11.3 Artifact Identification & Systematic Troubleshooting
Key Takeaways
OR artifact (Bovie/electrosurgery, 60 Hz line noise, pumps, warmers, electrode faults) can mimic or mask true neural change—identify the signature before calling an alert
Electrosurgery typically produces high-amplitude broadband interference during active cutting/coagulation; pause averaging and resume when the field is electrically quiet
60 Hz artifact is best reduced by fixing the source (impedance, cable routing, unnecessary devices)—notch filters are a last resort because they can distort EP morphology
Use a stepwise checklist: reproduce → technical integrity → environment/noise → physiology/anesthesia → surgical correlation—without relying on proprietary 'black box' algorithms
Document artifact periods and troubleshooting steps so postoperative review can separate technical downtime from true neurologic events
11.3 Artifact Identification & Systematic Troubleshooting
Quick Answer: Artifacts are electrical or mechanical contaminants that look like—or hide—neural signals. Learn the signatures of electrosurgery, 60 Hz interference, and electrode faults, then troubleshoot in a fixed order: confirm the finding, verify the recording chain, quiet the environment, check physiology/anesthesia, and only then attribute change to surgery.
Intraoperative recognition fails when noise is mistaken for injury or when real injury is dismissed as "just artifact." Domain II expects a disciplined approach that any competent technologist can explain—no secret vendor algorithm required.
Common Artifact Signatures in the OR
Electrosurgery (Bovie / bipolar / monopolar)
Active electrosurgery injects broadband, high-amplitude noise into nearby electrodes and often saturates amplifiers. Typical behavior:
- Noise appears only while the pedal is down (or while bipolar is active)
- Traces may flat-line or show hash across many channels
- Averaging during active Bovie contaminates the average with non-neural energy
Response: Pause acquisition/averaging during heavy electrosurgery; resume when the field is quiet. Do not interpret a Bovie-contaminated average as an SSEP alert. If the surgeon must coagulate continuously during a critical window, say so: monitoring sensitivity is temporarily reduced.
Line-frequency (60 Hz) interference
In North America, 60 Hz (and harmonics) couples from AC power, warming blankets, blood warmers, surgical lights, microscopes, and poorly routed cables. Appearance:
- Rhythmic sinusoidal contamination
- Often worse with high or unbalanced electrode impedances
- May preferentially affect long lead runs parallel to power cords
Response priority (source first):
- Recheck and balance impedances
- Separate electrode leads from power cords and electrosurgery cables
- Unplug or relocate unnecessary powered devices near the head/recording sites
- Verify ground/reference integrity and headbox connections
- Consider stimulus rate choices that avoid exact locking to line frequency when your lab protocol allows
- Notch filter last — a 60 Hz notch can reduce artifact but may distort EP morphology near the notch; treat it as temporary damage control, not a substitute for finding the source
Electrode and cable artifacts
| Clue | Likely technical issue |
|---|---|
| One channel wild, neighbors quiet | Loose electrode, dried gel, bent needle, bad lead |
| Intermittent spikes with movement | Partial disconnect, strain on cable, patient motion |
| All channels die at once | Headbox unplugged, stimulator off, power loss, master reference lost |
| Stimulus artifact huge / polarity wrong | Stim cable reversed, broken return, intensity unexpectedly high |
| EMG "popping" with saline irrigation | Fluid bridge, electrode lift, or true irritation—correlate with field |
Other environmental noise
C-arm, intraoperative MRI/CT, ultrasonic aspirators, drills, and somatosensory stimulators from other teams can inject interference. Time your averages between bursts of equipment use when possible. Document unavoidable downtime.
Artifact? → Does it track a device (Bovie on/off, warmer, C-arm)?
→ Is it one channel (electrode) or all (systemic/power)?
→ Fix source before changing interpretation rules
Systematic Troubleshooting Sequence
Use the same order every time so you do not skip technical causes or delay surgical communication.
Step 1 — Reproduce and characterize
Is the change present on the next average/trial? Unilateral or bilateral? Which peaks/muscles? Did it begin abruptly or drift? Characterization prevents treating a single noisy sweep as cord injury.
Step 2 — Technical integrity (recording and stimulation chain)
- Impedance check on affected channels
- Visual inspection of electrodes, tape, and connectors
- Confirm stimulator is delivering (peripheral twitch or known peripheral potential)
- Confirm correct montage, polarity, and that the intended channel is displayed
- Swap a suspect lead with a known-good spare if available
If technical failure explains the change, fix it, re-baseline if needed, and document—do not leave a false surgical alert hanging.
