Neuroscience, oriented

Think of a neuron as a cell that holds a voltage and spends it. The sodium-potassium pump uses energy to keep potassium concentrated inside and sodium outside. At rest, potassium crosses the membrane far more easily than sodium does, so the voltage settles close to potassium's equilibrium value, nudged a little less negative by a small sodium leak, at around minus 70 millivolts. Inputs arrive mostly on the dendrites as graded potentials, local changes of about 1 to 50 millivolts that scale with the input.

A spike usually begins just beyond the axon hillock, in a stretch called the axon initial segment. Threshold, typically near minus 55 millivolts, is where enough voltage-gated sodium channels open that incoming sodium keeps pushing the voltage up by itself. What you get next is about 100 millivolts and over in 2 to 4 milliseconds, whatever the push. Sodium channels inactivate during the rise, potassium flows out, and the voltage undershoots rest. While those channels are inactivated, no stimulus can fire a second spike; during the undershoot, only a stronger one can. That caps firing rates and keeps a spike from turning back along the stretch of axon it just left.

A spike travels by regenerating, not by flowing. Each patch of axon brings the next to threshold, which is slow when sodium channels are spread evenly. Myelin, wrapped on by glial cells that support neurons rather than carry impulses, insulates the axon, and regeneration happens only at the gaps called nodes of Ranvier, so the spike seems to hop from gap to gap. Wider axons conduct faster too: large myelinated touch fibers exceed 30 meters per second, and the unmyelinated fibers tuned to gentle stroking are about 50 times slower. How much myelin itself adds is still debated, since some reports suggest clustered sodium channels may be enough.

The trap is picturing a spike as a signal that grows with its cause. Say you tighten your grip on a heavy bag. No spike to your muscle gets larger. Each motor unit, one motor neuron and its muscle fibers, fires faster as its drive rises, and more units are recruited.

Rule: spikes come in one size, so look for strength in how often cells fire and how many take part.

At a chemical synapse, a spike opens voltage-gated calcium channels at the axon terminal, and the calcium triggers vesicles to empty transmitter into the cleft. The transmitter binds receptors on the next cell, then is cleared by transporters or enzymes. Blocking either route prolongs the signal; cocaine, for one, interferes with transporters. Not all release runs this way. Amphetamines drain vesicles into the axon, after which the transmitter leaves either passively or through a membrane transporter running in reverse. An electrical synapse skips the chemistry, joining two cells through gap junction pores so current passes straight across, usually in either direction.

What a transmitter does depends more on its receptor than on the molecule itself. An ionotropic receptor is itself a channel and responds fast. A metabotropic receptor sets off slower chemical steps inside the cell. Glutamate, the main excitatory transmitter, acts through both kinds, and dopamine can excite or inhibit. Say you compare two people given naloxone, an opioid antagonist, which occupies a receptor and blocks it without switching it on. In the person with opioids on board, the effects reverse. In the person with none, nothing happens.

A neuron takes input from thousands of others and sums it. An excitatory postsynaptic potential moves the membrane toward threshold, and an inhibitory one makes firing less likely, often by letting chloride in. Inputs add across space when several synapses are active at one moment, and across time when a burst arrives before earlier inputs fade. Gamma-aminobutyric acid is the brain's main inhibitory transmitter. Serotonin, dopamine and norepinephrine also act as neuromodulators, often diffusing to distant targets and adjusting how circuits respond rather than passing a specific message.

The trap is giving a transmitter one personality, such as pleasure or mood. Dopamine now looks more tied to repeating rewarding behavior than to producing pleasure. Its bursts are often interpreted as reward prediction error, meaning the difference between the reward you expected and the one you got, though one mouse study found movement signals that reading does not explain. Low serotonin is commonly tied to depression, yet one umbrella review found no consistent support for that account, and the question stays disputed.

Rule: before you say what a transmitter does, ask which receptor it reached and in which circuit.

Start with two divides. Your brain and spinal cord form the central nervous system, and the nerves and ganglia outside them form the peripheral one. Gray matter holds cell bodies and unmyelinated fibers, while white matter is bundles of myelinated axons.

The cerebral cortex is a gray layer about as thick as two or three stacked dimes, folded to fit more surface inside your skull. Frontal lobes do much of the work of planning and reasoning, and their back edge holds the motor cortex. Parietal lobes begin with the somatosensory cortex and also support reading and arithmetic. Temporal lobes receive information from the ears, and their undersides matter for forming and retrieving memories. Occipital lobes process images from the eyes. Sensory and motor cortex carry body maps scaled by receptor density or fine control, not by body size, which is why your fingers and the area around your mouth take up so much room.

Below the cortex, feedback and arousal change how the visual thalamus transmits input from the eye. The hypothalamus maintains homeostasis and links nerves to hormones. The basal ganglia help initiate movement and form habits, the cerebellum coordinates movement and takes part in learned movements, and the amygdala is central to detecting threats and learning conditioned fear. The hippocampus is described as an indexer that sends new memories to cortex for long-term storage, though how that transfer works is debated. The brainstem controls heart rate and breathing, which is why an opioid overdose can depress breathing and cause death.

