Added clinical value of ESI

in MRI-negative epilepsy

Cliniques Universitaires Saint-Luc (Brussels, Belgium)

Medical history and disease course

16-year-old, right-handed, female patient with unrevealing personal and familial past history. Epilepsy started at the age of 11 years with a stormy onset of frequent and brief focal aware and unaware seizures with pleiotropic but stereotyped motor manifestations: asymmetric tonic posturing (extension of the right upper limb and left head deviation) and hyperkinetic non-lateralized movements with gelastic features. Seizures were mainly nocturnal leading to frequent awakenings but could occur also during the day. Rarely, the patient experienced focal to bilateral tonic-clonic seizures. Several anti-seizure medications (ASMs) were tried (VPA, OXC, LEV, TPM, PER, LCM, BRV, CLB) with little improvement in seizure frequency and important side effects. At the time of presurgical evaluation, the patient reported on average 4 days of seizures/month while taking 5 different ASMs.

Preoperative workup

Long-term video-EEG Monitoring (LTEM) showed interictal spikes and sharp waves on the left mesial fronto-central regions. Eighteen seizures were recorded with ictal onset in the same regions. A high-density EEG (HD-EEG) recording to exclude false lateralization on low-density EEG (LD-EEG) completed the non-invasive EEG monitoring. 3T-brain MRIs were repeatedly negative and voxel-based morphometry (VBM) analysis was inconclusive. 18F-fluorodeoxyglucose PET (FDG-PET) revealed a discrete hypometabolic cortical area on the left fronto-parietal region. The neuropsychological evaluation did not observe any lateralizing cognitive deficit, but the patient had a global slowing of mental performance in relation to ongoing seizures and current poly-therapy.

Interictal Electrical Source Imaging (ESI) was realized on both LD- and HD-EEG recordings. Successively, the case was discussed in an epilepsy surgery meeting in a two-step approach.

    1.      The multidisciplinary team (MDT) formulated, blinded to ESI results, the following hypotheses:

        • left frontal superior gyrus (first hypothesis)
        • left dorsolateral frontal gyrus, and left anterior cingulate cortex (second hypotheses).
        • An stereo-electroencephalography (SEEG) plan was defined accordingly using eight depth electrodes.

    2.      In a second step, ESI results were presented at the same meeting and the MDT was asked to interpret them. Both LD and HD-ESI showed a principal cluster in the left middle/posterior cingulate cortex. Results are shown in Fig. 1. The spike cluster was considered genuine and the corresponding epileptogenic zone (EZ) location at the middle/posterior part of cingulum highlighted by ESI was clinically plausible, taking into account the whole presurgical data. Therefore, the MDT retained an additional hypothesis about EZ location, and the SEEG plan was changed accordingly. Practically, one depth electrode was added for better coverage of the middle and posterior cingulate cortex intending to delimitate the posterior margin of the presumed EZ.

Fig. 1: Results of automated ESI of the first (in red) cluster, at the peak of the averaged waveform (maximum localized at the crosshair), and of the individual spikes in the cluster. Note that both the maximum of the averaged waveform, and the majority of the individual spikes (represented as red circles in the lower image “ESL of singles”) indicate a source in the left middle/posterior cingulate cortex, which is in concordance with the SEEG-seizure onset zone (SOZ) and the resection site. ESI results of the other clusters and at half-rise time are similar and are not shown there.


Based on this information, an invasive EEG-monitoring session was planned. The modified SEEG plan is shown in the Fig. 2

Fig. 2: Modified SEEG plan suggested by ESI results, including an additional depth electrode (CMP), to cover the middle/posterior cingulate cortex (yellow circle). CingA: anterior cingular cortex; CMA: anterior/middle cingular cortex; CMP: middle/posterior cingular cortex; CingP: posterior cingular cortex; PreF: prefrontal; AMS: supplementary motor cortex; LP: paracentral lobule; PrCus: precuneus (superior part); PrCui: precuneus (inferior part).

Invasive EEG monitoring confirmed the EZ location suggested by ESI with concordant sublobar result in the left middle/- posterior cingulate cortex. An example of a seizure recorded during SEEG is shown in Fig. 3.

Fig. 3: SEEG recording showing a seizure with clear onset of ictal discharges in the CMP electrode (violet arrow), followed by rapid and multidirectional propagation to AMS (dark yellow arrow), Precui (red arrow) and CMA (green arrow).

A revision of the brain MRI was performed focusing on the brain area found to be responsible for generating the ictal activity (SOZ) and a left middle/posterior cingular gyrus dysplasia was finally suspected.

Resective epilepsy surgery

Based on the findings during the SEEG session, the patient underwent surgery with resection of the posterior cingular gyrus and part of the middle cingular gyrus. The margins of resection are depicted in Fig. 4. Histopathology revealed a focal cortical dysplasia type IA (FCD IA). There were no perioperative complications and the patient is still seizure-free at the 24-month follow-up after surgery (ILAE class 1).

Fig. 4: Postoperative MRI showing a limited posterior cingulate cortectomy and sparing of the front half. Margins of resection are bordered in red. ESI cluster is included in the resection zone. The patient is seizure-free (ILAE class I) at 2-year postoperative follow-up.


The current case nicely illustrates the added value of using electrical source imaging in the preoperative work-up. MRI-negative refractory epilepsies are the most difficult surgical candidates. Although the tests performed during pre-surgical work-up guided the MDT to establish several hypothesis being blinded to ESI as an initial step, additional information came from electrical source imaging allowing targeted implantation of intracranial depth electrodes for invasive EEG-monitoring and later respective surgery. Findings from the intracranial EEG were concordant with the results of non-invasive electric source localization. The patient was seizure-free (ILAE class I) at 2-year postoperative follow-up, providing additional evidence that the epileptogenic zone was indeed resected.


We would like to thank Dr. Roberto Santalucia from Cliniques Universitaires Saint-Luc (Brussels, Belgium) for providing the clinical and imaging data of this case.

Read more about the full publication in this link.