Electronic Theses and Dissertations





Date of Award


Document Type


Degree Name

Doctor of Philosophy


Biomedical Engineering

Committee Chair

Eugene Eckstein

Committee Member

Robert Ogg

Committee Member

Amy DeJongh-Curry

Committee Member

Erno Lindner


Children treated for brain tumors are at increased risk for developing cognitive deficits. The self-ordered search (SOS) is a computerized neuropsychological test used to investigate working memory, a cognitive system whose function is integral to many high level cognitive processes. Functional-MRI (fMRI) provides important opportunities to characterize neural correlates of SOSperformance non-invasively. Implementation of the SOStask presents challenges in the unique environment of the MRI scanner. First, SOSrequires participants to select a single stimulus from a set. Second, SOSis a behaviorally driven task that entails variable event timing among participants which complicates group analysis of fMRI data. The work presented here consists of the implementation, validation and application of the SOSfor fMRI and associated analysis techniques. Eye-tracking with a MRI-safe response device was used as an interface for the fMRI task, allowing the participant to select an individual stimulus from a two-dimensional array. Performance information was used to generate individual subject design matrices for fMRI analysis, preserving important behaviorally measures (time to completion). Healthy volunteers and patients treated for childhood brain tumors performed the SOS task and N-back task, a commonly used working memory task for fMRI. The eye-tracking interface performed well after initial problems with equipment and calibration routine were solved. Activation patterns identified by general linear model (GLM) analysis were similar between SOS and N-back tasks and included dorsolateral prefrontal cortex, ventral prefrontal cortex, dorsal cingulate, bilateral premotor, and parietal areas. Independent component analysis identified task-correlated components that were consistent with the GLM. Increasing activation across the general network was associated with fewer errors during the N-back task. Differences in activation between patient group and healthy group were identified in the parietal and retrosplenial cortex. Analysis of the performance data suggests differences between the healthy and patient groups. Our novel eye-tracking interface provides a natural interface that controls for movement and motor planning associated with complex response devices. The SOS for fMRI provides a new tool that will allow us to investigate deficits of working memory in children treated for brain tumors.


Data is provided by the student.

Library Comment

dissertation or thesis originally submitted to the local University of Memphis Electronic Theses & dissertation (ETD) Repository.