OBJECTIVES: To investigate the effect of increasing acoustic-phonetic difficulty in children with cochlear implants (CI) by means of auditory event-related potentials (AERPs). DESIGN: AERPs were recorded from a group of ten 9- to 14-year-old prelingually deafened children who exhibited open-set speech recognition, using the Nucleus 22 CI for at least 5 years. AERPs were recorded in sound field while children were performing oddball discrimination tasks with increasing acoustic-phonetic demand. The tasks consisted pairs of naturally produced stimuli that differed by one phonetic feature: vowel place (/ki/ versus /ku/), vowel height (/ki/ versus /ke/), voicing (/ka/ versus /ga/), and place of articulation (/ka/ versus /ta/). Using a repeated measure design, the effect of increasing acoustic-phonetic difficulty on P3 latency, amplitude, and scalp distribution as well as on the simultaneously obtained behavioral measures, performance accuracy, and reaction time was evaluated. RESULTS: AERPs elicited in the range of 350 msec poststimulus onset were contaminated by the CI stimulus artifact, thus enabling reliable identification of the P3 component only. Increasing acoustic-phonetic difficulty was manifested in all measures in a hierarchical manner: P3 latency and reaction time increased, whereas P3 amplitude and performance accuracy decreased. The correlations, however, between behavioral and electrophysiological measures were not significant. Further support for P3 sensitivity to increasing acoustic-phonetic demand was its absence in four of the 10 children, but only in the most difficult place of articulation task. P3 amplitude was maximal at the midline parietal cite, with equal amplitudes over the right and left scalp regardless of side of implant. CONCLUSIONS: The results underscore the significant value of the P3 potential as a sensitive neural index of speech-sound processing in children with CI. The similar hierarchy of acoustic-phonetic demand manifested in both behavioral and electrophysiological measures suggests that speech perception performance relates to neurophysiologic responses at cortical levels of the auditory system. Thus, recording the P3 potential to distinct phonetic contrasts may be useful for studying accessibility and neural encoding at the cortical level in CI recipients.