Artifact Recognition & Troubleshooting

Artifact — electrical activity recorded from sources other than the nervous system — is the single largest source of error in clinical neurophysiology. Over-reading muscle artifact as epileptiform activity falsely diagnoses epilepsy; mistaking 60-Hz line noise for high-frequency cortical rhythms misleads ICU teams; failing to recognize stimulus artifact in NCS distorts latency and amplitude measurements. The skilled electrophysiologist treats artifact recognition not as a nuisance but as a core diagnostic skill: every borderline finding gets tested against alternative explanations before it gets labeled pathology. This page covers the major artifact categories, the patterns that distinguish artifact from biology, and the troubleshooting steps that fix recordings rather than over-interpret them.

How to Approach Any Suspicious Finding

1. Is it reproducible? Artifacts often vary breath-to-breath, movement-to-movement; biology is more stereotyped.
2. Does it follow a recognized field pattern? Biological dipoles produce predictable polarity gradients; artifacts often have no neuroanatomic sense.
3. Does it appear on only one channel? True intracerebral activity usually has a measurable field; activity confined to one channel is almost always artifact (electrode pop) or local muscle.
4. What time course? Brief monophasic spikes <20 ms usually muscle or artifact; epileptiform discharges typically biphasic 70–200 ms.
5. What filter sensitivity? A finding that disappears with wider filter settings is suspicious for filter ringing.
6. Does the patient’s behavior correlate? Movement, swallowing, talking, sweating — log them and look for time-locked artifact.

EEG Artifacts

60 Hz (or 50 Hz) Line Noise

• Source: alternating current power lines coupling into the recording system.
• Pattern: pure sinusoidal at exactly 60 Hz (or 50 Hz), often on multiple channels simultaneously.
• Worsens with high electrode impedance, unequal active/reference impedances, or absent grounding.
• Fixes: re-prep skin, lower impedances, check ground electrode, move away from electrical equipment, use notch filter as last resort.
• The notch filter creates ringing artifact around true sharp waves — avoid if borderline epileptiform activity is being evaluated.

Electrode Pop

• Source: brief loss of electrode-skin contact; impedance change creates a transient voltage shift.
• Pattern: a sudden monophasic deflection (often square-wave shaped) limited to ONE channel, with abrupt baseline shift.
• Key feature: confined to a single electrode, no anatomic field around it.
• Often misread as focal sharp wave in inexperienced hands.
• Fix: re-prep skin, re-apply electrode, lower impedance.

Eye Movement Artifact

• Source: the eye is a dipole (cornea positive, retina negative); ocular movement displaces the dipole and creates voltage at frontal scalp electrodes.
• Pattern:
  • Vertical eye movement (blink): symmetric downward (positive) deflection at Fp1 and Fp2.
  • Horizontal eye movement: opposite polarity at F7 vs F8 (or LFp7 vs LFp8 in 10-10).
  • REM sleep on PSG: vertical eye movements visible in EOG channels.
• Identify with frontal predominance, smooth waveform shape, and concordance with patient behavior.
• Can suppress with patient instruction to “look straight ahead” or with EOG channels to mark.

Muscle Artifact (EMG Contamination)

• Source: scalp muscles (temporalis, frontalis, occipitalis); contamination from neck movements.
• Pattern: high-frequency (>30 Hz) bursts, irregular shape, often spiky.
• Frontal regions and temporal/occipital regions most affected.
• Distinguished from cortical activity by >30 Hz frequency, irregular pattern, and concordance with patient agitation or grimacing.
• Fixes: ask patient to relax jaw and neck; head support; sedation in some ICU settings.
• “Polyspike” appearance in muscle artifact is one of the most common over-reads as epileptiform.

ECG Artifact

• Source: cardiac dipole picked up by scalp electrodes, especially in patients with short, broad chests.
• Pattern: rhythmic at heart rate (60–100 bpm); often visible at vertex, mastoid, or ear electrodes.
• Identify by comparing with ECG channel (if recorded) or by noting strict regularity at cardiac rate.
• Important: ECG artifact at the vertex is a common cause of “sharp wave at Cz” misread.
• Fix: use mastoid reference rather than ear reference; or use averaged ears.

Sweat Artifact

• Source: sweat under electrodes creates slow ionic potentials.
• Pattern: slow undulating waves, frequency below 1 Hz, often on multiple channels.
• Worse in warm rooms or febrile patients.
• Fix: cool room, dry electrodes, re-prep skin.

