The cartoon archetype of a drunk person is a disheveled mess, with droopy eyelids, an erratic gait, and bubbles coming out of his head – for whatever reason. If only it was so easy to tell if someone’s lost their mind. If a cop stops you, he can objectively determine the poisoning on the side of the road only with a breathalyzer: in the lungs, ethanol is transferred blood in the air, so that the device can detect alcohol in your exhalation. Even then, one person on the U.S. federal legal driving limit of 0.08 Breath Alcohol Concentration (BrAC) could act perfectly normal, while another person would unsuccessfully attempt to fish for a slice of pizza. in a gutter.
But scientists are working on what may be a new way to determine intoxication, reverting to a stereotypical feature of the drunkard that is actually true: that soused walk. No matter how hard you think you walk when inebriated – especially if you compare yourself to your friend in the gutter – subtle and not-so-subtle changes in your gait could betray your alcohol levels. And if you’re carrying a smartphone, its built-in accelerometer can pick up on these changes. In fact, scientists at the University of Pittsburgh just published research showing that, at least in the lab, they can use smartphone motion data to detect whether a subject is intoxicated, with an average accuracy of 93%. It sounds like fun and games – charging people up and watching them stumble for science – but the work could have some serious use.
Their experiment began with a drink surprisingly sophisticated for the subjects: vodka with lime juice and simple syrup, loaded with vodka. The subjects had an hour to finish the drink steep, as the researchers wanted to get them to a peak of .20 BrAC fairly quickly, of course, safely. “We really wanted to have enough data points above the legal alcohol limit, which is 0.08,” said lead study author Brian Suffoletto, then at medical school in the University of Pittsburgh and now Stanford University. “So if you give them a lower dose, you may only get a point or two in time before the body metabolizes the alcohol.” Plus, there’s only a limited number of hours in a lab day – it doesn’t make sense to wait around for subjects to quietly create a buzz.
Before drinking, the researchers strapped regular smartphones to the subjects’ lower backs to get a feel for their basic gaits. They asked participants to walk 10 steps forward, turn around, walk and 10 steps back. The subjects then repeated this drunk leg, doing it once an hour for 7 hours as their BrAC went up and down. All the while, the smartphone was recording motion data.
Because each participant had their own restrained gait and unique intoxicated gait, Suffoletto and her colleagues used individualized mathematical models that allowed them to compare each person to themselves. “We’ve found that, really, what drives the model is the mid-lateral sway,” says Suffoletto. It means moving side by side; your typical sober gait is more or less across the board. “Which makes sense when you think of the drunk cartoon caricature, a character that sort of swayes back and forth,” he continues.
The individual gait model of each person allowed researchers to correctly identify more than 90% of the time a subject’s BrAC exceeded 0.08. Of course, in the real world, most people don’t strap their smartphones to the lower back, but researchers plan to do other experiments with subjects carrying the devices in their hands or pockets.