Controlling human brain networks has aroused wide interest recently, where structural controllability provides powerful tools to unveil the relationship between its structure and functions. In this article, we define the optimal control signal path where the external control signal flows from one node to other nodes in the network. The control signal path not only shows the connections of some specific nodes in the brain network and the functions but also helps us to have a better understanding of how the control signals select and pass through the nodes to enable the brain functions with the minimum control energy. In common cases, as the control signal located on different nodes and the possible permutations of the nodes en route, there are enormous numbers of potential control signal paths in the network. The efficiency of a control signal path is defined to evaluate the most important path of the network based on the control energy. We propose the algorithms using control centrality to find the most effective control signal paths under several cases of prerequisites. As the human brain functional networks could be divided into several subnetworks to accomplish different cognition tasks (such as visuality and auditory), by the local control centrality of nodes, we could select the control signal path more efficiently, which might lead to unveiling the potential neural pathway to accomplish cognition progress.