Your Brain Is a Magnet (Sort Of)

The Quietest Room on Earth, Listening to the Loudest Organ

Somewhere in the basement of a university hospital, there's a room unlike anything you've ever been in.

The walls are made of multiple layers of mu-metal, a nickel-iron alloy that absorbs magnetic fields the way a sponge absorbs water. The door weighs hundreds of pounds and seals with the precision of an airlock. Inside, the Earth's magnetic field, the one that makes your compass needle point north, has been reduced to nearly nothing. Stray electromagnetic signals from cell phones, power lines, passing cars, and the building's own wiring have been suppressed by a factor of a thousand or more.

This room exists for one purpose: to make it quiet enough, magnetically quiet, to hear your brain think.

Because here's the thing nobody tells you in biology class. Every thought you have, every memory you recall, every word you read on this screen, generates a magnetic field. Your neurons don't just produce electrical signals. Those electrical currents also create magnetic fields, the same way current flowing through a wire creates a magnetic field around it. (This is basic Maxwellian physics. If you run current through anything, you get a magnetic field. Neurons are no exception.)

The problem is that these brain-generated magnetic fields are unbelievably faint. We're talking about 50 to 500 femtotesla. A femtotesla is 10 to the negative 15 tesla. To put that in perspective, the Earth's magnetic field is roughly 50 microtesla, which is about a billion times stronger. A refrigerator magnet is about a hundred billion times stronger. The magnetic field from a single neuron firing is so weak that it makes a whisper in a hurricane sound deafening by comparison.

And yet, we can measure it. That's what magnetoencephalography does. MEG listens to the magnetic whispers of your brain, and what it hears is extraordinary.

Same Source, Different Physics: The EEG Connection

Before we get into how MEG works, you need to understand something fundamental. MEG and EEG are not competing technologies measuring different things. They're two different windows into the same room.

When a large group of pyramidal neurons in your cortex fires in synchrony, the post-synaptic currents flowing through those neurons create both an electrical field and a magnetic field simultaneously. It's one event producing two signals. EEG picks up the electrical side by measuring voltage differences between electrodes on your scalp. MEG picks up the magnetic side by detecting the magnetic fields those same currents produce.

Same neurons. Same moment. Same underlying neural event. Two different measurement physics.

So why bother with MEG at all, if EEG already captures the electrical side? The answer lies in what happens to each signal on its way out of your skull.

Electrical fields are volume-conducted. That means they spread out through brain tissue, cerebrospinal fluid, skull bone, and scalp skin, and each of those layers has a different electrical conductivity. The skull, in particular, is a terrible conductor. It smears and distorts the electrical signal, like trying to read a newspaper through frosted glass. By the time the signal reaches an EEG electrode on the scalp, its spatial detail has been significantly blurred. This is why EEG has relatively poor spatial resolution. The signals from different brain regions get mixed together on their way to the surface.

Magnetic fields, on the other hand, pass through biological tissue almost completely unaffected. The skull is essentially transparent to magnetic fields. They don't get smeared. They don't get distorted. They emerge from your head carrying much more precise information about where they originated.

This single physical fact, that magnetic fields aren't distorted by the skull, is the entire reason MEG exists as a technology. It's the reason someone decided it was worth building rooms lined with mu-metal and cooling sensors to near absolute zero. Because if you can measure the magnetic field instead of the electrical one, you can figure out where in the brain the signal came from with much greater precision.

How MEG Actually Works: SQUIDs and Superconductors

Now let's talk about the engineering, because this is where it gets genuinely wild.

To detect magnetic fields a billion times weaker than the Earth's, you need the most sensitive magnetic sensor ever invented. That sensor is called a SQUID. It stands for Superconducting Quantum Interference Device, and it is exactly as cool as that name implies.

A SQUID works by exploiting a quantum mechanical effect called the Josephson effect. Here's the basic idea. You take a loop of superconducting wire (a material that conducts electricity with zero resistance when cooled to extremely low temperatures). You introduce two thin insulating barriers into the loop, called Josephson junctions. In normal physics, current shouldn't be able to cross those barriers. But at superconducting temperatures, quantum tunneling allows Cooper pairs of electrons to pass through anyway, creating a measurable current.

