METHOD & APPROACH

The procedure to be performed on each patient includes 3 steps:

Step 1. Patients will be first carefully selected based on presurgical evaluation protocol.
         12 patients will be explored during 3 years. Patients will get all the information regarding  the study objectives, method and risks, they will have the opportunity to ask questions and finally they will be included only after signing the informed consent. The patients reserve the right to withdraw their consent at any time.
         Patients history and clinical examination will be carefully taken into consideration. Clinical data, electroencephalographic studies and imagistic assessment (magnetic resonance imaging , metabolic studies as positron emission tomography) as well as neuropsychological evaluation will be analyzed.  All patients will perform surface video EEG monitoring with the placement of electrodes according to the 10-20 international system, including the low basal temporal line.
Eligible patients will be included in the study based on criteria that generally recommend invasive recordings during the presurgical work up:

• surface video-EEG monitoring revealed interictal abnormalities and ictal episodes concordant with focal epilepsy with temporal lobe origin but neuroimaging procedures (high resolution MRI) fail to get the evidence of a lesional abnormality or the lesion is not concordant with electroclinical  findings.
• patients that could not afford standard surgical procedures in the temporal lobe because of the risk for inacceptable functional deficit postsurgery and might benefit from tailored, more limited resections.

Will be excluded patients:

• suitable for standard resections or lesionectomies that could be safely performed without invasive recordings

Step 2. All eligible patients will be subjects of invasive monitoring using stereotactic implantation of intracranial macroelectrodes (SEEG method) as part of presurgical work up to further outline the epileptogenic area for the resection that will take place during the third phase of the procedure.
Strategy of intracranial implantation and sampling of temporal structures will be decided based on detailed analysis of clinical ictal semiology and scalp EEG distribution of recorded interictal and ictal discharges, as well as data obtained from imaging studies.
Intracranial electrodes that will target temporal structures will be implanted orthogonal to the midsagittal plane or oblique using the Leksell frame. The SEEG exploration will be highly individualized and a number ranging from 7-13  electrodes will be used to explore different temporal lobe structures in order to delineate the ictal onset zone, spreading patterns and establish the surgical resection limits and the functional cortex that should be spared. According to this strategy, apart from temporal lobe structures, perisylvian operculum, insular cortex, frontobasal structures and temporo-parieto-occipital junction will be constantly explored based on individual case electroclinical characteristics. (Kahane  P., 2001)
Commonly used macroelectrodes (DIXI Besancon, France) have a diameter of 0.8 mm 10 to 18 contacts of 2 mm length, 1.5 mm spacing between contacts. The implantation procedure will take into account the individual anatomical restraints and vascular network disposition as will be displayed by angiography. (Szickla 1977, Talairach, and Tournoux, 1988)
The anatomical location of electrode placement will be verified on CT scans coregistered with MRI image and  also based on the analysis of EEG signal initially recorded at each side.
 All the intracranial recordings will run in chronic conditions from 5 to 12 days aiming to obtain synchronized video-EEG’s for patient’s habitual ictal episodes. A systematic stimulation protocol will be applied on contiguous contacts of the intracranial electrodes to confirm the involvement of a particular temporal structure in seizures initiation and delineate the eloquent functional cortex that should be spared from surgical resection.
The morphology of the discharge obtained at the beginning of the ictal episodes and the anatomoelectroclinical correlations will help to clearly delineate the seizure onset zones (SOZ) for every particular case explored with implanted macroelectrodes according to the SEEG method.
Risk of intracerebral electrode implantation varies from 1-8% being is mostly related to hemorrhages . Saint Anne Hospital reports a complication rate of 1% (0,17% per implanted electrode) and the center from Grenoble 1,4% (Kahane P., 2004).

