Averaged single trials (AST) allowed the functional magnetic resonance imaging (fMRI) response to auditory stimuli to be measured at high temporal (1 s) and spatial (0.1 cm3) resolution. Using this paradigm we investigated the transient signal response to 100-ms tone bursts in trains of between 100 ms and 25.5 s in total duration. We have demonstrated that the fMRI response to such auditory stimuli is approximately linear for trains of 6 s and longer, but that shorter stimuli produce signals that are larger than might be expected from the response to the longer stimuli. This nonlinear behavior can be modeled if an adaptive response to each stimulus is assumed. A study using a novel paradigm was also performed in order to study the influence of scanner noise during fMRI experiments on the auditory system response to tones. This study demonstrated that the temporal response to 700-ms tone stimuli is modified when performed in the presence of scanner gradient noise, the modification being a small but significant increase (P < 0.05) in the magnitude of the response. Finally the ability to measure the onset of functional activation using the AST method was examined. It was found, with the aid of computer simulation that a sampling rate of one image per second is adequate to distinguish temporal responses. Using the data acquired in this study, onset times were calculated for the auditory cortex, and these results are consistent with current models of functional activation.