Here's the key part. The current flowing through a SQUID is exquisitely sensitive to any external magnetic field passing through the loop. Even the tiniest magnetic field changes the quantum interference pattern of the tunneling electrons, which changes the measurable current. This makes SQUIDs the most sensitive magnetometers in existence, capable of detecting field changes as small as a few femtotesla.

But there's a catch. SQUIDs only work when superconducting, and that requires cooling them to around 4 Kelvin. That's negative 269 degrees Celsius. That's 4 degrees above absolute zero.

So a conventional MEG system contains a helmet-shaped structure called a dewar, filled with liquid helium, with roughly 300 SQUID sensors arranged around the inside surface. The patient sits with their head inside this dewar, and the SQUIDs, chilled to just above absolute zero, detect the magnetic fields seeping out of the brain.

The whole setup is, to put it gently, not portable. A MEG system weighs several tons. The magnetically shielded room alone costs hundreds of thousands of dollars. The liquid helium needs constant replenishment. You need dedicated technicians. You need a facility designed around the equipment.

This is a fundamentally different world from putting a device on your head and opening an app.

What MEG Sees That Other Methods Miss

So what do you actually get for all that infrastructure? What can MEG do that justifies a multimillion-dollar price tag and a room built to block the Earth's magnetic field?

The answer comes down to a rare combination: millisecond timing with good spatial precision.

Let me put this in context with a comparison table.

Notice something interesting in that last row. MEG and EEG don't even see exactly the same neural sources with equal sensitivity.

MEG is most sensitive to currents flowing tangentially to the scalp surface. These tend to be neurons in the sulci, the folds and grooves of the cortex. EEG picks up both tangential and radial sources (neurons on the crests of the cortical gyri pointing straight out toward the scalp). This means MEG and EEG, despite measuring signals from the same underlying neural events, actually have partially complementary sensitivities. They "see" different subsets of the cortical surface best.

This is why many research labs run MEG and EEG simultaneously. The combination gives you a more complete picture than either one alone.

What Researchers Actually Use MEG For

MEG isn't just a prettier version of EEG. Its unique combination of temporal and spatial resolution makes it the tool of choice for several specific problems that other imaging methods struggle with.

Pre-Surgical Brain Mapping

This is one of MEG's most important clinical applications. Before brain surgery, particularly for epilepsy or tumor removal, surgeons need to know exactly where critical functions live in that specific patient's brain. Where is the language cortex? Where is the motor cortex? How close are those regions to the tissue that needs to be removed?

MEG can map these functional areas non-invasively, with enough precision to guide the surgeon's hand. The technique is called magnetic source imaging, and it combines MEG data with structural MRI scans to produce a three-dimensional map of brain function overlaid on anatomy. Compared to the alternative, which involves placing electrodes directly on the exposed brain during surgery, MEG is obviously preferable.

Epilepsy Source Localization

For patients with epilepsy that doesn't respond to medication, the critical question is: where are the seizures starting? If you can find the seizure focus, you can potentially remove it surgically and cure the epilepsy.

MEG excels at this. The interictal spikes (abnormal electrical discharges between seizures) produce magnetic fields that MEG can detect and localize, often identifying seizure foci that scalp EEG cannot pinpoint. Several studies have shown that MEG-guided epilepsy surgery leads to better outcomes than surgery guided by EEG alone.

Cognitive Neuroscience Research

When researchers need to track the millisecond-by-millisecond flow of information through cortical networks, MEG is often the best tool available. It has been used to study language processing (tracking the cascade of neural activity from hearing a word to understanding it, which unfolds over roughly 400 milliseconds), visual perception, auditory processing, motor planning, memory formation, and social cognition.

The field of MEG research has produced some genuinely fascinating findings. Researchers have used MEG to show that your brain begins processing a word's meaning within 200 milliseconds of hearing it, that motor cortex begins preparing a movement about 500 milliseconds before you're consciously aware of deciding to move (echoes of Libet's famous experiment, but with much better spatial resolution), and that the resting brain cycles through distinct network states roughly every 200 milliseconds even when you're doing nothing at all.

The Crown captures brainwave data at 256Hz across 8 channels. All processing happens on-device. Build with JavaScript or Python SDKs.

Explore the Crown

The "I Had No Idea" Part: Why Your Brain's Magnetic Field Is So Absurdly Weak

Here's something that might rewire your intuition about the brain.

Your heart generates a magnetic field that's roughly a million times stronger than your brain's. A million times. Your cardiac magnetic field is strong enough to be detected several feet away from your body with relatively simple equipment. Magnetocardiography, the heart version of MEG, doesn't even require a magnetically shielded room in some implementations.

