Addressing biological barriers in in vivo human brain MRI acquisitions

TRD 3解决体内人脑MRI采集中的生物屏障 摘要 在这个项目中，我们提出了一个生物工程发展计划，以改善人类 功能性和解剖性MRI。我们的目标是在宏观和微观之间架起桥梁 通过提高功能磁共振成像的时空分辨率， 生物极限我们在采集编码方面面临的障碍，然后再一次在MRI的能力， 在中尺度（500 μ m或更小）下进行解剖成像， 功能磁共振成像图本身主要是生物学的。首先，我们广泛地影响MRI编码， 专注于解决外周神经刺激（PNS）屏障的方法， 快速梯度线圈我们将通过利用详细的外围设备来改进梯度线圈技术 神经刺激模型来预测神经刺激和优化减少策略。在目标2 我们通过以下方法解决了解剖成像中呼吸和患者运动的混淆： 使用数据一致性驱动的解剖MR图像的运动鲁棒图像重建 approach.我们联合估计来自呼吸和患者的生物滋扰调制 多通道kspace数据内的运动。成功的联合估计的图像和 综合正演模型中的干扰变量有效地消除了这些混淆 从图像中产生运动和呼吸鲁棒采集。最后，在目标3中，我们开发了 通过预测大血管系统对fMRI施加的空间分辨率限制 并通过模型将它们从平坦皮质表面上的激活图中移除， 结合来自高分辨率血管图的先验知识。总体而言，我们的工具将广泛 在保留整体的同时，推进介观尺度上人类大脑回路的研究- 脑和MRI的非侵入性特征。

TRD3 Addressing biological barriers in in vivo human brain MRI acquisitions Abstract In this project, we propose a program of bioengineering development to improve human functional and anatomical MRI at its acquisition stage. We aim to bridge the macro and micro scales of brain architecture by improving the spatio-temporal resolution of fMRI down to its biological limits. The barriers we face in acquisition encoding and then again in the ability of MRI to perform anatomical imaging at the meso scale (500 um or less) and finally the spatial fidelity of fMRI maps themselves are primarily biological. Firstly, we impact MRI encoding broadly by focusing on methods to address the Peripheral Nerve Stimulation (PNS) barrier in application of fast gradient coils. We will improve gradient coil technology through utilizing a detailed peripheral nerve stimulation model to predict nerve stimulation and optimize reduction strategies. In Aim 2 we address the confounds of respiratory and patient motion in anatomical imaging through motion robust image reconstruction of anatomical MR images using a data-consistency driven approach. We jointly estimate biological nuisance modulations from respiration and patient motion within the multi-channel kspace data. Successful joint estimate of the image and the nuisance variables within a comprehensive forward model effectively removes these confounds from the image yielding a motion and respiratory robust acquisition. Finally, in aim 3 we develop methods to address spatial resolution limits imposed on fMRI by large vasculature by predicting and removing them from the activation maps on the flattened cortical surface by models incorporating prior knowledge from high resolution vascular maps. Overall, our tools will broadly advance the study of human brain circuits at the mesoscopic scale while retaining the whole- brain and non-invasive features of MRI.