Abstract
Introduction. Optical imaging techniques have demonstrated that the cortical response to a moving visual stimulus appears to have an anticipatory leading edge component to the representation (Jancke et al., 2004). We sought to temporally and spatially resolve a moving stimulus on the cortical surface in humans using functional magnetic resonance imaging (fMRI). While the hemodynamic signal is sluggish, its response characteristics are highly reliable, and the ultimate resolving power is an issue of signal and noise. Our experimental goal was to determine the limits of the fMRI technique to resolve the path of a moving stimulus in the retinotopic human visual cortex.
Methods. Subjects' brains were scanned with a 3 T MRI scanner and a 32-channel head coil. Standard retinotopic mapping and cortical flattening procedures were performed. We experimented with EPI sequences with different k-space trajectories as well as reconstruction techniques to optimize the spatial and temporal resolution limits. The stimulus was a high contrast flickering checkerboard with a circular aperture that moved through the visual field with a constant velocity in polar angle at a fixed eccentricity.
Results. For each stimulus velocity, we were able to determine the amount of data required to achieve the same precision in the estimation of spatiotemporal position. We noted asymmetries between the leading and trailing edges, as a function of velocity.
Conclusions. We have demonstrated the limits of resolving moving stimuli along the human cortical surface. Being able to image the complete cortical representation of an object's trajectory allows us to test a number of hypotheses in areas of visual perception, including attentional object tracking and the properties of objects as the disappear behind occluders.