Neurotechnology for Intelligence Analysts

Today, I met with a Project Manager (Dr. Todd Hughes) and Science and Engineering Technical Adviser (Dr. Sean Deeny) from the Defense Advanced Research Projects Agency (DARPA), which is the main / public research branch / science and technology funding arm of the Department of Defense. We discussed the Neurotechnology for Intelligence Analysts (NIA) program, field deploy-able derivations of that program, and the Cognitive Technology Threat Warning System (CT2WS).

These programs all measure and utilize the P300 wave response in the brain, which was discussed in the previous post. The P300 wave response is an electrical signal that rises in the brain in response to recognition of 'objects of significance'. The response is uber fast - the 300 in P300 stands for 300 milliseconds. Dr. Larry Farwell is credited as the 'discoverer' of the practical applications of this response. His Brainwave Fingerprinting technology uses P300 diagnostics for criminal justice and credibility assessment applications.

The NIA program uses the P300 wave to accelerate Imagery Intelligence (IMINT) analysis. IMINT analysts pour through large amounts of photographic imagery, and scan this imagery for 'objects of significance', which can mean Improvised Explosive Devices (IEDs) or other. Usually, the IMINT process is slow and tedious. An analyst has to crawl through piles of data, scanning every bit more or less slowly, virtually in search of needles in a haystack. The NIA program speeds up this process, a lot. An analyst wears an EEG cap and this cap measures the P300 response while the analysts are doing their IMINT tasks. If the analyst sees an 'object of significance' then software automatically tags that IMINT bit as relevant, and it is later given further scrutiny. Using this methodology, analysts can course through 10-20 images per second . There is no conscious scanning of the information - the entire process occurs subliminally. Pretty amazing!

One derivation of this technique is its use in the field for live aerial surveillance. Basically, an analyst is hooked into an aircraft surveillance feed and uses the same subliminal procedure to scan the environment for said objects of significance. This kicks the shit out of the conventional manual 'eyeball' method and bootstraps cognitive video analysis (real-time video analysis that looks out for pixel-related changes).

Another derivation is use in conjunction with night vision, binoculars, or other visual-field enhancing devices. The unconscious mind of the warfighter can rather intuitively sense danger. CT2WS harnesses this 'intuitive' impulse (which is characterized by the P300 response) to alert warfighters when there is potential danger in the dojo. Maybe there is an IED or a sniper over there. The warfighter's (un)conscious mind(s) pick up on that danger. The warfighter's personal level of sensitivity to that very quick threat perception impulse can determine life or death. By integrating P300 neurotechnology with visual field devices, lives can be saved and battlefield intuition can be enhanced.

The major challenges for this technology is a problem of cumbersome hardware. Standard EEG systems do not work so well in 'noisy' (i.e. physical movement-heavy) environments, and they also require the application of conductive liquid (i.e. EEG gel) in order for the electrodes to pick up on the electrical signals of the brain. While companies are working on 'dry EEG' technology, as well as algorithms for processing EEG signals in noisy environments, the technology is just not there yet. This is the major obstacle from research and development to the actual field deploy-ability of such devices.

The minor challenge for this technology is user-friendliness. For instance, in the NIA program analysts get bored. Basically, they are staring at an apparently blank screen (images are being played at the rate of 10-20 per second, which more or less gives the appearance of a subtle strobe light, hardly offering any viewing pleasure). As far as the functioning of this technology depends on maintaining the attentional capabilities of the analyst, the process has to be a lot more engaging than it is now. Attempts have been made to get at this 'engagement factor' by having the analysts perform some type of 'routine' task during 'break time'.

This brings up an interesting point that the P300 response is very fast, so as to be subliminal, but its occurrence is relative to the brain functioning of the subject in question. For instance, an expert's brain fires the P300 a lot faster than a non-expert. Somebody with ADHD is not going to do very well at a P300 task. (Elsewhere, P300 response has actually been used as a measurement of ADHD). So the unconscious mind may be doing the light lifting, but it can only do such lifting so far as there is adequate 'cognitive infrastructure' in place.

Therefore, at DARPA, I posed the question: perhaps there is a signature in the brain (e.g. cross-cortical alpha-wave coherence), which denotes 'being in the zone'. Such a signature is representative of the 'cognitive infrastructure' necessary to keep the subject mind-machine interfaced (engaged in the P300 task). If you know the electrical signature that needs to be maintained, then why not engage the conscious mind of the subject in 'zone maintenance' via neurobiofeedback while his/her unconscious is going to town on the analysis of imagery intelligence.

That comment raised a couple eyebrows, so don't be surprised if you see this in the next DARPA NIA RFP at FBO.GOV! ;;)