Mapping Propagating Epileptiform Activity in the Pediatric Brain with Advanced Multimodal Neuroimaging

Epilepsy is a neurological disorder characterized by recurring, unprovoked seizures that result from abnormal brain activity. These seizures are caused by a sudden and intense discharge of cortical neurons, which is commonly referred to as an epileptic focus. Although many patients can manage their seizures with medication, approximately 20-30% are unable to control their seizures with such medications and suffer from drug resistant epilepsy (DRE). For these patients, there are various types of treatments available to achieve seizure freedom, such as resective surgery, neuromodulation devices or a balanced dietary approach. In particular, neurosurgery offers ~50% chance of achieving seizure freedom. Yet, it remains an underutilized treatment, particularly for children who may benefit the most since their brain possesses extensive neuroplasticity and, thus the opportunity to be rewired after surgery, preventing adverse neurodevelopmental and psychosocial outcomes. The primary objective of epilepsy surgery is to completely delineate and remove the brain region responsible for generating clinical seizures, which is known as the epileptogenic zone (EZ), while minimizing damage and negative consequences for the patient. To date, there is no clinical exam to delineate the EZ unambiguously. Instead, the EZ is defined as an approximation based on multiple non-invasive and invasive tests. The best estimation of the EZ is the cortical area where the seizures originate on the invasive electroencephalogram (iEEG), which is called the seizure onset zone (SOZ). Unfortunately, since seizures are unpredictable and difficult to record, the SOZ delineated during the invasive monitoring may not be the complete SOZ, which can lead to the patient experiencing seizures even after surgery. Spikes, or interictal epileptiform discharges (IEDs), are regarded as the most established interictal biomarkers of epilepsy and can be seen in brief time interval between seizures. Yet, spikes define the irritative zone (IZ) that suffers from low specificity to the EZ because the IZ is often larger than the EZ and thus can overlap with eloquent areas that should be preserved. In recent years, a new biomarker called high-frequency oscillations (HFOs) has emerged as a promising indicator to determine epileptogenic tissue. These oscillations, which occur at frequencies above 80 Hz, can be observed in intracranial and extracranial signals. HFOs are characterized in the time domain by at least four oscillations above the signal’s baseline. While HFOs detected with iEEG can be categorized into ripples (80-250 Hz) and fast-ripples (250-500 Hz), non-invasive methods such as scalp electroencephalography (EEG) and magnetoencephalography (MEG) are so far only able to detect ripples. The removal of tissue that generates high-frequency oscillations (HFOs) during epilepsy surgery is associated to good postsurgical outcome, i.e. the patient experiences no seizures after the surgery. However, patients with DRE show interictal ripples in large brain areas uncorrelated with the EZ and often epileptic HFOs overlap with physiological HFOs both in ripples and fast-ripples band. Several studies have suggested that interictal biomarkers do not occur as isolated events but instead appear across multiple iEEG contacts in a temporal sequence with a spatial displacement. These biomarkers propagate rapidly from onset to spread areas through neural pathways, similar to seizures. Previous iEEG studies have shown that spikes show early peaks, which travel to later propagated activity, traversing both adjacent and far brain areas following the same path and direction as ictal discharges during seizures. The regions firstly showing spikes are considered as located in the EZ, while areas where activity propagates late are considered less epileptogenic and therefore do not require excision. Similar to spikes, ripples also propagate across iEEG contacts; resection of ripple onset (but not spread) is associated with good outcome. Although there is a significant amount of literature on the propagation of interictal biomarkers, the underlying pathophysiological mechanism remains largely unexplored. On these pathophysiological bases, the ultimate goal of this thesis is to better understand the epileptic propagation phenomenon inside the brain and develop reliable biomarkers estimating the EZ. This was achieved by exploring the interictal spike and ripple propagation phenomena, pursuing three specific aims: the first aim was to reveal the relationship between the spatiotemporal propagation of spikes and the information flow among propagating areas and to assess their surgical prognostic value. Based on the investigated literature, the spike propagation phenomenon has never been analyzed using a combination of electric source imaging (ESI), virtual signals (VSs) reconstruction and effective connectivity (EC - i.e., information flow) in iEEG signals. Thus, these approaches were integrated in an innovative, semiautomated pipeline to define a brain region capable of invasively identifying the EZ. This new indicator will be referred to as spike-onset zone; the second aim was to reconstruct this zone via alternative non-invasive methods, overcoming the iEEG limitations remaining reliable across modalities; the third and last aim explored the reconstruction of the ripple propagation phenomenon with a virtual implantation, mimicking the iEEG implant, and at the whole brain level via non-invasive methodologies and the possibility of approximating the EZ with another brain area called the ripple-onset zone.