Understanding the epileptogenicity of lesions: A correlation between intracranial EEG and EEG/fMRI

Intracranial electroencephalography (EEG) remains the “gold standard” for the definition of seizure origin, the so-called seizure onset zone (SOZ). It measures directly neuronal activity with an excellent temporal resolution, but with a spatial resolution that depends on a priori clinical hypothesis. The method is very specific but invasive and hence can be used in a limited number of patients with severe intractable focal epilepsies. EEG–functional MRI (EEG/fMRI) is another method to evaluate epileptic activity (EA) generation. It measures the blood oxygen level-dependent (or BOLD) changes related to an epileptic event, reflecting indirectly the neuronal activity through metabolic and hemodynamic modifications. EEG/fMRI measures the EA with a good temporal and spatial resolution providing that the electrophysiologic signal is detectable, but EEG/fMRI is noninvasive and easily repeatable. Nodular and laminar gray matter heterotopia and focal cortical dysplasia (FCD) are three relatively frequent forms of malformations of cortical development that often cause pharmacoresistant focal epilepsy. We are still debating on the mechanisms that underline the generation of EA in these lesions. Periventricular nodular heterotopia (PNH) consists of normal-appearing neurons in an abnormal position with an abnormal synaptic organization. They can, themselves, generate normal and epileptic activity, but the overlying or distant cortex may also be reorganized and epileptic. Patients with PNH form a heterogeneous group, and seizures or spikes seem to result from complex interactions between heterotopia and allo- or neocortex. iEEG has been pivotal in showing that gray matter heterotopia can generate electrical activity. Human studies using iEEG recordings showed that heterotopia can generate normal EEG activity, spikes, and seizures (Kothare et al., 1998; Aghakhani et al., 2005; Tassi et al., 2005), usually synchronous with, but sometimes independent from, the surrounding cortex. These studies have demonstrated connectivity between nodules and the overlying cortical regions, a concept that has influenced the significant change in the management of the epilepsy in patients with PNH. Most of the studies, however, had minimized the role of the heterotopia in seizure generation; interictal epileptic activity can be generated by heterotopia, but seizures are more likely to be generated by the overlying cortex. Results obtained from EEG/fMRI studies appear to be generally in good agreement, with iEEG providing information on the generation EA in patients with PNH that overlaps (Kobayashi et al., 2006; Tyvaert et al., 2008). EEG/fMRI showed that BOLD changes inconsistently involved the lesion during interictal EA. During seizures, however, the maximum BOLD responses were always in the overlying cortex and not in the heterotopia. Interictal EA can be generated by nodular heterotopia, but seizures are more likely to be generated by the overlying cortex (Fig. 1). The connections between nodules and with adjacent or distant structures may play a role in amplification and synchronization of EA and explain the widespread epileptogenic networks often found in these patients. Complex periventricular and subcortical nodular heterotopia. (A) Anatomic magnetic resonance imaging (MRI) showed periventricular gray matter heterotopia combined with subcortical nodules and abnormal gyration visible in the right posterior quadrant (white arrowheads). (B) Maximal activation in the right posterior quadrant in the lesion, and in the cerebellum during T4-T10-P10 polyspikes. Deactivation (blue) was predominant in the left hemisphere. (C) Activation in the lesion and deactivation in the pre-cuneus and areas surrounding the lesion during T6-P10 slow waves bursts. (D) Activation involved the entire lesion and spread to the thalami (deactivation was observed in the pre-cuneus and bifrontal areas) during a seizure starting with a rhythmic spike and slow wave discharge over the right posterior quadrant, followed by a short attenuation and bilateral spiking activity with a clear right predominance (persistence of slow waves on the right side). The double cortex, or gray matter laminar heterotopia, is usually a diffuse cortical malformation consisting of a more or less normal cortex and of a subcortical heterotopic band of gray matter. The band itself contains clustered small or medium-sized rounded or pyramidal neurons, haphazardly oriented