Electroencephalographic responses and brain fingerprinting—we have a warrant to search your brain.

On the leading edge of forensic science, experts are arguing whether ear prints (the impression eavesdroppers might leave when pressing their ears against a door or window) are a useful way of identifying criminal suspects. In the only U.S. court test so far, an appeals panel in Washington state said last year that there's insufficient scientific backing for the technique. Another topic of debate is lip prints; the science jury is still out. But another kind of oral evidence is well accepted: matching bite marks to the teeth of suspected assailants.

Meanwhile, DNA has become investigators' tool of choice both to identify suspects and to clear the wrongly accused. Researchers also continue to press forward with new tools in other well-established areas of forensic science. One, an off-the-shelf package that combines the capabilities of an electron microscope and a computer workstation, runs fast and intensely detailed studies of gunshot residues, the trace elements that mark someone who has recently fired a gun.

There are always new frontiers and novel techniques. One gaining increasing attention concerns the value of memory as objective evidence. A psychophysiologist who has set up shop in Fairfield, Iowa, Larry Farwell, has devised a technique that he contends can tell whether a criminal suspect's brain harbours details of an offense. He calls it brain fingerprinting.

For those who think memory is a notoriously undependable record of anything, the notion seems unlikely. But Farwell's technique is backed by serious science (it has undergone peer review), and it's getting a serious hearing in court.

This past fall, Farwell used brain fingerprinting to determine whether Terry Harrington, convicted in 1977 of murdering a night watchman at a Council Bluffs, Iowa, car dealership, really committed the crime. Or, to use Farwell's terminology, he used the technique to see whether Harrington's brain stores a record of the killing. His conclusion: Harrington's brain does not carry information about the crime that only the killer would know, but it does include data that tends to support his alibi. A judge is studying Farwell's conclusions and is expected to rule this year on whether Harrington should get a new trial.

Farwell's technique involves seating subjects in front of a computer monitor and fitting them with a headband equipped to pick up brain waves. A stream of words and images are then flashed on the screen, including crime details that only the perpetrator should know. Farwell's apparatus is designed to pick up reactions that scientists believe occur only when a subject recognises a significant memory. Farwell says the activity he looks for "memory and encoding related multifaceted electroencephalographic responses" (MERMER)is proof that a record of the event is stored in the subject's brain.

How can anyone be so confident that a single tool can universally tell the guilty from the innocent? Farwell is winningly optimistic in his answer. Imagine, he suggests, sitting in a room with a friend. Then an elephant walks in. "You are going to notice that an elephant walked in because it's information we need to function properly-- we all need to take note of significant information as it arises, and this is true of all sorts of cultures," he says. "Our responses may be different, but the thing we don't have a choice about is taking in the information."