Why '10' and '20'? The Math Behind EEG Electrode Placement
The Coordinate System Your Skull Didn't Know It Had
Here's a question that probably never occurred to you: how do you put something in the exact same spot on two completely different heads?
Not approximately the same spot. The exact same spot, relative to the brain tissue underneath the skull. On a baby's head. On a football player's head. On your grandmother's head. Every time, the electrode needs to sit over the same cortical region, or the data you collect is useless.
This is not a trivial problem. Human skulls vary enormously. Head circumference in adults ranges from about 52 centimeters to 62 centimeters. The distance from the front of your skull to the back can differ by several centimeters from one person to the next. Your head is almost certainly a different shape than mine.
And yet, the brain inside those differently-shaped skulls follows a remarkably consistent layout. The motor cortex is always in the same relative position. The visual cortex is always at the back. The auditory cortex is always near the ears. The brain's geography is proportional, not absolute.
This single insight is the key to everything that follows. And it's the reason the most important standardization system in neuroscience is built on percentages, not centimeters.
The Problem That Almost Sank a Whole Field
To appreciate why the 10-20 system matters, you need to understand the chaos that preceded it.
When Hans Berger published the first human EEG recording in 1929, he ignited a wildfire of interest. Within a decade, labs across Europe and North America were recording brainwaves. This was genuinely exciting science. For the first time in history, researchers could watch the living brain's electrical activity in real-time.
But there was a problem nobody had anticipated.
Every lab was placing electrodes in different spots. One researcher in London might put a sensor "about two inches above the left ear." A researcher in Boston might place it "three centimeters lateral to the midline." A third researcher in Paris might describe it as "over the temporal region." They were all recording something, but nobody could compare their results because nobody could be sure they were recording from the same brain area.
This is like trying to build a map of a continent where every cartographer invents their own coordinate system. London is at position (3, 7) according to one mapmaker and (blue, Thursday) according to another. You end up with a pile of data and no way to assemble it into knowledge.
For nearly three decades, EEG research was stuck in this trap. Thousands of recordings, no common language.
Herbert Jasper's Elegant Solution
In 1958, the International Federation of Electroencephalography and Clinical Neurophysiology asked Herbert Jasper, a Canadian neuroscientist working at the Montreal Neurological Institute, to solve the problem once and for all.
Jasper's insight was deceptively simple. He realized that you couldn't use fixed distances (centimeters or inches) because heads come in different sizes. But you could use proportional distances. Percentages.
Here's how it works.
First, you find four landmarks on the skull that every human has. The nasion, which is the little dip between your forehead and the bridge of your nose. The inion, which is the bony bump at the base of the back of your skull (you can feel it with your fingers right now). And the left and right preauricular points, which are the small depressions just in front of each ear canal.
These four landmarks are unambiguous. You can find them on any human head, regardless of age, sex, or ethnicity. They don't move. They don't vary with hairstyle or hat size.
Second, you measure two arcs. One runs from nasion to inion, straight over the top of the head (the anteroposterior line). The other runs from the left preauricular point to the right preauricular point, also over the top of the head (the coronal line). These two arcs cross at the vertex of the skull.
Third, and this is where the magic happens, you divide each arc into segments using two specific percentages: 10% and 20%.
Why 10 and 20, Specifically?
This is the question in the title, and the answer is more interesting than you might expect.
The outermost electrodes on each arc are placed at 10% of the total distance from the landmark. The remaining electrodes between them are spaced at 20% intervals.
So on the anteroposterior line (nasion to inion), the first electrode sits at 10% from the nasion. The next at 30% (10% + 20%). Then 50% (the vertex). Then 70%. Then 90%. And the final reference is at 10% from the inion. That's 10 + 20 + 20 + 20 + 20 + 10 = 100%. The whole arc is covered.
But why these specific percentages? Why not divide the whole thing into equal segments of, say, 16.67%?
The 10% edges aren't arbitrary. Jasper discovered that the outermost electrodes needed to be closer to the anatomical landmarks because the cortical regions near the base of the skull (the frontopolar and occipital poles) are compressed into a smaller physical space. A 20% spacing at the edges would overshoot the brain entirely and land on muscle or bone with no useful cortical tissue beneath. The 10% near-landmark positions keep the electrodes over actual brain, while the 20% spacing through the middle provides even coverage where the cortex is more uniformly distributed.
There's a second reason, too. The 10% edge positions act as a buffer. Electrode signals near the edges of the head tend to be noisier due to muscle artifacts from the jaw (temporalis) and neck (occipitalis) muscles. By pulling those edge electrodes inward slightly, from where 20% would place them, Jasper improved signal quality without sacrificing coverage.
The result is 21 positions (19 recording sites plus two reference electrodes at the earlobes) that cover the entire scalp in a grid that is anatomically proportional, functionally meaningful, and mathematically reproducible on any human head.
