animals became oddly tame, approached snakes without fear and put inedible objects in their mouths. In 1887 the Russian psychiatrist Sergei Korsakoff described alcoholic patients who could not form new memories and confabulated to fill the gaps; their mammillary bodies, starved of the vitamin thiamine, were later found shrunken. In 1954 James Olds and Peter Milner at McGill University in Montreal found that rats would press a lever thousands of times an hour to stimulate the septal area of their own brains.

Did you know? Smell is the only sense wired almost directly into the limbic system, which is why a scent can return a forgotten memory with full emotional force, as Proust described with his madeleine.

Neuroscientists now argue about whether the limbic system is a system at all. Joseph LeDoux of New York University has pointed out that no one agrees which structures belong in it, that the hippocampus is mostly about memory rather than feeling, and that emotion also depends on regions far outside the ring. Most researchers keep the term as a convenient map label rather than a unit with one job. Its deeper lesson stands: memory, motivation and emotion are wired into the same neighbourhood and constantly shape one another.

That overlap is something you live with every day. You remember the frightening, the joyful and the rewarding far better than the neutral, because the amygdala tags those moments for the hippocampus to keep. Chronic stress, heavy drinking and poor sleep all wear on these structures. Feeling and memory are not separate departments; they share the same wiring.

The Basal Ganglia

Buried beneath the cortex lie the basal ganglia, a cluster of nuclei that decide which of the many actions you could take actually happens. The name is a historical slip: in modern anatomy a ganglion lies outside the brain, so these are really nuclei. The main input is the striatum, made of the caudate and the putamen; the main outputs are the internal globus pallidus and part of the substantia nigra; and in between sit the external globus pallidus and the small subthalamic nucleus.

The circuit works by taking off a brake. At rest, the output nuclei fire constantly and inhibit the thalamus, holding every possible movement in check. When the cortex proposes an action, the direct pathway through the striatum silences the output nuclei for that one action, releasing the thalamus to drive the cortex onward: GO. The indirect pathway, routed through the external pallidus and the subthalamic nucleus, strengthens the brake on everything else: NO-GO. This model, set out by Roger Albin, Anne Young and John Penney in Michigan in 1989, and by Mahlon DeLong at around the same time, still guides surgeons and drug designers.

Dopamine from the substantia nigra tilts the balance. It excites the GO cells of the striatum and quietens the NO-GO cells, making action easier. In 1957 Arvid Carlsson in Sweden drained dopamine from rabbits with the drug reserpine and found they lay motionless until given L-DOPA. In 1960 Oleh Hornykiewicz in Vienna found that the striatum of people who had died with Parkinson's disease had lost most of its dopamine. The first L-DOPA treatments followed within a decade. In 1982 a group of young drug users in California froze rigid overnight after injecting a homemade heroin substitute contaminated with MPTP, which kills those same dopamine cells.

The opposite failure produces too much movement. In 1872 the Long Island doctor George Huntington described a hereditary disorder of uncontrollable jerking, which destroys the NO-GO cells of the striatum first. Families around Lake Maracaibo in Venezuela, studied by Nancy Wexler from 1979, helped locate the gene in 1993. The tics of Tourette syndrome, described in Paris by Georges Gilles de la Tourette in 1885, are also traced to these loops.

Did you know? The output cells of the basal ganglia fire dozens of times a second all day long, which means the brain's default answer to almost every possible movement is no.

Since the 1990s, surgeons such as Alim-Louis Benabid in Grenoble have placed stimulating electrodes in the subthalamic nucleus to relieve Parkinson's symptoms. The same loops form habits and weigh rewards, which is why they matter for addiction. Every action you take begins by lifting a brake.

The Cerebral Cortex

The cerebral cortex is the wrinkled outer rind of the brain, a sheet only two to four millimetres thick. Spread flat, the two hemispheres together would cover roughly a quarter of a square metre, about the area of a pillowcase; crumpled into the skull, around two-thirds of the sheet lies hidden in the grooves, or sulci, between the ridges, or gyri. It contains around 16 billion neurons, and it is where perception, language, planning and much of memory take shape.

Four large lobes divide each hemisphere: frontal for action and planning, parietal for touch and space, temporal for hearing, language and memory, and occipital for vision, with a fifth region, the insula, folded deep inside. The folds are not random decoration. In 2016 a team at Harvard showed with a layered gel model that a sheet growing faster than the layer beneath it buckles into folds much like a human brain's. Folding mainly follows size, so dolphins and elephants have more folds than we do, and more wrinkles do not mean more intelligence.

Under the microscope the newest part, the neocortex, has six layers. Layer IV receives input from the thalamus; layers II and III talk to other areas of cortex; layer V sends the cortex's commands to the brainstem and spinal cord; layer VI feeds back to the thalamus. In 1874 the Kyiv anatomist Vladimir Betz described giant layer V cells in the motor cortex, now called Betz cells, whose axons reach down the spinal cord. In 1909 the German neurologist Korbinian Brodmann used differences in layering to divide the cortex into 52 numbered areas, a map neuroscientists still quote. In 2016 Matthew Glasser and colleagues, combining several kinds of MRI scan, counted 180 areas in each hemisphere.

There is also a vertical order. In 1957 Vernon Mountcastle at Johns Hopkins recorded from the touch cortex of cats and found that neurons in a narrow vertical column all responded to the same patch of skin. The cortical column suggested a repeated circuit, the same recipe applied to sight, sound or touch depending on its inputs.

Did you know? In the visual cortex the input layer is so thick that it shows as a pale stripe visible to the naked eye, first noticed in 1776 by Francesco Gennari, a medical student in Parma.

That uniformity has practical meaning. When one input is lost, its cortical territory can be recruited by another; in people born blind, the visual cortex responds to Braille. The cortex keeps thinning slowly from early adulthood, and exercise, education and good sleep are linked to keeping it healthier for longer. Six layers, one recipe, repeated across a crumpled sheet: that is the hardware of human thought.

The Frontal Lobes

The frontal lobes fill the front of the skull behind the forehead and make up about a third of the human cortex. The back strip controls movement; the large region in front of it, the prefrontal cortex, handles the jobs grouped under the name executive function: holding a goal in mind, planning the steps, switching strategy and stopping yourself from doing the obvious but wrong thing.

Much of what we know came from damage. Phineas Gage, the railway foreman whose skull was pierced by an iron rod in Vermont in 1848, survived but was famously changed in temperament. In 1935 the Lisbon neurologist Egas Moniz began cutting the connections of the prefrontal cortex to treat severe mental illness, earning a Nobel Prize in 1949. In the United States Walter Freeman popularised the cruder "ice-pick" lobotomy. Many patients were left apathetic and childlike, among them Rosemary Kennedy, sister of the future president, after her operation in 1941. By the late 1950s new drugs and growing revulsion ended the practice.

More careful studies followed. In 1963 Brenda Milner in Montreal gave patients a card-sorting test in which the rule for matching cards changes without warning. People with frontal damage kept sorting by the old rule even while saying it was wrong. In 1986 François Lhermitte in Paris described utilisation behaviour: patients with frontal lesions would pick up and use any object placed in front of them, putting on several pairs of glasses in a row or drinking from every glass. Without the frontal lobes, the environment takes charge of behaviour.

The prefrontal cortex is also the last region to mature. In 2004 Nitin Gogtay, Jay Giedd and colleagues at the US National Institute of Mental Health published MRI scans of children followed for up to a decade, showing grey matter maturing from the back of the brain to the front, with the prefrontal cortex last. The popular claim that the brain is "finished at 25" is a simplification: there is no finishing line, and wiring keeps changing well into the twenties and beyond.

Did you know? Egas Moniz's Nobel Prize for lobotomy has never been withdrawn, despite campaigns by the families of patients.

These findings shape daily life in plain ways. Teenagers take more risks in front of friends partly because control systems lag behind reward systems. Sleep loss, alcohol and stress all hit prefrontal function early, which is why tired people make impulsive decisions. Exercise, sleep and practice at planning help keep it sharp in old age. The region that says "not yet" is the last to arrive and the first to tire.

Two Hemispheres

The brain is divided down the middle into two hemispheres, joined mainly by the corpus callosum, a thick band of around 200 million nerve fibres. Each hemisphere controls and feels the opposite side of the body and sees the opposite half of the visual world. Normally the callosum shares everything between them so quickly that you experience one seamless mind.

In 1940 the American surgeon William Van Wagenen began cutting the callosum to stop epileptic seizures spreading from one side to the other. To everyone's surprise, patients seemed almost unchanged. The hidden effects emerged two decades later, after Philip Vogel and Joseph Bogen operated on a new series of patients in Los Angeles from 1962. At Caltech, Roger Sperry and his graduate student Michael Gazzaniga devised tests that sent information to only one hemisphere, by flashing a word to one side of a fixed gaze or placing an object in one hand out of sight.

The results were startling. When "key ring" was flashed so that "key" fell in the left visual field and "ring" in the right, each word reached only one hemisphere. The patient, speaking with his left hemisphere, said he had seen "ring", yet his left hand reached under a screen and picked out a key. Sperry concluded that each separated hemisphere had its own perceptions and memories, and won a Nobel Prize in 1981. Later Gazzaniga showed a patient known as P.S. a snow scene on one side and a chicken claw on the other; his hands chose a shovel and a chicken, and his talking left hemisphere promptly invented a reason. Gazzaniga called this storyteller the interpreter.

The hemispheres are specialised but not opposites. Language is based in the left hemisphere in around 95 per cent of right-handers and most left-handers, a fact tested before surgery with the Wada test, developed by the Japanese-born neurologist Juhn Wada in 1949, which puts one hemisphere briefly to sleep. The right hemisphere is stronger at recognising faces, reading tone of voice and attending to space; damage to it can leave people ignoring the left half of their world. Children who have had an entire hemisphere removed for severe epilepsy can still walk, talk and attend school.

Did you know? Some early split-brain patients described one hand undoing what the other had just done, such as unbuttoning a shirt the other hand was buttoning.

The popular idea of "left-brained" logical people and "right-brained" creative people is a myth. In 2013 a University of Utah team analysed brain scans from more than 1,000 people and found no sign that anyone favours one whole hemisphere. Ordinary thinking, from reading a joke to catching a ball, needs both sides talking. One brain, two hemispheres, and a conversation between them that never stops.

The Connectome

A connectome is a complete wiring diagram of a nervous system: every neuron and every connection between them. The word was coined in 2005, independently, by Olaf Sporns and colleagues in Indiana and by Patric Hagmann in Lausanne. The idea is older. Since Santiago Ramón y Cajal, neuroscientists have suspected that what a brain does depends on how its cells are joined.

The first connectome took more than a decade to make by hand. At the Laboratory of Molecular Biology in Cambridge, John White, Sydney Brenner and colleagues cut the tiny worm Caenorhabditis elegans into thousands of ultrathin slices, photographed each with an electron microscope and traced every neuron with felt-tip pens. Their 1986 report, subtitled The Mind of a Worm, mapped 302 neurons and around 7,000 synapses. In 2023 teams in Cambridge and at Johns Hopkins published the brain of a larval fruit fly: 3,016 neurons and about 548,000 synapses.

Then came automation. In 2024 the FlyWire consortium, led by Sebastian Seung and Mala Murthy at Princeton, published the whole brain of an adult fruit fly, 139,255 neurons joined by about 50 million synapses. Artificial intelligence traced the cells first; hundreds of scientists and volunteer gamers then corrected its mistakes. The same year Jeff Lichtman's lab at Harvard and Google published a cubic millimetre of human cortex, removed during epilepsy surgery: 57,000 cells and 150 million synapses. In 2025 the MICrONS project reported a cubic millimetre of mouse visual cortex with around 200,000 cells and more than 500 million synapses.

Did you know? The Harvard–Google sample of human cortex, smaller than a grain of rice, produced around 1.4 petabytes of image data.

At a coarser scale, the Human Connectome Project, launched by the US National Institutes of Health in 2010, uses MRI to trace the major cables between regions in living people. It cannot see single synapses, but it can compare thousands of brains, and its maps of the major pathways between regions are now used to plan surgery.

A map is not a mind. The worm's wiring has been known since 1986, yet scientists still cannot fully predict its behaviour, because synapses vary in strength, change with experience and are bathed in chemicals such as dopamine and serotonin that reshape the circuit from moment to moment. A whole human brain, with 86 billion neurons and somewhere around 100 trillion synapses, would need storage on a scale that does not yet exist. Still, the fly's map has already revealed circuits for walking, steering and choosing, and it shows that every brain is a particular, inherited and constantly edited web. Your connectome is, in part, the record of your life.

Vision: Eye to Cortex

Light enters the eye through the cornea and lens, which together throw a small, upside-down image onto the retina, a sheet of tissue about a quarter of a millimetre thick lining the back of the eyeball. Johannes Kepler worked out the inverted image in 1604, and René Descartes later saw it for himself on the scraped-back retina of an ox's eye. The flip is no problem for the brain: nothing inside the head is looking at a picture, so there is nothing to turn the right way up.

The retina is brain tissue pushed outwards during development, and it does a good deal of computing before anything leaves the eye. Around 120 million rods handle dim light and around 6 million cones handle colour and fine detail, crowded into the fovea, a pit about 1.5 millimetres across. Layers of bipolar and ganglion cells compress this flood of signals, contrasting each spot with its surroundings, into about a million fibres of the optic nerve. Where those fibres leave the eye there are no receptors at all: the blind spot, described by Edme Mariotte in the 1660s, which the brain quietly fills in.

At the optic chiasm, just in front of the pituitary gland, the fibres from the inner half of each retina cross over, so the left half of the visual world ends up in the right hemisphere and the right half in the left. Most fibres relay through the lateral geniculate nucleus of the thalamus and then fan out to the primary visual cortex, V1, at the back of the head. A smaller branch runs to the superior colliculus in the midbrain, which swings the eyes towards anything that moves.

We know how the map is laid out partly from war. In 1909 the Japanese ophthalmologist Tatsuji Inouye charted the blind patches of soldiers shot through the head in the Russo-Japanese War and showed that the centre of gaze claims a disproportionately large share of V1; Gordon Holmes confirmed it with British casualties of the First World War. In 1959 David Hubel and Torsten Wiesel, recording single cells in cat cortex at Johns Hopkins, found neurons that fired only to edges at a particular angle, and a Nobel Prize followed in 1981. In the 1970s Lawrence Weiskrantz in Oxford studied a man known as D.B., blind on one side after surgery to V1, who could nonetheless point to objects he insisted he could not see. This blindsight runs along the older routes through the midbrain.

Did you know? Each eye's blind spot is large enough to make a person's head vanish at a distance of around two metres, yet you never notice the hole.

Vision is a reconstruction rather than a camera feed, which is why a stroke at the back of the brain can wipe out half the world while both eyes stay perfect. The eye is also shaped by how it is used: children who spend more time outdoors in daylight are less likely to become short-sighted, a finding repeated from Sydney to Guangzhou. Seeing begins in the eye, but it is finished in the brain.

Seeing Colour and Motion

Colour is not in the light. A ripe tomato reflects mostly long wavelengths, but red is something the brain constructs from the replies of three kinds of cone, tuned to short, medium and long wavelengths with peak sensitivities near 420, 530 and 560 nanometres. Thomas Young proposed three receptors in 1802 and Hermann von Helmholtz developed the idea in the 1850s. Every screen, which mixes red, green and blue dots, relies on it.

Ewald Hering spotted a puzzle in 1878: nobody can picture a reddish green or a yellowish blue, and staring at red leaves a green afterimage. He proposed that colour is coded in opponent channels, red against green, blue against yellow and light against dark. Both men were right, at different stages. Cones measure wavelengths; the cells behind them compare the cones. In 1966 Russell De Valois and colleagues recorded cells in the monkey thalamus that were excited by red and silenced by green. Further up, Semir Zeki in London found in the 1970s an area now called V4 where many cells respond to colour, and damage near it in humans causes cerebral achromatopsia, a world drained to grey even though the cones are healthy.

Inherited colour blindness is far commoner. Around 8 per cent of men of northern European descent and about 0.5 per cent of women have a red-green deficiency, because the genes for the medium and long cone pigments sit on the X chromosome. The chemist John Dalton described his own in 1794 and asked for his eyes to be examined after his death. In 1995 researchers extracted DNA from his preserved eye in Manchester and found that he lacked the medium-wavelength pigment: a diagnosis 150 years late.

Motion has its own specialist, area V5, also called MT, on the side of the occipital lobe. In 1983 Josef Zihl and colleagues in Munich described a patient, L.M., who after strokes damaged this region on both sides could no longer see movement. Pouring tea, she saw a frozen arc like a glacier and the cup overflowed; cars seemed to jump from far away to suddenly near. The motion aftereffect shows the same machinery tiring. In 1834 Robert Addams stared at the Falls of Foyers in Scotland and found that the rocks beside the water then seemed to drift upwards.

Did you know? The 2015 photograph of "the dress", seen as blue and black by some and white and gold by others, divided viewers largely by the lighting their brains assumed it was in, studies suggest.

