In simply guilty (SG), countermeasure-using guilty (CM), and innocent (IN) subjects, a new concealed information test, the P300-based Dual Probe Complex Trial Protocol was tested in a mock crime scenario. It combines an oddball protocol with two stimuli (probe, irrelevant) and another with three stimuli (probe, irrelevant, target) into one trial, doubling detected mock crime information per unit time, compared to previous protocols. Probe-irrelevant amplitude differences were significant in SG and CM, but not IN subjects. On a measure from both two and three stimulus protocol parts of the Dual Probe Complex Trial Protocol trial, accuracy was 94.7% (based on a .9 bootstrap criterion). The criterion-independent area (AUC) under the receiver operating characteristic (from signal detection theory) measuring SG and CM versus IN discriminability averaged .92 (in a range of 0.5-1.0). Countermeasures enhanced irrelevant (not probe) P300s in CM groups.
Read full abstract