Digital filtration with zero and standard-phase shift characteristics was performed on unfiltered auditory evoked potentials recorded from 10 adult subjects. Standard-phase shift filters were seen to distort the response. Mild phase shift distortion often augmented the IV/V-Na1 amplitude (auditory brain stem response). When recording the IV/V-Na1 amplitude on a narrow timebase (20 msec), we recommend using a phase shift filter that has a high-pass cutoff frequency approximating 15 Hz and a slope of 12 to 24 dB/octave, in order to take advantage of the observed phase shift augmentation. The augmentation of the IV/V-Na1 amplitude is at the expense of the IV/V-Na2 amplitude. Thus, this phase shift distortion is not desirable if a long timebase (50 msec) is used. If a high-pass standard-phase shift filter of 100 Hz is used to eliminate muscle artifact, a reduction in the IV/V-Na1 amplitude from phase shift distortion is seen compared to the zero-phase shift control. In site-of-lesion testing, a nondistorted waveform produced by digital zero-phase shift filtration is also desirable over a distorted analog response. It is recommended that the slow wave activity of the response be recorded with a timebase of at least 40 to 50 msec and with a filter that produces minimal or no phase shifting. The major recorded response will be the Na trough and Pa peak. Evoked response units with steep filters will phase shift the response and produce a reduction in the IV/V-Na amplitude and a concomitant augmentation of the latter middle latency response waveforms.(ABSTRACT TRUNCATED AT 250 WORDS)