Slashdot Mirror
NASA Plan to Read Brainwaves at Airports
cascino writes: "In one of the more bizarre (and intrusive) spinoffs of the Government's 'crackdown on terrorism,' Officials of the National Aeronautics and Space Administration have told Northwest Airlines security specialists that the agency is developing brain-monitoring devices in cooperation with a commercial firm, which it did not identify. Space technology would be adapted to receive and analyze brain-wave and heartbeat patterns, then feed that data into computerized programs 'to detect passengers who potentially might pose a threat,' according to briefing documents obtained by The Washington Times." This is the second story recently that gives me second thoughts about flying Northwest.
...to test brainwaves, because it's obvious that normal travelers (being delayed by extensive security measures) are never stressed-out or homocidal. Especially if they're made to stop for one more scan by minimum wage federal employees that aren't doing jack squat ANYway.
GREAT IDEA. I feel safer already.
-Styopa
Of course you are correct.
To suggest we have any technology today that can infer a person's thoughts is ludicrous. Even at a coarser level, to suggest that a momentary detection of brainwaves can be reliably correlated with some "bank" of known EEG signatures which indicates the disposition or identity of the subject is fantasy.
The weakness and noisiness of scalp potentials cannot be overstated. The devices we use in our lab are state of the art, but even these require a sophisticated multi-electrode cap, each electrode carefully primed with an electrolytic gel, and fed into an extremely sensitive amplifier, while the subject sits in a completely electrically isolated room (basically, a glorified Faraday cage).
And even when *all that* goes well, the data you collect is extremely noisy due to the inherrent conflation of *billions* of neurons all contributing to the recorded potentials. The solution is multi-event averaging. We give subjects 100s of trials, and only after tedious signal processing and averaging can we extract the gross electrical activity associated with a particular cognitive act ("event related potential").
And to suggest that we (cognitive scientists) have some sort of repertoire of electrical signatures mapped to "thought patterns" is preposterous. The best is the suggestion that particular waveforms are associated with "orienting" or "error-making" or "perception" or "novelty." Most serious scientists work hard to localize these signatures to particular brain structures (a whole industry unto itself) rather than wonder if these tiny effects can tell us about a person's hidden agenda.
Much has been made in the popular press about a particular waveform called the P300...a characteristic "positive-going" wave occurring around "300" milliseconds post stimulus onset. This waveform has been associated with attending to a novel stimulus. Some people have suggested using this waveform as a sort of ERP "lie-detector" using the following flawed thinking: If you show a suspect scenes from a crime, if they are novel (new to the suspect), they'll elicit a P300. If they are not surprising (indicated by a *lack* of P300), then the guy's seen the scene before and is guilty. I won't even begin to address all the problems with this "guilt by failure to disconfirm" approach...I'm sure you all are bright enough to see the logical holes, much less the technical and cogntive-theoretical problems.
Anyway...no, some guy passing through a gate, and some gee-wizardry fingering him as a terrorist-like baddy? Only in Ashcroft's wet dreams for now.