Of course, neither the families of older people nor their doctors should have to be glued to their computer screens, watching and evaluating every bit of data coming from the motes in order to care for the elderly. "You don't want to get an e-mail every time a cabinet door opens in your mom's kitchen," says Dishman. "But you do want it when she's not drinking enough, after 30 years of having three cups of coffee or tea a day." It would take a large degree of intelligence, but software in a home PC could analyze the information coming from the motes, recording the routine and acting on the alarming.

Prevention of disease comes from unglamorous things like encouraging people to eat properly and to exercise, changes in habit that are famously hard to effect. The single best aid to such behavior modification is a support group, and there is some evidence that such groups can be bound more tightly together with technology. "[Motes] could let your walking buddy know you've got your walking stuff out and maybe suggest opportune times to go walking together," says Dishman.

Using wireless sensors to track the routine activities of daily life—how people dress, what they cook in the morning, how well they drive their cars—might make for a mundane set of data. But such data can greatly help in the diagnosis of neurological disorders such as Parkinson's or Alzheimer's disease. Parkinson's can so far be diagnosed only through behavioral changes, principally changes in gait.

Yet to get a really early diagnosis would require noticing quite subtle distinctions. Experts have studied videos from Michael J. Fox's sitcom "Spin City" to see how the actor behaved in the period before he was diagnosed with Parkinson's. These videos, which constitute a rare treasure trove, appear to show the length of Fox's stride getting shorter over time. Dishman contends that a set of behavioral markers for diagnosing Alzheimer's should emerge from such data as well. "Most people are not diagnosed until they have already had the disease for five or even 10 years before they either acknowledge or realize it," he says. The markers might show the onset of Alzheimer's well before that, in time for drug therapies to have an effect.

Wireless networks can also bring a measure of independence to the lives of people suffering from cognitive decline. Intel set up a prototype network of sensors inspired by the needs of Barbara, a patient with Alzheimer's disease whom Dishman recently met in a field study. The network used off-the-shelf motion sensors to detect Barbara's movements, pressure sensors in chairs to tell whether she was sitting in them, contact and magnetic switches to sense the opening of drawers and cabinets in the kitchen, radio frequency identification tags in her shoes, and antennas to sense the tags when she entered the kitchen.

All this was for a seemingly simple task: to make sure Barbara stayed hydrated, a common problem for those with Alzheimer's. The system, connected to a PC, can tell how long it has been since anyone was in the kitchen. If it's been too long, the unit can prompt Barbara to go get a cup of tea through a text message on the TV in whatever room she may be in. It then follows Barbara's progress as she tries to make tea. If the system notices she is having trouble—say, if she is in the kitchen but is taking an unusually long time to open the cabinet—it prompts a TV in the kitchen to ask her if she needs help. If she says she does, the system can then walk her through whatever problem she is having, monitoring her progress on the way.

A similar system can help patients comply with doctors' orders. It has been estimated that in the United States about 45 percent of people over 75 take multiple drugs. And according to a study at Johns Hopkins University, in Baltimore, as patients take more drugs, they make more errors in taking those drugs.