Mindless Computing Technology for Health
by Caitlin Hayes
Bad habits are stubborn, and our best attempts to change our behavior often fail us. The technology sector has tried to help with a slew of apps and devices, but in many cases the use of these gadgets or programs requires as much or more effort as any other intervention. We may not even try. If we do, we may give up after a few weeks or months.
Tanzeem Choudhury, Cornell Computing and Information Science, designs technology that can track our behavior almost invisibly, without user effort, and provide feedback that either gives us more awareness or actually prompts instinctive changes. She wants to create mindless computing technology that taps into our psychology and biology to improve wellbeing — a device that affects the perception of our heart rate to make it seem lower when it becomes elevated, or a program that tracks our social activity, sleep, and phone usage.
“How we live our lives has so much bearing on our health, but there are not great measures of lifestyle,” Choudhury says. “We look at how we can use sensors and mobile devices to study people’s behaviors related to health. Then we ask, how can we design technology to influence those behaviors, to improve quality of life?”
A main focus of Choudhury’s People-Aware Computing Lab and her startup, HealthRhythms, is to design technology for those suffering from mental health issues. This population may be most sensitive to disruptions in behavior and lifestyle. “We’re looking at how technology can help in both tracking and management and also clinical care,” says Choudhury. “The idea with HealthRhythms is really, how do we get these technologies into hospital systems and to health providers and to the people who need them?”
DoppleSleep, BodyBeat, Nutrilyzer — Tracking the Quality of Food and Sleep
Before Choudhury’s technologies can impact behavior, they first have to be able to track it. In her early research, Choudhury developed technology to track activity. She showed how measuring barometric pressure and acceleration could be used in a wearable device to determine whether a user is going up or down stairs or biking up or downhill, a method Apple has recently incorporated into a number of its devices.
Choudhury has now moved on to bigger challenges. “We have a pretty good handle on people’s locations, activity, a little less on the quality of sleep and very little sensor capability around food,” she says. “So some of our projects are asking how we can measure these things better, and make measurement almost effortless for the user.”
Two of Choudhury’s students used the concept of Doppler radar to develop DoppleSleep, a bedside device that can measure heart and breathing rate changes and movement in a completely contactless way. The device provides more objective information than self-reporting and eliminates the work of keeping a sleep journal or log.
A number of Choudhury’s projects also try to address the dearth of sensing methods surrounding our eating habits. BodyBeat is a wearable device fitted with a specially designed microphone that rests on the throat. The microphone amplifies internal body sounds. “You don’t get details of exactly what someone’s eating, but you can look at whether someone’s gulping, sipping, crunching — the type of food they’re eating and how many times they’re eating,” Choudhury says.
Choudhury’s group also worked with the lab of David C. Erickson, Mechanical and Aerospace Engineering, to develop Nutrilyzer, which uses photoacoustic principles to determine the quality of foods. The device characterizes food by shining light onto it and measuring how it vibrates. This could be used to detect pesticide residues, the concentration of alcohol, or whether food has begun to spoil.
Choudhury is even developing a watch-like device that uses infrared light to measure alcohol concentration just under the skin, a tool that could have an impact on college campuses. “Where the measurement hasn’t been great, we’re asking how we can develop it?” she says. “So you get the measurement,” she continues, “how can you then design technology to influence behavior?”
EmotionCheck for Healthful Behavioral Change
Choudhury is particularly interested in technology that can promote effortless behavioral change, or instinctive change. “We have instinctive and sometimes biological responses to things, and then we have conscious responses,” she says. “Very few technologies actually exploit these instinctive and biological responses.”
EmotionCheck, a technology that Cornell University’s Center for Technology Licensing is helping to patent, is a watch-like device that monitors and modulates heart rate. When you become stressed and your heart rate increases, the wristband provides a subtle vibrating pulse that’s slower than your actual heart rate. “Because it’s mimicking a slower heart rate, we saw that actually people become less stressed,” Choudhury says. “We found that to be true in a study where we induced stress, and now we’re going to have people use it in their daily lives.”
Choudhury’s group also designed a plate with special sensors and LED lights. When food is placed on it, the plate can make it look like there’s more or less food than there actually is. “We can have these technologies where instinctively our body or brain reacts to them in a way that helps us make positive changes,” Choudhury says.
Technology for Monitoring Mental Health
When Choudhury was in graduate school at Massachusetts Institute of Technology, she was already interested in improving human health and wellbeing by combining technology with an understanding of people. When a fellow graduate student and friend took his own life, she extended her interests to helping those who are suffering. “It was an issue nobody really talked about,” she says.
For those with mental health disorders or diseases, keeping a set routine — getting enough sleep, eating meals around the same time each day, and socializing enough — can reduce the severity of symptoms. Choudhury’s group has developed multiple smartphone applications that help patients and clinicians track these behaviors.
“We’ve actually deployed some of our sensing in people who are suffering from bipolar disorder, schizophrenia, and depression,” Choudhury says. The feedback Choudhury has received from these trials has been invaluable to the evolution of the technologies.
“I often say we have this adoption gap. The health sector has this clinical need, and we have technology, and we think we can throw it over, and they’ll run with it. But it requires so much back and forth,” Choudhury says. For instance, during a yearlong clinical trial with schizophrenic patients, clinicians were resistant to receiving the huge amount of data Choudhury’s group could provide. It was overwhelming. “They didn’t know what to do with it or how to take clinical action from the data,” she says.
“There’s no blood test for mental health, and it’s such a big issue with huge costs. These technologies can change things.”
Choudhury’s team then developed output clinicians could use, analyzing data from the applications to calculate scores for the Brief Psychiatric Rating Scale, an established metric that measures psychiatric symptoms. The score is typically collected using a self-reporting survey. “Now clinicians can get this information much more easily,” Choudhury says. “They don’t have to rely on the patients reporting it, which is key because often patients who are doing poorly are less likely to respond. This is how we’ve made life easier.”
Similarly, with bipolar patients, data from Choudhury’s smartphone sensors can be used to determine a Social Rhythm Metric score, which measures the stability of a patient’s routine and is also determined through self-reporting. “We’re trying to make the measurement more objective and continuous,” Choudhury says.
“There’s no blood test for mental health, and it’s such a big issue with huge costs,” she continues. “These technologies can change things. But working together is so important. Showing the potential of the technology and how it can transform things, and learning where the roadblocks are for the patients or clinicians will allow us to find solutions that are actually useful.”
Building hardware, performing data analytics, working with clinicians and patients — Choudhury says this work requires time, and the flexibility and freedom that Cornell offers. “In Information Science, we each have our area of expertise, but it’s interdisciplinary. We have the freedom to pursue the problem we’re solving with whatever disciplinary knowledge we need to bring to bear,” she says. “That’s very intellectually stimulating. That’s very exciting.”