Biomarkers to localize seizures in epilepsy
Epilepsy is the world’s most common, serious brain disorder, affecting nearly 50 million people worldwide, and accounting for 1% of the global burden of disease. For one-third of epilepsy patients, seizures remain poorly controlled despite maximal medical management. For these patients, approximately half have focal epilepsy, where seizures arise from a localized region in the brain and neurosurgical interventions are the most effective treatment option. For each of the available neurosurgical approaches, accurate identification of the pathologic brain tissue responsible for generating the seizures, the epileptogenic zone, is critical for a successful outcome. We are working to identify biomarkers to improve localization of the epileptogenic zone. We have previously identified the spike-ripple, a waveform that combines classic interictal epileptiform discharge with high frequency oscillations (HFOs) as a potentially improved biomarker compared to spikes and HFOs alone to predict seizure risk. We are now testing the sensitivity and specificity of this biomarker to localize the seizure onset zone in patients undergoing invasive EEG recordings. To do so, we have developed and validated semiautomated and fully automated detectors for ourselves and other groups to use to reliably find and quantify these events.
Neurostimulation to treat seizures in epilepsy
Rapid advances in brain stimulation in the last decade, in which focused electrical stimulation is used to terminate seizures and retrain pathologic seizure networks, now offer new promise to patients who are either not candidates for traditional resective brain surgery or prefer less invasive neurosurgical options. That neurostimulation is effective to reduce seizure burden in patients has been demonstrated in multiple studies. However, how neurostimulation works and the optimal stimulation parameters to use to impact pathological seizure networks remains largely unknown. With collaborators in biomedical engineering, neurosurgery, mathematics, and statistics, we are studying the cell populations involved in seizure networks and the impact of varying neurostimulation parameters on these cell populations. We are also using computational models to test varying stimulation parameters and testing the optimized sets on humans for improved outcomes.
Noninvasive Biomarker Discovery
Benign epilepsy with centrotemporal spikes (BECTS) is the most common childhood epilepsy syndrome, accounting for 20% of all childhood epilepsy and characterized by a transient period of seizure susceptibility of uncertain duration in school-age children. Despite extensive clinical experience with this disease, it remains a challenge to determine who will benefit from antiepileptic drug (AED) treatment and when it is safe to discontinue. Current clinical practice requires a trial-and-fail method for AED initiation and discontinuation with wide variability and controversy in treatment strategies across practitioners. Although non-treatment or premature taper may result in seizures and injury, chronic AED exposure is also not benign and may cause attentional deficits, aggression, hostility, nervousness, and somnolence in 30-70% of exposed children. A biomarker to isolate which children are at risk for ongoing seizures is needed in order to avoid the unnecessary consequences of over- or under-medication during critical years of cognitive, psychosocial and behavioral maturation in this large cohort of children. In this work, we leverage advanced, safe, non-invasive multimodal recording techniques from quantitative neurophysiology, structural, and functional neuroimaging and implement innovative approaches to integrate these technologies to develop objective measures of seizure risk and cognitive function in children with BECTS. By focusing on this unique, well-characterized, though poorly understood patient populations of the course of this transient disease, we also aim to identify candidate biomarkers of seizure risk which may have a broader relevance to other epilepsy syndromes.
We are also interested in evaluated biomarkers of cortical physiology in healthy development and other disorders of cortical physiology, such as autism, infantile spasms, coma, and disorders of consciousness.
Functional Brain Networks
The developing brain is an immensely complex system. During this critical period of cortical development, the neurological exam can be a poor assay of cortical function or long-term prognosis. However, EEG recordings offer opportunities for improved measures of subtle cortical function and organization. The scalp EEG correlates with known stages of normal neurodevelopment, provides information on local circuitry, and reveals age-specific functional networks throughout childhood representing large-scale cortical organization. We have established robust methods for inference of functional networks from scalp EEG. Through this work, we have improved techniques to mitigate limitations of EEG analysis due to the reference effect, spatial blurring, skull thickness, head size, and muscle artifact. Using these techniques, we have shown that stable EEG networks can be extracted from short samples of scalp EEG, that EEG networks rapidly mature in a stereotyped fashion over normal development, and that they reflect underlying white matter anatomical connectivity. Working with collaborators, our ongoing investigations explore whether these techniques can provide improved measures cortical disease in a variety of neurological disorders in children and adults, including autism, epilepsy, disorders of consciousness, and stroke.
Neonatal Seizures and Early Onset Epilepsy
Neonates and young children experience a disproportionately high burden of epilepsy. However, the unique presentations, treatment efficacy, and other epidemiological characteristics of early onset epilepsy remain poorly characterized. For several years, we have participated in multi-institutional collaborations to characterize the natural history and contemporary clinical practices and observational outcomes in neonatal and early childhood epilepsies. We have utilized these data to develop predictive models for seizure in neonates. We are now evaluating promising EEG biomarkers for seizure risk and long term developmental outcome in young populations by adapting tools we have developed in older populations to these age groups.