Movement Artifact

• Source: body or head movement displacing electrodes.
• Pattern: irregular high-amplitude slow waves correlating with movement.
• Fix: secure electrode head bands; patient instruction.

Ventilator Artifact

• Source: ventilator cycling produces small rhythmic voltage shifts.
• Pattern: regular at ventilator rate (typically 10–20/min), often subtle.
• Easily mistaken for periodic discharges; verify by correlation with ventilator settings.

IV Pump Artifact

• Source: pump motor or drip-counter cycling.
• Pattern: rhythmic regular electrical interference at pump cycle rate.
• Fix: relocate pump, change to different model, or use battery-operated pump.

Chewing, Swallowing, Tongue Movement

• Source: mouth and throat muscle activation; tongue is a strong dipole.
• Pattern: rhythmic temporal artifact during chewing; sharp deflections during swallowing.
• Glossokinetic artifact (tongue): bilateral synchronous deflections at temporal regions during talking or swallowing.

“Pulse Artifact”

• Source: a scalp electrode placed over a pulsating artery (often temporal).
• Pattern: rhythmic deflection at heart rate, limited to one or two channels over the pulsing vessel.
• Fix: relocate electrode 1–2 cm.

EEG Patterns Commonly Mistaken for Pathology

Normal Variants vs Pathology

• Wickets: temporal arches at 6–11 Hz, sharply contoured. Benign. Look for absence of sharp epileptiform discharge in surrounding activity.
• BETS (benign epileptiform transients of sleep): brief small sharp spikes in light sleep, mostly temporal. Stereotyped morphology, occur in isolation, no aftercoming slow wave.
• 6-Hz phantom spike-wave: small generalized spike-and-wave, especially during drowsiness. Female predominance, no aftercoming slow wave.
• SREDA (subclinical rhythmic electrographic discharge of adults): a benign rhythmic theta build-up in older adults, often during drowsiness. Can mimic non-convulsive seizure.
• Hypnagogic hypersynchrony: high-amplitude rhythmic theta in children at sleep onset. Normal.
• Photoparoxysmal response without clinical correlate: photoparoxysmal discharges in normal individuals during photic stimulation; only the photoconvulsive (full sustained 3-Hz spike-wave) variant is clinically significant.

Mu Rhythm

• Central arciform alpha-frequency (8–13 Hz) rhythm.
• Blocks with movement or imagined movement of the contralateral limb.
• Often unilateral or asymmetric.
• Normal — sometimes misread as focal central abnormality.

NCS Artifacts and Errors

Stimulus Artifact

• Source: the stimulating current itself coupling to recording electrodes.
• Pattern: large brief deflection at stimulus onset; can obscure short-latency responses.
• Fix: optimize ground placement (between stimulator and recording electrode), low electrode impedances, proper grounding circuit.

Volume-Conducted Response

• Source: another muscle activated by current spread.
• Pattern: response appears slightly different in shape, often inverted polarity (initial positive deflection) from the expected target muscle.
• Identifies as cross-talk; fix by repositioning electrodes or using a more selective stimulator.

Submaximal Stimulation

• Source: stimulator output too low to activate all axons.
• Pattern: amplitude does not plateau; raising intensity increases amplitude.
• Always raise stimulation until amplitude plateaus (supramaximal) — usually 20–25% above the intensity at which amplitude stops growing.

Anodal Block

• Source: stimulator polarity reversed (anode where cathode should be).
• Pattern: response shrinks or disappears at expected stimulation.
• Fix: verify cathode placement distal to anode for orthodromic motor studies.

Cold Limb

• Cooler nerve fibers conduct slower (~ 1.5 m/s per °C drop).
• An apparently demyelinating pattern can be cold-induced.
• Always warm distal extremity to 32°C before NCS.
• Measure skin temperature and adjust if needed.

Misplaced Recording Electrode

• Active electrode not over motor point: produces an initial positive deflection on CMAP (instead of expected initial negative).
• Adjust electrode position until clean initial negative deflection appears.

EMG Artifacts

Needle Motion

• Brief high-frequency burst as needle moves through tissue.
• Distinguish from fibrillation potential: fibrillations are stereotyped, repetitive, occur with the needle stationary.

End-Plate Noise

• Small (10–50 μV), high-frequency noise heard as “sea-shell” sound through speaker.
• Recorded when needle is in the end-plate region.
• Distinguish from fibrillations: end-plate noise is irregular, fibrillations are regular.
• Move needle slightly to avoid; can be painful for patient if needle is in end-plate.

