Students Track Frustration, Fear and Suspicion in Computer Use
The U.S. Air Force is funding a research project at Hamilton to help measure the neurological responses to fear, frustration and suspicion of humans as they interact with computers. This summer students Matthew Farrington ’12, Spencer Gulbronson ’12 and Diane Paverman ’13, led by Stuart Hirshfield, the Stephen Harper Kirner Professor of Computer Science, and Leanne Hirshfield '02, a Hamilton research associate as well as a member of the U.S. Air Force research team, have designed and conducted an experiment to support this effort.
The experiment uses fNIRS, or functional Near-Infrared Spectroscopy, to measure the physiological responses to frustration, fear and suspicion. fNIRS reads brain activity non-invasively through a cap that participants wear. The cap controls lasers and sensors to measure levels of oxygenated and deoxygenated blood in many different areas of the brain, which are considered as indicators of brain activity.
The group designed a novel experiment to provoke each of the emotions of interest in participants over a four-day period. As they performed a straightforward task -- Internet shopping -- the group members observed from another room and the fNIRS machine recorded the subject’s brain data.
On the first two days participants were given a list of bicycle models and were instructed to shop for the lowest price for each model online. On the third day the participants were asked to do the same task, but Hirshfield’s group controlled the internet connection to the computers so that it slowed down significantly, triggering feelings of frustration in the participants, who were working under a time limit.
On the fourth day the group asked participants to find a bike model that the researchers invented. The search for the bicycle led participants to a website that Farrington, Gulbronson and Paverman had built themselves. The site was programmed to automatically download a fake, though convincing, virus which opened pop-ups on the screen and showed the infamous “blue screen of death,” a well known indication of a crash. Group members exacerbated the feelings of fear that this “virus” provokes by pretending when questioned by the subjects not to understand the cause of the problem.
Participants’ brain activity was monitored throughout using the fNIRS, as was the galvanic skin response or the changes in skin’s ability to conduct electricity, using a different device. After collecting their data, they re-formatted it to apply machine learning algorithms to analyze their findings.
The group says that the project has been overall enjoyable, especially because the students were given a lot of freedom in the experiment design. They conceptualized and implemented the experiments with a large degree of independence, though Professor Hirshfield approved the project and provides support throughout the process.
Farrington, Gulbronson and Paverman’s summer research has taught them more about cutting-edge aspects of computer science, including advanced programming and using state-of-the-art neurological equipment. In fact, they developed from scratch all of the software to induce, control and record the “disruptions” they created on the subjects’ machines while recording and synchronizing the brain data from the fNIRS. At the same time, their research is helping to advance current knowledge on brain responses to several common emotions.
Matthew Farrington is a graduate of Averill Park High School (N.Y.); Spencer Gulbronson is a graduate of Thousand Oaks High School (Calif.); Diane Paverman is a graduate of Byram Hills High School in Armonk (N.Y.)