Long-term monitoring study fails when it stops at waveform memorization, because the credential's core work is reviewing hours of video-EEG and deciding, event by event, what actually happened. This guide organizes CLTM preparation around full-event review: pick one recorded event, reformat the montage, correlate the video, write a one-page report, and score it against a fixed rubric. Along the way it contrasts bipolar and referential localization, sets electrographic seizures against rhythmic look-alikes, and introduces intracranial reading at the contact-map level. Start this week with a single de-identified event from your own lab and finish its report.
Why electroclinical correlation anchors every CLTM study session
Long-term monitoring exists to pair EEG with recorded behavior, so make electroclinical correlation the spine of your study: for every event you review, name the EEG onset, the observed behavior, and whether the two agree.
Define electroclinical correlation precisely: a time-locked review in which the electrographic description (earliest involved electrodes, morphology, evolution) and the semiology description (first visible movement, automatisms, posturing, duration) are written as one linked account. Routine EEG reading asks what the background contains and which abnormalities are present. Long-term monitoring asks what happened and when, and whether the brain recording explains the behavior. A reported event without semiology, or semiology without EEG, is only half a finding.
To apply this, build a semiology glossary: terms such as automatisms, versive head turn, tonic posturing, and hypermotor behavior, each with a one-line note about what you saw on video alongside it. Then practice paired description. Play the video muted and describe the EEG alone; watch the video alone and describe the behavior alone; then check the two accounts for agreement. Disagreement is not failure. It is the finding that drives revision of the event log and sharpens your descriptions.
Montage choice: using phase reversal and amplitude gradients correctly
Bipolar montages localize through phase reversal between adjacent electrodes in a chain; referential montages localize through amplitude comparison against one common reference. Reformat the same event in both before committing to a localization.
The longitudinal bipolar (double banana) montage suits screening long stretches of recording: trace activity down each chain to the point where deflections reverse phase between adjacent electrode pairs, and that point approximates the maximum. Transverse chains let you compare homologous left-right regions in one view. Referential montages, using ipsilateral ears, an average reference, or another common electrode, preserve the full signal at each electrode, so the discharge peaks at the electrode nearest the generator. Bipolar derivations cancel shared signal; referential derivations keep it but import reference problems.
Those reference problems are the study point. An ear reference sits close to temporal generators, so a left temporal discharge can contaminate the left ear reference and distort every left-sided channel. An average reference can be dominated by one large focal field, flipping apparent polarity elsewhere. Build the habit of cross-checking: when a referential finding looks unusual, confirm it with a bipolar reformat, and vice versa. In review software this costs seconds; in a written report it separates a defensible localization from a guess.
Electrographic seizure versus rhythmic look-alike: a three-feature check
Call an event electrographically ictal only when an abrupt rhythm evolves in frequency, amplitude, or spatial field over seconds. Look-alikes are rhythmic but static. Check evolution, field, and video agreement every time.
Run every flagged event through three checks. First, evolution: frequency typically slows while amplitude grows, or the rhythm decrements. Second, field: activity spreads into adjacent electrodes beyond the earliest one. Third, clinical agreement: behavior on the time-locked video matches the timing. Know the benign rhythmic patterns for contrast, including rhythmic temporal theta of drowsiness, wicket spikes, and subclinical rhythmic electrographic discharges of adults, plus brief static rhythmic discharges. The distinguishing discipline is always the same: compare the event page against the page immediately before onset.
In practice, give each flagged event a one-line written justification against the three features. If two of the three are absent, label the finding as a rhythmic interictal or borderline pattern rather than an ictal one, and state why in the log. This habit matters because event labels in a monitoring study feed directly into seizure-frequency counts that clinicians use to judge treatment response. A static rhythm labeled as a seizure overstates what happened, and the correction is far easier at review time than after the report is signed.
Worked scenario: the 3 a.m. flag that was not a seizure
A detection algorithm flags rhythmic right temporal activity during sleep. The tempting call is 'focal seizure' from the EEG page alone. The better decision is to demand evolution, field spread, and video agreement before labeling.
Setup: during overnight review, a flag marks roughly 7 Hz activity maximal over the right anterior temporal electrodes. The reviewer's first draft reads 'electrographic seizure, right temporal onset.' The mistake is that the description rests on a single page: no baseline comparison, no referential reformat, and the video was never opened. The plausible error is treating the detection software's label as a diagnosis instead of as a pointer to a page that deserves examination.