The trap is reading a location as a complete explanation. Say you meet two patients with language trouble. Frontal damage can leave one with short, effortful phrases, and temporal damage can leave the other with fluent sentences carrying little meaning, yet neither result puts all of language in that spot. The popular version goes further. Some functions do lean to one hemisphere, but in one study of 1,011 people's resting scans the lean belonged to particular networks, and nothing marked a person as left-brained or right-brained overall.

Rule: learn each region as a contributor to a named job, and distrust any claim of one region, one function.

Every sense begins with transduction, turning a physical stimulus into an electrical signal. In your ear, channels open at the tips of a hair cell's stereocilia when they bend. In your eye, light closes sodium channels, so photoreceptors respond by growing more negative and cutting their transmitter release, not by firing. Some pathways keep maps: hair cells at the cochlea's base respond to high pitches and those at its apex to low ones, and that layout holds all the way to auditory cortex.

Two organizing ideas stay debated. Beyond primary visual cortex, a ventral stream is described as recognizing objects and a dorsal stream as guiding action, but one brain-stimulation study points to a dorsal role in recognizing object shape too. The labeled-line idea ties a sensation's quality to the pathway carrying it. Taste researchers argued over it for a long time, and one view accepts both labeled lines and patterns read across many fibers.

Output runs the other way. Commands leave the brain on upper motor neurons and reach muscle through lower motor neurons in the brainstem and spinal cord. A reflex arc need not wait for the brain, so a withdrawal reflex moves a body part before you are aware of the pain. The somatic system runs skeletal muscle under your conscious control, while cardiac muscle, smooth muscle and glands belong to the autonomic system. Many organs receive both its sympathetic and parasympathetic branches, and those are not simple opposites: one study of young women drinking cold water interpreted their heart-rate changes as the two activating together. The gut's own network can also run local reflexes without input from the brain or spinal cord.

The trap is treating pain as a direct readout of damage. Nociception, the detection of tissue-damaging stimuli, is only part of it. Neuropathic pain comes from nerve damage, descending pathways can both inhibit and facilitate signals on their way up, and the brain systems for pain overlap those for emotion, attention and memory. Location can mislead as well. Say a person with an inflamed pancreas feels the pain in the back; a likely though unsettled explanation is that fibers from one spinal segment supply several organs.

Rule: ask what was transduced, which pathway carried it and where it was adjusted before you treat a sensation as a fact about the world.

Synaptic plasticity means a connection can grow stronger or weaker with use. Hebb's idea is that connections strengthen between cells active together. Taken alone that rule runs away. One review argues that homeostatic plasticity, at the speeds experiments measure, is too slow to rein it in within current models, so faster compensatory processes have to work beside it. Long-term potentiation, usually induced in the lab with brief high-frequency bursts, is lasting strengthening, and long-term depression is its counterpart. In timing experiments, input just before the cell fires tends to strengthen a synapse, and firing just before the input tends to weaken it.

Change is physical too. Dendritic spines form and disappear, and young brains carry many more than adult brains. A critical period is a window of unusually strong reshaping. Deprive one eye of vision early in life and the acuity loss lasts, while the same deprivation in an adult does not have that effect.

Memory splits by system. A patient known as H.M. had medial temporal lobe structures removed on both sides to treat severe seizures. He could hold information for about 30 seconds but could not turn it into long-term memory, while older knowledge stayed largely intact. Skills and habits depend largely on the striatum: after extensive training, animals keep pressing a lever even once the food reward is devalued. A new memory first stabilizes through a cascade inside neurons that ends within hours. The standard model then has it come to rely on cortex rather than the hippocampus, but one critical review argues that detailed memories keep needing the hippocampus however old they are. Retrieval can reopen a memory to disruption, though not always.

The trap is treating plasticity as limitless rewiring. In adult mice most spines persist for long periods, recovery from amblyopia in adulthood is limited, and how much new-neuron production continues in adult humans is disputed. After a stroke, compensating with the less-affected side can lay down movement patterns that hold recovery back. Say you read that amputation rewires the brain and so causes phantom pain. One imaging study of people missing an upper limb saw only small map shifts and no tie to pain, so that story is contested.

Rule: ask what changed, on what timescale and in which species before you accept a claim that the brain rewires.

The central nervous system begins in the embryo as a flat plate that folds and fuses closed, and failed closure can cause spina bifida or anencephaly. Newborn neurons then migrate, and scientists think only about a third arrive. A growth cone at each axon's tip steers by reading cues, and one molecule can draw in some axon types while repelling others. Maturation runs long. One MRI study measured a myelin-linked marker in 484 people aged 8 to 85, and most regions leveled off only in the 30s. The prefrontal cortex keeps maturing into the mid-to-late 20s. Dual Systems models blame teen risk-taking on a quick socioemotional system that outruns slower control, but one review argues that this imbalance has not been rigorously tested. Later in life, word finding often slows, though many studies report that older adults know more words than younger ones.

