In Vivo Mapping of Human Brain Network Using GPU Parallel Computing: an Example in Limbic System

Cognitive functions are orchestrated and integrated through the interactions between large-scale neural assemblies. The temporal dynamics and anatomical pathways between multiple brain regions are regarded as the functional and structural connectivity, respectively. Recently, studying the relationships between functional, structural connectivity and cognitive performances becomes popular in the field of neuroscience since it provides a better understanding about how the brain works. At the current stage, the brain connectivity in the neocortex is under intense investigations, but the limbic system, the most essential part of the brain in charge of emotion and memory, is seldom addressed in terms of connectivity. Therefore, the current project is applying the brain connectivity technologies to investigations of emotion-related brain networks in the limbic system. We will adopt resting-state functional MRI and high angular resolution diffusion imaging (HARDI) for assessing the functional connectivity and structural connectivity, respectively. Nevertheless, the graph theoretical analysis of brain network would be tremendously time-consuming with large number of defined nodes and numerous fiber tracts extracted from HARDI-based tractography. Hence, we will develop the algorithms for minimizing the computational loads with higher computing power based on the graphic processing units (GPUs) and compute unified device architecture (CUDA). A standard template of limbic networks based on the integration of functional connectivity and structural connectivity will be estimated in this project. This connectivity template of limbic systems will be firstly addressed worldwide and facilitate investigations of limbic networks in the field of neuroscience, brain sciences and psychiatry. In summary, fulfilling this project would shed a light on presenting following advantageous features: (1) increasing the efficiency of brain network analysis on the basis of parallel computing; (2) providing a standard template of limbic systems for free download; (3) exploring the brain connectivity in limbic systems and its associations with the degrees of depression and anxiety.

Project ID:PB10207-0362
External Project ID:NSC102-2221-E182-004