The week of November 1, 2015

How the Internet is changing the face of psychological research

By Cynthia McKelvey

In May 1996, just as the Internet was becoming mainstream, Bill Choisser began posting to online forums for patients and families with neurological disorders like Asperger’s syndrome, asking if anyone else had trouble recognizing faces, even of friends and family. He got a reply, which convinced him of a long-held suspicion: He wasn’t just bad with faces—he was blind to them. And he wasn’t the only one.

Choisser has a condition called prosopagnosia—or face blindness (a term he claims to have coined). Where most people have no trouble finding their friends in a crowded restaurant, Choisser would walk right past them. Through the Web, Choisser became a contact point for face-blind people.

His online community caught the eye of a young researcher, Brad Duchaine. Looking for a topic for his dissertation, Duchaine reached out to Choisser and together they developed a new, more effective test for the condition. They called it the Cambridge Face Memory Test and did something peculiar for the era: They put it online.

The test was one of the first Web-based neuropsychological tests that researchers used to collect data and find study subjects with unusual disorders. At the time, it was a bit of a gamble; initially, the scientific community didn’t embrace this new, online way of doing things. There were questions and doubts—in fact, those questions and doubts remain. But over the past two decades, Web-based experiments have proven a legitimate research tool and, especially regarding rare and little-understood diseases, have led to advancements in knowledge that otherwise might never have happened.

• • •

Choisser says he realized his condition only gradually. It started by realizing his memory for faces was worse than normal. After cultivating his online community, he created a website called Face Blind! that collected everything he and his online friends had learned about their condition, through the sparse medical literature on the subject and through talking to one another.

Before Choisser and Duchaine teamed up, face blindness was thought to be an exceedingly rare symptom of a brain injury. “If you look back in the neurological literature, you can find reports of acquired prosopagnosia [dating] to [the] mid-19th century where people suffered some kind of brain damage and were no longer able to recognize faces,” Duchaine said. “Up until the ’60s or ’70s, there were no reports with people with developmental [i.e., not acquired] prosopagnosia.”

Choisser’s online group proved face blindness was more common than researchers thought. But it’s still relatively rare, and face-blind people are often difficult to find since many aren’t aware anything is wrong. Choisser and Duchaine both said many face-blind people feel they’re just bad with faces. To help recognize their friends and family, they compensate by looking for other cues, like hair and clothing.

“In the neurological literature you can find reports of acquired prosopagnosia [dating] to [the] mid-19th century where people suffered some kind of brain damage and were no longer able to recognize faces.”

That coping mechanism made prior tests for face blindness less accurate. Previous exams showed subjects photos of people and then asked them to recognize the people in the photos later. But the photos were always of the full bust, including clothing and hair. With the help of Choisser, Duchaine made a simple fix: take out the hair and clothing and show people only their faces. That made it easier to spot and diagnose the problem.

Face Blind! links to the Cambridge Face Memory Test and its cousin, the Famous Faces Test. Through them, Duchaine finds new test subjects. People who score poorly on the face recognition tests may receive an invitation to the lab for further testing. Then the diagnosis can be confirmed through rigorous testing that includes brain scans.

One of Duchaine’s research assistants, Laura Germine, advanced the online platform one step further. Instead of finding people on the Internet to bring into the physical lab, she brought the lab to the Internet.

Initially, it began as a way to follow up with some cognitive tests for people who had come through the face blindness lab, Duchaine said.

“You don’t want to bring 400 people into the lab and test them. [Germine] had all the emails for the people who had been scanned, so she made an online test,” Duchaine recalled. But the link got shared around and nearly 10,000 people took the test, Germine said. She realized people had an appetite for taking cognitive tests online, so she co-founded the research website Test My Brain.

It wasn’t always smooth sailing for Germine and her website. One of the tests she offered on Test My Brain’s early days involved asking people to recount their experiences with childhood trauma. Germine said that she got some pushback from the university ethics committee at first, because they were concerned the test might trigger traumatic memories for some people.

“Initially, I think people were very nervous about the Web because it was this strange no man’s land,” Germine said. She said the review board expected people would not tell their full stories, if they even came forward at all. This would bias her data set, wasting everyone’s time and money. But they were surprised by what actually happened.

“We found people were very forthcoming with that information,” Germine said. That the test was anonymous and involved entering their stories into a form rather than talking to a person on the phone or in person seemed to make people more willing to share information, she posited. Another possibility is that sharing the information anonymously online, especially if it can help others through research, may be cathartic to individuals who suffered deep trauma.

“People react surprisingly positively to what you would expect to be a mostly negative solicitation.”

“People react surprisingly positively to what you would expect to be a mostly negative solicitation.”

• • •

Germine thinks that as online testing has proven itself over time, people’s attitudes about the Internet have changed as well. It’s no longer the no man’s land it once was but is now a regular part of many people’s daily lives. And as Fitbits, Lumosity, period-tracking apps, and BuzzFeed quizzes have demonstrated, people are really interested in cataloging and defining themselves through their data.

But that doesn’t mean everyone is on board with finding data online.

“There is controversy [about online experiments] but it’s of the anonymous variety—it’s in reviews and papers and grants,” Germine said. She said this resistance has manifested in peer review especially. She recently submitted a paper about the usefulness of online data sets. Half of the peer reviewers—the gatekeepers of publication—said that using the Internet is a nonissue these days. The other half weren’t so convinced.

She thinks this may reflect a generation gap. