posted 02-16-2008 02:31 PM            

Here's the abstract:

quote:

---

Few data are available to address whether the use of ERP-based deception detection alternatives have sufficient validity for applied use. The present study was designed to replicate and extend J. P. Rosenfeld, M. Soskins, G. Bosh, and A. Ryan's (2004) study by utilizing a virtual reality crime scenario to determine whether ERP-based procedures, including brain fingerprinting, can be rendered less effective by participant manipulation by employing a virtual reality crime scenario and multiple countermeasures. Bayesian and bootstrapping analytic approaches were used to classify individuals as guilty or innocent. Guilty subjects were detected significantly less frequently compared to previous studies; countermeasures further reduced the overall hit rates. Innocent participants remained protected from being falsely accused. Reaction times did not prove suitable for accurate classification. Results suggested that guilty verdicts from ERP-based deception detection approaches are likely to be accurate, but that innocent (or indeterminate) verdicts yield no useful interpretation in an applied setting.

---

Jumping to the discussion, guilty participants (no CMs) were correctly classified 27 and 47% of the time, depending on what type of analysis was used. With CMs, 7 to 27%.

I haven't read what the CMs were - and it'll be a while. My wife has a list.