The Naming Convention: Letters, Numbers, and One Exception
Every 10-20 position gets a name consisting of one or two letters followed by a number or the letter "z." The letters indicate which brain region the electrode sits over:
| Letter | Brain Region | Location |
|---|---|---|
| Fp | Frontopolar | Very front of the forehead |
| F | Frontal | Front third of the skull |
| C | Central | Top center of the skull |
| P | Parietal | Behind center, before the back |
| T | Temporal | Sides of the skull, above the ears |
| O | Occipital | Back of the skull |
The numbers follow a simple rule. Odd numbers are on the left hemisphere. Even numbers are on the right. Lower numbers are closer to the midline. So F3 is left frontal, near the midline. F7 is also left frontal, but farther out toward the temple.
The letter "z" stands for zero and marks the midline. Fz is frontal midline. Cz is central midline (the vertex of the skull). Pz is parietal midline.
There's one position that breaks the pattern: "A1" and "A2" are the reference electrodes at the left and right earlobes ("A" for auricular). They aren't over brain tissue at all. They provide a neutral electrical reference against which the other electrodes' signals are measured.
Why Percentages Scale and Centimeters Don't
Here's a thought experiment that makes the brilliance of Jasper's system click.
Imagine you have two people: one with a nasion-to-inion distance of 30 centimeters and another with a distance of 36 centimeters. If you placed an electrode at a fixed 6 centimeters from the nasion on both heads, that electrode would sit over different cortical regions. On the smaller head, 6 centimeters is 20% of the way back, placing it over the premotor cortex. On the larger head, 6 centimeters is only about 16.7% of the way, placing it over the prefrontal cortex. Same measurement in centimeters, different brain area.
But if you place the electrode at 30% of the nasion-to-inion distance on both heads, you get 9 centimeters on the smaller head and 10.8 centimeters on the larger one. Different measurements in centimeters, same brain area.
This works because the brain itself scales proportionally with skull size. A person with a larger head doesn't have a motor cortex that's shifted forward or backward. They have a proportionally larger brain with the same functional areas in the same relative positions. The 10-20 system exploits this biological fact.
It's the same principle behind how maps work. A map of a state doesn't say "the capital is 4 inches from the left edge." It uses longitude and latitude, proportional coordinates that work regardless of whether the map is printed on a napkin or a billboard. The 10-20 system is, in essence, a latitude-longitude system for the human skull.
From 21 Positions to 300: The Extended Systems
Jasper's original 21 positions were sufficient for clinical EEG in the 1950s. But as research questions became more precise and technology improved, scientists wanted more spatial resolution.
In the 1990s, the 10-10 system (sometimes called the modified combinatorial nomenclature) was developed. It adds intermediate positions between every pair of original 10-20 sites by subdividing the 20% intervals into 10% intervals. This produces 81 positions, giving four times the spatial resolution.
The naming convention extends logically. Between F3 and F7 in the original system, the 10-10 system adds F5. Between Fz and F3, it adds F1. The letters stay the same, and the numbers fill in the gaps.
The 10-5 system, introduced in 2007, goes even further, adding positions at 5% intervals for a total of over 300 defined locations. This level of density is used in high-density EEG research with 128-channel or 256-channel caps, where researchers are trying to reconstruct the three-dimensional sources of brain activity with maximum precision.
Here's the interesting thing: each extension is backward-compatible with the original. Every 10-20 position is also a valid 10-10 position. Every 10-10 position is also a valid 10-5 position. Jasper's original framework was so well-designed that it has scaled gracefully across nearly seven decades of technological advancement.
What Each Position Actually "Sees"
An EEG electrode doesn't measure the activity of one specific neuron or even one specific brain region. It picks up the summed electrical activity of all the cortical tissue within roughly 2 to 3 centimeters of it. This means each 10-20 position corresponds to a neighborhood of brain function rather than a single address.
But those neighborhoods are remarkably consistent across people, which is what makes the system useful. Here's what the major positions tend to capture:
Frontal positions (Fp1, Fp2, F3, F4, F7, F8, Fz): Executive function, decision-making, working memory, planning. The prefrontal cortex is where your brain does its most sophisticated reasoning. Frontal electrodes pick up the electrical signature of deliberate thought.
Central positions (C3, C4, Cz): Motor planning and execution, somatosensory processing. C3 and C4 sit right over the primary motor cortex and are the positions most commonly used in brain-computer interfaces that decode movement intentions.
Parietal positions (P3, P4, Pz): Spatial awareness, attention, sensory integration. The parietal cortex is where your brain stitches together information from multiple senses into a coherent model of where things are in space.
Temporal positions (T3, T4, T5, T6): Auditory processing, language comprehension, memory retrieval. The temporal lobes house Wernicke's area (language understanding) and the hippocampus (memory formation, though this is deep enough that surface EEG can only pick up indirect effects).
Occipital positions (O1, O2, Oz): Visual processing. alpha brainwaves are strongest at occipital sites when your eyes are closed, because the visual cortex is idling. Open your eyes and that alpha drops immediately, a phenomenon called alpha blocking.
