What Is Seizure Onset? Interictal, Preictal/Ictal Patterns, and the Epileptogenic Zone

Introduction

The epileptogenic zone (EZ) concept arises from the stereoelectroencephalography (SEEG) method introduced by Talairach and Bancaud. ^, The term “EZ” was proposed by Bancaud and defined as “the regions of the brain involved in the primary organization of the ictal discharge”, which refers to the cortical areas, not necessarily anatomically contiguous, bound together through an excessive synchronization at seizure onset and “early spread”. Thus, the concept of “network” in EZ has been introduced since 1965, recognizing both spatial and temporal aspects of seizure dynamics. The EZ definition mentioned above is practical yet challenging as the term “primary organization” seems arbitrary, leading to difficulty in accurately delineating the EZ’s boundaries. With the spread of the SEEG methodology to North America, where the subdural grid (SDG) intracranial (IcEEG) recording dominated, an additional definition of the EZ was introduced. The EZ was defined as “the area of cortex indispensable for generating seizures” and “total resection or disconnection is necessary and sufficient for seizure freedom”, implying that the concept of EZ is a theoretical one. ^, Thus, the seizure onset zone (SOZ) was introduced as the “area of cortex from which clinical seizures are generated” as a “practical” counterpart of the EZ. This new definition doesn’t eliminate the challenge of accurately delineating the EZ’s boundaries in SEEG. In fact, the interchangeable usage of the terms “EZ” and “SOZ” makes the fellow reader more muddled as the SOZ in SDG doesn’t mean the same as the SOZ in the SEEG, as SDG and SEEG are two different methodologies. This chapter aims to demonstrate the important interplay between the lesional zone (or network), based on the background activity in SEEG, the irritative zone (or network), based on the interictal epileptiform discharges (IEDs), and the EZ (or network) based on seizure onset structures. Please be mindful that the main concept of SEEG methodology is based on anatomo-clinico-electrical correlations. Due to the scope of our chapter, we will not delve deeper into this concept. In this chapter, we aim to illustrate that defining the SOZ as a specific temporal time point proves impractical within the framework of the SEEG methodology and to describe the importance of the spatiotemporal dynamics transition of the interictal phase to the ictal phase as the key to localizing the EZ in the SEEG methodology, together with the main concept being anatomo-clinico-electrical correlations.

Interictal SEEG activities

Before the era of SEEG and EZ concept, Penfield and Jasper used the concept of the epileptogenic lesion based on interictal spikes localization using intraoperative electrocorticogram (ECoG) together with intraoperative electrical cortical stimulation, in which ECoG’s interictal epileptiform discharges (IEDs) extended far beyond the structural lesion in patients with tumors. Jasper et al. reported a landmark paper on a strong correlation between the complete removal of IEDs guided by the ECoG and seizure freedom in patients with lesional epilepsy, but not in all the cases, illustrating that IEDs alone were insufficient to define an epileptic focus and IEDs could be seen to be more extensive than the epileptogenic lesions. ^, Thus, Jasper et al. concluded that not all the IEDs had the same significance in acute intraoperative ECoG recording. With further studies, the importance of the background activity was stressed in relation to the spikes. It was documented that the IEDs seen in abnormal cortical regions, defined based on abnormal background activity in ECoG, needed to be surgically treated, but not those IEDs seen at the variable distances from lesional borders, which were defined as transmitted or propagated IEDs.

The concept is further advanced with the SEEG methodology, which allows subacute prolonged extra-operative recording of interictal and seizure activities. In the late 1950s, Bancaud and Talairach, equipped with the capacity to record seizures, documented for the first time recording of seizures using stereoelectroencephalography (SEEG). Bancaud introduced the concepts of the irritative zone (IZ) and lesional zone (LZ), emphasizing that abnormal interictal activities and lesions were not consistently colocalized with the epileptogenic zone (EZ) as a rule. ^, ^, IZ is defined as cortical regions with increased cortical excitability indicated by the presence of spontaneous spikes or spike-wave activity. LZ is defined as cortical regions of underlying cerebral dysfunction with or without structural anomaly characterized by the presence of continuous or sub-continuous slowing. Subsequent studies have consistently substantiated that the IZ could be entirely colocalized, partially localized, or even independent of the EZ. Subsequently, IZ can further be divided into either primary IZ (confined within the EZ) or secondary IZ (extending beyond the EZ). ^, ^, Notably, a meticulous analysis of IZ is crucial in SEEG analysis to estimate the dynamic changes of an IZ network or multiple IZ networks, thereby enhancing the accuracy in delineating and mapping the EZ. ^, ^,

Before identifying IZ and LZ, it is important to recognize baseline physiological SEEG activity to define abnormality. Distinct physiological rhythms are observed across distinct cortical regions of the brain. A multicenter ICEEG atlas ( https://mni-open-ieegatlas.research.mcgill.ca ) is available to reference neurophysiological awake activity in various cortical areas. The significant differences in the spectral density distributions between different frequencies are noted across the different brain regions ( Fig. 6.1 ). Hence, it is crucial to emphasize the significance of electrode reconstruction. If the designated electrodes are positioned incorrectly, there is a risk of misinterpretation as pathological slowing originating from white matter SEEG activity. ^, As illustrated in Fig. 6.2 , C′ mesial contacts (1–4) were initially targeted to the posterior hippocampus, but the SEEG activities in C′ one to four showed low amplitude delta and alpha activity similar to the C′ 4–10 mid-electrode contacts, which were not expected to be in the posterior hippocampus. Was it an abnormal lesional activity? In fact, the electrode anatomical reconstruction confirmed that electrode C’s mesial contacts were in the white matter outside the hippocampus ( Fig. 6.2C ). Thus, before concluding about the abnormality in SEEG analysis, one must understand the three-dimensional anatomical regions where the electrodes are in the individual brain. After the confirmation of electrode location, one needs to rule out the effect of anesthesia or sedation as the physiological SEEG signal can be mistaken as slowing if analysis and confirmation is made with a first few days. Thus, it is better to analyze and study at least 24–72 h following SEEG placement before considering the background activity as being abnormal. ^,