In 1982 Leslie Ungerleider and Mortimer Mishkin proposed that vision divides into two streams: a lower ventral stream into the temporal lobe that asks what something is, and an upper dorsal stream into the parietal lobe that asks where it is and how to reach it. The practical upshot is design. Maps, charts and warning lights that rely on red against green fail roughly one man in twelve, while adding shape, pattern or blue-yellow contrast fixes the problem. Colour and motion are separate calculations, stitched by the brain into a single seamless scene.

Hearing

Sound is a ripple of pressure in the air, and hearing is the art of turning those ripples into nerve impulses. The eardrum vibrates, and three tiny bones, the hammer, anvil and stirrup, pass the motion to the oval window of the inner ear, concentrating the force around twenty times so that it can push fluid rather than air. The stirrup, about 3 millimetres long, is the smallest bone in the body.

The fluid-filled cochlea, coiled two and a half turns like a snail shell, holds the basilar membrane, around 35 millimetres long if unrolled. It is narrow and stiff at the base and wider and floppier at the tip, so each frequency makes a wave that peaks at its own place: high notes near the entrance, low notes near the apex. Georg von Békésy watched these travelling waves in cochleae from cadavers and won the Nobel Prize in 1961. On the membrane sit roughly 3,500 inner hair cells, which send almost all the information to the brain, and around 12,000 outer hair cells, which act as amplifiers, lengthening and shortening with each cycle to boost faint sounds. In 1978 David Kemp in London found that a healthy ear actually emits faint sounds of its own, and newborn hearing screening now listens for these echoes.

From the cochlear nerve, signals climb through the brainstem. Locating a sound starts in the superior olive, which compares the two ears. A sound on your left reaches the left ear up to around two-thirds of a millisecond sooner and slightly louder, and neurons register timing differences of about a hundredth of a millisecond. Lloyd Jeffress proposed in 1948 that delay lines and coincidence detectors could do this, and Masakazu Konishi later found such a circuit in the barn owl, which can catch a mouse in total darkness. The signal then passes through the inferior colliculus and the medial geniculate nucleus of the thalamus to the primary auditory cortex on Heschl's gyrus, laid out like a piano keyboard by pitch.

Did you know? Mammals cannot regrow lost cochlear hair cells, but birds can, which is why researchers study chickens in the hope of reversing deafness.

The cochlear implant bypasses dead hair cells by stimulating the nerve directly. William House implanted a single-channel device in Los Angeles in 1961, and in 1978 Graeme Clark's team in Melbourne implanted a multichannel version in Rod Saunders, who had lost his hearing in an accident. More than a million people now use implants, and children implanted young usually learn spoken language.

Hair cells die from loud noise and from age, and they do not come back. The Lancet Commission on dementia identifies untreated hearing loss in midlife as one of the largest modifiable risk factors, perhaps because a straining brain withdraws from conversation. Ear protection at concerts and an early hearing aid are cheap insurance for the mind. You listen with your ears, but you hear with your brain.

Balance

Close your eyes and nod, and you still know which way is up. That knowledge comes from the vestibular system, a cluster of fluid-filled tubes and sacs in each inner ear, next to the cochlea and using the same kind of hair cell. Three semicircular canals, set at roughly right angles like the corner of a box, detect rotation: when the head turns, the fluid inside lags behind and bends a jelly flap called the cupula, tipping the hair cells. Two otolith organs, the utricle and the saccule, sense straight-line acceleration and the pull of gravity, using a layer of calcium carbonate crystals that slide like a weighted mat over the hair cells. Otolith simply means "ear stone".

The idea that dizziness could come from the ear was new in 1861, when the French doctor Prosper Ménière argued that attacks of vertigo, deafness and ringing began in the inner ear rather than the brain. In Vienna, Robert Bárány noticed that syringing a patient's ear with warm or cold water made the eyes flick, and turned it into the caloric test, still used today. His Nobel Prize, awarded for 1914, reached him in a Russian prisoner-of-war camp.

The signals travel to the vestibular nuclei in the brainstem and to the cerebellum, which act on them almost instantly. The vestibulo-ocular reflex turns the eyes against every head movement within around 10 milliseconds, which is why you can read a sign while walking. A doctor known as J.C., who lost his vestibular function to the antibiotic streptomycin in the 1940s, described in 1952 how the world jiggled with every step and every pulse, and how he braced his head between the bars of his bed to read. Over months his brain learned to lean on vision and the feel of his feet instead.

Motion sickness is best explained as a conflict between the senses. Below deck on a ship your eyes report a still cabin while your inner ears report heaving; reading in a car does the same. One theory holds that the brain treats the mismatch as a sign of poisoning, and responds by being sick. In orbit, where the otoliths lose their usual reference, many astronauts feel ill for the first few days.

Did you know? People with no working vestibular system cannot get motion sick however rough the sea, a fact US Navy researchers used in the 1960s to pin the condition on the inner ear.

Loose crystals cause the commonest vertigo of all, benign paroxysmal positional vertigo, when otolith fragments drift into a canal. The Epley manoeuvre, devised by John Epley in Portland, Oregon, in 1980, rolls the head to tip them back out and often works in minutes. Balance also fades with age, and around one in three people over 65 has a fall each year; exercise such as tai chi measurably cuts that risk. The quietest sense is the one you miss most when it fails.

Smell

Every breath carries molecules into the top of the nose, where a patch of tissue about the size of a postage stamp, the olfactory epithelium, holds millions of receptor neurons. Each one makes a single type of odorant receptor. In 1991 Linda Buck and Richard Axel at Columbia University found the huge gene family that codes them, the largest in the genome; humans have around 400 working versions. The pair shared the Nobel Prize in 2004. Because each odour activates its own combination of receptors, a few hundred types can tell apart an enormous number of smells.

The receptor neurons send their axons up through tiny holes in the skull's cribriform plate to the olfactory bulb, which lies just under the frontal lobes. There, all the neurons carrying the same receptor converge on the same few round knots called glomeruli, so each smell becomes a pattern of lit spots. From the bulb, signals go straight to the piriform cortex, the amygdala and the entorhinal cortex beside the hippocampus. Every other sense is relayed first through the thalamus, while smell reaches the brain's emotional and memory centres directly. That helps explain why a whiff of school dinners or a grandparent's soap can bring back a whole scene. Marcel Proust's madeleine dipped in tea, written in 1913, gave the experience its literary name.

The receptor neurons are among the few in the body that are constantly replaced, living for weeks to months before new ones grow from stem cells. That is why the loss of smell caused by COVID-19 in 2020, which mainly damaged supporting cells in the epithelium, usually recovered. For people whose smell stayed distorted, so that coffee smelled of sewage, smell training with a few strong odours twice a day is now a common treatment.

The belief that humans have a poor sense of smell is a myth with an author. In 1879 Paul Broca, struck by the small size of our olfactory bulbs relative to the rest of the brain, classed humans among the non-smellers. In 2017 John McGann reviewed the evidence in Science and concluded that our bulbs hold about as many neurons as those of other mammals, and that people can follow a scent trail across grass like a dog, if more slowly. A much-reported 2014 estimate that we can distinguish a trillion odours is debated, but the true number is certainly very large.

Did you know? Joy Milne, a retired nurse from Perth in Scotland, noticed a new musky smell on her husband years before his Parkinson's diagnosis, and has since helped scientists find chemical markers of the disease in skin oil.

Smell is also an early warning for the brain itself. Loss of smell often precedes the movement symptoms of Parkinson's disease by years and is common early in Alzheimer's disease, so a sudden loss is worth mentioning to a doctor. Smell is the oldest sense, wired straight into the parts of the brain that decide what matters.

Taste and Flavour

Taste is a narrow sense with a big job: deciding what goes into the body. The tongue, palate and throat carry several thousand taste buds, each a cluster of 50 to 150 receptor cells that are replaced roughly every ten days to two weeks. Together they report just five qualities. Sweet signals energy, salty signals sodium, sour signals acid in unripe or spoiling food, and bitter warns of possible poison; humans have around 25 different bitter receptors but only one sweet one. The fifth taste, umami, was identified in 1908 by Kikunae Ikeda in Tokyo, who extracted glutamate from the kombu seaweed in a bowl of dashi broth. Western science took most of the century to accept it, until matching receptors were found around 2000.

The famous tongue map, with sweet at the tip and bitter at the back, is wrong. It grew from a 1901 German thesis by David Hänig, which found small differences in sensitivity, and was hardened into rigid zones in a 1942 textbook by the Harvard psychologist Edwin Boring. Every part of the tongue with taste buds detects all five tastes.

Taste signals travel along three cranial nerves to the nucleus of the solitary tract in the medulla, then through the thalamus to the gustatory cortex in the insula, folded deep inside the side of the brain. The orbitofrontal cortex, just above the eyes, then combines taste with smell, texture and temperature, and judges how pleasant the result is. In experiments with monkeys, Edmund Rolls in Oxford found that its neurons stopped responding to a food once the animal had eaten its fill: a cellular version of having no room left for pudding.

What we call taste is mostly flavour, and flavour is mostly smell. Chewing pushes aroma molecules from the back of the mouth up into the nose, a route called retronasal smell, which Paul Rozin described in 1982 as a sense in its own right. Pinch your nose and a strawberry jelly bean tastes merely sweet; let go and strawberry floods in. Texture and chemical sensations join too: the burn of chilli and the cool of mint come from pain and temperature receptors, not taste buds.

Did you know? In 1931 the chemist Arthur Fox spilled a powder called PTC in his laboratory; a colleague complained of its bitterness while Fox tasted nothing, revealing that bitter sensitivity is inherited.

Genes give each of us a slightly different menu. People sensitive to PTC-like compounds tend to find Brussels sprouts and broccoli more bitter, and Linda Bartoshuk at Yale coined the term supertasters for those with especially intense taste. Age dulls smell more than taste, so older people often find food bland and reach for the salt. Cutting salt gradually works, because preferences adjust within weeks. The tongue reports what food is made of; the nose tells you what it is.

Touch

Skin is the largest sense organ, and it holds several kinds of mechanoreceptor, each tuned to a different part of feeling. Meissner corpuscles, packed into the ridges of the fingertips, respond to light stroking and slip, which is how you notice a glass beginning to slide. Merkel discs report steady pressure and fine edges, the sense a Braille reader relies on. Pacinian corpuscles, onion-like capsules deeper in the skin, respond best to vibration at around 200 to 300 cycles a second, which is why a phone buzzing in a pocket is unmistakable. Ruffini endings sense the skin stretching as fingers bend. Inside each ending, ion channels open when the membrane is deformed; the key one, Piezo2, was identified by Ardem Patapoutian's laboratory, earning him a share of the 2021 Nobel Prize.

The signals travel up the spinal cord in the dorsal columns, cross to the opposite side in the medulla, relay in the thalamus and arrive in the primary somatosensory cortex, the strip just behind the central sulcus. In the 1930s and 1940s Wilder Penfield, operating on awake epilepsy patients at the Montreal Neurological Institute, touched this strip with an electrode and asked what they felt. Patients reported tingles in a hand, a lip or a toe, in an orderly sequence that Penfield drew as the sensory homunculus: a little figure with huge lips and hands and a tiny trunk, scaled by how many receptors each part has.

Density sets sharpness. In 1834 Ernst Weber in Leipzig measured the two-point threshold, the gap at which two compass points feel like two: around 2 millimetres on a fingertip, several centimetres on the back. The map also changes with use. In 1993 Alvaro Pascual-Leone found that the cortical area for the reading finger of Braille readers was enlarged, and similar expansions appear for the left-hand fingers of violinists.

There is also a slow, social kind of touch. Unmyelinated C-tactile fibres in hairy skin fire most to gentle stroking at around the speed of a caress, a few centimetres a second. Håkan Olausson and colleagues described a woman, G.L., who had lost her large touch fibres; she could not feel a soft brush on her arm as touch, yet reported a faint, pleasant warmth. Harry Harlow's experiments of 1958, in which infant monkeys clung to a soft cloth mother rather than a wire one that gave milk, showed how deep the need for contact runs.

Did you know? In a 2013 study, people stroking surfaces with a fingertip could tell apart wrinkles only around 10 nanometres high, smaller than most viruses.

Touch is among the first senses to develop in the womb and among the last to fade. Skin-to-skin contact steadies the breathing and heart rate of premature babies, and children raised with little touch in understaffed orphanages have shown lasting delays. A hand on the shoulder is not a small thing: it is information delivered by its own dedicated wiring.

Pain

Pain begins with nociceptors, nerve endings that fire when tissue is damaged or threatened by heat, crushing or chemicals. Charles Sherrington named them in 1906. They come in two kinds, which is why a stubbed toe hurts twice: thin myelinated A-delta fibres carry a sharp first pain at around 15 metres a second, and slower, unmyelinated C fibres deliver the dull, spreading ache a second or so later. In 1997 David Julius and colleagues in San Francisco used capsaicin, the hot chemical in chilli, to find the receptor TRPV1, which also opens above about 43 degrees Celsius. That is why chilli feels hot, and it earned Julius a share of the 2021 Nobel Prize.

The signals meet their first checkpoint in the dorsal horn of the spinal cord. In 1965 Ronald Melzack of Montreal and Patrick Wall of London proposed the gate control theory: touch fibres can close a gate on pain transmission there, which is why rubbing a knocked elbow helps. Beyond the gate, pain travels to the thalamus and then to several areas at once, the somatosensory cortex for where and how intense, the insula and anterior cingulate for how unpleasant. The brain can also close the gate from above. In 1969 David Reynolds performed abdominal surgery on rats with no anaesthetic other than electrical stimulation of the periaqueductal grey in the midbrain. Part of the reason became clear in 1975, when John Hughes and Hans Kosterlitz in Aberdeen discovered the enkephalins, the brain's own morphine-like molecules.

Context changes pain enormously. During the Second World War the anaesthetist Henry Beecher noticed that badly wounded soldiers at Anzio asked for far less morphine than civilians with similar surgical wounds; for the soldiers, the injury meant a ticket home. Placebo pain relief is real and partly chemical: in 1978 Jon Levine and colleagues in San Francisco showed that it can be blocked by naloxone, a drug that blocks opioid receptors.

Rare families reveal the machinery. In 2006 Geoffrey Woods and colleagues in Cambridge studied families in northern Pakistan, among them a boy in Lahore who performed street stunts, walking on hot coals and pushing knives through his arms, who felt no pain at all. They carried mutations that disable SCN9A, a sodium channel gene. The boy died after jumping from a roof, a stark lesson in what pain is for.

Did you know? Jo Cameron, a woman from the Scottish Highlands, felt little pain through childbirth, burns and surgery; in 2019 scientists traced it to a rare genetic variant they named FAAH-OUT.

Pain can also outlast the injury, as the nervous system becomes sensitised and keeps sounding the alarm. Around one adult in five lives with chronic pain, and a phantom limb can hurt for decades after amputation. Modern treatment combines medicines with movement, good sleep and therapies that reduce the sense of threat. Pain is not a measure of damage but the brain's verdict on danger.

The Body Sense

Close your eyes and touch your nose. You can do it because of proprioception, the sense of where your limbs are and how they are moving. Charles Bell called it a sixth sense in 1826, and Charles Sherrington gave it its name in 1906. Its sensors lie in the muscles, tendons and joints. Muscle spindles, tiny capsules wrapped in sensory nerve endings, sit among the muscle fibres and report how long a muscle is and how fast it is stretching. Golgi tendon organs, where muscle meets tendon, report how hard it is pulling. Joint receptors and stretching skin add the rest.

The signals take two main routes. Fast pathways run to the cerebellum, which uses them to smooth movement without conscious thought, while the dorsal columns carry them up to the somatosensory cortex, which builds a felt sense of the body. Spindles also drive the knee-jerk reflex, a loop that stays inside the spinal cord and takes only a few hundredths of a second.

Losing proprioception shows how much it does. In 1971, aged 19, Ian Waterman, a butcher's apprentice, had an illness that destroyed the large sensory nerve fibres below his neck. His muscles still worked, but without watching his limbs he could not control them, and when the lights went out he collapsed. Over years he taught himself to stand, walk, dress and even drive by watching every movement, a feat the physiologist Jonathan Cole described in Pride and a Daily Marathon in 1991. Oliver Sacks wrote of a similar patient, "the disembodied lady", in 1985.

The brain's map of the body is surprisingly easy to fool. In the rubber hand illusion, reported by Matthew Botvinick and Jonathan Cohen in 1998, a person watches a rubber hand being stroked while their own hidden hand is stroked in time. Within minutes many feel the rubber hand is theirs, and flinch when it is threatened.

Did you know? A vibrating rod held on the tendon of your biceps fools the muscle spindles, and your motionless arm feels as if it is slowly straightening, sometimes past the point a real elbow can bend.

A second body sense looks inward. Interoception tracks heartbeat, breathing, hunger, a full bladder and temperature, carried by the vagus nerve and spinal pathways to the brainstem and then to the insula, which the neuroanatomist Bud Craig argued builds a moment-by-moment picture of the body's state. People differ in how well they can count their own heartbeats without feeling their pulse, a test introduced by Rainer Schandry in 1981, and researchers are exploring links between such differences and anxiety.

