Mesoskopic Tracking of White Matter Fibers in the Human Brain

纤维追踪或纤维束成像是一种使用弥散加权磁共振成像在体内重建神经元通路的方法。它在神经科学中的巨大潜力目前受到数据处理复杂性的限制。跟踪结果往往不准确，缺乏量化和比较方法。该项目的目的是开发定量纤维束成像方法，以便能够比较结果，特别是为了验证而使用的侵入性技术。最近创建的吉布斯跟踪被认为是该项目的核心方法。该算法已经产生了从神经学角度来看有希望的结果。作为这个项目的第一个任务，它将被加速大约两个数量级，以便变得实际可用。在这种方法中，纤维束被表示为一组三维曲线。这种结果的比较是一个普遍的问题，这是要解决的模式识别方法在计算机科学中发展。此外，由重建的纤维束介导的连接的概率将使用物理学中的路径积分的方法来量化。这种方法也表明吉布斯跟踪和广泛使用的概率跟踪之间的等价性。当建立这种等价性时，将能够根据绝对连接概率校准连接图。所有结果都应在软件中实施，并提供给科学界应用。

Fibre tracking or tractography is a method for the reconstruction of neuronal pathways in vivo using diffusion-weighted magnetic resonance imaging. Its high potential in neuroscience is currently limited by the complexity of data processing. The tracking results are often inaccurate, lack quantification and comparison methods. The aim of this project is to develop quantitative tractography methods in order to enable comparison of results, especially with the invasive techniques for the sake of verification. The recently created Gibbs tracking is considered as the core method of this project. This algorithm already yields results that are promising from the neurological perspective. As the first task of this project, it is to be accelerated by about two orders of magnitude in order to become practically usable. In this method, the fascicles are represented as a set of threedimensional curves. Comparison of such results is a general problem that is to be solved using methods of pattern recognition developed in computer science. Further, the probability of connection mediated by reconstructed fibre bundles is to be quantified using methods of path integration in physics. This methodology also suggests an equivalence between the Gibbs tracking and the widely used probability tracking. This equivalence when established would enable a calibration of connectivity maps in terms of absolute connection probability. All results should be implemented in software and made available for applications in the scientific community.