Hemodynamic-based functional magnetic resonance imaging (fMRI) techniques have proven to be extremely robust and sensitive methods for noninvasive detection and mapping of human brain activation. Nevertheless, limitations in temporal and spatial resolution as well as interpretation remain because hemodynamic changes accompanying brain activation are relatively sluggish and variable and therefore imprecise measures of neuronal activity. A hope among brain imagers would be to possess a technique that would allow direct mapping of brain activity with spatial resolution on the order of a cortical column and temporal resolution on the order of an action potential or at least a postsynaptic potential. Recent efforts in understanding the direct effects of neuronal activity on MRI signal have provided some degree of hope for those who want a more precise noninvasive brain activation mapping technique than fMRI as we know it now. While the manner in which electrical currents influence MRI signal is well understood, the manner in which neuronal firing spatially and temporally integrates on the spatial scale of an MRI voxel to produce a magnetic field shift and subsequently an NMR phase and/or magnitude change is not well understood. It is also not established that this field shift would be large or long enough in duration to be detected. The objective of this paper is to provide a perspective of the work that has been performed towards the direction of achieving direct neuronal current imaging with MRI. A specific goal is to further clarify what is understood about the theoretical and practical possibilities of neuronal current imaging. Specifically discussed are modeling efforts, phantom studies, in vitro studies, and human studies.