But the brain? The organ that controls literally everything you do, think, and feel? Its magnetic signature is barely a whisper in a thunderstorm.

Why the enormous difference?

It comes down to geometry. Your heart is a thick slab of muscle tissue where millions of cardiac cells depolarize in a coordinated wave, all roughly aligned in the same direction, producing magnetic fields that add up constructively. Think of it like a million people all pushing in the same direction. You get one big, strong push.

Your brain is the opposite. Cortical neurons point in all directions. Currents in neighboring sulci may flow in opposite directions, canceling each other's magnetic fields. Even within a single brain region, the neural architecture is far more geometrically complex than heart muscle. Only when a sufficiently large population of pyramidal neurons, typically tens of thousands at minimum, fire synchronously with their currents roughly aligned do you get a magnetic field strong enough for MEG to detect.

This means MEG only "sees" coordinated population activity. A single neuron's magnetic field is around 0.1 femtotesla, far below even SQUID sensitivity. What you're measuring is always the summed contribution of thousands of neurons acting in concert. The brain's magnetic field is weak not because the brain is doing less than the heart, but because its complexity works against magnetic coherence.

There's something poetic about that. The most complex object we know of in the universe is nearly magnetically silent, precisely because of its complexity.

Why MEG Is Still Rare (And Getting Less So)

There are roughly 200 MEG systems in the entire world. For comparison, there are over 40,000 MRI scanners and millions of EEG devices. MEG sits in an odd position: too expensive and complex for most hospitals, too specialized for most research labs, and completely inaccessible to consumers.

The barriers are substantial:

Cost. A SQUID-based MEG system costs $2M to $4M. The magnetically shielded room adds another $500K to $1M. Annual operating costs for liquid helium, which evaporates continuously and needs regular refilling, run $100K to $200K. You're looking at a total cost of ownership that rivals a small MRI installation, but for a tool with a narrower range of clinical applications.

Infrastructure. You can't just plug in a MEG machine. You need a custom-built shielded room, often in a basement or ground floor to minimize vibration. The room needs its own ventilation system. The building's electrical wiring may need modification to reduce interference. The MEG system itself weighs several tons and requires dedicated floor space.

Helium dependency. Liquid helium is a finite resource. Global helium shortages in recent years have driven up costs and created supply uncertainty. Some MEG labs have had to reduce operating hours or delay experiments because of helium availability. This ongoing consumable cost, and the logistical headache of managing cryogenic fluids, is one of the most cited barriers to MEG adoption.

Immobility. The patient must keep their head extremely still inside the dewar. Head movements of even a few millimeters can compromise data quality. This makes MEG studies with young children, patients with movement disorders, or anyone who has trouble sitting still for extended periods particularly challenging. The rigid sensor helmet doesn't accommodate different head sizes well, which further limits pediatric and population-diverse research.

The OPM Revolution

There's a technology on the horizon that could change several of these limitations. Optically pumped magnetometers, or OPMs, are a new type of magnetic sensor that works at room temperature. No liquid helium. No cryogenic cooling. No rigid helmet.

OPMs use a small glass cell filled with rubidium or cesium vapor. A laser beam polarizes the atomic spins in the vapor. When an external magnetic field hits the cell, it disrupts the spin alignment, which changes how the vapor interacts with the laser. By measuring this change, you can infer the strength of the magnetic field with sensitivity approaching that of SQUIDs.

Because OPMs don't need cooling, they can be mounted directly on the scalp, much closer to the brain than the SQUID sensors inside a helium dewar (which sit 2 to 3 centimeters away because of the thermal insulation). Closer sensors mean stronger signals. And because OPM sensors are small and lightweight, they can be arranged in a flexible helmet that fits different head sizes and even allows some movement.

Several research groups, including teams at the University of Nottingham, have demonstrated full-head OPM-MEG systems where participants can actually move during scanning. One famous demonstration had a subject bouncing a ping-pong ball while their brain activity was being recorded, something physically impossible with conventional MEG.

OPM technology is still in the research phase. These systems aren't yet commercially available at the scale of conventional MEG. But they represent a genuine shift in what's possible. Within the next decade, OPM-MEG could make the technique more accessible, more versatile, and significantly less expensive. It still won't be a consumer technology (you'll still need magnetic shielding and specialized expertise), but it could move MEG from "ultra-rare" to merely "uncommon."

