陣列信號處理於大腦功能性磁振逆影像重建之應用

Abstract

磁振造影(Magnetic Resonance Imaging，簡稱MRI)為一種非輻射及非侵入性之醫學影像。因其影像可提供多樣之生理訊息，近年來被廣泛應用於不同之研究領域與臨床診斷。MRI其中一種影像方式為功能性磁振造影(functional MRI，簡稱fMRI)，可利用大腦受刺激前後血氧濃度(Blood-Oxyen-Level-Dependent，簡稱BOLD)之不同產生影像對比，來提供大腦功能性影像。因MRI目前是唯一可同時提供非侵入性之人腦結構性與功能性的醫學影像，故被用於許多神經認知科學相關的研究領域，如醫學、生理學、經濟學、甚至社會學。此外，在生理上，因為血氧濃度變化僅是暫態之反應，為了要完整記錄大腦功能之變化，良好的時間解析度便是fMRI影像很重要的需求之一，目前普遍使用的fMRI 可在1~3秒內獲得一張全腦之功能性影像。最近所提出之動態磁振逆影像(MR Inverse Imaging，簡稱InI)可大幅提升fMRI全腦影像之時間解析度至毫秒的等級。其原理為於擷取功能磁振影像時僅取全腦二維之投射影像，再利用平行影像與陣列信號處理之方式，將陣列接收線圈中，結合不同位置線圈所獲得不同位置之空間影像，產生非良置逆運算問題(ill-posed inverse problem)，進而使用數學方式解出三維空間中人腦活動訊號源之位置(source localization)。先前研究發現可利用線性限制最小變異(linearly constrained minimum variance，簡稱LCMV)之空間濾波器(spatial filtering，又稱為波數集成beamforming)的方式來改善InI影像之空間解析度，並提升重建訊號源之顯著統計值。本研究提出一種利用本徵空間最小L1範數波數集成(eigenspace minimized L1 norm beamforming)之全新方式，命名為本徵空間線性限制最小振幅(eigenspace linearly constrained minimum amplitude，簡稱eLCMA)空間濾波器，來估計InI訊號源。本研究經由模擬，與真人之視覺刺激手指活動實驗發現，eLCMA可進一步提升大腦活動訊號源之空間解析度與偵測BOLD對比之靈敏度。 本篇論文在第一章簡介MRI，fMRI，InI 成像原理，第二章介紹fMRI 資料之擷取與重建。第三章介紹利用陣列信號處理來重建InI活動訊號源之理論。第四章則介紹模擬與真人實驗之結果。最後於第五章討論本研究之結果與總結。

Magnetic Resonance Imaging (MRI) is a noninvasive and nonradiative medical imaging modality. It has been used in many different fields for research and clinical applications. One of its imaging applications is called functional MRI (fMRI) which can acquire functional brain imaging dynamically with the help of the blood-oxygen-level-dependent (BOLD) contrast generated by external-body stimuli. fMRI has been applied to many areas related to the cognitive neuroscience, ranging from medicine, psychology, economics, and even sociology. Such popularity comes from the advantage that MRI is by far the only noninvasive imaging method acquiring both anatomical and functional information for human brain. The temporal resolution is crucial to fMRI since the fast transient behavior of the hemodynamic response needs to be captured correctly. Conventional fMRI can achieve the temporal resolution of one to three seconds. Recently-proposed dynamic magnetic resonance (MR) inverse imaging (InI) is a novel parallel imaging reconstruction technique capable of improving the temporal resolution of BOLD contrast fMRI to the order of milliseconds at the cost of moderate spatial resolution. Volumetric InI reconstructs spatial information from projection data by solving ill-posed inverse problems using simultaneous acquisitions from a RF coil array. Previously a spatial filtering technique based on linearly constrained minimum variance (LCMV) beamformer was suggested to localize the hemodynamic changes of dynamic InI data with improved spatial resolution and sensitivity. Here we report an advancement of the spatial filtering method, which combines the eigenspace projection of the measured data and the l1-norm minimization of the spatial filters’ output noise amplitude, to further improve the detection power of BOLD-contrast fMRI data. Using numerical simulation and in vivo data, we demonstrate that this eigen-space linearly constrained minimum amplitude (eLCMA) beamformer can reconstruct spatiotemporal hemodynamic signals with high statistical significance values and high spatial resolution in event-related two-choice reaction time visuomotor experiments. In this dissertation, we will provide the brief review for MRI, fMRI, InI in Chap.1, and introduce fMRI data acquisition reconstruction in Chap. 2. We then provide the InI reconstruction theory in Chap. 3. We will show the simulation and experimentresults in Chap. 4, and discuss and conclude our study in Chap. 5.