Step 3 — Environmental / artifact control
- Ask whether electrosurgery is active
- Move cables; silence unnecessary devices
- Wait for imaging bursts to finish
- Avoid interpreting during known noise epochs
Step 4 — Physiologic and anesthetic review
- MAP, temperature, EtCO₂, hematocrit/oxygenation as available
- Recent boluses, volatile changes, neuromuscular blocker redosing
- Correlate with Section 11.2 bilateral/systemic patterns
Step 5 — Surgical correlation and communication
- What maneuver is occurring?
- Communicate the pattern and your differential
- Request pause, reversal, or inspection as appropriate
- Re-acquire after intervention and report recovery or persistence
This sequence is deliberately generic. Vendor auto-alerts, proprietary SNR scores, or black-box "significance" flags may assist workflow but are not a substitute for understanding impedance, averaging, artifact signatures, and clinical correlation. On the CNIM exam, defend the physiologic and technical reasoning—not a brand-name feature.
Averaging, Filters, and Gain: Tools With Side Effects
More averages improve signal-to-noise for random noise but help less against continuous 60 Hz or Bovie hash that is time-locked or overwhelming. Do not "average away" electrosurgery.
Filter changes can clean a display while altering latency/amplitude appearance. If you change filters mid-case, document it; large unexplained morphology shifts after a filter tweak are technical until proven otherwise.
Gain/display scale changes how big a wave looks, not how big it is. Always judge alerts from measured amplitude relative to baseline, not from a zoomed screenshot.
Rejecting bad sweeps (artifact rejection) is appropriate when your system and protocol support it—rejection should remove obvious non-neural contaminants, not silently discard the very averages that show a real surgical change. Know your lab's rules.
Documentation During Troubleshooting
Annotate:
- Time artifact began and ended (for example, "Bovie 10:12–10:18—averaging paused")
- Impedance values before/after electrode replacement
- Devices relocated or unplugged
- Anesthetic/vital-sign interventions
- Surgical pauses and whether signals recovered
Postoperative questions often hinge on whether a flat period was technical downtime or a true neurologic event. Your annotations are the record.
Worked mini-scenario
Lower-extremity cortical SSEPs become unreadable during pedicle work. Free-run EMG shows intermittent spikes. Steps: (1) confirm Bovie is active—pause averages; (2) when quiet, impedance OK bilaterally; (3) MAP stable, no volatile change; (4) during screw stimulation, triggered EMG threshold is low on one screw—communicate possible breach. The "lost SSEP" during Bovie was artifact; the EMG finding was the surgical signal. Separating those threads is the skill this section trains.
Master artifact literacy and a repeatable checklist, and you will protect patients from both missed injury and unnecessary alarm fatigue—exactly what intraoperative signal-change recognition demands.
The most appropriate response when monopolar electrosurgery is saturating SSEP averages is to:
Pause averaging during active electrosurgery and resume interpretation when the field is electrically quiet
Lower the high-frequency filter to 1 Hz so Bovie noise disappears permanently
Treat every Bovie-contaminated average as a 50% amplitude alert
Increase stimulus intensity until the noise is buried
A 60 Hz sinusoid contaminates multiple SSEP channels. Which action is the best first-line approach?
Immediately enable a 60 Hz notch filter as the only intervention
Stimulate at exactly 60 Hz so the artifact cancels
Improve impedances, separate leads from power cords, and remove unnecessary nearby powered devices before relying on a notch filter
Declare a surgical emergency because 60 Hz proves cord ischemia
In a systematic troubleshooting sequence after a sudden signal loss, which check belongs earliest—before attributing the change to surgical injury?
Asking the surgeon to implant a permanent paddle stimulator
Verifying reproducibility plus technical integrity (impedance, connections, stimulator output, electrode security)
Changing the institutional alert criterion from 50% to 10% amplitude
Turning off all documentation to save time
Why should proprietary device "auto-alert" scores not replace the technologist's differential diagnosis?
Auto-alerts are illegal in all US operating rooms
CNIM candidates are forbidden to look at computer screens
Amplitude criteria never matter in IONM
Exam and clinical reasoning require explainable checks of artifact, physiology, anesthesia, and surgical timing—not unexplained vendor algorithms alone
Sections you finish are checked off in the contents.