Regarding LZ, the relationship between the MRI lesional cortex and SEEG-defined lesional cortex needs to be well-defined, as the lesional zone defined by SEEG from neurophysiological data can be far more extensive than the structural lesion noted in the MRI. After confirming the abnormal background, one needs to rule out the physiological EEG activity as the physiological SEEG signal can be mistaken as a spiking activity. ^, Thus, extra caution is required in order to avoid overcalling the normal physiological activity or missing the abnormal spike activity. Also, one should keep in mind that physiological and pathological activities can co-exist in the same region. ^, Thus, the following step-by-step analysis is recommended before defining the lesional or irritative zone ( Table 6.1 ). In elucidating the aforementioned practical observation, we present a case exemplified in Fig. 6.3 . The SEEG exploration was conducted in a patient diagnosed with bilateral occipital epilepsy. Anticipated within this context were normal physiological activities, specifically lambda waves, given the placement of SEEG electrodes in the occipital cortices. During the side-to-side eye movement, lambda activities were noted to be co-occurring over the right and left occipital electrode contacts (right occipital (L) and left occipital (L′) mesial and lateral contacts) ( Fig. 6.3A and C ). Although the L and L′ were symmetrically placed, the activation of lambda activity was more pronounced over the left lateral occipital cortex (L′ and V′ electrodes) compared to the right occipital cortex (L and V electrodes) ( Fig. 6.3C ). Following the reduction of ASM, the emergence of polyspikes and sharp waves in the lateral occipital cortices (L and L′ lateral contacts) was observed ( Fig. 6.3D and E ). This underscores the imperative of a thorough analysis of SEEG data and emphasizes the need for vigilant scrutiny of dynamic changes in SEEG IED activities, particularly in response to modifications in factors such as alterations in ASM.

Table 6.1

Step-by-step Approach to Analyze and Differentiate SEEG Background Physiological or Pathological Activities.

The following steps are recommended to analyze baseline SEEG physiology or pathological activities.
1.	Understanding and expectation of normal activity based on the cortical regions that are being explored by the SEEG electrodes
2.	Individual effects on the anesthesia are different, and thus, baseline activity needs to be studied at least 24–72 h following SEEG electrode placement.
3.	The changes in the background activities between stage changes, such as awake and sleep, should be analyzed.
4.	Routine baseline activation procedures are recommended if exploring functional cortex (photic stimulation, hyperventilation, eye closure, eye saccades (side to side or up and down movement, unilateral motor movements as well as sleep and awake)
5.	The changes in the background SEEG activities with antiepileptic (ASM) medication as well as sleep deprivation

Thus, in summary, after the abnormal SEEG activities were confirmed, as illustrated in Fig. 6.3 , one needs to identify interictal epileptiform discharges (IEDs) morphology and frequency, their prevalence or abundancy, their evolution over temporal distribution (i.e., occurrence earlier in the recording right after the SEEG implantation, occurrence during sleep stage changes, occurrence after medications changes, occurrence before or after seizures) and the evolution over spatial distribution (i.e., occurrence between different regions and in relation to each other) in order to map the IZ (or network (s)). ^, ^, ^,

Regarding morphology, a spectrum of interictal epileptiform discharges (IEDs) presenting diverse patterns is evident ( Fig. 6.4 ). These patterns encompass (1) spikes, characterized by a large-amplitude rapid component lasting 50–100 ms, typically followed by a slow wave spanning 200–500 ms; (2) sharp waves, featuring a rapid component lasting between 100 and 300 ms; (3) bursts of spikes; (4) fast oscillations; and (5) repetitive, paroxysmal slow waves. Various IEDs are known to be mediated by distinct neurobiological mechanisms, playing divergent roles in ictogenesis. ^, While the occurrence of different IED types in the same patients or regions is not uncommon, these distinct patterns hold clear diagnostic value. For instance, continuous high-frequency spikes and polyspikes strongly suggest Taylor-type II focal cortical dysplasia. Therefore, the manifestation of IEDs can vary based on cortical regions, structural dysfunction, and electrode placement relative to the sources of epileptiform discharges. Amidst the diverse IEDs, the prevalence of fast activities, particularly high-frequency oscillation activities (HFOs 80–400 Hz), associated with the IEDs is of paramount significance. Such activities have been demonstrated to correlate with interictal biomarkers in specific types of focal epilepsy. ^, Additionally, Guth and colleagues demonstrated that IEDs with HFOs are prominent in the epileptogenic zone and exhibit a tendency to increase during the transition to seizure activity. Moving beyond morphology and frequency, the signal amplitude in SEEG is contingent on the source orientation relative to the recording electrode. As such, amplitude lacks absolute significance; for example, amygdala spikes generally exhibit smaller amplitudes compared to hippocampal spikes (see also chapter in this book on The SEEG Signal). Moreover, the amplitude of spikes can vary based on regions of structural dysfunction, such as encephalomalacia.