Hemodynamic response function of functional magnetic resonance imaging and its application to epilepsy and carotid stenosis

Abstract

Different functional neuroimaging modalities, as Functional Magnetic Resonance Imaging (fMRI), have been contributing significantly to the best understanding of some cerebral diseases. However, the correct pattern of the hemodynamic response function (HRF) that follows neuronal activity in an fMRI exam, are not well known. Therefore, the first aim of this project is the development of an alternative analysis strategy of the time series generate in those exams, since all usual methods demand a model of a standard HRF. For such, we will adapt a method previously developed by us. We intended to extend that method turning to extract specific attributes from the HRF. Once developed, we intended to apply it to two specific problems: carotid stenosis and epilepsy. The study of the physiological mechanisms of the carotid stenosis is covered of great relevance by its impact in daily clinic, as well as for their possible population repercussions. A deficit of cerebral vasoregulation is a well established parameter in the determination of the hemodynamic commitment of patients with obstructive cerebrovascular diseases. Therefore, we intend to apply those analyses in the evaluation of the brain vascular reservation and vasoreactivity in patients with severe carotid stenosis or unilateral carotid occlusion and to correlate those alterations with the existence of collateral patterns, appraised by magnetic resonance angiography, and transcranial doppler sonography. Moreover, in the last few years the possibility of combining the good spatial resolution of fMRI with the excellent temporal resolution of electroencephalography (EEG) became feasible. Currently, one of the main possible clinical applications of such technique is the localization of epileptic event, as spikes. In such studies, it is even more difficult to precisely localize the BOLD signal involved in the electrophysiological event, since the pattern of HRF is not well characterized. Therefore, as the last aim of this project we propose to apply the analysis methods that will be developed as to analyze image time series, generated in a typical EEG-fMRI acquisition. (AU)