Energy sets hard limits. Your brain makes up about 2 percent of body mass but uses 20 to 25 percent of your body's energy, much of it to restore ion gradients and run synaptic transmission. Glucose is its main adult fuel, and the small store in astrocytes is used up within a few minutes if supply stops. So blood flow matters: ischemic stroke, caused by a blockage in the head or neck, makes up about 80 percent of strokes. The glymphatic hypothesis proposes that cerebrospinal fluid carries solutes out through tissue, and mouse tracer work reported far less entry awake than asleep. One detailed review challenges much of that, arguing the flow is too slow to outrun diffusion.

The trap is treating a sound mechanism as a treatment. Too much glutamate can kill the neurons it excites, and brain ischemia raises it within minutes. Say you read that blocking this damage should therefore protect the brain after a stroke. Most of the blocking drugs tried failed to show benefit in stroke trials. Most Parkinson's symptoms arise as nigrostriatal cells die, and when symptoms first show, most people are already down 60 to 80 percent or more of those cells. Levodopa can help movement symptoms, but it neither replaces those cells nor halts the disease. Telling typical aging from disease in one person takes a clinician.

Rule: treat a mechanism as an account of part of a disease, and ask what happened when someone tried to change it.

Noninvasive methods trade timing against location, so you pick by whether your question is when or where. Electroencephalography follows electrical activity on a millisecond scale, but skull and scalp smear the signal, so locating a source takes a model of the head. Functional magnetic resonance imaging (fMRI) tracks blood oxygen, a proxy for activity with roughly millimeter resolution in one review's table, and its signal peaks only about 4 to 6 seconds after activity. Implanted electrodes do well on both axes but need surgery. Positron emission tomography (PET) follows a tracer, and the tracer sets the question: a glucose analog tracks energy use, while others estimate receptor density. Diffusion imaging builds pathways from the way white matter constrains molecular motion, so its tracts are reconstructions, not traced axons.

Causal tools change activity instead of watching it. Transcranial magnetic stimulation induces weak currents that can act as a temporary virtual lesion in a healthy brain. Optogenetics engineers chosen neurons to respond to light. A double dissociation, where one injury harms task A but spares task B and another injury does the opposite, points to separable functions. Lesion reasoning stays an inference, since injury can also impair distant areas through lost connections.

Statistics add errors of their own. A whole-brain analysis can test over 200,000 voxels, the small units a scan is divided into, so leaving out correction can send false positives extremely high. With only 20 to 30 people, a study can have very low power for weak effects, meaning a poor chance of detecting an effect that exists. Results that do clear significance tend to exaggerate the effect's size, though how severe this is for fMRI is still debated.

The trap is reverse inference. Say a news story reports that a region lit up when people saw their phones, so they must love them. That reasons back from activity to a mental state, when an active region may reflect a wide variety of processes. It also treats correlation as need, since imaging records activity without manipulating it. Extra brain detail should earn the story no credit: in one preregistered study of 430 people, superfluous neuroscience raised ratings mostly among those with less expertise.

Rule: ask four things of any brain claim: what the method measures, how big the sample is, how large the effect is, and whether the inference runs backward.

The neural correlate of consciousness is usually defined as a minimal neural system sufficient for an experience. Minimal matters: the causes of an experience, its effects, and background needs like oxygen all track it without accounting for it. Global workspace theory says a state is conscious when its content becomes available to memory, motor, evaluative and attentional systems, and one preregistered adversarial test of 256 people backed some of its predictions and not others. Integrated information theory identifies consciousness with how much a whole carries beyond its parts, and in a 2025 adversarial test the sustained synchronization it predicted in posterior cortex did not appear. Chalmers's hard problem asks why physical processing comes with experience at all, and some philosophers deny there is such a problem.

Clinical cases are striking evidence, and how to read them is disputed. With the corpus callosum cut to treat epilepsy, patients in the classic finding denied seeing a stimulus shown to the left visual field, yet their left hand could pick it out. Several studies, though, find these patients can report a stimulus's location anywhere in the visual field. People with primary visual cortex damage may report no sight in part of the field yet score far above chance when forced to guess, and whether that vision lies outside awareness is contested.

Interfaces read or write neural signals, within limits. One trial that gave implanted arrays to three people with paralysis reached nearly 8 words per minute for its top participant, a research result and not a product you can buy. As of December 2019 about 736,900 cochlear implants were registered worldwide, and hearing through one still differs from normal hearing.

The trap is letting one striking experiment settle the big question. Say you read that Libet's wrist-flexing task proved your brain decides before you do. Reported intention came about 200 milliseconds before movement but lagged the readiness potential, a negative electrical signal recorded before a voluntary movement, by about 300 to 500 milliseconds. That reading has been severely criticized, and what the timing reports reflect is still debatable. Some claims do fail outright: one systematic review answers the claim that you use only 10 percent of your brain by noting that no single area is completely idle, even in sleep.

Rule: separate what a case shows from the grand conclusion attached to it, and keep disputed readings disputed.