Page 1891 · Distinguishing psychogenic non-epileptic seizures (PNES) from true epileptic events via semiology, video-EEG correlation, and clinical feature pattern recognition
Semiology — the descriptive study of the clinical features of an event — remains the starting point of every seizure differential. Long before an EEG electrode is placed, a careful description of duration, motor pattern, eye behavior, and responsiveness can already tilt the differential strongly toward epileptic seizure or psychogenic non-epileptic seizure (PNES). No single semiological feature is diagnostic alone, but the pattern, taken together, is highly informative.
Bedside observation and witness/video description offer the first differential clues:
• Duration: epileptic seizures are typically brief (30–120 seconds); PNES events are frequently longer (often >2 minutes, sometimes 10–20 minutes) • Onset: epileptic seizures usually begin abruptly; PNES onset is often gradual, building over seconds • Course: epileptic seizures tend to evolve in a stereotyped, unidirectional pattern; PNES often show a fluctuating, waxing-and-waning intensity • Eye state: eyes are typically open in epileptic seizures (tonic eye deviation possible); eyes are frequently forcefully closed in PNES, sometimes resisting passive opening by the examiner • Motor pattern: epileptic tonic-clonic activity is synchronous and stereotyped across repeated events; PNES motor activity is often asynchronous, out-of-phase limb movement, or pelvic thrusting • Vocalization: ictal cry at onset favors epilepsy; crying or speaking during the event favors PNES
None of these are individually pathognomonic — video-EEG remains the reference standard — but a structured semiology checklist, applied consistently, substantially narrows the differential before any electrode is placed.
A witnessed history alone is documented to be inaccurate in up to 20–30% of cases — even experienced neurologists misclassify seizure type on history and video review alone. Semiology narrows the differential; it does not close it.
A structured capture protocol typically documents:
1. Trigger/context: time of day, sleep state, emotional stressor immediately preceding 2. Onset characteristics: sudden vs. gradual; presence of aura or prodrome 3. Motor phenomenology: tonic, clonic, myoclonic, atonic, or non-classifiable movement; symmetry; synchrony 4. Eye and facial features: open/closed, deviation, blinking rate, resistance to opening 5. Responsiveness during event: can the patient follow commands, react to painful stimuli, or maintain partial awareness during bilateral motor activity 6. Duration (timed, not estimated) 7. Post-event state: immediate reorientation vs. prolonged confusion (postictal state)
This record becomes the baseline against which the EEG correlation in Stage 2 is interpreted — the two data streams together, not either alone, drive the diagnosis.
If semiology narrows the differential, simultaneous video-EEG monitoring closes it. The central neurophysiological fact underlying this entire diagnostic pathway: a true epileptic seizure is, by definition, an abnormal, excessive, hypersynchronous discharge of cortical neurons — and that discharge is visible on scalp EEG as rhythmic spike-and-wave or fast rhythmic activity time-locked to the clinical event. A PNES event, in contrast, occurs over a background EEG that remains normal throughout — no ictal correlate, because the event is not driven by abnormal cortical electrical activity.
During video-EEG monitoring, the electroencephalographer reviews the EEG channel-by-channel, time-locked exactly to the video-documented clinical event:
Epileptic seizure — expected ictal pattern: • Rhythmic evolution in frequency, amplitude, and/or spatial distribution • Spike-and-wave, polyspike, or rhythmic fast activity emerging from and returning to background • Onset typically precedes or coincides with clinical onset; may show a clear focal onset zone • Followed frequently by postictal slowing (diffuse delta activity) correlating with postictal confusion
PNES — expected pattern: • EEG background remains awake, reactive, and unchanged throughout the entire clinical event • Muscle and movement artifact may obscure the trace, but careful review (or artifact-reduction techniques) confirms no underlying rhythmic ictal discharge • No postictal slowing — because there was no ictal state to recover from
This single distinction — presence vs. absence of a time-locked ictal discharge — is why the toggle in this simulator is weighted more heavily than any other single variable in the composite likelihood scores.