Both senses fade with age, which is part of why older people fall and why balance training, tai chi and simply standing on one leg help. Alcohol dulls proprioception too, which is why police in many countries ask drivers to touch their nose with eyes closed. Without your noticing, the brain is always sensing you from the inside.

The Predictive Brain

Your senses deliver information that is noisy, ambiguous and slightly late, yet the world appears clear and immediate. The reason, many neuroscientists now argue, is that the brain does not wait for evidence. It predicts. Hermann von Helmholtz suggested in 1867 that perception is unconscious inference, a conclusion drawn from incomplete data using past experience. A modern version, predictive coding, was set out by Rajesh Rao and Dana Ballard in 1999 and extended by Karl Friston in London: higher areas send predictions down to lower ones, and mainly the mismatch, the prediction error, travels back up to revise the guess.

Illusions are where the guesses show. In Edward Adelson's checker-shadow illusion of 1995, two squares of identical grey look strikingly different because the brain discounts the shadow it assumes is falling on one of them. In the hollow-face illusion, the inside of a mask looks like a normal face pointing outwards, because the brain's strong expectation that faces bulge overrides the depth cues. In the McGurk effect, reported by Harry McGurk and John MacDonald in 1976, a video of lips saying "ga" dubbed with the sound "ba" is often heard as "da": the brain settles on the best compromise between eye and ear.

Prediction also explains why you cannot tickle yourself. In 1998 Sarah-Jayne Blakemore, Daniel Wolpert and Chris Frith in London showed that the brain, with the cerebellum playing a key part, predicts the sensations caused by your own movements and dampens them. When a robot inserted a delay of a fifth of a second between a volunteer's hand movement and the touch, it felt more ticklish again. The same machinery lets you ignore the feel of your own clothes and tells self-made sensations from the world's.

When predictions run unchecked, the brain can see what is not there. Charles Bonnet syndrome, named after the Swiss naturalist who in 1760 described the visions of his nearly blind grandfather, gives people losing their sight vivid, silent hallucinations of faces, patterns or tiny figures. One explanation is that with little input from the eyes, the visual brain's predictions go unchallenged. Similar ideas are being tested as accounts of some symptoms of psychosis, though this remains debated.

Did you know? In the flash-lag illusion, a light flashed level with a moving dot seems to trail behind it, possibly because the brain projects moving objects forward to make up for its own delays.

The predictive view has everyday consequences. Expectations shape pain, the taste of wine and the effect of placebo pills. Experts see more in an X-ray or a chess position because their predictions are better tuned, and a confident first impression can be hard to shift. Perception is a controlled hallucination, constantly checked against the world, and most of the time the check is good enough.

The Motor Cortex

Every deliberate movement you make, from signing your name to raising an eyebrow, begins as a pattern of activity in a narrow ridge of cortex running from the crown of the head towards each ear. This is the primary motor cortex, on the precentral gyrus just in front of the central sulcus. Its proof came in 1870, when Gustav Fritsch and Eduard Hitzig in Berlin passed weak currents across the exposed brain of a dog and watched its opposite legs twitch. The cortex, long thought to be the seat of pure thought, turned out to give orders.

The map was drawn properly in Montreal. Operating on epileptic patients under local anaesthetic in the 1930s, Wilder Penfield touched points along the strip with an electrode and noted what moved. With Edwin Boldrey he published the result in 1937: toes at the top of the brain, then knee, hip, trunk, arm and hand, then face, lips, tongue and throat down by the ear. Drawn to scale, this motor homunculus is a grotesque little figure with enormous hands and lips, because cortex is shared out by the precision a body part needs, not by its size.

Planning happens just in front. The premotor cortex and the supplementary motor area assemble sequences and get the body ready. In 1965 Hans Kornhuber and Lüder Deecke in Freiburg found a slow electrical build-up, the readiness potential, that starts a second or so before a voluntary movement. Commands then leave the motor cortex from giant Betz cells, described by Vladimir Betz in Kyiv in 1874, and travel down the corticospinal tract, about a million axons on each side. At the base of the brainstem most of them cross over, which is why a stroke in the left hemisphere weakens the right side of the body.

Did you know? In 2023 a team at Washington University found that the classic strip is interrupted by patches that light up for whole-body actions and planning, not for any single body part.

The neat little man is also something of a myth. Penfield's figure suggests one fixed spot per muscle, yet stimulating a monkey's motor cortex for half a second, as Michael Graziano showed in 2002, produces whole actions such as bringing the hand to the mouth. The map also redraws itself with use: musicians and recovering stroke patients both reshape it.

Reading the motor cortex has become medicine. In a study published in 2006, Matthew Nagle, paralysed below the neck, moved a cursor using an implant in this strip, and in 2012 Jan Scheuermann in Pittsburgh fed herself chocolate with a robotic arm. For everyone else the lesson is simpler: the strip devotes most space to what you practise, so skilled hands and a clear voice are partly built in the cortex.

Spinal Cord and Reflexes

The spinal cord is a column of nerve tissue about 45 centimetres long and roughly as thick as a finger, running inside the backbone from the brainstem to the small of the back. Thirty-one pairs of spinal nerves leave it, each carrying sensation in through a dorsal root and movement commands out through a ventral root. In cross-section it shows a grey butterfly of nerve cell bodies wrapped in white matter, the bundled cables that carry traffic up to the brain and down from it.

It is far more than a cable. When a doctor taps the tendon below your kneecap, the tap stretches tiny sensors called muscle spindles in the thigh. Their sensory fibres run into the cord and connect directly with the motor neurons that drive the same muscle. With just one synapse in the loop, the quadriceps contracts and the leg kicks in a few hundredths of a second, before the news of the tap has reached your awareness. The knee jerk was described in 1875 by Wilhelm Erb and Carl Westphal, independently, in Germany. Another neuron in the loop quietly relaxes the hamstring so the two muscles do not fight.

Charles Sherrington, working in Liverpool and Oxford, turned such reflexes into a science. He named the synapse in 1897, and his 1906 book The Integrative Action of the Nervous System showed how reflexes combine, so that stepping on a pin makes one leg withdraw while the other stiffens to take your weight. He shared the 1932 Nobel Prize with Edgar Adrian. In 1911 his colleague Thomas Graham Brown showed that circuits within the cord alone can produce the alternating rhythm of walking, now called a central pattern generator.

Did you know? Clenching your hands together and pulling hard, the Jendrassik manoeuvre described in 1883, makes a weak knee jerk much brisker, because the brain turns up the gain on spinal reflexes.

That trick explodes a myth. Reflexes are fast but not brainless: descending pathways constantly tune them, which is why their changes reveal damage higher up. In 1896 Joseph Babinski in Paris noticed that stroking the sole of the foot makes the big toe curl down in healthy adults but rise in people with damage to the corticospinal tract. Neurologists still check it every day.

Injury to the cord cuts the cables. The actor Christopher Reeve, paralysed by a fall from a horse in 1995, campaigned for research until his death in 2004. In 2018 Grégoire Courtine and Jocelyne Bloch in Lausanne used pulsed electrical stimulation of the lower cord to help David Mzee and other paralysed men take steps again, and in 2023 a wireless brain-to-spine bridge let Gert-Jan Oskam walk by thinking. The cord also explains everyday reflexes, from yanking a hand off a hot pan to keeping your balance on a bus: much of the body's quickest wisdom never needs the brain at all.

Learning a Skill

Watch anyone learning to drive. At first every action is spoken aloud in the head: mirror, signal, clutch down, find first gear. A few weeks later the steps blur together, and within months the driver can hold a conversation while changing lanes. In 1967 the psychologists Paul Fitts and Michael Posner described this progression as three stages: cognitive, when the learner thinks through each step; associative, when errors shrink and the movements link up; and autonomous, when the skill runs almost by itself.

Brain scans show the work moving house. Early practice leans heavily on the prefrontal cortex and attention networks, which is why a beginner is slow and easily distracted. The cerebellum then compares what was intended with what actually happened and corrects the next attempt; people wearing prism goggles that shift their view sideways miss a target at first, then adjust within a few dozen throws. As the skill matures, the basal ganglia and motor cortex take over. In rats learning a maze, Ann Graybiel's group at MIT found striatal neurons firing at the start and end of the run but going quiet in the middle, as if the whole route had been packed into one chunk.

The clearest proof that skills are stored separately from ordinary memories came from Henry Molaison, the patient H.M. In 1962 Brenda Milner in Montreal had him trace a star while watching his hand only in a mirror, a fiendish task. Over three days his errors fell steadily, yet each day he insisted he had never done it before. Skill learning, it turned out, does not need the hippocampus.

Did you know? Simply imagining piano practice for two hours a day changed the hand area of volunteers' motor cortex in a 1995 study by Alvaro Pascual-Leone, though real practice still did better.

Practice also changes the brain's structure. In 2004 Bogdan Draganski and colleagues in Regensburg taught volunteers to juggle for three months; scans showed grey matter growing in areas that track visual motion, then shrinking back when they stopped. The improvement never quite ends: in 1959 the engineer E. R. F. W. Crossman found cigar rollers still speeding up after years and millions of repetitions.

Two popular ideas need adjusting. Muscle memory for a skill lives in the brain, not the muscles. And practice works best spread out and followed by sleep: in 2002 Matthew Walker and colleagues found that people who learned a finger-tapping sequence were around 20 per cent faster the next morning, with no extra practice at all. Short, regular sessions, honest feedback on mistakes and a good night's rest are the closest thing the brain has to a shortcut.

The Autonomic Nervous System

While you read this, something is adjusting your heart rate, the width of your pupils, the squeeze of your gut and the sweat on your palms, and it has not asked your permission. This is the autonomic nervous system, named by the Cambridge physiologist John Newport Langley in 1898. Its control centres lie in the hypothalamus and brainstem, which read the body's state and send orders out along two branches that usually pull in opposite directions.

The sympathetic branch leaves the spinal cord from the chest and lower back and mostly uses noradrenaline. It widens the pupils, speeds and strengthens the heartbeat, opens the airways, releases glucose and slows digestion. In 1915 the Harvard physiologist Walter Cannon called its full-blown version the fight-or-flight response, and showed that the adrenal medulla, itself a kind of modified sympathetic nerve, pours adrenaline into the blood to spread the alarm.

The parasympathetic branch leaves from the brainstem and the base of the spine and uses acetylcholine. Its great highway is the vagus nerve, which wanders from the brainstem to the heart, lungs and gut. It narrows the pupils, slows the heart, gets saliva and stomach acid flowing and empties the bladder: rest and digest. Its role in the heart produced one of neuroscience's great experiments. In 1921 Otto Loewi in Graz stimulated the vagus nerve of a frog's heart, collected the fluid around it and dripped it on a second heart, which also slowed. He had shown that nerves talk with chemicals; his Vagusstoff was acetylcholine, and it won him a Nobel Prize in 1936.

Did you know? A transplanted heart, cut off from its nerves, beats at around 100 beats a minute at rest; the normal slower pace is the vagus nerve constantly holding the brakes.

The popular picture of a seesaw, with one branch on and the other off, is too simple. Both branches are active all day, each organ is tuned separately, and some targets hear only one: sweat glands get sympathetic nerves that, unusually, release acetylcholine. A sudden surge of vagal activity can even overshoot, dropping heart rate and blood pressure so fast that people faint at the sight of a needle.

Because these responses are hard to fake, people have tried to read them. John Larson's polygraph, built in Berkeley in 1921, tracks heart rate, breathing and sweating, but it detects arousal, not lies, and its verdicts are rarely admissible in court. The system can be nudged, though. Breathing in raises the heart rate slightly and breathing out lowers it, so slow breaths with a long exhale lean on the vagal brake. Calm is not only a thought; it is also a nerve.

The Gut–Brain Axis

The gut has more neurons than the spinal cord. Some 500 million of them form the enteric nervous system, a mesh woven through the wall of the oesophagus, stomach and intestines that can run digestion on its own, even when its links to the brain are cut. In 1998 the American anatomist Michael Gershon called it The Second Brain. It cannot think, but it does talk, and the conversation between gut and head is now called the gut–brain axis.

The main line is the vagus nerve. Around 80 per cent of its fibres carry messages up from the organs rather than down, reporting stretch, nutrients and irritation to the brainstem. In 2018 Diego Bohórquez's team at Duke University found gut lining cells they called neuropods, which form synapses with nerve fibres and pass on news of sugar within milliseconds. A second route is hormonal: cells in the gut wall release GLP-1, CCK and PYY after a meal, and ghrelin before one, all of which the brain reads. The gut also makes about 90 per cent of the body's serotonin, though this stays in the body, since serotonin cannot cross the blood–brain barrier.

The third and fourth routes involve the trillions of gut microbes and the immune cells that watch them. Bacteria ferment fibre into short-chain fatty acids and alter the supply of tryptophan, while inflammation sends cytokines through the blood. In 2004 Nobuyuki Sudo and colleagues in Japan showed that mice raised without any microbes overreacted to stress, and that giving them a single species of bacterium early in life calmed the response.

Did you know? In the 1820s the army surgeon William Beaumont watched digestion through a hole left in the stomach of a wounded fur trader, Alexis St Martin, and noticed that anger and fear changed what he saw.

Traffic runs downwards too. Stress, the smell of food or even the thought of a meal can speed or stall the gut through the autonomic nerves, which is why nerves can bring on a dash to the toilet before an exam. Irritable bowel syndrome is now thought of partly as a disorder of this two-way conversation. Some researchers also suspect that Parkinson's disease may start in the gut in some people, travelling up the vagus, though the evidence is debated.

The hype needs trimming. Studies in mice are striking, but evidence that probiotics treat depression or anxiety in people remains thin, and no single microbe has been shown to set human mood. What is solid is more ordinary: a varied, fibre-rich diet feeds a varied microbial community, regular meals keep gut hormones in rhythm, and a calm mind makes for a calmer gut. A gut feeling is a real signal, even if it is not a second opinion.

Hormones and the Brain

Neurons speak in milliseconds to their immediate neighbours. Hormones broadcast: released into the blood, they reach every cell in the body and act on any that carries the right receptor, over minutes, days or years. The brain sits at both ends of this system. The hypothalamus, through the pituitary gland beneath it, sets the release of most hormones, and the brain itself is studded with receptors for them, so they reshape mood, appetite, sleep and behaviour.

The first experiment was almost farmyard simple. In 1849 Arnold Berthold in Göttingen castrated cockerels, which then lost their combs and their interest in hens and fights. When he transplanted a testis into the abdomen of a castrated bird, the cockerel behaviour returned, even though no nerves were connected. Something in the blood was doing the work. That something, in this case, was testosterone, identified in 1935.

Oxytocin, made in the hypothalamus and released from the pituitary, triggers contractions in labour and the let-down of milk, and Vincent du Vigneaud synthesised it in 1953, earning a Nobel Prize. Its fame as a cuddle hormone is overblown. A 2005 Zurich study reported that a nasal spray of oxytocin made people more trusting with their money, but later attempts often failed to repeat it. The more solid behavioural evidence comes from voles: monogamous prairie voles have oxytocin and vasopressin receptors in reward areas of the brain where promiscuous meadow voles have few, and in 2004 boosting vasopressin receptors in meadow voles made the males pair-bond.

Did you know? In a 2010 study, women given testosterone made fairer offers in a bargaining game, while those who merely believed they had received it behaved more selfishly.

Other hormones shape the brain for life. Thyroid hormone is essential for brain development, which is why newborns in many countries have had a heel-prick test for an underactive thyroid since screening began in Quebec in 1974; iodine deficiency remains a leading preventable cause of intellectual disability worldwide. Oestrogen and testosterone remodel the brain at puberty, and the falling oestrogen of menopause disturbs a cluster of hypothalamic neurons that sets body temperature, producing hot flushes. A drug that calms those neurons was approved in 2023.

Some hormones keep time. The pineal gland releases melatonin in darkness, signalling night to the body, while the pituitary releases most of its growth hormone in pulses during deep sleep. Cortisol, the stress hormone of the next chapter, peaks soon after waking. The practical lesson is that the chemistry inside your head follows light, sleep, food and touch. Keep a steady rhythm of days and nights and you keep your hormones, and so your brain, in tune.

Stress

In 1936 a young researcher in Montreal, Hans Selye, noticed that rats injected with all sorts of noxious substances developed the same trio of problems: swollen adrenal glands, a shrunken thymus and stomach ulcers. The details of the insult hardly mattered; the body's response was generic. He later borrowed a word from engineering to name it, and stress entered every language on Earth.

The response runs on two tracks. The fast one starts when the amygdala flags a threat and the sympathetic nerves trigger adrenaline from the adrenal glands within seconds. The slower one is the HPA axis. Neurons in the hypothalamus release a hormone called CRH, identified by Wylie Vale's team in California in 1981; this tells the pituitary to release ACTH into the blood; ACTH tells the outer layer of the adrenal glands to release cortisol, which peaks some 20 to 30 minutes later. Cortisol frees up glucose, sharpens attention and helps lay down memories of what went wrong. Then it switches itself off: receptors in the hippocampus and hypothalamus sense the rising cortisol and shut the loop down.

