Lauren Burrell – ICSL

Phone: 404.894.9356
Fax: 404.894.4130

Room 218
Manufacturing Research Center
Georgia Institute of Technology
813 Ferst Drive, N.W.
Atlanta, GA 30332-0560 USA
Research Group: Bioengineering


Lauren earned a Bachelor of Computer Engineering from the University of Delaware in 2001 and a Master of Science in Electrical and Computer Engineering from Georgia Institute of Technology in 2002. She joined ICSL in the summer of 2003 and is currently a Ph.D. candidate in Electrical Engineering at the Georgia Institute of Technology. Lauren’s research focuses on combining features of electroencelphalography (EEG) and functional magenetic resonance imaging (fMRI) to accurately localize the source of epileptic activity in the brain in order to facilitate surgical removal of the affected area and to direct placement of implantable electrodes for neurostimulation.


Epileptic Focus.png

Epilepsy, a chronic neurological disorder characterized by recurrent, unprovoked seizures, affects up to one percent of the world’s population. Antiepileptic drug therapies are ineffective in over 30% of epilepsy patients. In these cases, the medications either do not successfully control seizures or have unacceptable side effects. Approximately one-third of patients whose seizures cannot be controlled by medication are candidates for surgical removal of the affected area of the brain, potentially rendering them seizure free. Accurate localization of the epileptogenic focus, i.e. the area of seizure onset, is critical for the best surgical outcome. Currently the most widely used tool for localization of the epileptogenic zone is electroencephalography. While the electroencephalogram (EEG) has high temporal resolution, it suffers from poor spatial resolution. Combining EEG recordings with a diagnostic tool possessing higher spatial resolution, such as functional magnetic resonance imaging (fMRI), might allow for more precise localization of the epileptogenic focus. The primary objective of the proposed research is to develop a set of fMRI data features that can be used to distinguish between normal brain tissue and the epileptic focus. These features will then be fused using genetic programming (GP) to find a single feature capable of localizing the epileptic activity.

Functional Magnetic Resonance Imaging (fMRI)
Functional magnetic resonance imaging (fMRI) is a noninvasive neuroimaging technique that indirectly measures neural activity by measuring changes in blood flow, blood volume, and blood oxygenation. Through the analysis of fMRI data, which consist of sequences of images acquired over time, the spatiotemporal dynamics of brain activation can be explored. The contrast in these images is attributable to differences in tissue function rather than structure. High-resolution anatomical scans enable identification of any structural abnormalities that might be present, while functional scans provide useful information about brain activity. This information is especially helpful in the absence of any structural irregularities, which would explain the atypical brain function. In fMRI, the voxel size is limited by the tradeoff between small voxel dimension and high signal-to-noise ratio; however, even with this constraint, the spatial resolution of fMRI is still relatively high – on the order of a few millimeters. The temporal resolution of fMRI in human studies is generally in the range of one to three seconds per three-dimensional image. Electroencephalogram (EEG) recordings have temporal resolution on the order of milliseconds, so when EEG and fMRI data are combined epileptic activity can be precisely localized both spatially and temporally.

Genetic Programming (GP)
Genetic programming (GP) is a machine learning technique that employs an evolutionary algorithm to generate an optimal program to solve a given problem. Evolutionary algorithms apply genetic operators such as reproduction, mutation, crossover, and selection to find optimal solutions from populations of candidate solutions, also known as individuals or chromosomes. Individuals are assessed through evaluation of a fitness function, which measures the individual’s ability to solve the given problem. The result of the algorithm is the individual with the best fitness.