Electronic Thesis and Dissertation Repository


Doctor of Philosophy


Biomedical Engineering


Terry Peters


Intractable or drug-resistant epilepsy occurs in over 30% of epilepsy patients, with many of these patients undergoing surgical excision of the affected brain region to achieve seizure control. Advances in MRI have the potential to improve surgical treatment of epilepsy through improved identification and delineation of lesions. However, validation is currently needed to investigate histopathological correlates of these new imaging techniques. The purpose of this work is to investigate histopathological correlates of quantitative relaxometry and DTI from hippocampal and neocortical specimens of intractable TLE patients. To achieve this goal I developed and evaluated a pipeline for histology to in-vivo MRI image registration, which finds dense spatial correspondence between both modalities. This protocol was divided in two steps whereby sparsely sectioned histology from temporal lobe specimens was first registered to the intermediate ex-vivo MRI which is then registered to the in-vivo MRI, completing a pipeline for histology to in-vivo MRI registration. When correlating relaxometry and DTI with neuronal density and morphology in the temporal lobe neocortex, I found T1 to be a predictor of neuronal density in the neocortical GM and demonstrated that employing multi-parametric MRI (combining T1 and FA together) provided a significantly better fit than each parameter alone in predicting density of neurons. This work was the first to relate in-vivo T1 and FA values to the proportion of neurons in GM. When investigating these quantitative multimodal parameters with histological features within the hippocampal subfields, I demonstrated that MD correlates with neuronal density and size, and can act as a marker for neuron integrity within the hippocampus. More importantly, this work was the first to highlight the potential of subfield relaxometry and diffusion parameters (mainly T2 and MD) as well as volumetry in predicting the extent of cell loss per subfield pre-operatively, with a precision so far unachievable. These results suggest that high-resolution quantitative MRI sequences could impact clinical practice for pre-operative evaluation and prediction of surgical outcomes of intractable epilepsy.