Hardware for functional MRI

Abstract Task-driven changes in neuronal metabolism during a functional magnetic resonance imaging (fMRI) experiment result in relatively small signal modulations, typically only a few per cent at 1.5 Tesla (Ogawa et al. 1990). Analysis of the data generally compares (in a statistical sense) the signal levels resulting from a sensory or cognitive task with those obtained during a baseline condition. Changes in the signal that result from system drift, physiological functions, basal metabolism fluctuations, and other sources add noise to the comparison process and reduce the significance of the results. Because the neuronal signal modulations are small, it is important that the scan technique is optimized to maximize the signal levels and reduce these spurious sources of noise as much as possible.