Clinical monitoring of patients with neurological disease generates large volumes of data, including electrophysiology (e.g. EEG), heart rate, blood pressure, and other physiological measures. The use of these data to generate actionable prognostic measures is a long-held goal in clinical neurophysiology. In this regard, the use of engineering theory, including signal processing methods and computational modeling paradigms, is providing new ways of interpreting neurological data and yielding new insights into brain mechanisms and disease pathophysiology. This research focuses on the analysis and modeling of aberrant brain dynamical phenomena that occur over hours-long temporal epochs. Particularly, we consider time-scale separated electrophysiological modulation, in which brain electrical activity contains distinct harmonic components that differ by orders of magnitude in the frequency domain. We propose methods for detecting such modulation, apply these methods to characterize its incidence in multiple clinical populations, and develop a biophysical dynamical systems model to study its underlying physiological mechanisms. Our primary focus is on the spatiotemporal analysis of slow (millihertz range) narrowband modulation in electroencephalogram (EEG) recordings. We propose a method that constructs sparse spectral estimates of power envelope signals obtained from physiological frequency bands of EEG. The former are obtained through a regularized basis pursuit approach, solved using LASSO methods and applied to temporal windows of variable length. This approach successfully identifies modulation in brain electrical activity on much slower (<0.01Hz) time-scales than conventional power spectral analyses. We apply these methods in several large clinical cohorts: neonatal patients with encephalopathy and adult patients with hemorrhagic or ischemic stroke. We establish validity of the method and its ability to derive a predictive biomarker. In the neonatal encephalopathy patients, we observe correlations between our modulation index values and developmental outcomes measures. Finally we generate a biophysically-informed dynamical systems model of the phenomenon. We identify parameter sets which generate aberrant oscillatory regimes in intracranial pressure (ICP) and other physiological process variables. We describe bifurcations in the model dynamics, providing hypotheses and predictions regarding potential mechanisms underlying millihertz electrophysioligical modulation. We compare the model outputs and predictions against findings from patients with multimodal (EEG, hemodynamic, ICP) monitoring data.
If you are a member of the WashU community, login with your WUSTL Key to interact with events, personalize your calendar, and get recommendations.Login with WUSTL Key
If you are not a member of the WashU community, please login via one of the options below to interact with our calendar.