Better decision: scroll to a pre-event baseline, reformat into a referential montage, and open the time-locked video. The rhythm coincides with rhythmic head rocking visible on camera, occupies electrodes over the moving region, shows no frequency evolution, and the movement bursts outlast any EEG change. Relabel: 'rhythmic movement artifact, no electrographic correlate.' This matters because one mislabeled night, repeated across a monitoring admission, becomes a systematic error in seizure counts that clinicians use when weighing treatment response and surgical candidacy.
Intracranial monitoring: strips, grids, and depth contacts on paper
Intracranial recording reverses surface expectations: amplitudes are large, fields are narrow, and localization depends on knowing each contact's anatomy. Learn contact labeling and referencing schemes before studying any waveform.
Subdural strips and grids rest on the cortical surface, sample broadly, and show very large amplitudes with sharp gradients between adjacent contacts. Stereo-EEG depth electrodes sample along a trajectory through the brain, so each contact covers different tissue and 'onset at contact X' means nothing without the contact map. Referencing decisions, such as a quiet contact or bipolar pairs, change which contacts appear to be involved at onset. The practical consequence: an intracranial tracing is unreadable without its implantation scheme diagram in hand.
Study implantation schematics from published case reports alongside their traces. Draw the electrodes, label the contacts, and practice stating onset in anatomical terms, for example 'onset at the two most mesial contacts of the amygdala-hippocampal depth,' rather than by channel number. Recognize that high-frequency activity, largely attenuated at the scalp, is well visualized intracranially and may mark the earliest field. Keep functional mapping as a procedure the clinical team performs; your study target is reading labeled recordings and understanding how mapping results are documented in reports.
| Feature | Surface LTM | Intracranial monitoring |
|---|---|---|
| Electrode position | Scalp, standardized system | Strips and grids on cortex; depth electrodes along trajectories |
| Expected amplitude | Microvolt range, broad fields | Much larger, sharp gradients between adjacent contacts |
| Localization evidence | Phase reversal or amplitude versus a reference | Contact map plus anatomical labels |
| High-frequency activity | Largely attenuated by scalp and skull | Well visualized; may mark the earliest field |
| Review prerequisite | Montage and filter fluency | Implantation scheme diagram in hand |
Worked scenario: sweat artifact mimicking generalized slowing
Hours into a recording, broad slow swishing spans many electrodes. The tempting call is 'diffuse slowing.' The better decision is to test its filter dependence and its electrode-spanning field before describing the background as abnormal.
Setup: on the afternoon of a day-long recording, anterior temporal and frontal regions show large, slow, irregular waveforms that bridge electrode boundaries and give every channel a soggy baseline. The draft note says 'diffuse cerebral slowing, worse than the morning.' The mistake is copying the impression from the screen without comparing frequency, distribution relative to electrode boundaries, or earlier pages. Sweat is a per-electrode phenomenon: each electrode develops its own slow potential, so the pattern does not respect the field of any single cerebral generator.
Better decision, as a display exercise: compare morning pages from the same recording, note that the slow activity is maximal at individual electrodes rather than at cerebral regions, and raise the low-frequency filter on a saved segment. Sweat collapses while genuine cerebral slow activity largely persists. Check what the video shows: a warm room, hours of recording, visible perspiration. Relabel the segment as artifact and describe the true background separately. This matters because the background characterization is often the report's headline impression, and a misread there misleads every reader downstream.
A four-week practice loop with a report rubric and readiness checks
Run a weekly loop: review one full event, reformat montages, write a one-page report, and score it against a rubric after a day's delay. Readiness means stable rubric scores and fluent reasoning, not a predicted result.
A suggested adaptable sequence: weeks one and two, surface events only, one event per session, with montage reformats, video correlation, and a written report each time. Week three, artifact drills using de-identified segments containing sweat, pulse, and movement artifact, each labeled with a one-line justification. Week four, intracranial paper cases with hand-drawn contact maps, plus one timed full-day summary. If rubric scores stall in one row, extend that portion of the sequence rather than moving on. Adjust the totals to your calendar; the loop, not the schedule, is the mechanism.
Readiness checks: you can state the phase reversal rule and the referential amplitude rule from memory; you can produce a one-page event report satisfying every rubric row in a single sitting; you can justify rejecting an artifact in two sentences; and you can summarize a full day of monitoring in three sentences that answer the referral question. Treat your self-check scores as learning milestones for pacing your study. They measure the consistency of your review process, not a prediction of your examination result.
- Onset: the earliest involved electrodes and the first visible behavior are both named with times
- Evolution: frequency, amplitude, and field changes are described across sequential pages
- Corroboration: montage reformats and the EKG channel were checked and mentioned
- Rejection reasoning: each alternative explanation, artifact or look-alike rhythm, is addressed in one line
- Impression: the closing sentence answers the referral question in plain language
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.