A normal EEG during a typical clinical event, correctly time-locked and technically adequate, is currently the single most decisive piece of evidence for PNES over epilepsy — but it must be paired with a technically adequate recording, since some deep or frontal epileptic foci can occasionally produce subtle or absent scalp changes.
Ictal EEG interpretation during genuine motor activity is technically demanding:
• Movement and muscle artifact can mimic or mask rhythmic activity — dedicated artifact-reduction filtering and careful visual review are essential • Some frontal lobe epileptic seizures produce bizarre, hyperkinetic semiology (pelvic thrusting, bicycling movements) that closely mimics PNES, while scalp EEG may show minimal or no clear ictal change due to deep or rapidly-propagating onset • Conversely, some PNES events occur with prominent muscle artifact that can be misread as "possible" ictal activity by an inexperienced reader • Best practice: events are reviewed by an epileptologist with simultaneous video, and ambiguous cases are re-captured with additional electrodes or longer monitoring before a final determination is issued
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Onset | |||
| Eye state | |||
| Motor pattern | |||
| Duration | |||
| Awareness | |||
| Post-event state | |||
| Injury / incontinence / tongue biting | |||
| Ictal EEG |
No individual clinical feature reaches diagnostic certainty on its own, but validated composite checklists — combining several PNES-suggestive or epilepsy-suggestive features — substantially improve pre-EEG diagnostic accuracy. Structured checklists such as those built on features like eye closure, ictal crying, gradual onset, and preserved awareness during bilateral movement have been prospectively validated against video-EEG gold-standard diagnosis.
The "PNES-suggestive features" slider in this simulator represents a composite count across six commonly cited, semiology-based indicators:
1. Gradual onset over seconds rather than an abrupt start 2. Closed eyes with resistance to passive opening by the examiner 3. Side-to-side (asynchronous) head or body movement rather than stereotyped tonic-clonic evolution 4. Pelvic thrusting or opisthotonic (arc-de-cercle) posturing 5. Waxing-and-waning course with fluctuating intensity, rather than a single stereotyped evolution 6. Retained awareness or responsiveness during bilateral motor activity, or rapid reorientation post-event
Each additional feature present increases the weight toward a PNES classification in the composite score — but this checklist score is deliberately combined with, never substituted for, the ictal EEG finding from Stage 2, since features can co-occur or be ambiguous in individual patients.
Studies combining eye closure + ictal crying + gradual onset + preserved awareness report positive predictive values for PNES exceeding 90% when three or more features co-occur — but video-EEG capture of a typical event remains required for a certain diagnosis before major treatment changes.
The corresponding features that favor a true epileptic seizure, tracked qualitatively alongside the PNES checklist:
• Sudden, unheralded onset (sometimes with a stereotyped aura) • Eyes open, sometimes with tonic upward or lateral deviation • Stereotyped, synchronous tonic-clonic motor evolution reproducible across events • Lateral tongue biting (highly specific, though insensitive) • Urinary incontinence during the event • Marked postictal confusion, lethargy, or focal neurological deficit (Todd's paresis) lasting minutes to hours • Occurrence during sleep, confirmed by polysomnography-linked EEG
Both feature sets are probabilistic, not absolute — patients with confirmed epilepsy can occasionally show an atypical feature, and patients with confirmed PNES can occasionally injure themselves or bite their tongue. The composite pattern, not any single sign, drives the classification.
In selected, monitored video-EEG admissions, a suggestion-based activation procedure — historically performed with a saline injection alongside verbal suggestion that an event may occur — can reproduce a patient's typical clinical event under continuous EEG observation. When ethically deployed with informed consent as part of routine care (never framed as deception to "catch" the patient), it is a recognized technique to safely capture a representative event for diagnostic confirmation.