The Relationship Between MEG and EEG (Why They Need Each Other)

If you've been reading carefully, you might be asking a reasonable question: if MEG and EEG measure signals from the same neural sources, and EEG is a thousand times cheaper and infinitely more portable, why does MEG exist at all?

The answer is that they see the same brain from different angles, and those different angles matter.

MEG's advantage is source localization. When a clinical team needs to pinpoint the exact cortical patch generating an epileptic spike, or when a researcher needs to track the flow of information from primary auditory cortex to Broca's area in real time, MEG's spatial precision is genuinely superior to EEG's. The skull transparency of magnetic fields isn't a minor technical detail. It's a qualitative difference in the information available.

EEG's advantage is everything else. Portability. Cost. Accessibility. The ability to record for hours, or days, or in someone's home, or while they're walking around, or during sleep. The ability to build consumer devices that let ordinary people see their own brain activity in real time. The ability to create developer ecosystems where people write code that responds to brainwave data.

These advantages aren't trivial. They're the difference between a technology that advances neuroscience in 200 labs and a technology that puts brain data in millions of hands.

And the relationship between MEG and EEG has actually made both better. Many of the source localization algorithms and signal processing techniques developed for MEG research have been adapted and applied to EEG data, improving EEG's spatial resolution. The beamforming and dipole fitting methods that MEG researchers pioneered are now routinely used with high-density EEG arrays. MEG research has taught us which neural sources generate which brainwave patterns, knowledge that directly improves how we interpret EEG recordings from portable, affordable devices.

In other words, MEG's contributions to neuroscience benefit everyone who uses EEG, including people who have never seen a magnetically shielded room.

Where This Is All Going

The trajectory of brain measurement technology over the past century follows a clear pattern. First, a technique is developed in a research lab. It's expensive, immobile, and requires deep expertise to operate. Over time, the core principles are miniaturized, the costs drop, and the technology becomes accessible to more people.

EEG followed this arc. Hans Berger's original 1929 recordings required a string galvanometer and a laboratory. By the 1970s, EEG machines were standard in hospitals. By the 2010s, consumer EEG headsets existed. Today, a device like the Neurosity Crown packs 8 EEG channels sampling at 256Hz into something you can wear while working at your desk, with on-device processing, open SDKs, and the ability to stream your brainwave data to AI tools through MCP.

MEG is earlier on that arc. The technology is roughly where MRI was in the 1980s: powerful, proven, but still confined to specialized facilities. OPM technology is the first real step toward portability, but we're still years, probably a decade or more, from anything resembling a consumer MEG device.

In the meantime, EEG carries the flag. It's the accessible cousin. It measures the electrical side of the exact same neural events that MEG measures magnetically. It won't give you MEG-level source localization. But it gives you something MEG never will: the ability to measure your own brain activity right now, wherever you are, for a price that doesn't require a grant application.

That's not a consolation prize. That's the whole point. The most powerful brain measurement technology isn't the one with the best specs on paper. It's the one you actually use. And right now, the most advanced brain imaging happens in rooms that only a few hundred people on Earth will ever enter. Meanwhile, EEG is quietly putting brain data in the hands of developers, researchers, meditators, and the genuinely curious.

The magnetically shielded room is a marvel of engineering. But the brain is a marvel of biology. And it goes everywhere you go.

What to Remember About MEG

MEG is one of the most elegant measurement technologies in all of science. It reads the magnetic whispers of neural activity with millisecond timing and millimeter-level spatial resolution. It has advanced our understanding of epilepsy, language, perception, and the fundamental dynamics of cortical networks.

It is also a technology defined by its constraints. The absurd weakness of the brain's magnetic field demands sensors cooled to near absolute zero. The sensitivity of those sensors demands rooms built to block the Earth's own magnetic field. The cost of all that infrastructure limits access to a few hundred sites worldwide.

Understanding MEG, really understanding it, does two things. It gives you a deep appreciation for the physics of neural activity, for the fact that every thought you have is both an electrical and a magnetic event. And it gives you a clearer picture of why EEG, for all its spatial limitations, is the technology that will actually bring brain data into daily life.

The brain doesn't care which method you use to listen. It's broadcasting on both channels, all the time, right now, as you read this sentence. The only question is whether you're listening.