Method of Direct Cortical Stimulation for Triggering Seizures

Stereoelectroencephalography (SEEG)-based presurgical evaluation of epilepsies has a double objective: to localize the epileptogenic zone (EZ) and to specify a surgical plan. Direct electrical stimulations from the intracerebral electrodes are an essential technique to reach these goals. As such, they are an integral part of SEEG.

Stimulations have been performed to trigger seizures ever since SEEG has been practiced. In the beginning, the presurgical “stereotaxic functional investigation” subscribed to the Penfield principles. The earliest sign or symptom of the patient’s seizures induced by stimulation of a circumscribed area constituted a solid argument for localization. Also, until the mid 70ies SEEG was a 1-day investigation. If the patient did not present any spontaneous seizure during the time allotted for recording, it became necessary to provoke a seizure by stimulation or chemical activation. Fifty years later, electrical stimulation still represents an indispensable tool to define the EZ. ^,

Electrical simulation (ES) has been used in ECoG to identify “eloquent” regions related to motor and cognitive functions. This procedure has been called “functional mapping”. There is a growing body of evidence considering the usefulness of ES to map language cortex in patients undergoing resection of epileptogenic cortex in the language dominant hemisphere. Electrical simulation producing a transient and functional impairment might predict which functions will be disturbed if the stimulated cortex were to be removed. However, the dispersed patterns of SEEG implantations hamper the possibility of a real mapping, so that the most reliable functional evaluation comes from a comparative behavioral analysis of spontaneous and electrically induced seizures. This is one of the major differences between subdural grids and intracerebral depth electrodes (SEEG) methods.

Rationale

The concept of EZ in SEEG is quite different from that of seizure onset zone used in subdural grid recordings. Without entering into detail, one of its main ideas is that the EZ is defined by its specific capability to synchronize the epileptogenic network. Therefore, stimulation of any of its component areas can trigger the patient’s seizure. It works as a validation of a method which, by nature, must tolerate a sampling bias.

Another difficulty for EZ localization through spontaneous seizures only is because of neocortical seizure patterns . Seizure onset in the neocortex can occur like quasi-simultaneous high frequency activities throughout a vast territory, possibly more extensive than the EZ. Whatever the signal processing applied, a cut-off frequency for epileptogenicity is impossible to define. Only stimulation will help discriminating between the involved areas and determine which ones are essential to activate the epileptogenic network according to its proper spatial-temporal pattern and to trigger seizures ( Fig. 9.1 ).

Stimulation is the only way to decipher seizure organization in infra- and peri-sylvian epilepsies . Temporal or occipital epilepsies may present with different seizure subtypes depending on whether one or two (or more) elements of the ictal architecture are activated. For instance, in a patient one seizure can be only temporal, another one temporal/peri-sylvian, and another one with a peri-sylvian seizure onset remotely from the propagation network. A similar organization can be observed in occipital/temporal or occipital/temporal/peri-sylvian epilepsies. To determine what is the primary epileptogenic organization and discriminate it from spread, stimulation will unravel how its constitutive elements are linked and how each of them contributes to early and late clinical semiology. Contrary to the stimulation techniques used for functional mapping with subdural grids, after-discharges are utilized to differentiate nonepileptogenic from epileptogenic areas. Electrical and clinical criteria are analyzed to interpret the effect of the stimulation trials (see below).

Techniques

The optimal time to perform stimulations depends on multiple individual factors during SEEG. Generally, stimulation sessions occur after some spontaneous seizures have been recorded. However, they may precede spontaneous seizure recordings in cases where the patient did not have any during the first week. Stimulations are performed in several sessions, each of them rarely exceeding 1 hour.

Stimulation is applied between two adjacent leads of one electrode. Single bipolar pulse or train stimuli are used. Single pulses (0.3–0.5 ms; 0.5–5 mA) are delivered pseudo-randomly or periodically (for instance 1/s) to stimulate the motor cortex, the hippocampus, or Heschl’s gyrus because of their low threshold for after-discharges or seizures. The same is applicable to dysplastic cortex, especially when interictal activity is characterized by repetitive spiking intermingled or not with bursts of fast activities. ^, All other areas are stimulated by single pulses or by train (50 Hz; 0.5 ms; 0.5–4 mA; 1–5 seconds). Train stimulation (50 Hz) can also be used for motor cortex, hippocampus, and Heschl’s gyrus once the risk of triggering generalized seizure (motor cortex, Heschl’s gyrus) has been estimated or when single pulse stimulations were ineffective (hippocampus). Comparing the chance of obtaining after-discharges with stimulations of different frequencies and intensities, Motamedi et al. (2007) showed that higher frequency (100 vs. 50 Hz) and larger pulse width (1 vs. 0.2 ms) were more effective. This is in good agreement with the fact that single pulses up to 3 ms duration were successfully used to obtain electroclinical responses in nonprimary motor areas, especially Supplementary Motor Area. ^,

There is no need to apply increasing intensities to measure a clinical or electrical threshold, the main goal being to obtain after-discharges and/or seizures. Repetition of stimulation at the same site exhausts the area stimulated leading to a refractory period of variable duration. This is in contrast with what has been reported with subdural grids, where after-discharges are “likely to occur when an electrode pair showed after-discharge (ADs) and was stimulated again, especially after short intertrial intervals or for longer duration.”