Short bursts are healthy. Trouble starts when the brake fails and the loop runs for weeks. In 1992 Bruce McEwen's lab at Rockefeller University found that 21 days of daily restraint stress made the dendrites of rat hippocampal neurons wither, and later work showed the prefrontal cortex losing connections while the amygdala grew more. Robert Sapolsky, studying wild baboons in Kenya, found that low-ranking males, bullied without escape, had chronically raised stress hormones, and Michael Marmot's Whitehall studies of British civil servants found that people in lower grades with less control over their work had more heart disease.

Did you know? Stress was long blamed for stomach ulcers, until Barry Marshall and Robin Warren in Perth showed in the 1980s that most are caused by a bacterium; Marshall drank a culture of it to prove the point.

Myths cut both ways. Stress is not simply bad: performance improves with moderate arousal and falls off only when it climbs too high, the inverted U described by Robert Yerkes and John Dodson in 1908. And the same deadline can feel like a threat or a challenge depending on how much control a person has. What reliably damages health is stress that is severe, uncontrollable and never-ending.

That points to the everyday defences, and they are well tested. Exercise burns off the fuel the alarm releases, sleep resets the cortisol rhythm, and friends and a sense of control shorten the loop. The stress response is one of evolution's great inventions. It was built to switch off, and the harm comes when it never does.

Hunger and Appetite

Appetite feels like a matter of choice, but it is set by a cluster of cells smaller than a grain of rice. In 1940 Albert Hetherington and Stephen Ranson at Northwestern University damaged the ventromedial hypothalamus of rats, which then ate until they were grossly obese. A decade later at Yale, Bal Anand and John Brobeck found that damage to the lateral hypothalamus made rats refuse food altogether. Hunger, it was clear, had an address.

The key is the arcuate nucleus, at the base of the hypothalamus, where the blood–brain barrier is unusually leaky so that neurons can sample hormones in the blood. It contains two rival crews. AgRP/NPY neurons drive hunger: in 2011 Scott Sternson's team switched them on with light in well-fed mice, which immediately ate voraciously. POMC neurons do the opposite, releasing a signal that says enough. Each meal is a tug-of-war between the two.

Their inputs come from all over the body. Ghrelin, found by Masayasu Kojima and Kenji Kangawa in Japan in 1999, is released by the empty stomach and rises before meals. Leptin, made by fat cells, reports the long-term size of the energy store. Its discovery began with a fat mouse strain bred at the Jackson Laboratory in Maine in 1949; in the 1970s Douglas Coleman joined such mice to lean ones and showed a blood-borne factor was missing; and in 1994 Jeffrey Friedman at Rockefeller University cloned the gene. In 1999 Sadaf Farooqi and Stephen O'Rahilly in Cambridge gave leptin to a nine-year-old girl born unable to make it, and her relentless hunger subsided. Gut hormones such as GLP-1 and PYY add short-term fullness, and drugs that mimic GLP-1, such as semaglutide, approved for weight loss in 2021, act partly in the brain.

Did you know? The name leptin comes from the Greek leptos, thin, chosen because injecting it made leptin-deficient mice slim down dramatically.

The common myth is that obesity is simply a lack of leptin. Most people with obesity have plenty of it; their brains respond less, a state called leptin resistance, and the body defends its weight by raising hunger after dieting. Appetite is also not only about fuel. The reward circuit, driven by dopamine, can override fullness for something delicious, and sensory-specific satiety explains the famous space for dessert: we tire of one taste long before we tire of food.

All this changes how we might eat. Eating slowly gives gut hormones time to reach the brain, protein and fibre trigger stronger fullness signals, and short sleep raises ghrelin and lowers leptin, which is why tired people snack. Appetite is a body signal, not a character flaw, and working with it beats fighting it.

Mirror Neurons

In the early 1990s Giacomo Rizzolatti's team at the University of Parma was recording from single neurons in area F5, a region of the macaque monkey's premotor cortex involved in grasping. Some cells fired when the monkey picked up a piece of food. To the team's surprise, some of them also fired when the monkey simply watched an experimenter pick up food. In 1992 Giuseppe di Pellegrino and colleagues reported the finding, and by 1996 Vittorio Gallese, Rizzolatti and others had a name for these cells: mirror neurons.

The basic result is solid and has been repeated many times. Mirror-like cells have been found in the monkey's parietal cortex as well, and in 2005 Leonardo Fogassi showed that some of them fire differently depending on whether a grasp leads to eating or to placing an object, as if they encode the goal of the action. A popular story says the discovery came when a neuron fired as a researcher licked an ice cream; it is a good tale, but the published work describes experimenters grasping food.

Then came the leap. In 2000 the neuroscientist V. S. Ramachandran predicted that mirror neurons would do for psychology what DNA did for biology. They were proposed as the basis of empathy, imitation, language and culture, and the broken mirror theory suggested that autism arises from faulty mirror neurons. Popular books and articles repeated these ideas as settled fact.

Did you know? The first direct recordings of mirror-like neurons in humans came only in 2010, from 21 epilepsy patients at UCLA whose brains held electrodes for surgery planning.

That 2010 study, by Roy Mukamel, Itzhak Fried and colleagues, found cells that fired both for doing and for seeing actions, some of them in unexpected places such as the medial temporal lobe. Human brain scans had long shown overlapping activity for action and observation, but scans measure blood flow in areas containing millions of cells, so they cannot show that the same neuron does both. Critics, most prominently Gregory Hickok in his 2014 book The Myth of Mirror Neurons, pointed out that understanding an action does not require being able to perform it: people who cannot speak can still understand speech. Later reviews found little support for the broken mirror theory of autism.

The emerging view is modest but interesting. Mirror neurons are real, and they probably help the motor system recognise and anticipate actions, which may aid imitation and learning a movement by watching. They are not, on their own, the seat of empathy or the key to other minds. That makes them a useful lesson in how a genuine discovery can be stretched by enthusiasm: the cells map actions, and the rest is still being worked out.

The Brain and the Immune System

When you have flu, you feel feverish, sleepy, achy, uninterested in food and faintly miserable. For a long time these symptoms were blamed on the germs themselves. In 1988 the veterinarian Benjamin Hart of the University of California, Davis, argued instead that sickness behaviour is a strategy laid on by the brain: rest saves energy for the immune fight, fever makes the body hostile to microbes, and withdrawal keeps an infectious animal away from its group. The brain, in other words, takes orders from the immune system.

The messengers are cytokines, proteins such as interleukin-1 beta, TNF and IL-6 released by immune cells at the site of infection. They reach the brain by two main routes. One is neural: cytokines activate the vagus nerve, which reports to the brainstem within minutes, and in the 1990s Linda Watkins, Steven Maier and colleagues in Colorado showed that cutting the vagus below the diaphragm blocked several sickness responses in rats. The other is through the blood. Cytokines act on the walls of brain blood vessels, which release prostaglandin E2 into the hypothalamus and reset the body's thermostat upwards. Aspirin and ibuprofen bring down a fever by blocking the enzyme that makes it.

Inside the brain, the resident immune cells are the microglia, named in 1919 by Pío del Río-Hortega in Madrid and making up around a tenth of brain cells. They patrol constantly, sweeping their fine branches through the tissue and briefly touching synapses, and when infection or injury strikes they swell and release cytokines of their own. They are also gardeners. In 2007 Beth Stevens and colleagues showed that immune complement proteins tag weak synapses in the developing brain for removal, and that microglia engulf them. In 2016 variants of one complement gene, C4, were linked to the risk of schizophrenia, raising the possibility that over-pruning plays a part.

Did you know? In 2015 two teams, in Virginia and in Helsinki, independently found lymphatic vessels in the meninges around the brain, drainage channels that anatomy textbooks had missed for centuries.

That discovery buried an old myth. In 1948 Peter Medawar showed that tissue grafts survive unusually well in the brain, and the brain came to be described as immune-privileged and walled off. The barrier is real, but the traffic is constant. Immune attacks on the brain cause multiple sclerosis, and in people treated with the immune protein interferon-alpha for hepatitis C, around a third developed depression, a clear sign that inflammation can alter mood.

The two-way link matters for everyday health. Long-term inflammation from obesity, smoking or poor sleep is associated with low mood and faster cognitive decline in later life, and regular exercise, which reduces inflammation, is good for both. Next time you feel wretched with a cold, remember that the misery is partly your brain obeying your immune system, and rest is exactly what it is asking for.

Patient H.M. and the Hippocampus

On 1 September 1953, at Hartford Hospital in Connecticut, the neurosurgeon William Beecher Scoville tried to cure a 27-year-old man of crippling epilepsy by suctioning out the inner part of both temporal lobes. The seizures eased. Something else went badly wrong. Henry Molaison, known to science for 55 years only as H.M., could no longer make new memories. He would read the same magazine with fresh interest, grieve anew each time he was told that his uncle had died, and greet researchers who had worked with him for decades as polite strangers.

How we know what went wrong is down largely to Brenda Milner, a British-born psychologist at the Montreal Neurological Institute. In 1957 she and Scoville published his case, and her careful testing showed that his intelligence, language and personality were intact, as were most memories from his childhood. What he had lost was the ability to turn new experience into lasting memory: anterograde amnesia. The removed tissue included most of the hippocampus, a curled ridge of cortex named after the Greek for seahorse, along with the amygdala and the entorhinal cortex beside it. Scoville estimated that he had taken 8 centimetres on each side; MRI scans in the 1990s showed it was nearer 5.

Henry's world was not a blank. Holding information in mind still worked: in one famous session he kept the number 584 for fifteen minutes by rehearsing an elaborate arithmetic scheme, then lost it the moment his attention moved. In 1962 Milner had him trace a star seen only in a mirror. Over three days his tracing improved like anyone's, though each day he insisted he had never tried it before. Memory, it turned out, was not one thing. Skills could be learned without the hippocampus; new facts and episodes could not.

Did you know? After Henry died in December 2008, his brain was cut into 2,401 thin slices at the University of California, San Diego, in a 53-hour session that was streamed live online.

Suzanne Corkin at MIT studied him for 46 years, and her work refined the picture. Henry could slowly absorb a few new facts through sheer repetition, and he could sketch the floor plan of a house he moved into years after the operation. He had also lost memories from some years before the surgery, suggesting that the hippocampus stays involved while new memories settle into the cortex, a process called consolidation. Later cases sharpened the rule. Patient R.B., studied by Stuart Zola-Morgan, Larry Squire and David Amaral in San Diego, became amnesic in 1978 after a loss of blood supply destroyed just one small field of cells, CA1, in each hippocampus.

Henry's tragedy also ended a practice: surgeons stopped removing both hippocampi, and epilepsy surgery today takes tissue from one side only, after memory has been mapped. His case also explains why a stroke, a head injury or the early tangles of Alzheimer's disease, which strike the hippocampal region first, so often begin with forgetting this morning rather than 1960. Henry never knew he was famous, but he gave science its map of memory.

Kinds of Memory

Ask someone what memory is and they will usually describe one thing: a store of past events. Neuroscience sees a family of systems, each with its own job and its own address in the brain. The modern map grew from patients like Henry Molaison, who could learn a motor skill without remembering the lessons, and was drawn up formally by Larry Squire in San Diego in the 1980s. Its first split is between declarative memory, which you can consciously report, and nondeclarative memory, which shows itself only in what you do.

In 1972 the Estonian-born Canadian psychologist Endel Tulving divided declarative memory in two. Episodic memory is mental time travel: your last birthday, where you parked, the smell of a grandparent's kitchen, each tagged with a when and a where. Semantic memory is knowledge stripped of its origin, such as the fact that Paris is the capital of France, though you probably cannot recall learning it. Tulving's patient K.C., a Toronto man injured in a motorcycle accident in 1981, showed that the two can come apart: he knew facts about his own life, such as that his family owned a summer cottage, yet could not re-experience a single event that happened there. People with semantic dementia show the reverse pattern, losing the meanings of words and objects while recent episodes survive for longer.

Nondeclarative memory is a looser coalition. Procedural memory for skills such as cycling or typing depends on the basal ganglia and cerebellum. Priming makes a word or picture you met recently easier to process again; in 1968 Elizabeth Warrington and Lawrence Weiskrantz in London showed that amnesic patients completed word fragments they had seen before while denying that they had seen any words at all. Classical conditioning, studied by Ivan Pavlov in St Petersburg from the 1890s, links a signal to an outcome, through the cerebellum for reflexes such as the eye-blink and through the amygdala for fear.

Did you know? In 1911 the Swiss doctor Édouard Claparède hid a pin in his hand when greeting an amnesic patient; the next day she refused to shake his hand, though she could not say why.

The most moving demonstration is Clive Wearing, a British musician and conductor whose hippocampi were destroyed by herpes encephalitis in 1985. His memory for events spans seconds, and his diary records, again and again, that he has just woken up for the first time. Yet he can still play the piano, sight-read and conduct, and he greets his wife Deborah with joy every time she walks into the room.

The familiar split between short-term and long-term memory is real but incomplete; Squire's map is about kinds, not just durations. It also means that losing one kind rarely means losing all. For daily life the lessons are practical: facts stick better when tied to a vivid episode, skills need doing rather than reading, and routines learned by the body can outlast any memory of learning them. Memory is less a library than a household of specialists.

Working Memory

Read a phone number, turn to dial it, and you are using working memory, the small mental workspace that holds information for a few seconds while you do something with it. In 1956 the American psychologist George Miller published a celebrated paper on "the magical number seven, plus or minus two", the number of items people can typically repeat back. He opened it with a famous complaint: "My problem is that I have been persecuted by an integer." Later work, summed up by Nelson Cowan in 2001, argued that once rehearsal and grouping are stripped away the true limit is nearer four chunks. Either way, the workspace is tiny.

In 1974 Alan Baddeley and Graham Hitch, working in Britain, proposed that it is not a single box but a team. A central executive directs attention. A phonological loop holds about two seconds of speech, refreshed by an inner voice, which is why lists of long words are harder to keep than lists of short ones. A visuospatial sketchpad keeps images and locations, and in 2000 Baddeley added an episodic buffer that binds the pieces into scenes and links them to long-term memory. Language shapes the loop: Welsh number words take longer to say than English ones, and in 1980 Nick Ellis and Richard Hennelly found that bilingual children had shorter digit spans in Welsh than in English.

Patients show that this workspace is separate from long-term memory. Henry Molaison had a normal digit span but could not store anything new. In 1970 Tim Shallice and Elizabeth Warrington described the reverse case, K.F., a man with damage to the left hemisphere whose span for spoken digits was about two items, yet who learned new material normally. Recording in monkeys, Joaquín Fuster in 1971 and Patricia Goldman-Rakic in the 1980s found neurons in the prefrontal cortex that keep firing through a delay, holding a location or object in mind until the animal responds.

Did you know? The word "chunk" entered psychology through Miller's 1956 paper; he noticed that people remember about seven items whether the items are digits, letters or whole words.

Capacity matters. Working memory predicts reading comprehension, mental arithmetic and reasoning remarkably well, and it shrinks under stress, after sleep loss and with age. That has fuelled a market for brain-training games. The evidence is sobering: in 2010 Adrian Owen and colleagues tested 11,430 viewers of the BBC programme Bang Goes the Theory and found that six weeks of online training improved the trained tasks but not general abilities. Players got better at the games.

The practical lesson is to work with the limit rather than against it. Write things down, break tasks into steps, group digits the way phone numbers already are, and avoid stacking demands such as driving while arguing on the phone. The workspace is small, but it is where every thought gets done.

How a Memory Is Made

In 1949 the Canadian psychologist Donald Hebb proposed a simple rule: when one neuron repeatedly helps to fire another, the connection between them grows stronger. It is often paraphrased as "cells that fire together wire together", a slogan coined decades later by the neuroscientist Carla Shatz. Hebb could not see synapses change. Twenty-four years later, two researchers in Oslo did.

In 1973 Timothy Bliss and Terje Lømo stimulated a pathway into the hippocampus of anaesthetised rabbits with brief high-frequency bursts. Afterwards a single test pulse produced a much bigger response than before, and the boost lasted for hours; later experiments stretched it to days and weeks. They had found long-term potentiation, or LTP, still the leading candidate for how experience is written into the brain.

The mechanism hinges on the neurotransmitter glutamate and two kinds of receptor on the receiving spine. AMPA receptors open whenever glutamate arrives, letting in sodium. NMDA receptors are plugged by a magnesium ion and open only when glutamate arrives while the cell is already strongly excited, which makes them detectors of coincidence. When they open, calcium floods in and switches on enzymes such as CaMKII, which insert extra AMPA receptors into the membrane, so the synapse speaks louder next time. Lasting memories need more: genes switched on by the protein CREB build new proteins and even new spines. Eric Kandel traced this sequence in the sea slug Aplysia and shared the 2000 Nobel Prize for it. In 1986 Richard Morris in Edinburgh showed that rats given a drug that blocks NMDA receptors could not learn where a hidden platform lay in a pool of cloudy water.

