Here's some more info from Wikipedia:
Lie to Me* is an American television midseason replacement that premiered on the Fox network on January 21, 2009. The main character, Dr. Cal Lightman (played by Tim Roth), aided by his colleague Dr. Gillian Foster (Kelli Williams), detects deception by observing body language and microexpressions through the Facial Action Coding System, using this talent to assist clients (such as law enforcement). The character is based on Paul Ekman, notable psychologist and expert on body language and facial expressions. Dr. Ekman and his colleague, Dr. Maureen O'Sullivan, identified "naturals" in what is known as The Wizards Project, which identified 50 people with the ability to spot deception after testing 20,000 people from all walks of life. In real life, they call these "naturals" Truth Wizards, or wizards of deception detection.
The TV show is also fairly educational by the way, it was even a little too unnatural at times, in my opinion. But it's okay. Check it out if you're interested in human nature, as you should be.
Anyway, apart from that, I was actually at a presentation by Paul Ekman once. It was two years ago, at the FMX in Stuttgart. He was in fact just a guest if I remember correctly, but he still showed the audience some interesting footage about microexpressions. The presentation was more about a 3D project based on his work though: Facial Animation Toolset
The toolset basically implements the Facial Action Coding System (FACS) into 3D models. It's quite useful. It enables you to create complex facial gestures with ease. Well if you don't know anything about 3D it's hard to explain why it's as cool as it is, but it's cool. :) Part of its coolness comes from it being a free, open source software though. If you register over there you also get access to a Facial Expression Repertoire and things like that, it's interesting.
We even used the toolset on our music video thingy back then. We weren't able to finish it (long story) but you can still download the finished half of it here: Remembered for Whatever - Prevaricator
Or you can watch a slightly lower quality version too:
If you watch it you'll see a singing tear-like creature in it. To animate his lips and the rest of the facial movements during singing, at first we tried to film and track my face with infrared markers+cam while I pretended to sing, but that didn't work out well since we didn't really have any face tracking experience before. (The mouth was opening and closing but the calibration was too bad and the setup too simplistic. It just looked like a spastic 3D version of South Park style animation.) So we ended up having to hand-animate all the details. In the end the lipsync and the facial expressions of the singer-creature is good enough, I think. But it had other problems and it isn't finished anyway, so who cares... :)
Finally, remember that the best way to ensure that you won't get caught while lying is to be honest. Honesty is also more fun for everybody. :)