摘要:Functional MRI, a powerful method of neuroimaging, is expected to have profound and far-reaching consequences in the understanding of human brain function, a problem of central scientific interest at the present time. However, for higher cognition studies, the $time-resolution$ of present techniques has to be greatly improved. In this paper, we describe a fast functional MRI approach which we have developed, including an echo-volumar imaging sampling scheme, image reconstruction based on prolate-spheroidal kernels, optimality properties of these sampling and image reconstruction methods, and statistical analysis of fast functional MRI data. Results from real experiments are presented to demonstrate the feasibility of our methods.