Functional Controllability on Brain Networks


In recent years, both network neuroscience and cognitive science have developed vigorously. The network approach provides an analytical perspective of understanding the brain structures and functions, and uncovers the intrinsic correlation between them. However, this family of methods pays more attention on the discovery and pattern recognition in the phenomenon, thus lack a mechanistic explanation of why and how this correlation happens. In this work, focusing on the core scientific problem of modeling the human cognitive control, we proceed from the functional MRI signals and study the following subproblems: 1) building the control theoretical dynamics model based on the blood oxygen level neural signals; 2) developing the framework of control theory analysis of human brain functional network; 3) investigating the relationship between functional control measures and the descriptive cognitive measurement. For the first target, we fit the stochastic linear system and adapt the noise term in the model into the control part to build the functional control frameworks. Next, based on the model setting, we investigate the minimal control sets for each person and further quantify each network’s controllability, syncronizability and robustness w.r.t the control dynamics. Finally, we examine the distribution and relationships of these control measures, which are potential to be utilized as biomarkers correlated to the individual difference in cognitive performance.