–Written by Sarah Falkovic–
Welcome to the update on my lie detector project!
In my previous blog post, I discussed the modality of electrodermal activity, a.k.a our unconscious sweat response and its ties to stress. However, as a method of measuring lies it leaves much to be desired due to its ability to be swayed by nerves, increased blood pressure, and other factors such as the weather. This has led me to the world of EEG (electroencephalogram) responses- a type of test that measures electrical activity of the brain by placing electrodes on the head.
The specific type of EEG response I am working with is called a P300 response. Why this name? Because it is a positive reading that is initiated at around 300 ms after the stimulus. Known as an ‘I saw it!’ response, P300s are typically used in what is called an ‘oddball’ test, where a participant is shown images, and an unexpected ‘oddball’ image illicites a P300 response where neutral information may not.
So what does this have to do with the forensic goals of the polygraph and other lie detectors? P300s have been used to help determine neutral versus significant information relative to the subject. For example, you can show true and untrue information about a crime to a subject and potentially determine if those features elicit a P300 response compared to the neutral unknown details, which would indicate a special knowledge about the crime.
While this doesn’t determine false versus true information as a typical polygraph would, it is ideal for determining if a person finds personal significance in a crime beyond an average uninvolved innocent suspect.
The real problem this test can solve is the issue of what is known as saliency. Typical polygraphs only measure a stress response, which isn’t always tied to crimes- suspects may not really care if they have committed a crime and thus may not be as stressed while lying. However, the P300 is more about visual detail recognition, which tends to increase for more relevant stimuli.
An example of this would be to present similar answers to a question about a crime where only one is true- a person implicated in the crime would elicit a P300 for that true detail alone. It can be much more difficult to hide this sort of recognition signal compared to the general stress exam that is the polygraph. Here is what an average P300 looks like compared to a normal neutral EEG response:
Ideally, there is quite a dramatic increase in EEG in the P300 compared to the non P300. To test our equipment’s abilities, I put together a base experiment activity I like to call ‘Find the Dot’. Subjects were shown a randomly flashing screen of blue and yellow dots using the Wio Terminal, and were asked to count the number of yellow dots they saw until they reached a count of 50 dots. Here’s what a subject would see while in the experiment:
Despite this being a simple experiment, there is certainly evidence of the “I see it” response! I took about 60 minutes of data from 3 trials to produce the graph below averaging responses where we would expect a ‘oddball’ P300 response (i.e. when the yellow dot is shown) compared to the neutral stimulus (blue dot). The blue line is the average response for P300, and there is certainly a large difference between it and the neutral response, especially directly after the 300 ms line.
This experiment helped me realize that there is likely a slight lag between the screen and my recording equipment that is shifting my P300 response into something more like a P600 response, taking about 600 ms to flash!
My current goals are to retroactively fix what data I have collected to take this ‘lag’ into account and collect some data from a more salient source. Where the P300 shines is in its testing of visual stimuli. Comparatively, P300s are much bigger for personally relevant visual information compared to irrelevant or neutral information.
To bring this subject of relevance into my experiments, I am preparing some tests involving the fellow’s names and their pets! (Hopefully some cute animal pics will be incoming in a future update!) Long term, if those tests can also elicit a P300, I should be able to move forward with some crime simulations- likely either mock theft or shooting a confederate (a specialist in on the research) with a Nerf play gun.
P300s are a new foray into the world of personal relevance testing, despite being discovered in 1965. Hopefully this reinvention will further cement the polygraph’s place as a moment in history and not an accurate form of lie detection!