We explored the feasibility of a subtraction technique described by Friesen and Picton to remove the cochlear implant (CI) artifact to long duration stimuli in the soundfield and using direct input all through the participant's preferred MAP. Friesen and Picton previously explored this technique by recording cortical potentials in four CI users with 1000 pulse per second (pps) stimuli, bypassing the speech processor. Cortical auditory evoked potentials (N1-P2) to 1000 Hz tones were recorded from a post-lingually deafened adult with three different stimulus presentation setups: soundfield to processor T-mic (SF), soundfield to lapel mic (SF-LM), and direct input (DI). Stimuli were presented at 65 dB SPL(A). The SF setup required stabilizing the head to minimize changes in magnitude for the CI artifact. The SF-LM and DI setups did not require head stabilization, but were evaluated as alternatives to the SF setup. Clear N1-P2 responses were obtained with comparable waveform morphologies, amplitudes, and latencies despite some differences in the magnitude of the CI artifact for the different stimulus presentation setups. The results of this study demonstrate that subtraction technique is feasible for recording N1-P2 responses in CI users, though further studies are needed for the three stimulation setups.