An intra-individual approach for detecting evaluation with event-related potentials

Abstract

The P3-based concealed information test (CIT) is an accurate indirect measure of non-evaluative memories (e.g., knowledge of an incriminating item). Less clear and established, however, is the accuracy of indirect measures that rely on the P3-like late positive potentials (LPPs) in discriminating evaluative (e.g., pleasant or unpleasant) memories. Using an LPP-based evaluative oddball paradigm in which participants were truthful on half of the trials about their evaluation toward pictures and concealed their evaluation on the other half of trials toward pictures, we applied an intra-individual Bayesian scheme to classify whether participants' evaluations were congruent or incongruent with a preceding context. LPPs were predictably larger to evaluatively incongruent than congruent pictures, and this LPP effect was reduced during misreporting presumably because of enhanced cognitive load. Notably, across two experiments the sensitivity (80%) was respectable during truth telling, but poor during concealment (sensitivity = 35%). Taken together, these data suggest that indirect measures such as the LPP-based evaluative oddball may be useful for detecting individual evaluation, but more work is warranted that explores conditions under which concealment of evaluation may be more accurately assessed.

Keywords: Deception; ERPs; Evaluation; LPP; Oddball.