The Washington Post investigates the neuroscience of lying in a recent article on whether new brain-scanning technologies will be able to separate facts from falsehoods.
This technology is of particular interest to governments interested in whether neuroscience can get more reliable information from suspects, and to companies willing to pay to ‘interrogate’ clients about their truthfulness.
The article mentions a company called No Lie MRI Ltd which claims to use “the first and only direct measure of truth verification and lie detection in human history”, which surely must violate any number of laws regarding truthfulness in product advertising – considering that the recent research on fMRI lie detection suggests a poor reliability with current methods.
Presumably, they took their own lie detection test and convinced themselves they were telling the truth.
This is not to say that this technology will develop in the future to be more reliable, though.
This prospect has sparked concern about the potential legal (pdf) and ethical issues of this technology and spurred the American Civil Liberties Union to submit a freedom of information request to the US Government earlier this year to see if they are already using fMRI ‘lie detection’ on terrorist suspects.
Some of the hype around brain-scan lie detection harks back to similar claims that were made for the polygraph tests in the past, despite evidence of their poor reliability and high levels of false positives.
Whether fMRI based lie detection turns out to be anything other than a similarly unreliable detection method (but with prettier pictures) remains to be seen.
Nevertheless, one method which does seem to be generating a lot of interest is the Guilty Knowledge Test (pdf), which relies on the fact that the brain tends to produce reliably different automatic responses for items that are recognised compared to items that aren’t.
The idea behind this is that you could show items to suspects that were taken from the ‘crime scene’ and look for the traces of successful recognition measured from the brain.
This technique is now reliable enough that it is starting to be admissible in court. The success of this technique has given researchers hope that successful lie detection may be possible for more than simple recognition situations.
Nevertheless, as every good conman knows, the best lies have a kernel of truth and it’s not clear how well these techniques will detect economies of truth when compared to outright whoppers.
Link to article ‘Brain on Fire’ from The Washington Post.