Invasive recordings

Comparison between cortical electrode and depth electrode

Cortical electrode	Depth electrode
More risky	Less risky
Likely to offer excision surgery upon removal of cortical electrode	Can remove the depth electrode and then remove it

Monitor changes in epileptiform activity during the surgery
Preoperative mapping of eloquent areas to avoid post-operative neurological deficits

Single pulse electrical stimulation (SPES) can be used as an active technique to demonstrate areas of cortical excitability
Use of such techniques is used to guide surgery in complex epilepsy cases.

Inform the tailoring of dominant temporal lobectomies to spare lateral cortical regions exhibiting speech arrest with stimulation (reducing postoperative language deficits while maximizing the extent of lateral temporal resection).

Unilateral craniotomy for placement of subdural electrode grids
Craniotomy for placement of subdural electrode grids and strips is frequently employed to guide tailored extratemporal cortical resections.
Craniotomy and grid placement can be combined with

Invasive (vs scalp electrode)
Risk of sampling error as a result of the limited spatial distribution selected for monitoring.
Bilateral temporal and frontal implantation of subdural electrode strips through enlarged bur holes, though less precise, are associated with a lower risk of intra-cerebral hemorrhage. (vs depth electrode)
Risks of craniotomy, mass effect from the grids and a higher infection risk,

Aka: SEEG (Stereoelectroencephalography)
In some centers, simultaneous EEG–fMRI are used to inform the placement of electrodes

Patients considering epilepsy surgery due to medication refractory seizures, + noninvasive evaluation has not led to an adequate localization of the epileptogenic zone, yet provides sufficient evidence to produce a plausible hypothesis for the epileptogenic zone’s location eg:0

To try to localize the area of seizure onset in individuals in whom a confident solution cannot be arrived at with noninvasive data, the functional data derived from [18F]FDG PET, ictal SPECT, MEG,

The placement of intracranial EEG electrodes should not be a blind fishing expedition; there must be a testable hypothesis, and the individual must understand the risk/benefit ratio of invasive EEG and the possibility that, after an invasive recording, resection will not be recommended

Provides unparalleled sampling from deep cortical anatomy not directly accessible by other means,
Allowing for the definitive lateralization and localization of mesial temporal lobe, insular region, mesial frontoparietal and pericingulate, orbitofrontal, and submerged perisulcal cortical onset epilepsy.

Depth Electrodes and Stimulation
Last investigation when concordance is not achieved from imaging
Correlation Neurosurgery, Neurophysiology, Neuroradiology (fMRI) as to position
Correlation Neurophysiology e.g. Wernicke's , Heschel 's
Post placement imaging to check position & to assess for complications
MR safety! Involvement of Physics. Ensure electrodes are (conditionally) MRI safe
Team work to ensure correct identification of position and labelling of electrodes

Where there is suspicion of mesial temporal lobe epilepsy but surface EEG is inconclusive,

Sphenoidal electrodes can be placed percutaneously under fluoroscopic guidance through the mandibular notch under the zygoma to sit anterior and lateral to the foramen ovale to record over the parahippocampal gyrus.
Risks include

Injury to middle meningeal artery entering the nearby foramen spinosum,
Injury to the contents of the foramen ovale (otic ganglion, CN V3, accessory meningeal artery, lesser petrosal nerve [CN IX], emissary veins).

The most common surface EEG pattern for temporal lobe seizures is rhythmic theta or delta activity that is maximal at F7/F8 or the sphenoidal electrodes and evolves over the temporal derivations with spread centrally and contralaterally (~ 80% of temporal seizures).
Epileptiform discharges are seen bilaterally in about 40% of cases during long-term scalp EEG recordings. Surface EEG recordings may also suggest independent seizure onset from both temporal lobes. Despite this, intracranial recordings may go on to find seizures arising only from one side.
When bitemporal abnormalities are observed during presurgical evaluation, potential surgical candidates are identified if there is a unilateral lesion on MRI, predominance of interictal discharges from one temporal lobe, and strongly lateralizing neuropsychological testing.