Stability, repeatability, and the expression of signal magnitude in functional magnetic resonance imaging

In 23 fMRI studies on six subjects, we examined activation in visual and mo tor tasks. We modeled the expected activation time course by convolving a t emporal description of the behavioral task with an empirically determined i mpulse response function. We evaluated the signal activation intensity as b oth the number of activated voxels over arbitrary correlation thresholds an d as the slope of the regression line between our modeled time course and t he actual data. Whereas the voxel counting was strikingly unstable (standar d deviation 74% in visual trials at a correlation of 0.5), the slope was re latively constant across trials and subjects (standard deviation < 14%). Us ing Monte Carlo methods, we determined that the measured slope was largely Independent of the contrast-to-noise ratio. Voxel counting is a poor proxy for activation intensity, with greatly increased scatter, much reduced stat istical power, and increased type II error. The data support an alternative approach to functional magnetic resonance imaging (fMRI) that allows for q uantitative comparisons of fMRI response magnitudes across trials and labor atories. J. Magn Reson. Imaging 1999; 10:33-40. (C) 1999 Wiley-Liss, Inc.