Neurophysiological Signature of Memory Performance during the DRM Task

Abstract

Background: The Deese-Roediger-McDermott (DRM) task is a typical experimental paradigm used to induce false memory production. Previous investigations using event-related potentials (ERPs) revealed differences in a late negative component, FN400, in healthy subjects performing the DRM. However, false memory production may be related to the earlier engagement of cortical circuits. So far, there is no study exploring the electrophysiological activation of both early and late neural networks DRM-related in healthy subjects. Methods: We compared the neuronal bases of good and bad performers in healthy subjects performing the DRM. Electroencephalography (EEG) was used to find differences in the event-related potential (ERP) components typical for the DRM, namely the early sensory N1 and P2, and the late FN400 component. Using independent component analysis (ICA), we extracted one functional component time-locked to the P2 component, and with source reconstruction (LAU-RA) we investigated the underlying sources of the FN400 component. Results: Compared to good performers, bad performers displayed increased latency of sensory N1 component and significantly lower amplitude of the P2 and the late FN400 components. They showed a lower activity of the P2 generator as revealed by ICA. Using source reconstruction, a different activation inparietal lobules and the left orbitofrontal cortex was also revealed in bad performers in the FN400 time-range. Conclusion: Altogether, our findings indicate that the performance in distinguishing false and true memories is correlated to an early activation of neuronal network. Such activation patterns suggest a more efficient activation of neuronal circuitry allocated to discrimination ability, selective attention and subsequent familiarity processing as well as more efficient prefrontal and parietal cortices activation.