Faculty Advisor

Karl G. Helmer, Ph.D.

Faculty Advisor

Fuhai Li, M.D.

Faculty Advisor

Christopher H. Sotak, Ph.D.

Faculty Advisor

Robert A. Peura, Ph.D.

Faculty Advisor

Alexander J. de Crespigny, Ph.D.


Magnetic resonance imaging (MRI) is a valuable research and clinical imaging modality for the non-invasive detection and characterization of cerebral ischemia. Specifically, diffusion-weighted imaging (DWI), which derives image contrast based on the diffusion of endogenous water molecules, is sensitive to cerebral ischemia within minutes of the onset of stroke. In combination with perfusion-weighted imaging (PWI) and T2-weighted imaging (T2WI), DWI can be used to characterize the temporal and spatial evolution of cerebral ischemia. The primary role of this dissertation is to outline several studies that investigate DWI, PWI, and T2WI changes in a rat stroke model of transient cerebral ischemia. Secondarily, this dissertation will introduce the method and results of an experiment designed to elucidate the relative roles of the intracellular (IC) or extracellular (EC) spaces to the water diffusion coefficient changes that occur as a result of cerebral ischemia. The use of MRI to detect cerebral ischemia is well established; however, the ability to distinguish between reversibly and irreversibly damaged tissues is limited. It has been shown in temporary focal ischemia models that the DWI abnormality (manifested as an image hyperintensity in the DWI) can be resolved if reperfusion is performed soon after the onset of the stroke. Initial studies suggested that the renormalization of water diffusion was associated with permanent restoration of cellular function (i.e., infarction was prevented). However, subsequent studies demonstrated that the disappearance of the acute ischemic lesion following reperfusion is not necessarily permanent and is related to the duration of the transient insult. Following short occlusions [e.g., 10 minutes in a rat middle cerebral artery occlusion (MCAO) model], there is complete tissue renormalization and restoration of normal neurological function. In contrast, following long periods of occlusion (e.g., 90 minutes), there are areas of the brain that do not recover and progress to infarction without delay. Intermediate durations of occlusion (e.g., 30 minutes) exhibit complete renormalization in all regions of ischemia; however, following several hours there is a gradual, secondary decline of the water diffusion coefficient values within the regions initially defined as abnormal. In this dissertation, the significant temporal and spatial heterogeneity in the secondary diffusion changes will be described and evaluated. Ultimately, MR techniques may provide valuable information regarding the response of tissue to transient ischemia as well as potential avenues for therapeutic intervention, which would have major clinical benefit. The significant changes in the apparent diffusion coefficient (ADC) of water that occur in ischemic brain are still not well understood. The leading hypothesis suggests that cellular swelling associated with the failure of the ionic gradient across the cell membrane results in an increase in EC tortuosity of the diffusion paths. Another theory suggests that the influx of fast-diffusing EC water, that occurs during cellular swelling, increases the proportion of water in the IC space, which is more restricted and viscous than the EC space. The final experiment presented herein demonstrates that significant cellular swelling remains in the regions of renormalized of ischemic ADC values that occur following reperfusion in transient ischemia. In short, the changes in the ADC values are not only the result of cellular swelling. Since conventional MR data contains the combined signals from the IC and EC spaces, it is difficult to determine the separate roles of these two compartments to the overall changes in water ADC. First, using a yeast-cell model, a method for separating the NMR signals is introduced. This method utilizes differences in the compartmental relaxation properties to isolate the MR signals from IC and EC spaces, and then secondarily the diffusion coefficients can be calculated. Using a modified version of this method, the experiment was performed in normal and ischemic rat brain. Intracerebroventricular (ICV) infusion of an MR contrast reagent (CR) was used to isolate IC T1, T2, and ADC values in vivo in normal and middle cerebral artery occluded (MCAO) rats using volume-localized, diffusion-weighted inversion-recovery spin-echo (DW-IRSE) spectroscopy and diffusion-weighted echo-planar imaging (DW-EPI). The presence of the EC contrast reagent (CR) selectively enhances the relaxation of water in the EC space and allows the IC and EC signal contributions to be separated based on T1-relaxation time differences between the two compartments. The results presented in this dissertation suggest that the IC ADC value is the major determinant of the overall ADC value measured in the normal rat brain. Further, the data suggests that the ADC decline experienced during acute ischemia is dictated largely by changes in the IC ADC, possibly due to failure of energy-dependent IC microcirculation (cytoplasmic streaming).


Worcester Polytechnic Institute

Degree Name



Biomedical Engineering

Project Type


Date Accepted





NMR, diffusion coefficient, cerebral ischemia, Nuclear magnetic resonance, Cerebral ischemia, Brain, Imaging