Magnetic Field Of Magnetic Resonance Detection Of Nerve: Contribution Of Dendritic Branching Model

Since the early 1990s,functional MRI (fMRI) has been widely used to study the human brain function. Currently used fMRI techniques, however, rely on measuring regional cerebral hemodynamics such as blood flow, blood volume, and blood oxygenation level to infer neural activation, rather than detecting the neuronal activity directly. This indirect measurement of neuronal activity bears several limitations.First, the coupling between regional cerebral hemodynamics and neuronal activity is complex and non-linear. Regional cerebral hemodynamics does not necessarily always reflect neuronal activity. Second, vascular geometry may not overlap with the area of neural firing. The observed BOLD contrast, which is mainly vascular origin, therefore, may not always be a reliable indication of the area of neuronal firing. Spatial localization of neural activity could be degraded. Third, the cerebral hemodynamics responses are much slower (in the order of a second) than neuronal firing (in the order of a millisecond). Temporal resolution of the hemodynamic measurement is limited and downgraded with respect to the underlying neural activation.All these pitfalls limit the spatial and temporal resolution of the current hemodynamics-based fMRI techniques and could be resolved if neural activity is directly measured with MRI.Here, we utilized a modified current-dipole model to compute magnetic field generated by neural firing and to calculate MRI signal changes resulting from the neuronal magnetic field (NMF). Each dendrite or each unmyelinated axon was modeled as a modified current-dipole. NMF were estimated based on a synchronized activity of multiple neurons. Sensitivity of using MRI phase and magnitude to measure effects of NMF was evaluated.The results show that the MRI signal changes depend on the strength of magnetic field as well as geometry (orientations and configurations), NMF can potentially generate up to a few percent changes in MRI magnitude signals. Phases of MRI signal are insensitive to NMF when the distribution of the activated dendrites is symmetrical. Phases could be detected when the distribution of the activated dendrites is asymmetrical and inhomogeneons. Our modeling implies that direct MRI detection of neuronal activity is possible.