Where does a whole memory live? The German zoologist Richard Semon coined the word engram in 1904. The American psychologist Karl Lashley spent decades cutting pieces out of rats' cortex without finding it, and in 1950 concluded, half in jest, that learning seemed impossible. Modern tools found it after all. In 2012 Xu Liu, Steve Ramirez and Susumu Tonegawa at MIT tagged the hippocampal neurons active while mice learned to fear a box, then reactivated just those cells with light: the mice froze in a completely different box. An engram, it seems, is a sparse, scattered set of cells, a small fraction of a region, bound together by strengthened synapses.

Did you know? In 2013 the same MIT team planted a false memory in mice, reactivating cells that held the memory of a harmless box while the animals received a mild shock elsewhere; afterwards the mice froze in the safe box.

Myths die here. Memories are not stored like files in single neurons or as recordings; they are patterns of connection, and recalling one briefly makes it changeable again. Because building new proteins takes hours, and repeated activation strengthens the trace, the biology favours spaced practice, rest between sessions and a good night's sleep. Every lasting memory is a small piece of construction work.

Forgetting

Between 1879 and 1885 the German psychologist Hermann Ebbinghaus memorised lists drawn from a pool of about 2,300 meaningless syllables, such as DAX, BOK and YAT, and then tested how quickly he forgot them. He was his own only subject. To measure memory precisely he used savings: how much less time it took to relearn a list than to learn it the first time. The results, published in Über das Gedächtnis in 1885, gave psychology its first great graph, the forgetting curve.

The curve falls steeply and then flattens. Twenty minutes after learning, Ebbinghaus saved about 58 per cent of the original effort; after an hour, 44 per cent; after a day, 34 per cent; after a month, 21 per cent. Most forgetting happens almost at once, and what survives the first day tends to linger. In 2015 Jaap Murre and Joeri Dros in Amsterdam repeated the experiment, with Dros spending some 70 hours learning syllables, and found almost the same shape.

Why do we forget? Some traces fade with disuse. Others are crowded out by interference from similar material learned before or after. In 1924 John Jenkins and Karl Dallenbach at Cornell showed that two students forgot far less after a night's sleep than after the same hours awake, because waking life piles on new experiences. Many memories are not lost at all but unreachable, as anyone who has felt a name on the tip of the tongue knows; Roger Brown and David McNeill captured that state in the laboratory at Harvard in 1966. And the brain forgets on purpose: in 2001 Michael Anderson and Collin Green showed that people told repeatedly to suppress a memory later recalled it less well.

Forgetting is a feature. The Russian journalist Solomon Shereshevsky, studied for some 30 years by the neuropsychologist Alexander Luria, could recall long tables of numbers years later, yet struggled to grasp metaphors or the gist of a story because every detail clung on. Jorge Luis Borges imagined the same fate in his 1942 story "Funes the Memorious", whose hero remembers so much that he can barely think. A memory that keeps the general and discards the trivial is better at predicting the future.

Did you know? People with highly superior autobiographical memory, first described in California in 2006, can say what they did on almost any date of their adult lives, yet score about as well as anyone else on ordinary memory tests.

Ebbinghaus also found the remedy: reviews spread over time flatten the curve. Modern research agrees. Spaced repetition, the principle behind flashcard apps such as Anki, and the testing effect, demonstrated by Henry Roediger and Jeffrey Karpicke in 2006, both beat rereading. Forget a little, retrieve with effort, and the memory comes back stronger.

False Memories

Memory feels like a recording you replay. It is closer to a story you retell. In 1932 the Cambridge psychologist Frederic Bartlett asked English students to read and later recall a Native American folk tale, "The War of the Ghosts". With each retelling the strange details dropped out or turned familiar: canoes became boats, and the supernatural twist was explained away. Remembering, Bartlett concluded, is reconstruction, with the gaps filled in by expectation.

Elizabeth Loftus turned the idea into experiments. In 1974 she and John Palmer in Seattle showed students film of a car crash and asked how fast the cars were going when they "smashed into" or "hit" each other. "Smashed" produced estimates of around 41 miles per hour, against 34 for "hit", and a week later more than twice as many in the smashed group remembered broken glass that was never there. In 1995 Loftus and Jacqueline Pickrell went further, persuading about a quarter of their volunteers that they had once been lost in a shopping mall as children. Kimberley Wade's team in New Zealand later used a doctored photograph to give half their volunteers some memory of a hot-air balloon ride they had never taken.

The brain helps explain why. Recalling a memory reactivates the network that holds it, and in 2000 Karim Nader, Glenn Schafe and Joseph LeDoux in New York showed that a reactivated fear memory in rats became fragile again: a drug that blocks protein synthesis, given just after recall, wiped it out. This reconsolidation means that each act of remembering is also an act of rewriting, open to new suggestions, moods and knowledge.

Did you know? The psychologist Jean Piaget vividly remembered his nurse fighting off a kidnapper when he was a baby in Paris, until she confessed years later that she had invented the whole story.

The stakes are real. In 1984 Jennifer Thompson, a student in North Carolina, studied her attacker's face carefully and confidently identified Ronald Cotton. He spent more than ten years in prison before DNA evidence freed him in 1995 and identified the real attacker, Bobby Poole. Mistaken eyewitness identification played a part in around seven in ten of the convictions overturned by DNA evidence in the United States. Confidence is no guarantee either. The morning after the Challenger shuttle exploded in 1986, Ulric Neisser asked students how they had heard the news; asked again about two and a half years later, many gave quite different accounts and held them with great certainty.

None of this makes memory useless. The gist is usually right, and reconstruction is the price of a system built for flexibility rather than archiving. But it argues for humility. Police are now advised to record a witness's first account and confidence at once, and therapists to avoid suggestive questioning after the recovered-memory scandals of the 1990s. Write important things down early. Every memory is a draft.

Habits

Around four in ten of the things people do each day are done habitually, in the same place and without much thought, according to diary studies published in 2002 by Wendy Wood and colleagues at Texas A&M University. A habit is a behaviour that has been handed over from deliberate choice to an automatic routine triggered by context, and its home is the basal ganglia, the cluster of nuclei buried beneath the cortex.

The handover was caught in action at MIT. In the late 1990s Ann Graybiel's team recorded neurons in the striatum of rats learning to run a T-shaped maze for chocolate at the sound of a tone. Early on, cells fired throughout the run. As the route became automatic, activity gathered at the start and the finish, as if the whole sequence had been packaged into a single chunk, a pattern Graybiel called task-bracketing. The popular summary, made famous by Charles Duhigg's 2012 book The Power of Habit, is a loop: a cue triggers a routine that ends in a reward, while dopamine signals teach the striatum which cues predict good outcomes.

Two brain systems compete for control. In 1996 Mark Packard and James McGaugh trained rats in a cross-shaped maze. Early in training the rats navigated by place, using the hippocampus; after more training they simply repeated the same turn, using the caudate nucleus, and numbing that region switched them back. In people, Barbara Knowlton, Jennifer Mangels and Larry Squire found in 1996 that patients with Parkinson's disease, whose striatum is starved of dopamine, struggled to learn a probabilistic weather-forecasting task, while amnesic patients learned it normally but could not remember the training sessions. Later studies traced a shift inside the striatum itself, from a goal-directed dorsomedial part to a habitual dorsolateral part.

Did you know? Psychologists test for a habit by spoiling the reward: in Anthony Dickinson's experiments at Cambridge in the 1980s, overtrained rats kept pressing a lever for food they had been taught to dislike.

How long does a habit take to form? The often-quoted 21 days traces back to a 1960 self-help book by the plastic surgeon Maxwell Maltz, who noticed how long patients took to get used to a new face. When Phillippa Lally and colleagues at University College London followed 96 volunteers adopting a new daily behaviour, published in 2010, the median time to automaticity was 66 days, ranging from 18 to 254. Missing a single day made little difference.

The same machinery underlies addictions and the compulsions of obsessive-compulsive disorder, where routines detach from any reward. For daily life the lesson is to design cues rather than rely on willpower: keep the running shoes by the door, the biscuits out of sight, and the new routine attached to an old one. Habits free attention for everything else, which is exactly why they are so hard to unlearn.

Neuroplasticity

To earn a licence to drive a black cab in London, trainees must master the Knowledge: about 25,000 streets and thousands of landmarks within six miles of Charing Cross, a feat that usually takes three to four years. In 2000 Eleanor Maguire and colleagues at University College London scanned the brains of 16 licensed drivers. The back of their hippocampus, a region tied to spatial memory, was larger than in men who did not drive taxis, and the longer they had been driving, the larger it was. The front part was slightly smaller.

A difference in a scanner could simply mean that people with big hippocampi become cabbies. So Katherine Woollett and Maguire followed 79 trainees through their years of study and reported in 2011 that the 39 who qualified showed growth in the posterior hippocampus, while those who failed or gave up did not. Experience had reshaped the adult brain. This capacity is called neuroplasticity, and it overturned the old belief that the brain is fixed after childhood.

Other studies make the same point. In 2004 Bogdan Draganski and colleagues in Regensburg taught volunteers to juggle three balls for three months; grey matter grew in areas that track visual motion, then shrank back once they stopped. In 1995 Thomas Elbert's group in Konstanz found that string players gave more cortex to the fingers of the left hand, the hand that presses the strings, especially those who began young. And in the early 1980s Michael Merzenich and Jon Kaas showed in monkeys that when the nerve from part of the hand was cut, neighbouring fingers took over its patch of cortex within weeks.

Did you know? Maguire's team found that retired London cab drivers had less grey matter in the posterior hippocampus than working ones, as if the advantage faded with disuse.

There are limits and windows. In the 1960s David Hubel and Torsten Wiesel, a Canadian and a Swede working at Harvard, closed one eye of kittens during their first months: the visual cortex rewired to favour the open eye, and the change was permanent, whereas the same closure in adult cats did little. Their work shared the 1981 Nobel Prize and is why a child's lazy eye is treated early. Plasticity is greatest in childhood, as children who have had an entire hemisphere removed for severe epilepsy show by relearning to walk and talk, but it never stops.

Plasticity cuts both ways. The same rewiring contributes to chronic pain, tinnitus and phantom limbs. Yet it is also the basis of stroke rehabilitation: Edward Taub's constraint-induced therapy restrains the good arm so the weak one must work, helping the brain reclaim movement. For everyone else the message is that practice leaves a physical mark at any age, and that what goes unused fades. The brain is less a finished machine than a path worn by walking.

Sleep and Memory

In 1924 John Jenkins and Karl Dallenbach at Cornell University asked two students to learn lists of nonsense syllables, then tested them after hours spent either asleep or awake. After sleep they remembered markedly more. Jenkins and Dallenbach thought sleep simply sheltered memories from interference. A century of research points to something more active: during sleep the brain replays the day and moves what matters into long-term storage.

The night is built from cycles of about 90 minutes that descend from light non-REM sleep into deep slow-wave sleep and then rise into REM sleep, when dreams are most vivid. Each stage seems to contribute. In stage N2 the thalamus generates sleep spindles, bursts of rhythm at 11 to 16 cycles per second, and people with more spindles tend to show bigger overnight gains in learning. In deep sleep the cortex swings between activity and silence roughly once a second, and these slow oscillations set the timing for everything else.

The key evidence came from rats. In 1994 Matthew Wilson and Bruce McNaughton in Tucson recorded place cells, hippocampal neurons that fire at particular spots in a maze, and found that cells active together during the day fired together again in later sleep, as if rehearsing the route. This replay rides on sharp-wave ripples, brief high-frequency bursts in the hippocampus that tend to coincide with spindles and slow oscillations, a tightly timed handshake thought to pass memories to the cortex. In 2009 a team led by Gabrielle Girardeau and Michaël Zugaro in Paris disrupted the ripples electrically while rats slept, and their learning suffered.

Did you know? Replay during sleep runs in fast-forward: a route that a rat took seconds to run can be replayed by its hippocampus within a fraction of a second.

People can be helped along. In 2006 Lisa Marshall and Jan Born in Lübeck boosted slow oscillations with gentle electrical stimulation and improved recall of word pairs. In 2007 Björn Rasch, Born and colleagues released the scent of roses while volunteers learned the positions of cards, then again during their deep sleep, and recall next morning was better. Sleep also seems to recombine knowledge: in 2004 Ullrich Wagner's team found that a night's sleep more than doubled the chance that people would discover a hidden shortcut in a number task. The role of REM sleep, perhaps in emotional memories and some skills, is still debated.

Sleep before learning matters too. In 2007 Seung-Schik Yoo, Matthew Walker and colleagues at Harvard found that after a night without sleep the hippocampus responded weakly during learning, and later memory was poorer. An all-nighter before an exam therefore fails twice, blunting what goes in and skipping the night that would have stored it. Learn, then sleep on it: the night is when the day is written down.

Expertise

In 1946 the Dutch psychologist and chess master Adriaan de Groot showed players a position from a real game for a few seconds, then asked them to rebuild it on an empty board. Grandmasters placed almost every piece correctly; weaker club players managed far fewer. Yet when he asked them to think aloud while choosing a move, the grandmasters did not search further ahead than the club players. They simply saw the good moves first.

In 1973 William Chase and Herbert Simon at Carnegie Mellon in Pittsburgh found out why. Shown a real middle-game position for five seconds, a master recalled about 16 pieces on the first attempt, a good amateur about 8 and a beginner about 4. But when the same pieces were scattered at random, the master's advantage all but vanished. The expert was not blessed with a better memory. He had built a vast library of familiar patterns, or chunks, estimated at around 50,000, each recognised at a glance. His working memory still held only a handful of items, but each item had grown enormous.

Chunking can be trained. In the late 1970s Anders Ericsson, William Chase and Steve Faloon worked with a Carnegie Mellon student known as S.F., a keen club runner. After more than 200 hours of practice his span for digits read aloud rose from 7 to 79, because he recoded strings of digits as running times, so that 3492 became three minutes 49.2 seconds, a near world-record mile. His span for letters stayed at about six. Expertise, it turned out, is narrow.

Did you know? Judit Polgár, schooled in chess at home in Budapest by a father who believed that geniuses are made, became a grandmaster in 1991 at 15, younger than Bobby Fischer had been.

How do experts get that way? In 1993 Ericsson, Ralf Krampe and Clemens Tesch-Römer studied violin students at a Berlin music academy and estimated that the best had accumulated about 10,000 hours of solitary deliberate practice by the age of 20, far more than less accomplished peers. Malcolm Gladwell's 2008 book Outliers turned this into a "10,000-hour rule", which Ericsson himself disputed: there is no magic number, and the quality of practice matters as much as the quantity. A 2014 meta-analysis led by Brooke Macnamara found that practice explained about a quarter of the differences in performance in games and about a fifth in music, and far less in professions. Starting age, teaching, temperament and working memory count too.

In the brain, expertise often looks like efficiency: experts recognise rather than reason, which is how a radiologist spots a shadow at a glance and a seasoned firefighter senses that a floor is about to give way. For the rest of us the recipe is unglamorous: focused practice just beyond current ability, quick feedback, and time for thousands of patterns to settle. Expertise is memory, organised.

Fear and the Amygdala

Step off a kerb as a bus swings past and your body has already lurched back before you know why. That reaction is organised by the amygdala, a pair of almond-sized clusters of nuclei buried in each temporal lobe, named from the Greek for almond by the German anatomist Karl Burdach in the early 19th century. It is the brain's smoke detector: cheap, fast and happy to raise a false alarm rather than miss a real fire.

The first clue came from Chicago in 1937, when Heinrich Klüver and Paul Bucy removed the temporal lobes of rhesus monkeys. The animals became strangely tame, handling snakes and approaching people they had once fled. In the 1980s and 1990s Joseph LeDoux, working with rats in New York, traced the wiring. Sound reaches the amygdala by two routes: a low road straight from the thalamus, crude but quick, and a high road through the auditory cortex, slower but detailed. The low road lets you jump at a stick that might be a snake; the high road tells you, a beat later, that it is only a stick.

Fear is also learned, and the lateral amygdala is where the lesson is written. In 1920 John Watson and Rosalie Rayner taught an infant known as Little Albert to fear a white rat by banging a steel bar behind him whenever he reached for it. Rats show the same thing: pair a tone with a mild shock a few times and the tone alone makes them freeze. Once triggered, the central nucleus sends orders outwards, to the hypothalamus for stress hormones and a racing heart, to the periaqueductal grey for freezing, and to the brainstem for a bigger startle.

Did you know? The patient known as S.M., whose amygdalae were destroyed by a rare disease, felt no fear of snakes, spiders or a haunted house, yet panicked when she breathed air laced with carbon dioxide in a 2013 study.

S.M., studied in Iowa by Antonio Damasio, Ralph Adolphs and Justin Feinstein, also struggled to recognise fear in other people's faces, as a 1994 paper showed. Her case corrects a popular myth. The amygdala is not simply a fear centre: it flags whatever matters, including rewards, and fear felt from inside the body, such as suffocation, can bypass it. LeDoux himself argues that it runs defensive survival reactions, while the conscious feeling of being afraid is built elsewhere in the cortex.

Unlearning fear is a separate act. When the tone keeps sounding without shocks, the ventromedial prefrontal cortex learns to hold the amygdala back, a process called extinction. The old memory survives, which is why fears can return under stress, and why exposure therapy for phobias and post-traumatic stress works step by step, in many sessions. Facing a fear repeatedly does not delete it; it teaches the front of the brain to outvote it.

