Biosensor System Detects Human Emotion

If someone asks whether you’ve been under a lot of stress lately, you probably don’t have much trouble answering.

Emre Ertin, a research assistant professor in electrical and computer engineering, leads an Ohio State team developing StressWare, a suite of sensors integrated into AutoSense, an unobtrusively wearable wireless sensor system for continuous assessment of humans in their natural environments. The system shows more than 84 percent accuracy compared to human subjects’ perceptions of their stress levels.But what if the question could be answered for you automatically — and help you decrease that stress in your life?

AutoSense, supported by the National Institute on Drug Abuse as part of the Genes Environment and Health Initiative at the National Institutes of Health, involves computer science, electrical engineering, behavioral science, physiology and biochemistry researchers from Ohio State and the universities of Memphis and Minnesota.

“AutoSense focuses on physiological signal measurements of the body’s cardiovascular, respiratory and thermoregulatory systems, since these are most affected by both psychologically and physically demanding conditions,” Ertin says.

StressWare, Ertin explains, has a comfortable, unobtrusive chest band that can be worn outside of the lab environment for extended periods of time. It will enable real-time, multimodal stress index calculations; have a long lifetime through its low-power sensor design; and reduce data bandwidth with its distributed signal processing.

Each sensor in StressWare, along with the associated bio-amplifiers and signal conditioning circuitry, is integrated into the chest band with two wireless motes. A third mote on the chest band transmits readings to a Bluetooth receiver on a mobile smartphone, which then sends data to an external database over Wi-Fi or a GSM cellular network. The sensors measure electricity of the heart and galvanic skin response, a measure of conductivity and heat in skin affected by nerves and sweat; relative lung volume and breathing rate; skin temperature; and general activity or rest information for the subjects.

Ohio State engineering alumnus Santosh Kumar, ’02 M.S. and ’06 Ph.D. CIS, a computer science assistant professor at the University of Memphis, leads the AutoSense project. While Ertin and his students are working on developing wireless sensors for AutoSense, Kumar and his colleagues at the University of Memphis do field studies and create the phone software that processes the behavioral inferences from the sensor measurements. Researchers at the University of Minnesota Medical School under the direction of Mustafa al’Absi, a professor of behavioral medicine and world renown stress expert, conduct laboratory testing and field studies.

The team envisions uses of the technology such as restricting potentially stressful phone calls at inopportune moments; enabling subjective sensing, perhaps recording the change in emotion associated with various activities and then making recommendations, such as restaurant choices; and making mobile devices become affective machines, even adjusting game levels based on real-time changes in emotion or playing music appropriate for or to change a person’s mood.

“In the long run,” Ertin says, “sensors like these will be integrated into the health care system to study factors affecting health and will be used in intervention for the early stage of cardiovascular disease.”

Contact:

Emre Ertin, (614) 688-3928, ertin.1@osu.edu

On the Web: AutoSense project, sites.google.com/site/autosenseproject/