Administrative Supplement to Physiology-driven seizure management post-cardiac arrest

PROJECT SUMMARY/ABSTRACT
Dr. Edilberto Amorim is a neurologist with subspecialty training in critical care and epilepsy who aims to employ
biomedical technology innovations in brain monitoring to personalize treatment for patients with hypoxic-
ischemic brain injury post-cardiac arrest. This career development award and its rigorous curriculum will
establish Dr. Amorim as a clinician-scientist with independent expertise in: 1) Deep learning applied to
physiology time-series, 2) Causal inference for observational data, and 3) Quantitative brain imaging. Every
year, more than 500,000 Americans have a cardiac arrest. Brain injury is the number one cause of death for
patients surviving initial resuscitation, and refractory seizures and other seizure-like brain activity are
diagnosed in up to 50% of patients. Despite being a common complication, outcomes are dismal and current
treatment strategies for seizures post-cardiac arrest are limited. Dr. Amorim aims to identify physiology-driven
biomarkers of resilience to hypoxic-ischemic brain injury by utilizing state-of-the-art computational methods and
a massive EEG and neuroimaging dataset with >1,500 subjects. His central hypothesis is that specific time-
dependent changes in spike and accompanying EEG activity during cardiac arrest treatment predict seizure
control and, ultimately, neurological recovery. The primary objectives of this proposal are: 1) Identify early
longitudinal epileptiform EEG phenotypes predictive of neurological recovery using interpretable and deep
learning algorithms; 2) Establish quantitative EEG biomarkers of seizure treatment response to anesthetics;
and 3) Estimate the causal effect of rapid seizure treatment with anesthetics in preventing structural brain
injury quantified with brain MRI. Dr. Amorim has generated preliminary data to demonstrate the feasibility of
modeling EEG phenotypes longitudinally for outcome prediction and has applied quantitative EEG biomarkers
to predict degree of brain injury on brain MRI. His primary mentor in this proposal will be Dr. Edward Chang, a
neuroscientist and leader in human neurophysiology research. His co-mentors will include Dr. Brandon
Westover, an authority in machine learning applied to critical care EEG, and Dr. Donna Ferriero, an
accomplished translational and neuroimaging investigator in hypoxic-ischemic brain injury. Additional
mentoring in quantitative neuroimaging (Dr. Srikantan Nagarajan) and biostatistics (Dr. Charles McCulloch) will
be essential components of his training. These aims are expected to establish early non-invasive predictive
biomarkers of neurological recovery and seizure control that may: 1) Guide patient selection for clinical trials
enrichment and 2) Serve as target to therapeutic interventions after hypoxic-ischemic brain injury. By
leveraging the deep expertise of a cross-disciplinary group of world-class mentors and the unparalleled
innovation environments of the University of California, San Francisco and the Bay Area, Dr. Amorim will be
ideally positioned to uncover fundamental knowledge about epileptogenesis after acute brain injury as well as
spearhead clinical trials focused on improving outcomes meaningful to cardiac arrest patients.

Grant Number: 3K23NS119794-05S1
NIH Institute/Center: NIH
Principal Investigator: EDILBERTO AMORIM