Walter Karlen and his colleagues in Lausanne, Switzerland have published an interesting article in this month's IEEE Transactions on Biomedical Circuits and Systems about a wearable device that could be used to monitor drowsiness. Basically, they took a shirt and mounted various detectors on it. A belt running around the ribcage measures respiration; gel electrodes fixed to the pectoral area read out an electrocardiogram; and other electrodes detect muscle activation. The whole thing runs off a battery that's also mounted right on the shirt.
The approach combines a lot of the techniques traditionally used to determine whether a person is awake or asleep. Normally, however, the subject would have to be hooked up to a bunch of external machinesâ''the kind we usually only find in hospitals. The shirt would make it possible to monitor people, like truckers and pilots, whose occupations put them at a high risk of becoming drowsy.
One interesting aspect of the article is the creative way which the experimenters chose to deal with data artifacts. Cardiorespiratory signals tend to contain a lot of errors that get introduced when the subjects makes random movements. Instead of finding ways to subtract this data from their results, the group used it as a measurement in its own right, reasoning that if the person is moving they are more likely to be awake.
When tested, the device was able to accurately distinguish between sleep and wake states after a period of calibration. The problem is that falling asleep is not like flipping a switchâ''especially when you are trying to fight it back. As it is, the system canâ''t detect a gradual descent from consciousness. But you can imagine that if it were hooked up to an alert mechanism that jolted drivers or pilots back into focus (such as this one), the idea could save some lives.