and separated by bundles of fibers. Almost every patient with SBH has epilepsy, and the epilepsy is often severe, focal, or generalized. The role of the heterotopic cortex in generating EA is again not well understood. The distinction is especially difficult because the normo- and heterotopic cortices are strongly interconnected, and the activity of the heterotopic cortex is highly correlated with that of the normal cortex. Animal models suggest that the normocortex is by itself sufficient to generate seizures (Chen et al., 2000). Human studies however, using iEEG recordings, showed variable results: no EA recorded from the bands (Bernasconi et al., 2001); propagation of EA to the bands (Russo et al., 2003); or independent EA in normo and heterotopic cortices (Bernasconi et al., 2001; Mai et al., 2003). EEG/fMRI human studies have shown hemodynamic changes in both the normo- and heterotopic cortices during ictal and interictal EA with the maximum BOLD response in the heterotopic band (Fig. 2), indicating that the EA can be generated in the malformation itself (Kobayashi et al., 2006; Tyvaert et al., 2008). In humans, the iEEG and fMRI data are also suggesting that more than one focus may exist in the same patient and that the generator involves an extended portion of both the normo- and heterotopic cortices. Subcortical band heterotopia. (A) Anatomic magnetic resonance imaging (MRI) showed a bilateral posterior band heterotopia (head of white arrows). (B) Activation in the left band heterotopia and overlying cortex during T5-P9-O1 polyspikes (deactivation, blue, posteriorly in temporooccipital junction). (C) Activation of the right band heterotopia during T6-P10-O2 polyspikes (deactivation, blue, in both occipital lobes). (D) Maximum activation during seizures seen in right and left band heterotopia and overlying cortex (deactivation in blue). Ictal electroencephalography (EEG) onset showed right posterior temporal fast activity followed by left posterior temporal fast activity. Both during interictal and ictal events, heterotopic and normotopic cortices are involved by BOLD changes. These BOLD changes are predominant in the heterotopic cortex. FCDs consist of abnormal-appearing neurons, glial cells, and lamination with an abnormal synaptic organization. The malformation is usually focal, affecting only a small portion of the cerebral cortex. The epilepsy is, however, generally severe and focal, with a great impact on the patients. The FCDs have “intrinsic epileptogenicity” and generate continuous or bursts of epileptic discharges with overlapping interictal and ictal EA and lesion. Intracerebral EEG recordings have clearly shown that the typical rhythmic spiking anomaly (Palmini et al., 1995; Chassoux et al., 2000) is generated in the lesion itself, and that the seizures are also generated in the same dysplastic tissue with again overlapping epileptogenic and irritative areas. In the few patients with FCDs studied so far, BOLD changes overlapped in the lesion during interictal and ictal EA (Fig. 3), supporting common neuronal generators (Federico et al., 2005; Tyvaert et al., 2008). In addition, EEG/fMRI showed distant metabolic and hemodynamic changes, implying a larger common epileptic network, which could explain the important impact of this apparently confined lesion on brain functioning. Focal cortical dysplasia. (A) Anatomic magnetic resonance imaging (MRI) showed a right dysplastic sulcus in the anterior frontal lobe (black arrows). (B) Activation in the dysplasia during interictal right frontal spike and waves (F4-C4). (C) Blood oxygenation level-dependent (BOLD) signal changes observed during clinical seizures with prolonged rhythmic discharge of right frontal spike and waves and synchronous contralateral activity. Maximum activation is again seen in the lesion with propagation over the right frontopolar area, thalami, and the midbrain (left frontopolar area and left anterior cingulate area were also involved but not shown on this picture). With the limited experience we have so far, we found that iEEG and EEG/fMRI are generally in good agreement providing information on EA that overlaps. Both methods can evaluate the epileptic network and neuronal connectivity involved during epileptic discharges. BOLD changes, however, are frequently observed at very distant areas, suggesting that EEG-fMRI can provide additional information about the generators and structures during EA, and maybe about their impact on brain function. Neither author has any conflicts of interest to declare.