Obtaining an after-discharge is the test for cortical excitability. Pulse width and train frequency generally are kept constant throughout the whole stimulation procedure. The initial intensity may vary according to several factors : level of current anti-seizure medication (ASM), history of generalization (particularly when stimulating lateral premotor or precentral cortex), area stimulated supposed to be part of the EZ (as indicated by already recorded spontaneous seizures), type of structure stimulated (see above). Triggering of after-discharges needs to be prevented only during electrocorticography (ECoG) functional mapping. Blum et al., studying properties of after-discharges from cortical electrical stimulation in focal epilepsies, stated that when ADs involved more than the stimulus site, they might inaccurately localize cortical function. Interestingly, in the same work, no consistent relationship was noted between the site of stimulation elicited AD and that of spontaneous seizures. This observation made during ECoG paradoxically serves as a rationale for triggering seizures in SEEG.

In fact, the main objective of stimulations in SEEG is to find the site(s) from which stimulation is capable of synchronizing the epileptogenic network (or part of it) to induce an electro-clinical seizure. Consequently, the site(s) from which a local after-discharge without any electrical or ictal clinical features can be triggered is (are) considered as being outside the EZ. The EZ being structured as a neural network, the site location from which a seizure can be triggered does not define by itself the localization of a “focus” at this site. The way synchronization is produced between the areas involved through stimulating this site critically informs about the organization of the EZ. This is the reason why analysis of stimulation effects does not simply consist in identifying the areas from where a seizure could be triggered and concluding that EZ lies there. Interpretation is less straightforward. Its local and remote electrical effects must be carefully analyzed before reaching any conclusion. Puzzling situations are encountered when an ictal discharge is generated from the efferent network of the stimulated cortex. According to the afferent-tract targeting principle (further detailed below), ^, the optimal electrode placement is upstream to the target structure.

Frontal and parietal (supra-sylvian) epilepsies on the one hand, temporal and occipital (infra-sylvian) epilepsies on the other hand have different EZ organizations. It turns out that results of stimulation differ from one to the other. In supra-sylvian epilepsies, except for early symptoms (“auras”) that can occur during stimulation time even without after-discharge, triggering a seizure is an all-or-none phenomenon. In the infra-sylvian epilepsies, parts of the typical seizure can be evoked from different sites of the network. The high excitability and the wide cortical efferent connectivity of the limbic system (especially hippocampus) adds another factor of complexity in interpreting results of stimulation. Chauvel et al. (1993) studied the conditions of obtaining seizures similar to the patients’ spontaneous ones. In medial temporal epilepsies, hippocampal onset seizures could be triggered by hippocampal or by amygdala stimulation as well, amygdala onset seizures by amygdala or hippocampus stimulation, and rarely by temporal neocortical stimulation. This should be paralleled with the fact that in lateral temporal epilepsies, except for superior temporal gyrus epilepsies, seizures are preferentially triggered from medial limbic regions. Therefore, any directly connected area can elicit a delayed after-discharge or ictal discharge in the hippocampus; conversely, the hippocampus can trigger a seizure in any efferent epileptogenic area. Hence, interpretation of stimulation in hippocampus-related areas must consider the network organization of the triggered discharge pattern, especially its frequency characteristics (fast activities) and its electrical and clinical similarity with spontaneous seizures.

According to the afferent-tract targeting principle, any electrical stimulation in the brain is more efficient for driving a downstream structure discharge than for eliciting a local excitatory effect. This fact explains a classical pitfall in stimulation interpretation in SEEG. Triggering a seizure from a given electrode does not necessarily mean that this electrode is in the epileptogenic zone. A usual observation is that the best way to trigger a seizure from the hippocampus is to stimulate amygdala and vice-versa (see Fig. 9.2 ). This is also true in the cortical motor system where a stimulation in a nonepileptogenic cortical motor area can remotely trigger a seizure from epileptogenic premotor and/or precentral cortex. The stimulation procedure in SEEG is not a binary approach. Interpretation of its effects requires a meticulous analysis of the frequency patterns and their spatial distribution in correlation with any induced behavior. Much emphasis has been put on the “reproducibility” of the spontaneous seizure to corroborate the EZ localization. Some of the earliest studies gave an estimation of the “concordance” between stimulation-induced and spontaneous seizures. ^, A total of 77%–90% were concordant in temporal epilepsies, and 86% in frontal epilepsies ^, (note that Chauvel et al. selected patients only with concordant seizures). Bernier et al. confirmed that obtaining an after-discharge is not a tool for localizing the EZ.