Electrophysiological correlates and predictive power of ERP components for false memories

Abstract

This study investigates the neural correlates of false memory using event-related potentials (ERPs) in a Deese-Roediger-McDermott (DRM) paradigm. The primary aim was to assess the predictive utility of three ERP components in distinguishing between true and false memories. Additionally, we examined the influence of semantic and orthographic similarity, as well as emotional valence, on memory accuracy and reaction times. EEG data were collected from 33 participants using a 128-channel system and analyzed through repeated measures ANOVA and logistic regression. Results indicated significantly heightened P3 and LPC amplitudes for false memories, as well as more negative peak amplitudes for the FN4 component. No significant ERP amplitude differences were found between semantically and orthographically related critical items. Logistic regression models demonstrated promising predictive power for P3 and FN4 components in distinguishing false from true memories, although not reaching forensic standards of specificity. These findings suggest that while certain ERP components show potential in identifying false memories, further refinement and inclusion of additional neural markers are necessary for robust forensic applications. Future directions include expanding the ERP model with additional components and exploring advanced EEG analyses, such as time-frequency and spectral density, to enhance predictive accuracy and deepen understanding of false memory mechanisms.