Intracranial EEG records electrical activity directly from the brain surface or within brain tissue, providing spatial and temporal resolution unmatched by scalp recordings. The two main forms are subdural grids and strips placed over cortical surface (electrocorticography, ECoG) and stereo-EEG (sEEG) depth electrodes inserted through small drill holes into deeper structures. Intracranial monitoring is the cornerstone of epilepsy surgery planning when scalp EEG is non-localizing, ambiguous, or when seizures arise from regions invisible to scalp recording (mesial temporal, insular, deep frontal, cingulate). This page covers the indications, techniques, recording principles, and clinical applications of intracranial monitoring.

Why Intracranial EEG

  • Scalp EEG has fundamental limitations: spatial smoothing, source inversibility, attenuation of deep sources.
  • Some seizures originate in regions invisible to scalp EEG (mesial temporal, insula, mesial frontal, deep parietal).
  • Intracranial recording captures activity not detectable at the scalp.
  • Spatial resolution improves from 6+ cm (scalp) to millimeters (intracranial).
  • Time resolution preserved at millisecond level.

Subdural Grid and Strip Electrodes

Approach

  • Craniotomy required for placement.
  • Grids placed over cortical surface (size: 4×4 cm to 8×8 cm or larger).
  • Strip electrodes for limited coverage in specific areas.
  • Multiple grids and strips can be combined.

Advantages

  • Broad surface coverage of cortex.
  • Excellent for mapping eloquent cortex (motor, language, sensory) via cortical stimulation.
  • Standard for ECoG during intraoperative epilepsy surgery.

Disadvantages

  • Cannot sample deep structures.
  • Requires craniotomy (significant procedure with morbidity).
  • Limited 3D sampling.
  • Adjacent grids can record overlap and confusion.

Stereo-EEG (sEEG) Depth Electrodes

Approach

  • Multiple depth electrodes inserted via stereotactic guidance.
  • Small drill holes only (no craniotomy).
  • Each electrode has multiple recording contacts along its length.
  • Sampling 3D regions of interest (mesial temporal, insula, cingulate, deep frontal, parietal).

Advantages

  • 3D coverage including deep structures.
  • Less morbidity than craniotomy.
  • Multiple regions can be sampled simultaneously.
  • Particularly suited for non-lesional or multifocal cases.
  • Allows detailed mapping of seizure onset zone and propagation.

Disadvantages

  • Limited surface coverage (relative to grids).
  • Sampling determined by planned trajectories — bias toward expected targets.
  • Requires careful pre-implantation planning.

Combined Subdural + Depth

  • Combine grids over cortical convexity with depth electrodes into mesial structures.
  • Maximizes coverage when both surface and deep regions are at risk.
  • More complex pre-operative planning.

Recording Principles

Bandwidth and Filtering

  • Wider bandwidth than scalp EEG: 0.1 Hz to 500+ Hz typical.
  • Captures high-frequency oscillations (HFOs).
  • Notch filter for line noise (60 or 50 Hz).

High-Frequency Oscillations (HFOs)

  • Ripples: 80–250 Hz oscillations.
  • Fast ripples: 250–500 Hz.
  • Pathological HFOs (often fast ripples) localize epileptogenic zone better than IEDs alone.
  • Emerging role in epilepsy surgery planning.

Reference Selection

  • Multiple options: averaged reference, common average, bipolar, Laplacian.
  • Different references reveal different aspects of activity.
  • Bipolar montages emphasize local activity.

Clinical Applications

Pre-Surgical Epilepsy Evaluation

  • Phase II of epilepsy surgery workup.
  • Indicated when:
    • Phase I (scalp EEG, MRI, etc.) is non-localizing.
    • Multiple potential ictal onset zones.
    • Non-lesional epilepsy.
    • Suspected mesial temporal vs lateral temporal origin.
    • Frontal lobe seizures (often invisible on scalp).
    • Insular epilepsy.
  • Goal: identify ictal onset zone for resection.

Functional Cortical Mapping

  • Cortical stimulation through grid electrodes:
    • Motor: stimulation produces movement; identifies primary motor cortex.
    • Sensory: stimulation produces tingling; identifies primary sensory cortex.
    • Language: stimulation during task disrupts speech, naming, comprehension.
  • Maps eloquent cortex to be spared during surgical resection.

Intraoperative ECoG

  • Records cortical activity during surgery.
  • Identifies residual epileptiform activity in resection cavity.
  • Can guide extent of resection.
  • Less commonly used now; emphasis on extra-operative mapping.

Pediatric Applications

  • Similar indications.
  • Special considerations for skull immaturity.
  • Increasingly used for complex pediatric epilepsies.

Pre-Implantation Planning

Pre-Surgical Data Review

  • Scalp EEG: identifies regions of IEDs and ictal onset.
  • MRI: identifies structural lesions.
  • PET: identifies regions of hypometabolism.
  • SPECT: identifies regions of ictal hyperperfusion.
  • Neuropsychological: lateralization.
  • MEG: source localization.
  • Semiology: lateralization clues.

Hypothesis-Driven Implantation

  • Implantation strategy based on hypotheses from non-invasive data.
  • Cover most likely seizure onset zones.
  • Include eloquent cortex if relevant.
  • Cover propagation regions for ictal network analysis.

Risks and Complications

  • Subdural grids: infection, CSF leak, hemorrhage (especially during placement and removal), seizures during/after placement.
  • Depth electrodes: hemorrhage along trajectory, infection, eloquent cortex injury.
  • Pneumocephalus.
  • Wound dehiscence.
  • Overall complication rates: 1–5% major, 5–10% minor.

Recording Duration

  • Typical: 5–14 days.
  • Goal: capture multiple habitual seizures (3+).
  • Anti-seizure medications tapered during admission.
  • Monitoring continued until enough seizures captured for localization.