Reward and Dopamine

In 1954, at McGill University in Montreal, James Olds and Peter Milner placed an electrode slightly off target in a rat's brain and noticed that the animal kept returning to the corner of its box where it had been stimulated. Given a lever, rats with electrodes in the right places pressed it thousands of times an hour, ignoring food. The pair had discovered a reward system, and the chemical at its heart turned out to be dopamine, made by a few hundred thousand neurons in two midbrain clusters, the ventral tegmental area and the substantia nigra.

For decades dopamine was called the pleasure chemical. The truth is subtler. In the 1980s and 1990s Wolfram Schultz, in Fribourg in Switzerland, recorded single dopamine neurons in monkeys learning that a light predicted a squirt of juice. At first the neurons fired when the juice arrived. Once the light had become a reliable signal, they fired at the light and ignored the juice. When the light came on and the juice did not, firing dipped below its normal rate at the moment the treat should have arrived. In 1997 Schultz, Peter Dayan and Read Montague showed that this pattern matched a reward prediction error from machine learning: better than expected, worse than expected, or just as expected.

Dopamine is therefore a teaching signal more than a feeling. It updates expectations and drives the striatum to repeat whatever paid off. Kent Berridge at the University of Michigan made the second split, between wanting and liking. Rats whose dopamine had been almost entirely destroyed still made the same contented tongue-licking faces when given sugar, but would no longer work to get it. Liking depends on tiny opioid hotspots in the nucleus accumbens, about a cubic millimetre in a rat; wanting is dopamine's job.

Did you know? Some people taking dopamine-boosting drugs for Parkinson's disease develop compulsive gambling or shopping, which usually fades when the dose is reduced.

The idea explains a great deal of ordinary life. Rewards that arrive on an unpredictable schedule, a pattern B. F. Skinner studied with pigeons, keep generating prediction errors, which is why slot machines, lottery scratchcards and refreshing a feed are hard to stop. Addictive drugs such as cocaine and amphetamine push dopamine directly, so the teaching signal shouts long after the pleasure has gone flat. The gap between wanting and liking helps explain why someone can crave what they no longer enjoy.

It also explains why the second slice of cake is never as thrilling as the first bite, and why a pay rise feels wonderful for a few weeks and then normal. Dopamine stops firing for what it has learned to expect. Keeping some surprise in life, a new route, a new skill, an unplanned treat, is how to keep the signal interested.

Love and Bonding

On the grasslands of the American Midwest lives a small rodent with unusual habits. The prairie vole pairs for life: male and female share a nest, raise pups together and snub strangers. Its close cousin the montane vole mates and moves on. In the 1990s Sue Carter, Thomas Insel and Larry Young traced the difference to receptors in the brain. Prairie voles have far more receptors for oxytocin in the nucleus accumbens and for vasopressin in the ventral pallidum, both parts of the reward system. In 2004 Young's team at Emory University used a virus to add vasopressin receptors to the ventral pallidum of male meadow voles, another promiscuous species, and the males began to prefer one partner.

Oxytocin is an old molecule with a short recipe of nine amino acids. It is made in the hypothalamus and released from the posterior pituitary, and Henry Dale in London found in 1906 that it made the womb contract. Vincent du Vigneaud synthesised it in 1953, earning a Nobel Prize. Its first jobs were labour and milk let-down; evolution seems to have borrowed the same signal to tie mother to infant, and in some species partner to partner, by making that one individual especially rewarding.

Attachment begins in infancy. In 1958 Harry Harlow at the University of Wisconsin raised baby rhesus monkeys with two dummy mothers, one of bare wire holding a milk bottle and one of soft cloth with none. The infants fed from the wire mother and spent almost all their time clinging to the cloth one, a blow to the theory that babies love whoever feeds them. John Bowlby in London and Mary Ainsworth, who studied mothers and babies in Uganda and Baltimore, built that finding into attachment theory. Children raised in Romanian orphanages under Ceaușescu, studied after 1989, showed how deep the need runs: those moved into foster families before about the age of two recovered markedly better.

Did you know? When Helen Fisher, Arthur Aron and Lucy Brown scanned people newly and intensely in love in 2005, a photo of the beloved activated the ventral tegmental area and caudate, the same dopamine circuits that cocaine excites.

Popular science calls oxytocin the cuddle hormone, which oversells it. A 2005 Zurich study suggested a nasal spray made people more trusting with money, but later attempts often failed to repeat it, and spray may barely reach the brain. Oxytocin also sharpens loyalty to one's own group and wariness of outsiders, as Carsten De Dreu reported in Amsterdam in 2010. It boosts attention to social signals rather than spreading goodwill.

Bonds, once built, act on the body. Holding a partner's hand reduced threat responses in the brain in a 2006 study led by James Coan, and loneliness is linked to poorer sleep and health. Love is not one chemical but a circuit trained by touch, time and reward, and it is kept by tending.

Empathy

Watch someone slam a finger in a car door and you wince. That flinch is the start of empathy, a word coined in English only in 1909, from the German Einfühlung, feeling into. Neuroscientists now split it into three overlapping skills: sharing another person's feeling, working out what they think, and caring enough to help. Each leans on different circuits, and each can fail on its own.

The sharing part was caught in a scanner in 2004. Tania Singer and colleagues at University College London scanned women while either they or their partners, sitting beside the scanner, received mild electric shocks to the hand. When the partner was hurt, the woman's anterior insula and anterior cingulate cortex became active, regions that also responded to her own pain. The areas that register where and how hard the body is touched stayed quiet. We share the unpleasantness of another's pain, not its sensation.

Reading minds is a separate skill, called theory of mind. In 1983 Heinz Wimmer and Josef Perner in Salzburg showed that most children younger than four cannot grasp that someone else holds a false belief. Simon Baron-Cohen, Alan Leslie and Uta Frith in London turned this into the Sally–Anne test in 1985: Sally hides a marble in a basket, Anne moves it to a box while Sally is out, and the child is asked where Sally will look. Brain imaging, notably by Rebecca Saxe at MIT, ties this reasoning to the temporoparietal junction, behind and above the ear.

Did you know? In a 2011 Chicago study, rats learned to open a little door to free a trapped cagemate, even when a pile of chocolate chips sat nearby, and often shared the chocolate afterwards.

Empathy has limits, and they are not flattering. In a 2006 study by Singer's team, men's pain-sharing responses shrank, and reward areas became active, when they watched someone who had cheated them in a game being shocked. People give generously to a single named child and far less to thousands of unnamed victims; when the toddler Jessica McClure fell down a well in Texas in 1987, donations poured in. The psychologist Paul Bloom argues that empathy works like a spotlight, intense but narrow and biased towards the familiar. People with psychopathy can read minds well yet show weak amygdala responses to others' distress.

Caring is the part that can be trained. Singer and Olga Klimecki found that training people simply to feel others' suffering increased their own distress, while compassion training, which fosters warmth and concern, increased positive feelings and helping. Nurses, doctors and carers risk burnout from the first and draw strength from the second. Feeling with someone is a reflex; acting kindly for them is a skill.

Making Decisions

A businessman known as Elliot had a tumour removed from just above his eye sockets. Afterwards he scored normally on intelligence and memory tests, yet his life collapsed. He could spend an afternoon deciding how to sort documents, lost his job and his savings, and when asked to pick between two dates for an appointment weighed the merits of each for half an hour. His neurologist, Antonio Damasio in Iowa, noticed that Elliot described his own misfortunes with no emotion at all. The damage lay in the ventromedial prefrontal cortex, a region that links knowledge with feeling.

Damasio's team built a game to capture the problem. In the Iowa Gambling Task, first published by Antoine Bechara and colleagues in 1994, players draw cards from four decks of play money. Decks A and B pay large wins but hide larger losses; decks C and D pay modestly and lose little, so they win in the long run. Healthy players drift towards C and D within a few dozen cards. In a 1997 study their palms began to sweat slightly, measured as skin conductance, before they reached for a risky deck, and they shifted strategy before they could explain why. Patients with ventromedial damage showed no warning sweat and kept choosing the bad decks.

Damasio called these signals somatic markers: bodily feelings that tag options as good or bad and prune the choice before reasoning begins. The idea has critics, and some argue that the healthy players knew more than they said, but the core finding has held: emotion is not the enemy of good judgement but part of its machinery. The orbitofrontal cortex keeps track of values, the dorsolateral prefrontal cortex holds options in mind and the anterior cingulate cortex flags conflict.

Did you know? In a classic 1981 experiment by Amos Tversky and Daniel Kahneman, most people preferred a plan to save 200 of 600 people, yet rejected the same plan when it was described as letting 400 die.

That framing effect is one of the systematic biases Kahneman and Tversky catalogued, starting from Jerusalem in the 1970s. Their prospect theory of 1979 showed that losses feel roughly twice as strong as equal gains, and that people judge outcomes against a reference point rather than in absolute terms. Kahneman received the 2002 Nobel Prize in economics, and his 2011 book Thinking, Fast and Slow popularised a fast, intuitive system and a slow, effortful one. The popular myth that good decisions are purely rational misses the point: the fast system is usually right, but it is easily nudged.

Everyday life tests the machinery constantly. Sleep loss makes people take bigger risks and weigh losses less, and stress pushes them towards habit. Default options, such as automatic enrolment in pension schemes, change behaviour hugely because most people stick with what is set. Good decisions come from listening to the gut, then checking it against the numbers.

Phineas Gage

On the afternoon of 13 September 1848, near Cavendish in Vermont, a 25-year-old railway foreman named Phineas Gage was preparing a blasting charge for the Rutland and Burlington Railroad. He was tamping gunpowder into a hole in the rock when the iron sparked and the charge went off. The tamping iron, 109 centimetres long, about 3 centimetres thick and weighing six kilograms, shot up through his left cheek, behind his left eye and out of the top of his skull, landing some 25 metres away.

Gage did not lose consciousness for long, if at all. He was carried to town in an ox cart, sat up on a hotel veranda and, according to Dr Edward Williams, greeted him with the words "Here is business enough for you." The local physician, John Martyn Harlow, cleaned the wound, fought off a serious infection over the following weeks and by the new year had him back home. Gage could walk, talk, remember and reason.

Yet something had changed. In 1868 Harlow wrote that the once efficient and well-liked foreman had become fitful, profane and impatient of restraint, unable to stick to plans, so that his friends said he was "no longer Gage". The railway would not take him back. He showed himself and his iron in New England and possibly at Barnum's museum in New York, worked with horses, then spent about seven years driving stagecoaches between Valparaíso and Santiago in Chile. He died in San Francisco on 21 May 1860 after a series of epileptic seizures. Harlow later obtained his skull and the iron, both now at Harvard's Warren Anatomical Museum.

Did you know? No picture of Gage was known until 2009, when two collectors, Jack and Beverly Wilgus, realised that their daguerreotype of a "whaler with a harpoon" actually showed a one-eyed Gage proudly holding his tamping iron.

Gage became the founding case for the idea that the frontal lobes shape personality and social judgement. In 1994 Hanna Damasio and colleagues used his skull to reconstruct the rod's path and argued that it damaged the ventromedial prefrontal cortex on both sides. A 2012 reanalysis led by John Darrell Van Horn at UCLA suggested that the damage was confined to the left frontal lobe but severed many of its long white-matter connections.

The legend has outgrown the man. The Australian psychologist Malcolm Macmillan, who spent decades on the sources, showed that many textbooks exaggerate Gage's decline, inventing drunkenness and brutality that no witness reported. Holding down a demanding coaching job in a foreign country suggests that he recovered a good deal, perhaps helped by the structure of the work. That makes his story more hopeful, not less: even after a terrible frontal injury, routine, purpose and time can help a brain rebuild a life.

Attention

At any moment your senses deliver far more than you can use: the hum of the fridge, the pressure of the chair, the words on this page. Attention is the process that picks a little and lets the rest go. William James wrote in 1890 that everyone knows what attention is, the taking possession by the mind of one out of several possible trains of thought. Explaining how it works has taken rather longer.

In 1953 the British engineer Colin Cherry named the cocktail party problem: how we follow one voice in a noisy room. He played different messages to each ear through headphones and asked people to repeat one aloud. They could say almost nothing about the other, not even when it switched to German or ran backwards. In 1958 Donald Broadbent in Cambridge proposed a filter that blocks unattended input early. Then Neville Moray found in 1959 that people often noticed their own name in the ignored ear, and Anne Treisman suggested that the filter merely turns ignored channels down rather than off.

Attention also moves through space like a spotlight. In 1980 Michael Posner showed that people respond faster to a target that appears where a cue had drawn their attention, even without moving their eyes. In the brain, a network linking the frontal eye fields and the parietal cortex steers the beam on purpose, while a second network, centred on the right temporoparietal junction, yanks it towards sudden surprises. In 1985 Robert Desimone and Jeff Moran recorded neurons in a monkey's visual area V4 that responded to a shape only when the animal was attending to it.

Did you know? In 1978 Edoardo Bisiach and Claudio Luzzatti asked stroke patients in Milan to imagine Piazza del Duomo from the cathedral steps; they described only the right side, then only the other side when imagining it from the far end.

Those patients had hemispatial neglect, usually after damage to the right parietal lobe, and may eat only the right half of a plate. Healthy people have a milder version. In 1999 Daniel Simons and Christopher Chabris at Harvard asked viewers to count basketball passes in a short film; about half failed to notice a person in a gorilla suit who strolled through the game and thumped their chest. This inattentional blindness is why drivers miss motorcyclists who were in plain view.

Two popular claims fail the test. Multitasking is mostly fast switching, and each switch costs time and accuracy: a 2006 Utah study by David Strayer found drivers on a phone, even hands-free, about as impaired as drunk drivers. The widely shared claim that human attention spans have shrunk to eight seconds, below a goldfish's, has no real study behind it. Attention is not a fixed span but a resource that is spent, so the simplest way to see more is to look at fewer things at once.

Consciousness

Every night you lose it and every morning it returns. Consciousness, the fact that there is something it is like to be you, is the most familiar thing in the world and the hardest to explain. In 1974 the philosopher Thomas Nagel asked what it is like to be a bat, and in 1994 and 1995 the Australian philosopher David Chalmers named the hard problem: even a complete account of what neurons do might not explain why any of it is felt from the inside.

Scientists split the question in two. The level of consciousness runs from coma to full alertness, and depends on arousal systems in the brainstem, first studied by Giuseppe Moruzzi and Horace Magoun in 1949, acting through the thalamus on the cortex. The contents of consciousness, what you are aware of at a given moment, are a separate puzzle. In the 1970s Lawrence Weiskrantz in Oxford studied a patient, D.B., who had lost part of his visual cortex. He insisted he saw nothing in part of his field of view, yet could point to objects there far better than chance. This blindsight proved that the brain can process what the mind never sees.

In 1990 Francis Crick and Christof Koch urged colleagues to hunt for the neural correlates of consciousness, the minimal brain activity that accompanies each experience. Two leading theories now compete. Global workspace theory, proposed by Bernard Baars and developed by Stanislas Dehaene in Paris, holds that a content becomes conscious when it is broadcast widely across the cortex in a sudden ignition. Integrated information theory, from Giulio Tononi in Wisconsin, holds that consciousness is the degree to which a system's parts work as an inseparable whole. A head-to-head test published in Nature in 2025 challenged key predictions of both.

Did you know? The cerebellum holds roughly four-fifths of the brain's neurons, yet people born without one, like a Chinese woman described in 2014, are fully conscious.

Measuring consciousness has saved lives. In 2006 Adrian Owen's team in Cambridge asked a young woman diagnosed as vegetative after a road accident to imagine playing tennis, then walking through her house, while she lay in a scanner. Her brain lit up in the same distinct patterns as healthy volunteers'. In Milan, Marcello Massimini's team zap the cortex with a magnetic pulse and measure how complex the echo is; a high score means someone is conscious, even if dreaming or paralysed, and a low one fits deep sleep, anaesthesia or coma.

These tools matter at the bedside, where some patients judged unaware are in fact present. The popular image of consciousness as a single place in the brain, like Descartes' pineal gland, is wrong; it seems to depend on how widely regions talk to one another. Each night in deep sleep that conversation fades, and each morning it returns, which may be the closest look at the mystery anyone gets.

The Sense of Self

Sit at a table with your left hand hidden behind a screen. In front of you lies a rubber left hand. Someone strokes both hands with paintbrushes at exactly the same moments, and within a minute or two, for most people, the rubber hand begins to feel like their own. This is the rubber hand illusion, published by Matthew Botvinick and Jonathan Cohen in Pittsburgh in 1998, and it shows that the feeling of owning a body is not given but constantly inferred.

When asked to point to their hidden hand with eyes closed, people aim towards the rubber one, a shift called proprioceptive drift. Threaten the fake hand with a hammer or a needle and their skin conductance rises in alarm. Henrik Ehrsson at the Karolinska Institute in Stockholm found in 2004 that the illusion tracks activity in the premotor cortex, where sight, touch and limb position are combined. The brain takes the most plausible story from the evidence: if a hand is seen being stroked exactly when one is felt, that hand must be mine.

