Imaging the seizure onset zone with stereo-electroencephalography

Abstract

Stereo-electroencephalography is used to localize the seizure onset zone and connected neuronal networks in surgical candidates suffering from intractable focal epilepsy. The concept of an epileptogenicity index has been proposed recently to represent the likelihood of various regions being part of the seizure onset zone. It quantifies low-voltage fast activity, the electrophysiological signature of seizure onset usually assessed visually by neurologists. Here, we revisit epileptogenicity in light of neuroimaging tools such as those provided in statistical parametric mapping software. Our goal is to propose a robust approach, allowing easy exploration of patients' brains in time and space. The procedure is based upon statistical parametric mapping, which is an established framework for comparing multi-dimensional image data that allows one to correct for inherent multiple comparisons. Statistics can also be performed at the group level, between seizures in the same patient or between patients suffering from the same type of epilepsy using normalization of brains to a common anatomic atlas. Results are obtained from three case studies (insular reflex epilepsy, cryptogenic frontal epilepsy and lesional occipital epilepsy) where tailored resection was performed, and from a group of 10 patients suffering from mesial temporal lobe epilepsy. They illustrate the basics of the technique and demonstrate its very good reproducibility and specificity. Most importantly, the proposed approach to the quantification of the seizure onset zone allows one to summarize complex signals in terms of a time-series of statistical parametric maps that can support clinical decisions. Quantitative neuroimaging of stereo-electroencephalographic features of seizures might thus help to provide better pre-surgical assessment of patients undergoing resective surgery.

What do patients with epilepsy tell us about language dynamics? A review of fMRI studies

“…Unlike research in stroke patients, in patients with epilepsy, it is possible to assess the multiple mechanisms of reorganization at different times, including 'chronic' reorganization induced by the epilepsy, 'acute' reorganization induced by surgery, and 'subacute' and 'secondary chronic' reorganization several months or years after surgery. This multimodal approach to evaluate the representation of language is standard for the assessment of language plasticity in epilepsy in our clinical environment (Baciu et al, 2001;Lachaux et al, 2007aLachaux et al, , 2012Jerbi et al, 2010;David et al, 2011David et al, , 2013Vidal et al, 2011;Perrone-Bertolotti et al, 2014). In practice, patients undergo extensive behavioral, neuropsychological, functional, and structural neuroimaging examination.…”

Mapping human brain networks with cortico-cortical evoked potentials

“…The grid and strip approach involves a craniotomy and durotomy followed by the placement of two-dimensional strips or sheets (grids) of electrodes (typically 3 mm diameter, 1 cm inter-electrode spacing), where neural activity can be recorded from the surface of the cortex (termed electrocorticography, or ECoG [53][54][55][56][57]). On the other hand, stereoelectroencephalography (SEEG) involves the placement of multi-contact electrode leads penetrating the brain [58][59][60]. While in this review we focus on the grids and strips approach, the principles discussed may be applied to both techniques.…”

Detection of recurrent activation patterns across focal seizures: Application to seizure onset zone identification

“…A great number of them have analyzed spectral features of SEEG signals (Bartolomei et al, 2008;David et al, 2011;Gnatkovsky et al, 2011;Gnatkovsky et al, 2014;Andrzejak et al, 2015), which mimic the patterns that are visually identified as epileptogenic during diagnostic EEG inspection (Gnatkovsky et al, 2014;Lagarde et al, 2016). In these works, epileptogenic biomarkers are typically built around two variables, the activation of signal's spectral properties (e.g., the relative amount of signal power in a frequency band (Bartolomei et al, 2008)) and the onset time of this activation, which measure respectively the amount of ictal activity and the degree of early participation of such structures during seizures.…”