Electrodes, montages and the 10-20 system
10-20/10-10 names and conventions, electrode versus channel versus reference versus ground, impedance, and referential versus bipolar montages.
Prerequisites: L0.1 · What EEG measures
3 claims on this page are unverified. TODO(confirm) marks a specific statement the author has not yet checked against a
primary source. Everything else on this page has been reviewed. Treat a marked claim as
provisional and go to the cited source rather than quoting the sentence.
Objectives
- Name 10-20/10-10 positions and conventions (odd = left, even = right, z = midline)
- Distinguish electrode, channel, reference and ground
- Explain impedance and why active electrodes relax the requirement
- Describe referential and bipolar montages
Why this matters
A channel called “C3” is only useful if it is over the left central cortex, is being compared with a reference you know, and was recorded through a contact good enough to pass the signal. Files arrive with channels in acquisition order, with labels in a vendor’s spelling, with the reference sometimes present as a flat trace and sometimes missing altogether, and with no record of which contacts were bad. This lesson gives you the vocabulary to sort that out before you look at a single trace, and it introduces the first “look at the data and decide” question of the curriculum: which channel is the reference?
Concepts
The 10-20 system and its extensions
The international 10-20 system places electrodes at fixed fractions (10% and 20%) of the distances between four skull landmarks: the nasion (bridge of the nose), the inion (bump at the back of the skull) and the two preauricular points. Its names carry three pieces of information:
- Letters name the region: Fp (frontal pole), F (frontal), C (central), P (parietal), O (occipital), T (temporal). The extended systems add intermediate rows: AF, FC, CP, PO, and the lateral FT and TP.
- Numbers give the side and the distance from the midline: odd numbers are on the left, even numbers on the right, and the numbers grow from the midline outward (C1 is next to the midline, C5 is far lateral).
- z (for zero) marks the midline: Fpz, Fz, Cz, Pz, Oz.
The names T7, T8, P7 and P8 are the modern names for the positions older papers call T3, T4, T5 and T6; you will meet both spellings in the literature and in files (TODO(confirm) whether the author wants the older names shown in the widget).
The 10-10 system halves the spacing and fills in the intermediate rows; the 10-5 system halves it again and defines several hundred positions, of which any given cap uses a subset (TODO(confirm) the exact position counts to quote for 10-20, 10-10 and 10-5). A “64-channel cap” is a particular selection of 10-10 positions, and two 64-channel caps from different vendors need not select the same ones. The widget’s montage view shows which positions a given montage actually contains.
Electrode, channel, reference, ground
Four words that are routinely confused, and whose confusion produces real errors:
- An electrode is a physical contact on the scalp.
- A channel is a recorded voltage difference between two electrodes. Every channel has two ends; the trace you see is one electrode minus another.
- The reference is the electrode (or combination of electrodes) subtracted from every channel in a referential recording. It is present in every trace, so anything wrong with the reference appears everywhere. In many systems the reference electrode’s own channel is recorded too, and since it is subtracted from itself it is flat, zero by construction. In
ds-iowapdthe online reference is Pz and the Pz channel is flat; inds-lemonthe online reference FCz is not stored as a channel at all and must be reconstructed for an average reference. Both are documented in the catalog and both are the same lesson: the reference is a choice made at recording time, and you must know it before comparing amplitudes across studies (L2.3). - The ground (also called the common electrode) is not the reference. It ties the subject’s body to the amplifier’s zero so that the amplifier can reject signals common to both inputs (L0.3). It appears in no channel. A recording with a bad ground has line noise everywhere; a recording with a bad reference has the reference’s artifacts everywhere.
Impedance, gel and dry contacts
The impedance of an electrode is the opposition of the electrode–skin interface to current flow. High impedance matters for two reasons: an unequal impedance at the two inputs of a differential amplifier converts common-mode interference (mains pickup) into a differential signal that looks like data; and high impedance makes the electrode more sensitive to movement and to slow skin-potential drift. With passive electrodes and older amplifiers the conventional target was a few kilohms per electrode (TODO(confirm) the figure to quote: 5 kΩ and 10 kΩ are both common targets in the literature). Modern amplifiers with very high input impedance tolerate more, and active electrodes, which carry a preamplifier at the contact, tolerate considerably more because the signal is buffered before it travels down the cable. That relaxes the requirement; it does not remove it, and it does not cure a contact that has no gel under it.
Gel electrodes (wet) bridge the skin with a conductive paste and give low, stable impedances at the cost of preparation time and the risk of bridging (below). Dry electrodes trade preparation time for higher and less stable impedance and greater motion sensitivity. ds-brain-invaders was recorded with 16 dry electrodes and no online filter, and the consumer-headset datasets in the directory are dry or semi-dry; the difference between device classes is a fact about impedance and channel count, not a judgment about products.
Montages: referential and bipolar
A montage is the set of derivations you display or analyse; a channel list is not a montage until you say what is subtracted from what.
- Referential montage: each electrode minus one reference (the recording reference, a mastoid or linked mastoids, Cz, or the average of all electrodes). It preserves the absolute distribution of a widespread potential and puts the reference’s problems into every trace.
- Bipolar montage: each trace is the difference between two neighbouring electrodes along a chain, front to back (“longitudinal”, the clinical double banana) or side to side (“transverse”). It cancels what neighbours share, so widespread activity and reference problems disappear and local gradients stand out; a focal feature shows as a phase reversal, appearing with opposite signs in the two derivations that share the electrode nearest to it.
