Rosalind Picard

Director of affective computing research, M.I.T. Media Lab

Cambridge, Mass.

How did this project start?

We built our first sweatband sensor to get emotional information from autistic children outside the lab. But then one of my undergrad researchers asked if he could borrow a wristband for his little brother who had autism. When he came back, we looked at the data and saw this weird peak on one side of [the boy's] body. It turned out that was 20 minutes before his little brother had a seizure.

How do they work?

These bands put a tiny current through the surface of the skin at the sweatband and measure changes in that electric current. The changes show how the brain is reacting to things in a much more detailed way than, say, a heart rate monitor. Sometimes someone's brain might be reacting, but the person doesn't get sweaty or have a higher heart rate. But he or she will have increased conductance. The bands are also easier to wear than the standard EEG [electroencephalogram] cap, which has lots of electrodes.

How can parents use these sweatbands?

Right now we cannot predict seizures. We also don't know why some seizures cause death and others don't. One hypothesis is that after a deadly seizure is over, the brain shuts itself down for too long. If I had a child with epilepsy, I would want to know: Is this a big surge or a little surge? It's hard to tell that just by looking. If the surge was big and the brain shut down, then I'd want to get my kid to a doctor.

How can doctors use the bands?

Before, there was no way to get lots of seizure data unless people were willing to wear an EEG all the time. The wristband could give us that information, and then we could try interventions