Localization Lab Research | Overview
Jurriaan Peters is an Associate Professor of Neurology at Harvard Medical School, and a staff member of the Department of Neurology, Division of Epilepsy and Clinical Neurophysiology at Boston Children's Hospital. He is a pediatric epileptologist with additional expertise in clinical neurophysiology and neuroimaging. He earned his MD at the Catholic University of Leuven, Belgium, and his PhD at the University of Utrecht. He did his Child Neurology training and Clinical Neurophysiology fellowship training at Boston Children’s Hospital. He is the principal epileptologist at the Multidisciplinary Tuberous Sclerosis Clinic at Boston Children’s Hospital.
The field of pediatric epilepsy and clinical neurophysiology is changing rapidly. Using Tuberous Sclerosis Complex (TSC) and focal cortical dysplasias (FCD) as model disorders, our laboratory studies early and aggressive intervention in childhood epilepsy to mitigate detrimental effects from seizures on neurodevelopment, and to improve long-term epilepsy outcomes. By investigating both the localization of epileptogenic lesions and the identification of networks responsible for seizure propagation, we will inform the rational design of trials aimed at modulation of the epileptogenic focus (e.g. TMS, minimally invasive epilepsy surgery), or, alternatively, at manipulation of the epileptic network (e.g. DBS).
At the Localization Laboratory, most of the work falls roughly into one of three lines of research:
1. Localization of the seizure onset zone, and of the epileptogenic lesion in multilesional epilepsy
We develop and apply novel MRI and EEG modeling techniques in the localization of the seizure onset zone in children with medically refractory epilepsy. In TSC, this comes down to identification of the epileptogenic tuber among many candidate lesions. We do this with an eye on epilepsy surgery, the success of which hinges on identification of the epileptogenic zone.
For example, we have recently proposed a TSC-specific adaptation of Electrical Source Imaging we named Lesion-Constrained Electrical Source Imaging (LC-ESI). This method of source localization is applied in the context of multilesional epilepsy. In these cases, sources are constrained to the lesions, and the lesions are rank-ordered (and subsequently color-coded) by their ability to explain the scalp EEG signal of the seizure. This work won with the Cosimo Ajmone-Marsan Award for best original paper in 2020, by the Journal of Clinical Neurophysiology and the American Clinical Neurophysiology Society.
Currently, we are working on automating this method, for validation and more widespread use. We are employing convolutional neural networks to automatically segment lesions in 3D MRI. Ultimately we'd like LC-ESI to become part of a family of online clinical tools, operating in a cloud and available to the world.
2. Expansion and propagation
We study the regional expansion of the epileptic network, as focal spike populations recruit larger volumes of cortex into the pathological activity over time, due to activity-dependent plasticity. In the past, the phenomenon of kindling was considered elusive: "Do seizures beget seizures"? Clinically and qualitatively, we aim to characterize this process in serial EEGs. In addition, can focal interventions (surgery) or systemic interventions (medications) alter the course of this process?
In addition, we seek to identify key hubs in seizure networks that allow focal lesions give rise to generalized seizures or, when left untreated, epileptic encephalopathies. How come that infantile spasms and tonic seizures are semiologically so similar across patients, i.e. a tightly defined phenotype that suggests a "final common pathway" in various lesional etiologies? What possible deep, widespread, and pre-exisiting network is present in humans that focal seizures tap into?
For example, we have recently applied Lesion Network Mapping (LNM) to demonstrate that tubers (lesions) in children with TSC-associated infantile spasms are more often connected to internal globus pallidus as compared to lesions in children who do not have infantile spasms.
We are currently confirming these findings in epilepsy surgical cases, and refining these methods to identify single tubers that place the patient at risk for spasms. We are also examining the presence of this network in other lesional epilepsies. Eventually, we may be able to use this method as a prior in a multi-modal model, i.e. we would assign probabilities to each lesion based on network properties.
3. Prediction and prevention
Our EEG and MRI work in young patients prior to epilepsy onset aims to predict impending epilepsy and localize the seizure focus early.
With early localization of the seizure focus, we may be able to render tailored surgical interventions at an early stage, and optimize the developmental outcomes by stemming the seizures. Recently, our paper on identification of structural MRI features in TSC-associated epilepsy demonstrated how hard this can be - the study was essentially negative. Other methods for early localization we are looking into include consistency of an epileptic focus in serieal EEGs as a marker of epileptogenicity, and longitudinal changes in diffusion tensor imaging (DTI) metrics.
In prior work, we have predicted the onset of epilepsy within a two or three month timeframe when the EEG from patients converts from normal to abnormal, i.e. when an EEG starts to show epileptiform discharges. Now we can identify patients at high risk for seizures that are about to start, and use this clinically available EEG biomarker to investigate whether preventative treatment yields benefits in terms of neurodevelopment and epilepsy. We are at the tail end of an NIH-funded multicenter randomized double-blind placebo controlled trial of preventative vigabatrin in TSC. The PI of the PREVeNT study is Dr. E Martina Bebin from UAB - stay tuned for updates!