A study published this week reports that a machine known as magnetoencephalography can correctly identify PTSD 90% of the time.  Researchers at the VA in Minneapolis, Minnesota used technology that has been studied for years: imaging the activity of the brain using magnetic sensors.  Wikipedia has a nice article about the technique, but it should be noted that the Wikipedia article about Magneto the comic book character is probably longer.

We already have the capability of measuring electrical activity in the brain: EEG.  

The journal was the Journal of Neuroengineering, according to Medgadget.com.  The journal's first and only issue is published February 2010.

To my knowledge, there has not been, ever, any biological correlate of PTSD (or for that matter, any psychiatric diagnosis).

To say this is outstanding news is an understatement--if we can truly use brain scans to diagnose PTSD, will we be using brain scans to diagnose depression, or schizophrenia?

The researcher they interviewed on Minnesota Public Radio throws around terms like magnetoencephalography and supercomputer.  He essentially says the patient sticks his head in a contraption, stares at a spot for a minute, the data are fed into a supercomputer, and then they get a diagnosis.

This sounds suspiciously like the researchers Petricoin and Liotta who went around the country a few years ago saying they could diagnose ovarian cancer with a drop of blood and a supercomputer.  This turned out to be unsupportable by the underlying science.

The FDA eventually warned LabCorp, who was gearing up for a launch in 2009, not to market the test.  As of January 2010, we still do not have a commercial test based on the original world-beating announcement in 2002.

So I love stories about weird contraptions that can save the world as much as the next guy, but extraordinary claims require extraordinary proof.  Can other researchers replicate the data?  Are we entering a new phase of psychiatry?  I hope so!