Voluntary Activation Contaminating Insertional Activity

• If patient cannot relax, voluntary motor unit activity contaminates the insertional and spontaneous activity recording.
• Fix: patient relaxation, position support, sometimes mild sedation.

EP/SSEP Artifacts

• Excessive trial averaging: low-amplitude EP requires many trials (≥500), but movement artifact contaminates averages.
• Baseline drift: very low frequency activity contaminating the average; high-pass filter (1 Hz) reduces.
• Inadequate stimulus intensity: weak peripheral stimulation produces small or absent EPs.
• Inadequate trial number: noisy average; rerun with more trials.

PSG Artifacts

• ECG contamination of EEG: prominent in obese patients; use mastoid reference.
• Respiratory artifact: chest belt or airflow movement coupling into EEG via volume conduction; pattern correlates with respiration.
• Snoring artifact: vibration from snoring picked up by chin EMG and microphone.
• Sleep-related leg movements that look like seizures: distinguished by stereotyped pattern, frontal-only EEG, and no EEG correlate.

Practical Artifact Troubleshooting

🔍 Did You Know?

The over-reading of EEG artifact as epileptiform activity is one of the most consequential errors in neurology — and one of the most common. Studies of “epilepsy” patients re-evaluated at tertiary centers consistently find that 25–30% of patients carrying an epilepsy diagnosis do not have epilepsy, and the most frequent reason is over-interpretation of normal variants or artifacts on prior EEGs. Common errors: wicket spikes called temporal IEDs, BETS called pathologic sharp waves, 60-Hz line noise called fast spike activity, electrode pop called focal sharp waves, eye-movement artifact called frontal epileptiform discharges, and SREDA called non-convulsive seizure. The consequences for misdiagnosed patients are substantial: years of ASM exposure, restricted driving, anxiety about a chronic disease, and missed diagnoses (the actual cause of the spells — syncope, psychogenic non-epileptic seizures, migraine variants — goes untreated). The lesson is clinical humility: a single equivocal EEG finding should never be the sole basis for an epilepsy diagnosis. Confirm with repeat recording (with sleep), correlate with semiology, and when in doubt, refer to a clinical neurophysiologist with epilepsy training before committing the patient to long-term therapy. The cost of being right about real epilepsy is high — but the cost of being wrong about pseudo-epilepsy is at least as high.

Pitfalls and Pearls

• Any single-channel sharp wave: suspect electrode pop until proven otherwise.
• “Polyspike” in agitated patient: probably muscle, not IED.
• Vertex sharp at heart rate: ECG artifact, not Cz IED.
• Slow undulating baseline: sweat, not focal slowing.
• Bilateral frontal slow waves: eye movement, not frontal pathology.
• Wickets: benign temporal arches; not IED.
• BETS: small sharp spikes in light sleep; benign.
• SREDA: benign rhythmic theta build-up; do not call non-convulsive seizure.
• Mu rhythm: central arciform alpha, blocks with movement; normal.
• 60 Hz line noise: re-prep skin first; notch filter last resort.
• Notch filter creates ringing: artifact-looking sharp waves around real transients.
• Filter ringing test: a sharp wave that disappears with wider filter is not real.
• Cold limb NCS: looks demyelinating; warm to 32°C.
• Initial positive CMAP: electrode off motor point.
• Anodal block: response disappears at high stim; check polarity.
• Multiple modalities for borderline findings: don’t diagnose epilepsy from a single EEG.
• Document patient behavior: time-lock artifact to events.
• Photoparoxysmal but not photoconvulsive: usually benign; consider only photoconvulsive (sustained 3-Hz SW) clinically relevant.

References

1. Ebersole JS, Husain AM, Nordli DR Jr, eds. Current Practice of Clinical Electroencephalography. 4th ed. Wolters Kluwer; 2014.
2. Klass DW, Westmoreland BF. Nonepileptogenic epileptiform electroencephalographic activity. Ann Neurol. 1985;18(6):627-635.
3. Benbadis SR. Errors in EEGs and the misdiagnosis of epilepsy: importance, causes, consequences, and proposed remedies. Epilepsy Behav. 2007;11(3):257-262.
4. Preston DC, Shapiro BE. Electromyography and Neuromuscular Disorders. 4th ed. Elsevier; 2021.
5. Kimura J. Electrodiagnosis in Diseases of Nerve and Muscle. 4th ed. Oxford University Press; 2013.
6. Benbadis SR, Lin K. Errors in EEG interpretation and misdiagnosis of epilepsy: which EEG patterns are overread? Eur Neurol. 2008;59(5):267-271.