Nowadays, polygraph machine are used commonly all over the world.
Abstract
Nowadays, polygraph machine are used commonly all over the world. The problem with existing polygraphs is, that when subject is over excited, the machine might be wrong and indicate that the person is telling a lie while in truth, he isn’t. Because of the common use of the polygraph, there is an increasing demand for a reliable machine that will solve the problem described above. In order to find a way to determine whether a subject lied, relying on the signals measured from his body, we must first characterize and explore the properties of the lie – that is the goal of this project.
The basic approach
In order to characterize a lie, we measure two signals from the body with the help of a sampling machine. The first signal will be taken from the right hand, and will indicate for the arousal of the left lobe of the brain. The second signal will be taken from the left hand, and will indicate the arousal of the right lobe. According to recent researches in that area, each of the hemispheres is supervising other parts in our behavior. The left hemisphere supervises the analytical, verbal, logical, sequential processes. It can make arguments, hypotheses, and also make things up – telling a lie. The right hemisphere supervises the holistic, non-verbal, relational processes and responsible for feeling and actual memories. The right brain sees the world as it is, it has no logical explanation for it, and therefore it cannot lie. In this project we’ll try to find a pattern in the behavior of the lobes while lying, and their behavior while telling the truth. We believe, that by performing that, we would be able to recognize lies without being wrong according to the subject state of mind.
Tools
The measurement of the brain arousals is one of the main subjects in this project, and one of the most sensitive too. We are using a sampling machine made by “Norav Medical”. This machine is made for measuring the ECG, but we use it in order to measure electrical signals from both hands. One of the side effects of using this machine is that the ECG signal is added to one of the wanted signals. Our first step was to filter this noise, so that, at the end of the process we’ll have a nice, neat, package of signals for processing. Each subject was presented with a questionnaire, which he had to answer. The questionnaire was consisted of two parts: calibration and examination. The examination part was given to the subject before the inquiry began. The subject had to answer it truthfully and keep it for himself. Then, the inquiry begins, the subject is being connected to the sampling machine, and the calibration part begins. In this part, the subject is asked 10 questions which he has to answer truthfully. Each answer is recorded and then entered to the pool of signals. In the examination part the subject tries to “trick” the machine by lying in some of the questions. By the end of this process, we have a pool of signals, while every one of them is tagged with “Truth” or “False”. These signals are fed to the Matlab for further observation.
Conclusions
The PCA algorithm gives us a fine way to characterize lies. As we saw, there is a mathematical way for distinguishing lies from truth by comparing the arousal of the left brain to the arousal of the right brain. The developed mechanism gave promising results that will have to be tested out on real subjects. We believe that this project will help further researches in the area, and will promote the progress for developing a better polygraph machine.
Acknowledgment
We are grateful to our project supervisor Dr. Danny Lange for his help and guidance throughout this work.
We are also grateful to all the lab stuff for supporting this project.