DoDPI Banner for Research Papers


MILLER, J. C. Cardiovascular indices of guilty knowledge. June 1994, Report No. DoDPI94-R-0016. Department of Defense Polygraph Institute, Ft. McClellan, AL 36205.

The project attempted to determine, through a multivariate analysis of cardiovascular function using impedance cardiography, what pattern of autonomic activity was reliably associated with concealed knowledge. The hypothesis was that (h1) reliable, meaningful changes in the autonomic control of cardiovascular function are associated with concealed knowledge. The subjects were ten non-obese males with no prior polygraph experience. Four trials of six questions each were used. All trials were constructed like the Numbers Test (Furedy and Ben-Shakar, 1991). The first trail was based upon playing cards. The second through fourth trials involved the subject's year, month and day of birth, respectively. Digitized physiological data were acquired for 30 sec following each stimulus. The breathing pattern, skin conductance responses, electrocardiogram, impedance cardiogram and fingertip pulse pressure were recorded. The SCR-based detection rates for lies, 62.5% and 67.5%, were well above the expected random guessing rate of 20%. These relatively high rates suggested that different autonomic responses were associated with lies than with true responses. In turn, this observation suggested that differences in cardiovascular dynamics between lies and true responses may have been triggered by autonomic changes. The responses evoked in the cardiovascular measures by pre-lie, lie, and post-lie answers appeared to occur in the context of tonic activity in the cardiovascular components of the two autonomic branches with the highest level of coactivation preceding the pre-lie truth and the lowest preceding the post-lie truth. Most of the interesting phasic cardiodynamic activity following truthful and untruthful answers occurred in the first 10 or 11 beats following the subject's answer. The results showed that:

  1. The pre-lie condition was associated more than the lie or post-lie conditions with a tonic "fight-or-flight" response.
  2. For the pre-lie true response, TPR was highly reactive immediately following the stimulus for half of the subjects. For the other half, TPR was reactive several seconds later. For both groups, Q reacted inversely to TPR.
  3. There were phasic cardiac indications (contractility, SV, IBI) of greater "relaxation," but a phasic, peripheral vascular indication (MAP) of greater "tension" for the lie compared to the post-lie response.

Key-words: impedance, plethysmography, psychophysiological detection of deception (PDD), skin conductance, vagal tone, electrocardiogram (ECG, EKG)

Director's Foreword

This study is one in a series of studies to explore a variety of sensors for recording cardiovascular physiological activities (CPA). The purpose of this line of research is to find a sensor for recording CPA that will: (1) be easier to quantify; (2) be more sensitive to basic CPA; (3) not cause discomfort to the examinee; and (4) allow lengthier test formats should circumstances require them. This study was designed to examine various non-invasive measures of cardiovascular activity, using impedance cardiography and blood pressure measures.

While there is some evidence here that subject veracity may be accurately ascertained using advanced technology and procedures, assumptions based on this study alone should be tempered with caution, because of the small subject population tested. While it is true that impedance cardiography is non- invasive, it is intrusive and may not be practical for the psychophysiological detection of deception (PDD) in the field. The demonstrated ability of technologically advanced instruments and processes to differentiate among multiple patterns of cardiovascular responses has been shown. This should encourage further investigations into the development of more sensitive and practical instruments and procedures for analyzing and categorizing theses patterns for PDD.

Michael H. Capps