Visual mapping on the brain - caveats for multifocal fMRI

While multifocal electroretinography has become a standard ophthalmological technique, its use in cortical neuroimaging has been lesser. Vanni et al. (2005) presented the first exploration of the multifocal visual mapping methodology with fMRI. This commentary confirms the utility of this method, but also presents empirical results which suggest caveats for the use of the technique. In the current study rapid multifocal fMRI was established using m-sequence pseudo-random binary stimuli applied to visual field mapping in six young adults with normal vision. Nine contiguous regions of visual field - two rings of 4 patches with a central patch, areas scaled for cortical magnification, were pseudo-randomly stimulated, with patterned or grey images. The decorrelation of stimulus patches allowed all 256 volumes to be used for the analysis of each of the nine stimulus areas. Strong localized activation was observed for each of the four peripheral regions with the location of the activation conforming to the expected visual field retinotopy. The inner regions, including the foveal patch, did not significantly activate. We propose, on the basis of a simple correlational model of simulated eye movements, that the loss of signal is due to gaze instability. Thus, while the rapid multifocal method can be successfully applied to fMRI, the results appear quite sensitive to eye movements, the effects of which may have been overlooked by smoothing evoked responses to achieve a retinotopic map.