EMG Electrode Placement — The Practical Guide

Why placement decides the measurement#

An electrode pair records whatever electrical activity reaches it through the tissue. Move it two centimetres and three things change at once:

• Amplitude. Over the innervation zone action potentials travel in opposite directions and partly cancel — the amplitude drops and becomes unstable. Over the tendon there is hardly any signal at all. Only the muscle belly between the two gives a stable, representative amplitude.
• Selectivity. Near the border of a muscle the pair also picks up the neighbour — crosstalk that no filter can remove afterwards, because it is real EMG, just from the wrong muscle.
• Reproducibility. A position defined as "on the biceps" cannot be repeated next week; a position defined as "at one third of the line from the acromion to the cubital fossa" can — by you or by a colleague.

That is why standardised placement is not pedantry. Without it, a difference between two sessions can be a difference in electrode position rather than in the muscle.

The seven rules of surface EMG placement#

These are the SENIAM sensor recommendations, condensed with the practical notes from Konrad and Cram's:

1. Electrode type and size. Pre-gelled Ag/AgCl electrodes, conductive area ≤ 10 mm in diameter. Smaller electrodes are more selective; larger ones average more of the muscle and more of its neighbours.
2. Inter-electrode distance 20 mm (centre to centre). Shorter for very small muscles — but never more than a quarter of the fibre length. Keep the distance identical between sessions and between sides.
3. Parallel to the fibres. The line through both electrodes follows the muscle fibre direction. Across the fibres the amplitude drops and the frequency content is distorted.
4. On the muscle belly, between innervation zone and distal tendon. SENIAM places most sites at a fraction of a landmark line for exactly this reason. Avoid the motor point region and the musculotendinous junction.
5. Defined by landmarks. Acromion, C7, epicondyles, iliac crest, greater trochanter, patella, malleoli — measure the line, mark the fraction, place the pair. Write the distances down.
6. Reference electrode on inactive tissue. Wrist, ankle, C7 spinous process, iliac crest or the patella — a bony spot away from the recorded muscle. A poor reference is the most common cause of a noisy baseline.
7. Fix everything. Cables taped or netted so they cannot pull; electrodes placed with the muscle in the most shortened position for dynamic tasks so the pair stays on the active mass (Konrad names biceps and vastus medialis as typical migrators).

Step by step: from bare skin to a valid signal#

1. Position the person in the SENIAM starting posture given on the muscle page (e.g. sitting, elbow flexed 90° for Biceps Brachii; prone for Longissimus Thoracis). Landmarks move with posture — palpate in the position you will measure in.
2. Prepare the skin. Shave if necessary, abrade lightly (abrasive paste or fine sandpaper, three to four gentle strokes), clean with alcohol, let it dry. The skin should be faintly reddened. For static clinical tests alcohol alone may do; for running, jumping or long sessions do the full routine.
3. Find the landmarks and mark the site. Measure the reference line with a tape, mark the fraction (e.g. two thirds of the way along the line from the anterior superior iliac spine to the lateral side of the patella for Vastus Lateralis), then mark the fibre direction with a skin pen.
4. Apply the pair. Centre the two electrodes on the mark, 20 mm apart, along the fibre line. Press the rims, not the gel centre. Wait — impedance keeps falling for several minutes.
5. Apply the reference electrode on the nearest bony site and connect all leads; check that each lead really belongs to the muscle you think it does.
6. Fix cables and sensor. Tape or elastic netting, no tension on the electrode. Never tape over the electrode itself.
7. Signal check — see next section. Only after it counts as "clean" does the recording start.

The five-step signal check (after Konrad)#

1. Validity. Ask for an isolated test contraction of that muscle (the test movement on every EMG Guide page). The channel that responds must be the one you labelled. Swapped leads are embarrassingly common.
2. Impedance. If your system shows it: 1–5 kΩ very good, 5–10 kΩ good, 10–30 kΩ acceptable for easy conditions, above 50 kΩ repeat the preparation.
3. Raw baseline at rest. Person fully relaxed, best lying down. Zoom in on the raw trace: a flat line with random spikes below 10–15 µV, mean rectified level around 1–3.5 µV, no offset, no slow waves.
4. Baseline behaviour around contractions. Every burst must return to zero within milliseconds. Shifts that persist longer point to cable movement, pressure or muscle wobble.
5. Frequency spectrum during a moderate static contraction: steep rise from 10 Hz, peak around 50–80 Hz, near zero by 200–250 Hz. A sharp peak at 50/60 Hz means mains hum — check grounding of nearby devices and the reference electrode before reaching for a notch filter.

Mistakes that ruin a recording#

Where to find the position for each muscle#

EMG Guide documents 106 muscles. For each of them the muscle page states the exact position, orientation, starting posture and test movement, and the glossary explains the terms used. The 3D viewer shows the same position on the model — open a muscle in the viewer and the electrode pair is highlighted where it belongs:

• Shoulder and neck: Upper Trapezius, deltoid, sternocleidomastoid
• Arm: Biceps Brachii, triceps, wrist flexors and extensors
• Trunk: Rectus Abdominis, obliques, Longissimus Thoracis
• Hip and thigh: Gluteus Medius, Rectus Femoris, Vastus Lateralis, hamstrings
• Lower leg: Tibialis Anterior, Gastrocnemius (Medial Head), soleus, peroneals

52 muscles (26 bilateral pairs) carry the official SENIAM recommendation; the remaining muscles use placements from peer-reviewed sources named on each page. Every position also lists the crosstalk neighbours and the artifacts typical for that site.

Beyond SENIAM: three things to know#

Not every position is equally specific. Cram's atlas distinguishes specific sites (predominantly one muscle) from quasi-specific sites (a group that surface electrodes cannot separate — wrist flexors, medial hamstrings, hip adductors). Both are legitimate; the interpretation must match. The practice block on each muscle page states the type.

Not every muscle is a surface muscle (check a muscle). Levator scapulae, rhomboids, supraspinatus, iliopsoas or tibialis posterior lie under other muscles and need fine-wire EMG. They are deliberately not in the EMG Guide catalogue — a surface pair over the trapezius records the trapezius, whatever the label says.

Standard beats individual optimisation. A position that gives you a bigger amplitude today is not automatically better; a position you can reproduce next week and compare with the literature is. Use the standard, note any deviation, and change nothing between sessions.