There's something wrong with this picture...
Problem summary:
As many as one in six innocent people would fail a polygraph test,
with false negative rates being even higher.  While not admissible in US courts, the polygraph and other lie detection tools are used frequently by law enforcement and the US government for investigation and even hiring and firing purposes. 
Traditional lie detection systems such as the polygraph lack scientific validity, yet they are used every day by many government agencies for very real investigations.
Because of that, our research was focused on how might we increase accuracy and bring scientifically valid methodologies to lie detection?
Project manager & data analyst
Collaborators & Advisors:
Dr. John Cacioppo, University of Chicago
Dr. Tyler Lorig, Washington and Lee University
Key insights:
1. The polygraph is actually better at intimidating people than it is at
detecting actual lies.
2. By relying on naturally occurring, multiply determined biological measures to detect deception, the polygraph fails to accurately identify deception without significant errors.
By asking how we might identify a unique physical response to deception, we realized that we could use classical conditioning methods to instill unique changes in peripheral blood flow in response to true and false statements. By using thermoelectric coolers/heaters were able to induce vasodialation and vasoconstriction in order to train participants' fingertips to show a unique response when true or false statements were presented to them. Using these techniques, we developed an alternative to the polygraph that worked in controlled environments. Ultimately, it was not able to be refined to be a field-grade diagnostic tool, but research continued to explore related directions. 
An experimental prototype sensor unit with a thermoelectric cooler and photoplethysmograph designed for fingertip mounting
Back to Top