Analyzing Intracranial EEG Data

Ictal Onset Identification

  • Identify the channel(s) where seizures begin earliest.
  • Often a rhythmic build-up of fast activity (low-voltage fast activity is classic pattern).
  • Sometimes preceded by sharp transients or focal slowing.
  • Spread pattern mapped over time.

Interictal Discharges

  • Spikes and sharp waves from epileptogenic zone.
  • Frequency, distribution, and morphology analyzed.
  • Higher frequency in epileptogenic vs spreading regions.

High-Frequency Oscillations

  • Ripples and fast ripples in seizure onset zone.
  • Predict epileptogenic zone better than IEDs alone.
  • Increasingly used in surgical planning.

Connectivity Analysis

  • Mapping which regions activate together.
  • Identifies functional networks in epilepsy.
  • Research and emerging clinical applications.

Decision Making After Intracranial Monitoring

  • Identify ictal onset zone and seizure propagation.
  • Map eloquent cortex (motor, language, sensory).
  • Determine surgical candidacy:
    • Resectable onset zone clear of eloquent cortex.
    • Single focal onset.
    • Patient acceptance of risks.
  • If resection not feasible:
    • Responsive neurostimulation (RNS): closed-loop implanted system.
    • Deep brain stimulation.
    • Continued medical management.
    • Vagus nerve stimulation.

Responsive Neurostimulation (RNS)

  • Implanted device that detects ictal onset and delivers stimulation to abort seizure.
  • For patients with focal seizures not amenable to resection.
  • Variable efficacy.
  • Long-term seizure reduction in selected patients.

Outcomes of Intracranial Monitoring

  • Successful localization in ~70% of cases.
  • Surgery offered to those with localized findings clear of eloquent cortex.
  • Seizure freedom rates after resection guided by intracranial monitoring: 50–70%.
  • Important for non-lesional cases where standard scalp EEG would not localize.

🔍 Did You Know?

The increasing use of stereo-EEG (sEEG) over subdural grids represents one of the most consequential shifts in epilepsy surgery over the past decade. Stereo-EEG, originally developed by Talairach and Bancaud in France in the 1960s but only recently adopted widely in North America, has substantial advantages over traditional subdural grids: less morbid (no craniotomy required), better 3D sampling of deep structures, ability to monitor multiple non-contiguous regions, and superior detection of seizures arising from deep mesial temporal, insular, and cingulate sources. Studies comparing sEEG to subdural grids have shown similar surgical outcomes with reduced complication rates. The shift is particularly important for non-lesional epilepsy and for cases where the seizure focus is suspected to lie in regions invisible to scalp EEG (insula, mesial frontal cortex, deep parietal). For practicing epileptologists and neurosurgeons, the implications: major epilepsy centers are increasingly transitioning to sEEG as primary intracranial monitoring approach, with subdural grids reserved for cases requiring extensive surface mapping. The same trajectory generalizes: less invasive yet more informative diagnostic techniques replace older approaches as the field matures. For patients facing intracranial monitoring for epilepsy surgery, the take-home is that modern sEEG offers a meaningful reduction in surgical risk while improving the precision of seizure localization — and ultimately better surgical outcomes. The integration of sEEG with stereotactic robot-assisted placement, advanced imaging registration, and computational analysis represents one of the most important areas of contemporary clinical neurophysiology.

Pitfalls and Pearls

  • Intracranial EEG: subdural grids and strips OR depth electrodes (sEEG).
  • Subdural grids: cortical surface; broad coverage; requires craniotomy.
  • Stereo-EEG (sEEG): 3D depth sampling; less invasive than craniotomy.
  • sEEG advantages: deep structures, multiple regions, lower morbidity.
  • Combined: grids + depth electrodes for some cases.
  • High-frequency oscillations (HFOs): ripples 80–250 Hz, fast ripples 250–500 Hz; localize epileptogenic zone.
  • Pre-implantation planning: hypothesis-driven based on scalp EEG, MRI, PET, SPECT, neuropsych, MEG.
  • Goal of monitoring: identify ictal onset zone for surgical resection.
  • Cortical stimulation mapping: motor, sensory, language; spare during resection.
  • Recording duration: 5–14 days; capture 3+ habitual seizures.
  • Risks: hemorrhage, infection, CSF leak; 1–5% major complications.
  • Ictal onset pattern: low-voltage fast activity classic.
  • Successful localization: ~70% of cases.
  • Post-resection seizure freedom: 50–70% in selected cases.
  • RNS: alternative for non-resectable focal epilepsies.
  • Modern shift to sEEG: superior for deep structures, less morbid.
  • Robot-assisted sEEG placement: improving precision.

References

  1. Engel J Jr. The current place of epilepsy surgery. Curr Opin Neurol. 2018;31(2):192-197.
  2. Iida K, Otsubo H. Stereoelectroencephalography: indication and efficacy. Neurol Med Chir (Tokyo). 2017;57(8):375-385.
  3. Cardinale F, Cossu M, Castana L, et al. Stereoelectroencephalography: surgical methodology, safety, and stereotactic application accuracy in 500 procedures. Neurosurgery. 2013;72(3):353-366.
  4. Jacobs J, Zijlmans M, Zelmann R, et al. High-frequency electroencephalographic oscillations correlate with outcome of epilepsy surgery. Ann Neurol. 2010;67(2):209-220.
  5. So NK, Lüders HO. Practical algorithm for surgical evaluation of patients with refractory focal epilepsy. Epilepsy Behav. 2002;3(5S):S2-S15.
  6. Kuzniecky R, Devinsky O. Surgery insight: surgical management of epilepsy. Nat Clin Pract Neurol. 2007;3(12):673-681.