NCCA Banner for Research Papers


STELMACK, R. M., HOULIHAN, M. and DOUCET, C. Event-related potentials and the detection of deception: A two-stimulus paradigm. November 1994, Report No. DoDPI93-R-0004. Department of Defense Polygraph Institute, Fort McClellan, AL 36205.

Event- related potentials (ERP) employed in two studies (exp. 1, n = 20 females; exp. 2, n = 20 males) attempted to distinguish subjects who had participated in a mock crime (guilty group) from innocent subjects who did not perform in the crime scenario. A two- stimulus paradigm was used in which neutral and crime relevant questions were followed by either a yes or a no target stimulus. Subjects were instructed to indicate their agreement or disagreement that the yes or no target stimulus was consistent with the truth of the preceding question. Guilty subjects were told to deny the truth of the crime relevant items. In both experiments, P3 amplitude was smaller for the guilty subjects in the crime relevant condition than comparison conditions when the task demanded an agree response. In experiment 1, P3 amplitude to the target stimulus was smaller for guilty than innocent subjects in the crime relevant condition when an agree response was demanded. This effect was also apparent in the ERP waveforms for experiment 2, but it was not statistically significant.

Key-words: lie detection, P3, N2, N4, stimulus congruity, response compatibility

Director's Foreword

This report describes the use of event-related potentials (ERP), recorded from the scalp, during the psychophysiological detection of deception (PDD). Its most unique feature is that a two stimulus paradigm is used during the questioning procedure. The first stimulus is the question. The second stimulus is a Yes or No response to the question. The subject is then required to indicated agreement or disagreement with the second stimulus. The authors suggest that this procedure provides physical and semantic simplicity and repetitive constancy that is not characteristic of other questioning techniques. The promising results of this preliminary study suggest that the average P3 wave amplitude recorded during deception is smaller than that recorded during truthful responses. As with any preliminary study, replication will be required to confirm the results.

This study is one of several recent investigations concerning ERP during the PDD. These investigations represent the application of relatively new technology and procedures to the PDD, and each one provides additional information towards improving the PDD instrumentation and procedures. The Institute will continue to encourage and support these productive studies.

Michael H. Capps