Synthetic aperture magnetometry (SAM) has been used to image source power or source signal-to-noise ratio from MEG. However, the locations of maximal event-related oscillatory activity (or changes from resting state) do not necessarily coincide with those sites that are phase-locked to external events (i.e., localized by dipole fit to the averaged evoked response). Since an estimate of the source time-series may also be obtained by applying the beamformer coefficients to the MEG signal, one can image event-related activity by mapping some function reflecting the reliability of the averaged source waveform at each location. We have devised a new analysis method, SAM(erf), for obtaining a functional image of event-related brain activity and revealing the corresponding waveforms for activated sites. The mapping function used is the ratio of RMS amplitude of the averaged source waveform to that of the +/- average waveform, for a selected time window. This function is computed at each coordinate on a three-dimensional grid in the head. In addition to the SAM(erf) functional image, the averaged source waveforms for each local maximum in the image can be computed and displayed. This procedure can reveal multiple locations and waveforms at sites in the brain engaged in event-related activities. When this method is applied to evoked response studies, phase-locked activity can sometimes be found in areas distant from primary sensory cortex. Given the sensitivity of this functional imaging method to areas outside primary sensory cortex, it has the potential for detecting subtle changes in brain activity in health and disease.