- Average reference: each electrode minus the mean of all electrodes. Common in research; it assumes reasonably even coverage of the head, and it silently includes bad channels unless you remove them first (L2.2, L2.3).
The same recording looks different under each montage, and none of them is “the data”: the data are the electrode potentials, and the montage is a linear transformation you choose.
Channel names in real files
Files do not arrive in 10-20 order and do not always use 10-20 spelling. ds-eegbci stores its 64 channels in acquisition order beginning with the frontal-central row, and its labels follow a Sharbrough-style convention (Fc5, Cp3, Af7, Po8, Fcz) that must be mapped onto the canonical names (FC5, CP3, AF7, PO8, FCz) before a standard montage will attach positions to them; MNE’s eegbci.standardize() does exactly that. ds-hbn names its channels E1–E128 plus Cz, which no 10-20 template recognises without a lookup table. Mapping names correctly is a prerequisite for every topography, every interpolation and every source model you will ever compute, and the notebook makes you do it by hand once.
Bridging: when two electrodes become one
If gel spreads between two neighbouring electrodes, they are electrically bridged: the two channels record nearly the same signal, including the same noise, and their difference is close to zero. Each looks perfectly healthy in isolation, so the recording proceeds and the analysis “works”; the montage has silently lost a channel and gained a duplicate. The check is to look at differences or correlations between neighbours rather than at channels one at a time. The pitfall entry below gives the detector; L2.2 puts it into the pipeline.
Before trusting a channel list, answer four questions from the data rather than from the header: which channel is flat (the reference, or a dead contact), which pairs are near-identical (bridged), which labels do not match their position (blinks on a channel that is supposed to be occipital), and whether the ordering is anatomical or acquisition order. The notebook’s Iowa PD file is your first case.
The data behind this lesson
ds-eegbci: 64-channel 10-10 cap, 160 Hz, Sharbrough-style labels; the notebook renames them and attaches thestandard_1005montage.ds-iowapd: 64-channel cap at 500 Hz, online reference Pz, whose channel is flat; the catalog records 60 analysable channels after excluding Pz, Iz, I1 and I2.ds-lemon: 62-channel cap (61 EEG plus VEOG), online reference FCz absent as a channel.
Explore
Start at 10-20 density and read the names aloud: letters for the row, odd left, even right, z on the midline. Switch to 10-10 and find the rows that appeared (AF, FC, CP, PO) and the lateral numbers that grew (C5, C6). Then choose the eegbci-64 montage and the lemon-62 montage and note which positions each omits. Positions are MNE’s standard_1005 coordinates projected onto a flat head with the nose up.
Practice
Electrodes, montages and the 10-20 system: standard montages, channel names and types, and where the reference is nb-0-2-montages
The notebook loads a ds-eegbci run, renames the Sharbrough-style labels, sets a standard montage and plots the sensors; sets channel types (EEG, EOG, misc) so that later steps treat them correctly; and loads one ds-iowapd recording whose online reference (Pz) is absent from the stored channels — add it back with add_reference_channels and it is the flat channel, because a channel that is its own reference records zero by construction.
Exercises
Exercise ex-0-2-scrambled-channels
Put in orderA ds-eegbci file lists its channels in acquisition order using Sharbrough-style labels. Below are seven channels of its frontal-central row, scrambled. Put them in anatomical order from the subject's left ear to the right ear (the widget's 10-10 view will help; remember odd = left, even = right, z = midline, and numbers grow outward).
- Fc3
- Ft8
- Fc1
- Ft7
- Fc2
- Fcz
- Fc4
Exercise ex-0-2-find-the-reference
Multiple choiceIn the ds-iowapd recording the notebook loads, which channel is the online recording reference, and what in the data tells you so?
Pitfalls
Electrode bridging
- Symptom
- Two neighbouring channels are near-identical; interpolation "works".
- Cause
Conductive gel spreads between two neighbouring electrodes and connects them electrically, so both record the same potential. It is most common with dense caps (short distances), with too much gel, with sweating, and with long recordings in which gel migrates. Because bridged channels are not noisy, the usual bad-channel detectors, which look for variance, amplitude or spectral outliers, do not s…
- Detect
- Compute the electrical distance between every pair of channels (the variance of their difference signal over the recording); bridged pairs stand out as near-zero values in an otherwise broad distribution. MNE provides mne.preprocessing.computebridgedelectrodes for this check (TODO(confirm) the citation for the electrical-distance method the author wants listed in the reading list). - Look at a…
- Fix
- At acquisition: use less gel, check impedances for suspiciously low neighbouring pairs, and re-prepare bridged sites before recording. - Offline: mark bridged channels as bad before computing an average reference, interpolating, or running ICA; if both channels of a pair are bridged to each other, keep one and interpolate the other only from unbridged neighbours, or drop the pair. - Keep the br…
In other tools
In other toolsEEGLAB · FieldTrip — names only
The equivalents of what this lesson does, for a reader who works in another toolbox. Function names only: their own documentation is the place to learn how to call them.
EEGLAB
pop_chaneditEEGLAB
FieldTrip
ft_read_sensFieldTrip
Names checked 2026-09-18 against EEGLAB 2026.0.0 (plugins at the versions in EEGLAB’s own plugin list) and FieldTrip 20251218.
Reading
- Nunez & Srinivasan (2006). Electric Fields of the Brain. unverified
- Kane et al. (2017). Revised glossary of clinical EEG terms. unverified