Abstract: Accurate deception detection is a central concern in psychology, forensic science, and national security. Although most existing studies have focused on detecting denied true information, research on fabricated information, such as false identities frequently used in criminal activities, remains scarce. To address this gap, the present study conducted three ERP experiments to examine whether the P300-based Concealed Information Test (CIT) can effectively distinguish true from fake identity information. We hypothesized that true identity names would elicit larger P300 amplitudes than fake identity names because of their greater self-relevance and stronger memory encoding. Experiment 1 tested the feasibility of detecting false identity information using a P300-based CIT. Forty-eight participants were randomly assigned three days before testing to either a fake-ID group or a real-ID group. Participants in the fake-ID group memorized a fake identity and used it as their own during the test, whereas those in the real-ID group responded using their true identity. Experiment 2 extended this design with 43 participants by adopting a Complex Trial Protocol (CTP) to separate recognition from response classification and enhance attention to probe stimuli. Experiment 3 further used a within-subject CTP design with 23 participants, enabling direct comparison between real and fake identities while minimizing individual differences in P300 responses. Experiment 1 showed that the probe-irrelevant P300 difference was significantly larger in the real-ID group than in the fake-ID group, with ROC analysis yielding 83% classification accuracy. Experiment 2 replicated this effect and improved accuracy to 86%. In Experiment 3, participants consistently showed larger P300 amplitudes to their real names than to their fake names, producing an individual-level detection rate of 90% (18/20). Across the three experiments, both real and fake names elicited larger P300 amplitudes than irrelevant names, but fake names evoked significantly weaker responses than real names. Across the three experiments, both real and fake names elicited larger P300 amplitudes than irrelevant names, but fake names evoked significantly weaker responses than real names. These findings demonstrate that the P300-based CIT/CTP paradigm can reliably distinguish true from fabricated identity information. The P300 provides a robust neural marker for false identity detection and may offer a useful tool for forensic and security applications.

Key words: P300, memory, fake information, conceal information test, lie detection