The same logic can move the whole self. In 2002 Olaf Blanke and colleagues in Geneva stimulated the right angular gyrus of a woman with epilepsy, and she reported seeing herself from above, lying in bed. Ehrsson's lab later used cameras and head-mounted displays to make people feel they were standing behind their own bodies, and in 2008 to swap bodies with a mannequin. Damage can break ownership too. After some right-hemisphere strokes, patients insist that their own paralysed arm belongs to someone else, a condition called somatoparaphrenia.

Did you know? In 1880 the Paris neurologist Jules Cotard described a woman who believed she had no brain, nerves or insides and could not die, a delusion now named after him.

Above the body sits the narrator, the voice that explains what we do. Working with split-brain patients from the 1960s onwards, Michael Gazzaniga showed one patient's left hemisphere a chicken claw and his right hemisphere a snowy scene. The left hand, steered by the right hemisphere, picked a shovel. Asked why, the speaking left hemisphere, which knew nothing of the snow, instantly explained that a shovel is needed to clean out the chicken shed. Gazzaniga called this the interpreter, a system that spins plausible stories and believes them.

Mirrors offer another window. Gordon Gallup showed in 1970 that chimpanzees use a mirror to inspect a mark on their foreheads, and human toddlers pass similar tests at around 18 to 24 months. The popular picture of a little self sitting in the head watching a screen does not survive this evidence: the self is assembled from body signals, memory and story, moment by moment. That is unsettling, but useful, because prosthetic limbs that feel like one's own, and therapies for phantom pain, depend on the brain's willingness to adopt a new part.

Free Will

Flex your wrist whenever you feel like it. That is all Benjamin Libet asked of volunteers in his laboratory in San Francisco, in a study published in 1983. Electrodes on the scalp recorded the readiness potential, a slow build-up of activity over the motor areas first described by Hans Kornhuber and Lüder Deecke in Freiburg in 1965. The volunteers watched a dot sweeping round a clock face once every 2.56 seconds and reported where it had been when they first felt the urge to move.

The result has been debated ever since. For spontaneous movements the readiness potential began around 550 milliseconds before the wrist moved, but the reported urge came only about 200 milliseconds before. The brain seemed to start preparing roughly a third of a second before the person felt they had decided. Libet himself did not conclude that free will is an illusion. He argued that consciousness keeps a veto, a last fraction of a second in which a starting movement can be stopped.

Later studies sharpened the challenge. In 2008 John-Dylan Haynes and colleagues in Berlin used brain scans to predict whether people would press a button with the left or right hand up to about seven seconds before they knew, though only with around 60 per cent accuracy. In 2011 Itzhak Fried's team at UCLA recorded single neurons in epilepsy patients and predicted a decision to move about 700 milliseconds ahead. Many headlines declared free will dead.

Did you know? In a 2012 reanalysis, Aaron Schurger, Jacobo Sitt and Stanislas Dehaene showed that a readiness potential can arise simply from random neural noise drifting towards a threshold, with no decision behind it.

That reanalysis is the heart of the modern answer. If background fluctuations nudge the timing of a meaningless wrist flick, the early signal reflects when, not whether, and not why. Libet's task was also odd: a choice with nothing at stake, timed by a clock and reported by a sense of urge that is itself hard to pin down. In 2019 a team led by Uri Maoz found the readiness potential present when volunteers made arbitrary choices but absent when they made deliberate decisions about real charity donations.

Philosophers remain divided. Some argue, with Daniel Dennett, that free will worth wanting means acting on one's own reasons, which a brain can do perfectly well. Others, like Sam Harris, call it an illusion. The law already lives with the question: after Charles Whitman shot people from a tower in Austin, Texas, in 1966, his autopsy found a brain tumour, and experts still argue how much it mattered. The popular claim that neuroscience has proved we have no choice goes beyond the data. What the experiments show is that choices are made by brains, partly before we notice, and that the slower, reasoned choices that shape a life are the ones we have most say over.

Broca and Wernicke

In April 1861 a 51-year-old man named Louis Victor Leborgne was carried into the surgical ward of the Bicêtre hospital outside Paris. He had lost the power of speech some twenty years earlier and could utter only one syllable, "tan", usually said twice, which is why the staff called him Tan. He understood questions, answered with gestures and grew angry when he was not understood. He died a few days after the surgeon Paul Broca examined him. At the autopsy Broca found a damaged patch on the left side of the frontal lobe, at the back of the third frontal convolution.

A few months later a second patient, an 84-year-old labourer called Lazare Lelong, arrived with a vocabulary of five words, among them "oui", "non" and a mangled version of his own name. His lesion sat in almost the same spot. By 1865 Broca had collected enough cases to make a bold claim: speech is produced by the left hemisphere. The region, roughly Brodmann areas 44 and 45, is now called Broca's area, and damage there leaves people with halting, effortful speech stripped of small grammatical words, while comprehension stays fairly good.

In 1874 a 26-year-old German neurologist, Carl Wernicke, working in Breslau, described the opposite problem. His patients spoke fluently, with normal rhythm and grammar, but the words were wrong or invented, and they could not follow what was said to them. Their damage lay further back, in the upper part of the left temporal lobe near the hearing cortex. Wernicke went further and drew a wiring diagram: sound images of words stored at the back, movement plans at the front, and a bundle of fibres, the arcuate fasciculus, linking the two. He predicted that cutting this link would leave a person who understood and spoke but could not repeat a sentence. Such conduction aphasia was duly found.

Did you know? Leborgne's brain is still preserved in Paris, and in 2007 it was put through an MRI scanner, nearly 150 years after Broca first held it.

That scan, by Nina Dronkers and colleagues, revealed that Tan's damage ran much deeper than the surface patch Broca described, into white matter beneath. In the 1960s the American neurologist Norman Geschwind revived Wernicke's diagram, and it became the textbook model. Modern imaging has made it messier. Language turns out to depend on two broad streams, one mapping sounds to meaning along the temporal lobe and one mapping sounds to speech movements through parietal and frontal cortex, with Broca's area also helping with grammar and working memory. Deaf signers with left-hemisphere strokes, studied by Ursula Bellugi at the Salk Institute in the 1980s, lose sign language in the same patterns, which shows these areas serve language, not just the ears and mouth.

For the hundreds of thousands of people who develop aphasia after a stroke each year, these maps still guide diagnosis and speech therapy. Two Parisian patients, one able to say only "tan", started the science of where words live.

The Reading Brain

Writing is around 5,000 years old, first appearing in Mesopotamia and Egypt not long before 3000 BC, and mass literacy is younger still. That is far too recent for evolution to have built a reading organ. Yet every fluent reader recognises a printed word in a fraction of a second, whatever its size, font or case. The French neuroscientist Stanislas Dehaene calls the explanation neuronal recycling: reading borrows circuits that evolved for something else, and bends them to a new job.

The first clue came in 1892, when the French neurologist Joseph Jules Dejerine described a Parisian businessman, known in the literature as Monsieur C, who woke one day unable to read. He could still write, fluently, but could not read back what he had just written, and he could still recognise numbers and faces. At his death Dejerine found damage at the back of the left hemisphere, cutting the visual system off from the language areas. The condition is called pure alexia, or alexia without agraphia.

In 2000 Laurent Cohen and Dehaene, in Paris, used brain imaging to pin down the spot. A small patch in the left occipito-temporal cortex, on the underside of the brain, responds to written words more than to faces, houses or objects. They named it the visual word form area. It sits in roughly the same place in readers of English, Hebrew, Chinese and Japanese, and it is wedged between regions for faces and objects, the kind of high-resolution shape-spotting that our primate ancestors used to identify things.

Did you know? Learning to read means unlearning a skill: the visual system treats mirror images as the same object, which is why young children often write "b" for "d".

Literacy literally reshapes the cortex. In a 2010 study, Dehaene and colleagues scanned adults in Portugal and Brazil who had learned to read in childhood, as adults, or never. The more literate the person, the stronger the word area's response to print, and the more the response to faces shifted slightly towards the right hemisphere. Reading also strengthens early visual cortex and the links to the sound system, because words must be turned into speech sounds before meaning arrives. The eyes do not sweep smoothly along a line: they make jumps called saccades of about seven to nine letters, pausing for about a quarter of a second at each stop.

When this system struggles, the result is often dyslexia, which affects something like one child in ten to twenty, depending on how it is defined. It is not seeing letters backwards, a popular myth, but chiefly a difficulty linking letters to speech sounds. In 2001 Eraldo Paulesu and colleagues showed that dyslexic readers in Italy, France and Britain shared the same reduced activity in the left temporal lobe, even though Italian spelling is far more regular and their problems looked milder.

Children who are read to, taught phonics clearly and given time to practise build this circuit faster. A brain that learned to read is a brain that has been rewired by culture.

The Number Sense

Glance at a plate holding three biscuits and you know there are three without counting. Glance at one holding seventeen and you have only a rough idea. That difference reveals two systems. The first, named subitising by the psychologist E. L. Kaufman and colleagues in 1949, grasps up to about four items instantly and accurately. The second, the approximate number system, estimates larger quantities, and its precision depends on ratios rather than differences, just as Weber's law predicts for brightness or weight.

This number sense is old and shared. In 1992 Karen Wynn showed five-month-old babies one toy, then a second placed behind a screen. When the screen dropped to reveal only one, the infants stared longer, as if surprised by bad arithmetic. In 2000 Fei Xu and Elizabeth Spelke found that six-month-olds could tell 8 dots from 16, a ratio of two, but not 8 from 12. Monkeys, crows, fish and even bees discriminate quantities, and in 2002 Andreas Nieder and Earl Miller recorded neurons in monkey frontal and parietal cortex tuned to particular numbers, firing most for three dots and less for two or four.

Did you know? Clever Hans, a horse who amazed Berlin crowds around 1904 by tapping out sums, could not count at all: Oskar Pfungst showed he was reading tiny unconscious cues in his questioners' faces.

In humans the key region is the intraparietal sulcus, a groove in the parietal lobe that becomes active whenever we compare or estimate quantities, whether shown as dots, digits or spoken words. Language and culture then add exact counting on top. The Munduruku people of the Brazilian Amazon, studied by Pierre Pica and Stanislas Dehaene and published in 2004, have number words only up to about five. They could compare and add large approximate quantities as well as French adults, but struggled with exact sums such as subtracting 4 from 6 when the numbers were large. Exact arithmetic beyond small sets seems to need symbols.

Damage shows the parts. In 1924 the Austrian neurologist Josef Gerstmann described patients with lesions near the left angular gyrus who could not calculate, could not name their fingers, confused left and right and could not write. Fingers and numbers are linked from childhood, which is why so many counting systems are based on ten. Dyscalculia, a lasting specific difficulty with numbers, affects an estimated three to seven per cent of children, about as many as dyslexia, yet receives far less attention. Some affected children cannot tell at a glance which of two dot clouds is larger, suggesting a weakness in the core number sense itself.

A wordless feel for quantity is the foundation, and school maths is built on it. Number games, board games with counters and finger counting all strengthen the bridge between symbols and quantities. The brain was counting long before anyone wrote a digit.

Music and the Brain

There is no music centre in the brain. A song engages almost all of it at once: the auditory cortex for pitch and timbre, motor areas and the basal ganglia for rhythm, the hippocampus for memory, and deep reward circuits for the shiver down the spine. That is one reason music survives brain damage in surprising ways, and why it has become a tool for research and therapy alike.

Pitch is analysed in the auditory cortex, in and around Heschl's gyrus in each temporal lobe, with the right side especially sensitive to fine differences in tone. In professional musicians this area can be markedly larger, as Peter Schneider and colleagues in Heidelberg reported in 2002. Rhythm recruits the motor system even when we sit still: the supplementary motor area, the cerebellum and the basal ganglia predict when the next beat will fall. In 2009 Aniruddh Patel showed that Snowball, a sulphur-crested cockatoo, adjusted his head-bobbing when the tempo of a Backstreet Boys song was changed, evidence that beat-keeping is not uniquely human.

The pleasure is real chemistry. In 2001 Anne Blood and Robert Zatorre at the Montreal Neurological Institute scanned people listening to pieces that gave them chills, and found activity in the same reward regions that respond to food and sex. A decade later Valorie Salimpoor, Zatorre and colleagues used PET to show that dopamine is released twice: in the caudate during the build-up, as the brain anticipates a climax, and in the nucleus accumbens at the peak itself. Music plays with prediction, delaying and then delivering what we expect, or surprising us just enough.

Did you know? About one person in a hundred or so has congenital amusia, a lifelong inability to tell tunes apart or notice a wrong note, despite normal hearing.

Isabelle Peretz in Montreal has studied amusia for decades and found that it often comes with weak connections between the right auditory cortex and frontal areas. The reverse also happens. The Russian composer Vissarion Shebalin kept writing music, including a symphony, after strokes in 1953 left him with severe aphasia. And absolute pitch, the ability to name a note without a reference, is rare in Europe and North America, perhaps one in several thousand people, but far more common among musicians who started early or speak a tonal language such as Mandarin.

Music also reaches places other cues cannot. People with Alzheimer's disease often remember songs from their youth long after names and faces have gone, perhaps because musical memory relies partly on regions that the disease reaches late. Rhythmic beats help people with Parkinson's disease walk more smoothly, and melodic intonation therapy uses singing to coax words back after stroke. A favourite song is not a luxury for the brain but a full workout of hearing, movement, memory and reward.

The Bilingual Brain

By many estimates around half of the world's people use two or more languages in daily life, from Mumbai to Montreal to Lagos. For much of the 20th century, though, psychologists warned that bilingualism confused children and lowered their intelligence. In 1962 Elizabeth Peal and Wallace Lambert in Montreal turned that view around: comparing French-English bilingual ten-year-olds with monolinguals of the same social background, they found the bilinguals scored as well or better on many tests. The earlier studies had mostly compared poorer immigrant children with richer native speakers.

How does one head hold two languages? In 1997 Karl Kim, Joy Hirsch and colleagues in New York asked bilinguals to describe their day silently in each language inside a scanner. In people who had learned both languages in infancy, the two activated overlapping tissue in Broca's area. In those who learned the second later, the activations sat side by side, a few millimetres apart. In Wernicke's area both groups showed overlap. The finding is debated, and proficiency matters as much as age, but it suggests that the brain files a language differently depending on when and how well it is learned.

Both languages are always switched on. Bilinguals reading in one language are slowed by words that look alike in the other, so the brain must constantly select one and suppress the other. A control network does this work, including the anterior cingulate cortex, the dorsolateral prefrontal cortex and the left caudate nucleus. In 2006 Jenny Crinion and colleagues showed that the caudate tracked which language was in use across speakers of German, English and Japanese. Learning a second language also leaves physical marks: in 2004 Andrea Mechelli and colleagues found denser grey matter in the left inferior parietal cortex of bilinguals, greatest in those who started young.

Did you know? Newborns whose mothers spoke both English and Tagalog during pregnancy were equally interested in both languages, while babies of English-only mothers preferred English.

Ellen Bialystok in Toronto argued that this constant juggling sharpens executive control, and a 2007 study suggested bilinguals developed dementia symptoms about four years later than monolinguals. A 2013 study of over 600 patients in Hyderabad found a similar delay. Yet the famous bilingual advantage is now one of psychology's liveliest disputes. Large studies since 2013, including work by Kenneth Paap, often find no benefit in everyday control tasks, and reviews suggest small or absent effects once background is accounted for. What is clear is that bilingualism does no harm.

Stroke can reveal the layers. Some bilingual patients recover their languages unevenly, and the French neurologist Albert Pitres noted in 1895 that the language used most just before the injury often returns first. For families, the message is simple: raising a child with two languages is safe, and it gives them two worlds rather than a muddle of one.

Intelligence

In 1904 the English psychologist Charles Spearman noticed something odd about schoolchildren's marks. Pupils who did well in Latin also tended to do well in mathematics, music and even judging pitch. He proposed a single underlying factor, which he called g, for general intelligence. More than a century of testing has confirmed the pattern: scores on very different mental tasks correlate positively, so a person good at one tends to be somewhat good at the others.

The tests themselves began with a practical problem. In 1905 Alfred Binet and Théodore Simon in Paris designed tasks to spot children who needed extra help at school. In 1912 the German psychologist William Stern suggested dividing mental age by actual age, the intelligence quotient, and in 1916 Lewis Terman at Stanford multiplied it by 100. Modern tests, such as those David Wechsler introduced in 1939, compare a person with others of the same age, scaled so the average is 100. The history has dark chapters: early tests were misused to justify immigration limits and forced sterilisation in the United States.

What do tests predict? Quite a lot, statistically. IQ scores correlate with school results, job performance and even health and lifespan, though many other things matter too. Heritability, the share of differences between people linked to genes, rises from roughly 20 per cent in infancy to perhaps 60 per cent or more in adulthood, as people increasingly choose their own environments. But environment shows its power in the Flynn effect: average scores rose by around three points a decade across much of the 20th century, as the New Zealand researcher James Flynn documented in the 1980s, far too fast for genes.