How Consumer EEG Devices Choose Their Positions
No consumer device uses all 21 original positions. That would require a helmet-like array with gel electrodes and a trained technician. Instead, consumer BCIs select a strategic subset of positions based on what they're designed to measure.
This is a design decision with real consequences. The positions you choose determine what cognitive states you can detect, what applications you can support, and how reliable your data will be.
The Neurosity Crown, for example, uses 8 channels at positions CP3, C3, F5, PO3, PO4, F6, C4, and CP4. These aren't random. They span the frontal cortex (F5, F6) for executive function and focus, the central cortex (C3, C4) for motor intention and mental imagery, the centroparietal cortex (CP3, CP4) for attention and cognitive processing, and the parieto-occipital cortex (PO3, PO4) for visual processing and spatial awareness.
This particular arrangement was chosen to maximize coverage across all four lobes of the brain using only 8 channels, giving the Crown whole-brain sensitivity in a device you can put on like a pair of headphones. It's enough to detect focus and calm states, measure cognitive load, track frequency-band power across the cortex, and serve as an input device for brain-computer interface applications.
Compare that to devices that cluster all their sensors on the forehead. They can measure frontal activity well, but they're blind to everything happening in the parietal, temporal, and occipital cortex. That's like building a weather station that only measures temperature in the kitchen and claims to know the climate of the whole house.
The 10-20 System Is a Language, Not Just a Map
Here's something that often gets lost in the technical description: the 10-20 system's most important contribution isn't the electrode positions themselves. It's the common language it created.
Before 1958, a researcher who discovered an interesting EEG pattern couldn't communicate exactly where it occurred. After 1958, they could say "increased theta power at Fz" and every other researcher in the world would know exactly what that meant, what brain region, what hemisphere, what functional significance.
This is why the 10-20 system appears in virtually every EEG study ever published since its adoption. It's not just a placement method. It's the vocabulary of brain electrical mapping. When a neurologist reports "focal slowing at T3-T4," that means something specific. When a BCI researcher reports "mu desynchronization at C3 during right-hand motor imagery," other researchers can replicate that finding because they know exactly where to look.
The system's longevity is a testament to how well Jasper identified the core problem. Brains are proportional. Skulls vary in absolute size. Use percentages. That's it. And 68 years later, that insight still underpins every EEG recording made anywhere on the planet.
What the 10-20 System Can't Do
No system is perfect, and the 10-20 system has real limitations worth understanding.
Spatial resolution is low. Each electrode captures activity from a cortical area several centimeters across. You can tell which general brain region is active, but you can't pinpoint the exact neural source. This is an inherent limitation of surface EEG, not the 10-20 system specifically. But the system's 20% spacing means you're sampling relatively coarsely. The 10-10 and 10-5 extensions help, but they require more hardware.
Deep brain structures are invisible. The hippocampus, amygdala, thalamus, and basal ganglia all sit deep inside the brain. Their electrical activity is too far from the scalp to be directly detected by surface electrodes at any 10-20 position. You can sometimes infer their activity from surface patterns, but you can't measure it directly.
Individual anatomy varies more than the system assumes. While the proportional approach handles differences in head size well, it doesn't perfectly account for individual variation in brain folding patterns (sulcal anatomy). Position C3 might sit squarely over the hand area of the motor cortex in one person and slightly off-center in another. For clinical and research applications where precision matters, this limitation is addressed through MRI-guided electrode placement or source localization algorithms.
Despite these limitations, the 10-20 system remains the universal standard because no better alternative exists that's as simple, reproducible, and widely understood. It solves the 80% problem brilliantly: getting electrodes over approximately the right brain region, on any head, without any imaging equipment, in about five minutes.
From 1958 to Your Desk: The System That Keeps Scaling
Herbert Jasper could not have imagined that his percentage-based coordinate system would one day be used in a wireless consumer device that sits on someone's head while they write code in their home office. He designed it for clinical neurophysiology labs with gel electrodes and analog amplifiers.
And yet, the Neurosity Crown's 8 channels at CP3, C3, F5, PO3, PO4, F6, C4, and CP4 are defined by the same proportional measurements Jasper published in 1958. The technology around the system has changed beyond recognition, but the coordinate system itself hasn't needed to change at all.
That's the mark of a truly elegant solution. It doesn't just solve the problem of its time. It solves the problem, period. The human skull is proportional. The brain's geography is consistent. Percentages capture both facts in a single framework.
Every time you see a study citing an EEG position, every time a neurologist reads a clinical EEG, every time a developer builds an application using the Neurosity SDK and accesses data from channel C3, they're using the system that Herbert Jasper formalized 68 years ago in Montreal. Two numbers. One insight. And the entire field of brain electrical mapping rests on top of it.
The next time someone tells you that math isn't practical, tell them about the 10 and the 20.