Spatial scale of correlated signals in 7T BOLD imaging

Abstract

The spatial distribution of signals from magnetic resonance imaging (MRI) using measures of Blood Oxygen Level Dependent (BOLD) activations presents a fundamental limit on the ability of MRI to resolve the neural signals from the brain. Here we show that the multiple samples of low-level BOLD activity comprise a form of neural “imaging dust” with distinct spatial characteristics. We apply the distance-dependent measurement of variance to spatial maps of BOLD signals to deliver a new approach to estimating the empirical point-spread function (PSF) of MRI. We show that these new estimates are similar to earlier measures of the PSF of high field 7-T imaging, but deliver the advantage that they are specific to each individual tested in a single scanning session. We explore various potential applications of this approach.