脑深部刺激与核磁共振成像在多种脑疾病的治疗、诊断以及脑科学研究中正扮演着越来越重要的角色，两者结合，无论在临床上还是在脑科学研究中，都具有重要意义。然而两者的不兼容问题，特别是磁共振射频场引起的脑起搏器电极的发热问题，限制了脑起搏器在核磁共振环境下的应用。核磁共振测温是近年来快速发展起来的一种无创测温手段，本文以此为基础结合生物传热仿真模型提出了一种评估脑起搏器电极在核磁扫描时发热安全的方法。首先对比分析了几种常用的核磁测温方法，选取基于质子共振频率转移的核磁测温方法作为本文的测温手段，并对该方法的影响因素进行了分析。由于金属伪影的存在，核磁测温无法测得电极触点周围的温度变化，为此本文提出以核磁测温得到的伪影外数据作为生物传热模型边界条件反推电极触点温升的方法。进一步介绍了生物传热理论，建立了生物传热仿真模型，最后基于已有的研究结果提出一种利用测温加热序列串评估脑起搏器电极在核磁扫描时发热安全的方法。其次在琼脂模型中采用外源交变电流模拟射频诱导产生的电流对电极触点加热，以此实验为基础建立了核磁测温结合生物传热模型反推电极触点温升曲线的算法。对比了三种不同幅值电流加热模式下反推方法的准确性，结果表明该方法能够在电极触点温升较高且保证患者安全的前提下获得准确的反推结果，平均误差百分比在10%左右。最后，本文设计了测温加热序列串，在体模中实现了利用核磁设备交替进行测温加热的功能。研究了三种不同的测温加热序列串扫描模式下反推方法的准确性，结果表明当序列串中测温时间较短时，本文提出的基于核磁测温与生物传热模型反推电极触点温升的方法准确性较高，平均误差百分比也在10%左右。

Deep brain stimulation (DBS), together with magnetic resonance imaging, is playing a growing important role in treating a variety of brain diseases and brain science research. Combining the two techniques will have great significance in clinic as well as in brian research. However, incompatibility between them, especially the DBS electrode heating induced by the RF field of MRI, makes DBS device forbidden in MR environment. MR thermometry is a highly developed non-invasive method to measure temperature in recent years. This thesis proposes a novel method based on MR thermometry and bioheat transfer model to evaluate the safety of patients with DBS while performing MRI.Firstly, by comparing several frequently-used MR thermometry methods, the one based on proton resonance frequency shift (PRFs) was chosen as thermometry in this thesis. Four influence factors of the method were analyzed. Among them, the existence of the metal artifact makes the MR thermometry based on PRFs unable to determine the temperature adjacent to the electrodes. Using the MR thermometry outside the artifact range as the boundary conditions of bioheat transfer model to determine the temperature within the artifact may be an alternative solution to this problem. Therefore, thermal modeling based on bioheat transfer theory was developed. In the last, based on previous studies this thesis proposed a method using thermometry-heating sequence train to evaluate the safety of RF-induced heating.Secondly, in agar model external alternating current was used to mimic the RF-induced current to heat the electrodes. Based on this experiment, An algorithm using MR thermometry and thermal modeling to determine the temperature changing curve adjacent to the electrodes was developed. Three kinds of current with different amplitudes was applied to study the accuracy of the method with respect to thermal power. The results showed that this method was able to acquire accurate temperature data near the electrodes under the condition of ensuring patients’ safety, with 10% average percentage error level.Finally, the thermometry-heating sequence train was designed and successfully applied in an phantom. Three different kinds of scanning model was used to study the accuracy of the method with respect to the sequence train. The results indicated that when the scanning time of the thermometry sequence in the train is short enough, the method based on MR thermometry and thermal modeling to measure the temperature near the DBS electrodes will have a relative high accuracy, which is 10% percentage error level averagely.