CHARLESTON, S.C. (AP) - Picture this: Your boss is threatening to fire you because he thinks you stole company property. He doesn't believe your denials. Your lawyer suggests you deny it one more time — in a brain scanner that will show you're telling the truth. Just the other day I lay flat on my back as a scanner probed the tiniest crevices of my brain and a computer screen asked, "Did you take the watch?"

The lab I was visiting recently reported catching lies with 90 percent accuracy.

And an entrepreneur in Massachusetts is hoping to commercialize the system in the coming months.

"I'd use it tomorrow in virtually every criminal and civil case on my desk" to check up on the truthfulness of clients, said attorney Robert Shapiro, best known for defending O.J. Simpson against murder charges.

Shapiro serves as an adviser to entrepreneur Steven Laken and has a financial interest in Cephos Corp., which Laken founded to commercialize the brain-scanning work being done at the Medical University of South Carolina.

That's where I had my brain-scan interrogation. But this lab isn't alone. Researchers at the University of Pennsylvania have also reported impressive accuracy through brain-scanning recently. California entrepreneur Joel T. Huizenga plans to use that work to start offering lie-detecting services in Philadelphia this July.

His outfit, No Lie MRI Inc., will serve government agencies and "anybody that wants to demonstrate that they're telling the truth," he said.

Both labs use brain-scanning technology called functional magnetic resonance imaging, or fMRI. It's a standard tool for studying the brain, but research into using it to detect lies is still in early stages.

Nobody really knows yet whether it will prove more accurate than polygraphs, which measure things like blood pressure and breathing rate to look for emotional signals of lying.

But advocates for fMRI say it has the potential to be more accurate, because it zeros in on the source of lying, the brain, rather than using indirect measures. So it may someday provide lawyers with something polygraphs can't: legal evidence of truth-telling that's widely admissible in court. (Courts generally regard polygraph results as unreliable, and either prohibit such evidence or allow it only if both sides in a case agree to let it in.)

Laken said he's aiming to offer the fMRI service for use in situations like libel, slander and fraud where it's one person's word against another, and perhaps in employee screening by government agencies. Attorneys suggest it would be more useful in civil than most criminal cases, he said.

Of course, there's no telling where the general approach might lead. A law review article has discussed the legality of using fMRI to interrogate foreigners in U.S. custody.

Maybe police will use it as an interrogation tool, too, or perhaps major companies will find it a cheaper than litigation or arbitration when an employee is accused of stealing something important, other observers say.

For his part, Shapiro says he'd switch to fMRI from polygraph for screening certain clients because he figures it would be more reliable and maybe more credible to law enforcement agencies.

In any case, the idea of using fMRI to detect lies has started a buzz among scientists, legal experts and ethicists. Many worry about rushing too quickly from the lab to real-world use. Some caution that it may not work as well in the real world as the early lab results suggest.

And others worry that it might.

Unlike perusing your mail or tapping your phone, this is "looking inside your brain," Hank Greely, a law professor who directs the Stanford Center for Law and the Biosciences, told me a few days before my scan. It "does seem to me to be a significant change in our ability ... to invade what has been the last untouchable sanctuary, the contents of your own mind," Greely said. "It should make us stop and think to what extent we should allow this to be done."

But Dr. Mark George, the genial neurologist and psychiatrist who let me lie in his scanner and be grilled by his computer, said he doesn't see a privacy problem with the technology.

That's because it's impossible to test people without their consent, he said. Subjects have to cooperate so fully — holding the head still, and reading and responding to the questions, for example — that they have to agree to the scan. "It really doesn't read your mind if you don't want your mind to be read," he said. "If I were wrongly accused and this were available, I'd want my defense lawyer to help me get this."

So maybe the technology is better termed a "truth confirmer" than lie detector, he said.

Whatever you call it, the technology has produced some eyebrow-raising results. George and his colleagues recently reported that using fMRI data, a computer was able to spot lies in 28 out of 31 volunteers.