Did you know? Albert Einstein's brain, removed by the pathologist Thomas Harvey in 1955, weighed about 1,230 grams, slightly below average for a man of his age.

The brain basis is a network, not a spot. In 2007 Rex Jung and Richard Haier reviewed dozens of imaging studies and proposed the parieto-frontal integration theory: intelligence draws on regions in the frontal and parietal lobes and on the white matter that connects them. Patients with damage in these regions, studied by Jan Gläscher and colleagues in Iowa, lose points on g. Overall brain size correlates only modestly with test scores, and more intelligent people sometimes use less energy on a task, as if their networks work more efficiently.

Popular ideas stretch the science. Howard Gardner's theory of multiple intelligences, from 1983, is beloved in schools but has little support from testing data. Brain-training games improve performance on the games themselves, not general intelligence. What reliably helps is education: studies of school-leaving age reforms suggest each extra year of schooling adds a few IQ points. Intelligence is real and measurable, but it is not fixed at birth, and it is not the whole of a mind.

Creativity

Ask people to list as many uses for a brick as they can in two minutes and you get a quick measure of divergent thinking: the ability to produce many, varied and unusual ideas. The American psychologist J. P. Guilford put creativity on the research agenda in a famous address to the American Psychological Association in 1950, complaining that psychology had almost ignored it. Most creative work, though, also needs convergent thinking, the narrowing down to the one idea that actually works.

The best-known account of how ideas arrive is older. In 1926 the English social psychologist Graham Wallas, in The Art of Thought, described four stages: preparation, hard conscious work on a problem; incubation, setting it aside; illumination, the sudden flash; and verification, checking that it holds up. He drew on the French mathematician Henri Poincaré, who described how the solution to a problem about mathematical functions came to him in 1880s France just as he stepped onto a bus, after days of fruitless effort.

Neuroscience has now caught the flash in the act. In 2004 Mark Jung-Beeman, John Kounios and colleagues gave volunteers word puzzles: find one word that links pine, crab and sauce. Sometimes people solved them step by step, sometimes with a sudden insight. Just before the insight solutions, EEG recorded a burst of high-frequency gamma activity over the right anterior temporal lobe, about a third of a second before the answer reached awareness. A moment earlier, alpha waves increased over the back of the right hemisphere, as if the brain were briefly closing its eyes to outside distractions to hear a faint inner signal. The answer, by the way, is apple.

Did you know? In a 2004 study at Lübeck, people who slept after practising a number task were more than twice as likely to discover its hidden shortcut as those who stayed awake.

That experiment, by Ullrich Wagner, Jan Born and colleagues, gives incubation a mechanism: sleep reorganises memories and lets loose connections form. Brain scans of creative thinking show cooperation between networks that usually work against each other. In 2018 Roger Beaty and colleagues found that people who produced more original ideas had stronger connections between the default mode network, which generates spontaneous thoughts, the executive control network, which evaluates them, and the salience network, which switches between the two.

This undermines a stubborn myth. Creativity is not a right-hemisphere gift that only some people possess; it draws on both hemispheres and on ordinary memory, knowledge and attention. Famous flashes such as Archimedes in his bath or August Kekulé dreaming of a snake biting its own tail are better as legends than as evidence, and historians debate Kekulé's account. The practical lesson from Wallas still stands. Work hard on the problem, then walk, shower or sleep on it, and keep a notebook handy. Insight favours the prepared mind that has stopped staring.

The Default Mode Network

For most of the 20th century, brain scientists treated the resting brain as idle. Experiments subtracted the scan taken during rest from the scan taken during a task, and studied the difference. But in 1997 Gordon Shulman and colleagues, pooling PET studies at Washington University in St Louis, noticed something strange: a consistent set of regions became less active whenever people started almost any focused task. In 2001 Marcus Raichle and colleagues named it the default mode of brain function.

The default mode network links a handful of hubs: the medial prefrontal cortex at the front, the posterior cingulate cortex and precuneus in the middle, the angular gyrus on each side and parts of the medial temporal lobe, including the hippocampus. It switches on when attention turns inward: daydreaming, recalling the past, imagining the future, thinking about yourself and working out what other people think. In 2003 Michael Greicius and colleagues at Stanford showed that these regions rise and fall together even when people lie quietly in a scanner, and in 2005 Michael Fox and colleagues found that they rise as the attention networks fall, like children on a seesaw.

The resting brain was never idle. It uses about 20 per cent of the body's energy while making up about 2 per cent of its weight, and a demanding task raises that consumption by only a few per cent. The German psychiatrist Hans Berger, who recorded the first human EEG in 1929, had suspected as much, noting that the brain hummed with activity even when its owner sat still.

Did you know? A 2010 smartphone study of 2,250 people found their minds wandering in about 47 per cent of sampled moments, and they were less happy when it did.

That study, by Matthew Killingsworth and Daniel Gilbert at Harvard, put a number on something everyone knows. Mind-wandering has costs: errors at work, rereading the same paragraph three times, and in depression a tendency to ruminate, which has been linked to overactivity and altered connections in this network. It also has benefits. The default mode network helps consolidate memories, plan ahead, rehearse social situations and incubate creative ideas, often while you shower or walk.

The network also matters for disease. Randy Buckner and colleagues showed in 2005 that the amyloid plaques of Alzheimer's disease build up preferentially in default mode regions, possibly because they are so metabolically busy throughout life. Studies of experienced meditators, and of people given psilocybin in research settings, report reduced default mode activity, which some link to a temporary loosening of the sense of self, though these findings are still being tested.

The practical lesson is balance. Focused work suppresses the default network; boredom, walks and unscheduled time let it run. A mind that never wanders loses its chance to plan and reflect, and a mind that wanders all the time loses the present. Daydreaming is not wasted time; it is the brain doing different work.

Self-Control

In the late 1960s and early 1970s, at the Bing Nursery School on the Stanford University campus, the psychologist Walter Mischel and his colleagues sat four- and five-year-olds at a table with a treat, often a marshmallow or a pretzel, and a bell. The child could ring the bell and eat one treat straight away, or wait until the researcher returned, up to 15 or 20 minutes later, and get two. Some ate at once; others covered their eyes, sang, kicked the table or turned the marshmallow into an imaginary picture, and held out.

The test became famous because of what came next. Follow-ups published in 1988 and 1990 found that children who had waited longer were rated by their parents as more competent teenagers and had higher scores on the American SAT college entrance tests. In 2011 B. J. Casey and colleagues scanned some of the original participants, then in their forties. Those who had struggled to wait as children still found it harder to hold back responses to tempting happy faces, with differences in the inferior frontal cortex and the ventral striatum. A much larger study from Dunedin, New Zealand, which followed about 1,000 children from birth, found that childhood self-control predicted adult health, wealth and criminal convictions at age 32, even after accounting for intelligence and family background.

The neuroscience fits a tug of war. The ventral striatum and its dopamine signals push for the reward on offer now. The prefrontal cortex, especially its lateral regions, holds goals in mind and can override the pull. The prefrontal cortex is among the last parts of the brain to mature, continuing into the mid-twenties, which helps explain why teenagers are more impulsive in front of their friends.

Did you know? In a 2013 Rochester study, children who had just been let down by an unreliable experimenter waited about three minutes for a second marshmallow; those who had met a reliable one waited about twelve.

That study, by Celeste Kidd and colleagues, changed how the test is read. Waiting is partly a rational bet on whether the promise will be kept. In 2018 Tyler Watts, Greg Duncan and Haonan Quan repeated the marshmallow test with about 900 children from more varied backgrounds. The link to later achievement was about half as strong as in Stanford, and it mostly disappeared once family background and early cognitive skills were taken into account. Another popular idea, that willpower is a muscle which runs out, called ego depletion, largely failed a 23-laboratory replication published in 2016.

What actually helps is strategy, not grit. Mischel's best waiters looked away, distracted themselves or thought of the marshmallow as a fluffy cloud. Adults do the same by keeping biscuits out of sight, setting automatic savings or putting the phone in another room. Trustworthy surroundings matter too. Self-control is less a fixed trait than a skill of arranging the world so temptation does not get a vote.

Brain Myths

Few organs attract as much nonsense as the brain. In 2012 Sanne Dekker, Paul Howard-Jones and colleagues surveyed teachers in Britain and the Netherlands and found that, on average, they believed about half of the common neuromyths presented to them, and teachers who knew more about the brain in general were not less likely to believe the myths. The errors are worth unpicking, because each hides a real and more interesting fact.

The most famous claim is that we use only ten per cent of our brains. Its origins are murky; it is often traced to a misreading of the American psychologist William James, who wrote around 1907 that people use only a small part of their mental resources, and it was repeated in a 1936 preface to Dale Carnegie's self-help bestseller. It survives in films such as Lucy in 2014. Brain scans show that over a day every region is active, and damage to almost any area causes some loss. An organ that is about 2 per cent of body weight but uses about 20 per cent of its energy would not be kept by evolution if it were mostly idle.

Second, the idea of left-brained logical people and right-brained creative ones. It grew from the genuine split-brain studies of Roger Sperry and Michael Gazzaniga in the 1960s, which showed that the hemispheres specialise. But in 2013 Jared Nielsen and colleagues at the University of Utah analysed resting scans from more than 1,000 people and found no sign that individuals rely more on one whole hemisphere. Language leans left and some aspects of attention lean right, but everyone uses both.

Did you know? In Dekker's survey, more than nine in ten British teachers agreed that pupils learn better when taught in their preferred learning style, an idea a 2008 scientific review found no good evidence for.

That review, by Harold Pashler and colleagues, found that matching lessons to visual, auditory or kinaesthetic styles does not improve learning. The Mozart effect is similar. A 1993 study in Nature by Frances Rauscher found a small, brief boost in a spatial task after college students heard a Mozart sonata. It was turned into claims that music makes babies smarter, and in 1998 the governor of Georgia proposed giving every newborn a classical CD. Later analyses found the effect small, short-lived and probably due to mood and arousal.

Other myths have been overturned more slowly. For decades textbooks said adults never grow new neurons; in 1998 Peter Eriksson, Fred Gage and colleagues found new cells in the adult human hippocampus, though how many persist into old age is still argued over. Brain-training apps promised sharper minds, but a 2016 review led by Daniel Simons found they mostly improve the games themselves. The truth is better than any myth: a whole brain, always busy, shaped by learning throughout life.

Why We Sleep

Every animal with a nervous system that has been carefully studied sleeps, from fruit flies to whales, and humans spend roughly a third of their lives doing it. For centuries sleep was assumed to be the brain simply switching off. The first crack in that idea came in 1937, when Alfred Loomis, a Wall Street financier with a private laboratory in Tuxedo Park, New York, recorded the electrical rhythms of sleepers through the night with the electroencephalogram and saw them change in distinct stages.

The bigger surprise came in Chicago in 1953. Eugene Aserinsky, a graduate student working with Nathaniel Kleitman, noticed that sleepers' eyes darted under their lids in bursts while their brain waves looked almost awake. This was REM sleep. Modern labs divide the night into light stages N1 and N2, deep slow-wave stage N3, and REM, run in cycles of about 90 minutes, four to six times a night. Deep sleep dominates the early hours; REM periods lengthen towards dawn, which is why alarm clocks so often interrupt a dream.

What drives the need? In 1982 the Swiss researcher Alexander Borbély proposed a two-process model that still holds up. A pressure to sleep builds the longer we are awake, tracked partly by a chemical called adenosine, while a separate body clock opens and closes the gate for sleep across the 24-hour day. When both line up, we drop off.

The question of what sleep is for has several answers, not one. In N3, large slow waves roll across the cortex, and in 2013 Maiken Nedergaard's team at the University of Rochester showed that in sleeping mice the spaces between brain cells widened by about 60 per cent, letting fluid flush out waste such as beta-amyloid. Sleep also replays and sorts the day's memories, tunes synapses, and restores hormone and immune balance. Deprivation shows how essential it is: in 1964 Randy Gardner, a San Diego schoolboy, stayed awake for about 264 hours, growing irritable, forgetful and briefly paranoid, and rats kept continuously awake in Allan Rechtschaffen's Chicago experiments died within a few weeks.

Did you know? Bottlenose dolphins can sleep with one half of the brain at a time, keeping one eye open and surfacing to breathe while the other hemisphere rests.

A popular myth says some people thrive on five hours. A rare few do, carrying gene variants such as one in DEC2 described in 2009, but most adults need seven to nine hours, and short sleepers tend to underestimate how impaired they are. Regular hours, a dark cool room and protecting the early deep sleep and late REM both matter. Sleep is not time stolen from life; it is the maintenance shift that keeps the waking brain running.

The Body Clock

In 1729 the French astronomer Jean-Jacques d'Ortous de Mairan shut a mimosa plant in a dark cupboard and watched its leaves keep opening by day and folding by night. The rhythm came from inside. Humans carry the same kind of circadian clock, from the Latin circa diem, about a day, and it times sleep, body temperature, hormones and alertness.

The master clock sits in the suprachiasmatic nucleus, a pair of tiny clusters holding around 20,000 neurons in all, in the hypothalamus, perched just above the point where the optic nerves cross. In 1972 two American teams, Robert Moore and Victor Eichler, and Friedrich Stephan and Irving Zucker, destroyed it in rats and abolished their daily rhythms. In 1990 Martin Ralph, Michael Menaker and colleagues clinched the case by transplanting the nucleus between hamsters: the hosts adopted the donor's rhythm, including a mutant clock that ran about four hours short.

The ticking itself is molecular. In fruit flies Jeffrey Hall, Michael Rosbash and Michael Young showed that clock genes such as period make proteins that build up through the day and then switch off their own production, a loop that takes about 24 hours; they shared the 2017 Nobel Prize for it. Left alone, the human clock runs slightly long, on average about 24 hours and 10 minutes, so it must be reset daily. Light does that, through special retinal cells containing the pigment melanopsin, described in 1998 and 2002, which measure brightness rather than form images. Their signal runs straight to the clock.

Did you know? In 1962 the French caver Michel Siffre spent two months alone in the Scarasson cave in the Alps without a watch, and when he emerged believed it was 20 August when it was actually 14 September.

From the clock a long chain of neurons runs down into the spinal cord and back up to the pineal gland, which releases melatonin in darkness. Melatonin is a signal that night has come, not a knockout drug; bright light, including indoor light in the evening, holds it back. When the clock and the world disagree, the result is jet lag. Flying east from London to Tokyo means advancing the clock, which most people find harder than delaying it on the way home, and it takes several days to catch up.

The clock also shapes daily life far from airports. Teenagers' clocks shift later at puberty, so early school starts collide with biology, and night shift work, which keeps the body clock at odds with daily life for years, was classed as a probable cause of cancer by the World Health Organization's cancer agency in 2007. Morning daylight, dimmer evenings and regular meal and sleep times are the simplest ways to keep the clock honest.

Dreams

People have read dreams as messages from gods, from the dead and, after Sigmund Freud's The Interpretation of Dreams, published in Vienna at the end of 1899, from the hidden wishes of the unconscious. The science of dreaming began more prosaically in Chicago in 1953, when Eugene Aserinsky and Nathaniel Kleitman found that sleepers woken during bursts of rapid eye movement usually reported vivid dreams. William Dement, then a student in the same lab, went on to map REM sleep across the night.

During REM a switch in the pons, in the brainstem, sends the cortex into an active, almost waking state while blocking the motor neurons of the spinal cord. The result is atonia: you can run in a dream without moving a leg. In Lyon in the 1960s Michel Jouvet damaged a small region of the pons in cats; in REM sleep the animals rose, stalked and pounced at nothing, apparently acting out their dreams. Brain scans of sleeping volunteers in the 1990s, by Pierre Maquet in Liège and Allen Braun in the United States, showed the emotional amygdala and visual areas busy in REM while parts of the prefrontal cortex were relatively quiet, which fits dreams' strong feelings, vivid scenes and cheerful disregard for logic.

Why we dream remains debated. In 1977 Allan Hobson and Robert McCarley at Harvard proposed that dreams are the forebrain's attempt to make a story from random brainstem signals. Mark Solms, studying patients in London and Johannesburg in the 1990s, found that damage to forebrain regions involved in motivation could stop dreaming altogether while REM continued, so dreaming and REM are not the same thing. Indeed, people woken from non-REM sleep also report dreams, usually shorter and plainer. Many researchers now think dreaming helps process emotion and rehearse memories, but no theory has won.

Did you know? People blind from birth dream without pictures, but their dreams are rich in sound, touch, smell and taste, as recordings of their dream reports have shown.

In a lucid dream the dreamer knows they are dreaming. In 1975 Keith Hearne at the University of Hull recorded a volunteer, Alan Worsley, sending agreed eye movements from inside a dream, and Stephen LaBerge at Stanford repeated the trick in the early 1980s. LaBerge's work also undermined the myth that dreams last only a second or two: counting to ten in a lucid dream took about as long as counting while awake.

Dreams are vivid because REM pairs a feeling-rich brain with a dimmed critic. Sleep loss, alcohol and some antidepressants trim REM, and nightmares often flare in stress and trauma. Every night, by the sleep laboratory's measure, you hallucinate for around two hours, safely paralysed.