作者: Lawrence A. Farwell , Emanuel Donchin
DOI: 10.1111/J.1469-8986.1991.TB01990.X
关键词:
摘要: The feasibility of using Event Related Brain Potentials (ERPs) in Interrogative Polygraphy (“Lie Detection”) was tested by examining the effectiveness Guilty Knowledge Test designed Farwell and Donchin (1986, 1988). subject is assigned an arbitrary task requiring discrimination between experimenter-designated targets other, irrelevant stimuli. A group diagnostic items (“probes”), which to unwitting are indistinguishable from items, embedded among irrelevant. For subjects who possess “guilty knowledge” these probes distinct irrelevants likely elicit a P300, thus revealing their possessing special knowledge that allows them differentiate irrelevants. We report two experiments this paradigm tested. In Experiment 1, 20 participated one mock espionage scenarios were for both scenarios. All stimuli consisted short phrases presented 300 ms each at interstimulus interval 1550 ms. set designated as “targets” appeared on 17% trials. Probes related also rest Subjects responded pressing switch following targets, other (and, course, probes). ERPs recorded Fz, Cz, Pz. As predicted, elicited large P300s all subjects. associated with given scenario P300 scenario. bootstrapping method used assess quality decision subject. algorithm declared indeterminate 12.5% cases. cases made. There no false positives negatives: whenever determination made it accurate. second experiment virtually identical first, results, except time 4 tested, had minor brush law. determine whether they possessed information own “crimes.” results expected; determined correctly information. implications data practice interpretation discussed.