Step 3. The third step will comprise the surgical intervention that will be tailored based on data obtained during the previous phase. Surgery will be scheduled at 4-6 weeks distance from the invasive exploration to allow alleviation of inflammation and edema that might have happen in the cerebral structures due to the electrode implantation procedure. The surgical procedure will be performed  in two stages:

Stage 1. Patients in the interictal period prepared for surgery will be initially explored using microelectrodes.
The use of the Leksell frame as opposed to Talairach system offers the ability to refine trajectories such that major blood vessels as evidenced on the contrast-enhanced MRI, as well and other anatomical structures are avoided.
The microelectrode implantation will particularly target the previous macroelectrode location that has been identified inside the SOZ. For the electrode insertion and single-unit recordings we will use clinical equipment commonly used for electrophysiological mapping for Deep Brain Stimulation (DBS) targeting.  A Leksell frame (Elekta Instruments AB, Stockholm, Sweden) will be mounted to the patient head followed by a CT scan that is co-registered with the initial MRI using FrameLink software (Medtronic Inc., Minneapolis, Mn). FrameLink software will also be used for planning the electrode trajectory to the SOZ. Special care will be taken to avoid blood vessels, as highlighted on the MRI based on contrast agents. Once the trajectory established and the stereotactic frame adjusted, a small initial craniotomy (15 mm diameter) will be performed, allowing for the insertion of electrodes while minimizing the air intake resulting in brain shift affecting the accuracy of the targeting. The set of electrodes (typically 3 electrodes) will be advanced in the tissue using a standard motorized microdrive used for stereotactic and functional neurosurgeries (mT Drive, FHC Inc, Bowdoin, ME). An electrophysiological workstation (Guideline LP+ FHC Inc, Bowdoin, ME) will apply electrical microstimulation and record data from the microelectrode. Typically, three electrodes in a linear arrangement will be inserted, out of the 5 possible locations in a so called Ben-gun configuration (Benabid et al, 1991). This will allow bipolar stimulation in between the distal electrodes while recording through the center electrodes.  Electrodes will be slowly advanced to detect single-units. Once a unit is isolated, the baseline activity will be recorded for 60 seconds, followed by constant current electrical stimulation at 10, 30, 60 and 130 Hz, typically. The electrical stimulation will be applied for 30 seconds, with 30 seconds inter-stimulation intervals. Recording will continue for another 60 seconds to evaluate the post-stimulation activity.  Different pulse waveforms (McIntyre and Grill, 2002), polarities and phases will be explored for maximal stimulation outcome, as the Guideline LP+ system allows delivering complex stimuli, including arbitrary waveforms. The pulse duration will typically be in the 75 microseconds range for the stimulation through the macro contacts of the electrodes and 300 microseconds range for stimulation through the micro contacts.  The amplitude of the electrical stimulation will be gradually increased until the threshold for eliciting spikes is reached, at any frequency.  An amplitude typically 50% larger than the threshold will be used for subsequent stimulation epochs. Effects of sub-threshold stimulation levels (typically 50%) will be used for testing a possible depolarization blockade mechanism, at high frequencies (60, 130 Hz).  The objective of this step is to test the excitability of the single neurons in the SOZ and compare it with the excitability of the neurons outside SOZ.

Stage 2. After removing the acute electrodes and the stereotactic frame, a larger craniotomy will be performed, allowing for the resection of the tissue in the epileptogenic area. The limits of resection will be planned and tailored according to the data obtained from invasive recordings to completely remove the epileptogenic zone and avoid functional eloquent cortex.

 

Data analysis. Data analysis will be performed using MATLAB (MathWorks, Natick, MA, USA).  To overcome the presence of large amplitude stimulus artifacts and recover the neural signal during stimulation epochs, we will use the stimulus artifact removal algorithm we proved to be effective in STN (Toleikis et al, 2007, Verhagen et al, 2011). The method is based on building a stimulus artifact template (Hashimoto et al, 2002) and subtracting it from the raw signal to recover the neural activity, as illustrated in Figure 2.  Single unit data acquired before, during, and after electrical stimulation at each stimulation frequency will be sorted using an autoclustering method as implemented in the KlustaKwik component of the MClust Matlab toolbox (http://www.cbc.umn.edu/~redish/mclust, David Redish). Results of the automatic clustering will be manually reviewed and edited, as needed, using MClust toolbox and/or Plexon Offline Spike Sorter (Plexon Inc, Dallas, TX, USA). Raster, time histograms and mean firing rates of the activity during the stimulation epochs will be compared with the results during the non-stimulation epochs. Also, characteristics of the neurons outside SOZ will be compared against those of the neurons in the SOZ.