Contemporary practice frames suggestion-based activation as an informed, disclosed component of care rather than covert deception:
1. The patient is informed in advance that a procedure exists which, in some patients, can trigger a typical event under safe monitored conditions, and gives explicit consent 2. A benign stimulus (e.g., saline infusion, alcohol swab, or a tuning fork) is paired with a verbal suggestion that an event may occur 3. Continuous video-EEG recording is running throughout 4. If a typical clinical event is reproduced, the recorded semiology and (absence of) EEG correlate provide strong, patient-specific diagnostic confirmation 5. Immediately following, results are discussed with the patient using non-judgmental, validating language — the event is real and distressing to the patient, even though its origin is not epileptic
Modern epilepsy monitoring units favor obtaining a typical spontaneous event whenever feasible, reserving suggestion-based activation for cases where spontaneous capture has failed after adequate monitoring time, and always with transparent consent.
Historically, saline-injection suggestion testing was sometimes performed without patient knowledge — an approach now considered ethically unacceptable. Contemporary guidelines require informed consent and transparent post-procedure disclosure as a condition of use.
Suggestion, placebo, and expectation effects are powerful, well-documented modulators of PNES events specifically because PNES episodes are generated by psychological rather than epileptogenic cortical mechanisms — voluntary and involuntary psychological processes can trigger the same motor and behavioral pattern the patient experiences spontaneously. True epileptic seizures, driven by abnormal hypersynchronous neuronal discharge, are not reliably reproducible by suggestion alone. This asymmetry is what gives the technique diagnostic power when a spontaneous event cannot be captured within a reasonable monitoring window, shortening length of inpatient stay and time to diagnosis.
A definitive diagnosis of PNES requires a typical event captured on video with simultaneous EEG showing no ictal correlate, interpreted by an epileptologist. Equally critical — and frequently underestimated — is how the diagnosis is communicated. Patients with PNES have a real, disabling condition; framing the diagnosis dismissively ("nothing is wrong," "it's not real," "it's just stress") damages trust and worsens outcomes, while a validating, collaborative disclosure improves engagement with psychiatric treatment and reduces unnecessary anti-epileptic drug (AED) exposure.
Once semiology, EEG correlation, and feature pattern data are integrated, the case resolves into one of three outcomes:
• Epileptic seizure confirmed — ictal EEG discharge present, semiology consistent with epilepsy → continue or optimize anti-epileptic drug (AED) therapy, evaluate for surgical candidacy if drug-resistant • PNES confirmed — normal ictal EEG background with a typical event and a convergent PNES-suggestive feature pattern → initiate psychiatry/psychology referral (commonly cognitive behavioral therapy adapted for functional seizures), and begin a gradual, supervised AED taper if the patient has no independent epilepsy history • Indeterminate — a captured event is atypical, technically inadequate, or the feature/EEG data are discordant → extended video-EEG monitoring is required before any treatment change; dual diagnosis (coexisting epilepsy and PNES) must also be considered, since 10–30% of confirmed PNES patients also have epilepsy
This three-branch logic is exactly what drives the "Classification" and "Recommended next step" tiles in this simulator as the sliders are adjusted.
Roughly 10–30% of patients with confirmed PNES also carry a genuine epilepsy diagnosis. A PNES diagnosis should never automatically trigger full AED discontinuation without individualized epileptological review — tapering is supervised, gradual, and reversible if seizures recur.
Evidence-based disclosure practice for PNES emphasizes:
1. Validate the reality of the events: "These episodes are real, involuntary, and not something you are consciously creating." 2. Show the evidence collaboratively: reviewing the video-EEG findings together helps patients understand the mechanism rather than feel dismissed 3. Reframe, don't just negate: explain PNES as a functional neurological disorder — a genuine disorder of brain network function, often connected to stress, trauma, or psychological factors, rather than "nothing wrong" or "fake" 4. Provide a concrete next step at the same visit: a specific psychiatry/psychology referral, ideally to a clinician experienced in functional neurological disorders, delivered before the patient leaves 5. Address AED tapering explicitly and gradually, with a clear medical rationale, to avoid the perception of abandonment
Outcome studies consistently show that how the diagnosis is delivered materially affects whether patients engage with psychiatric treatment, and therefore affects long-term seizure frequency and quality of life.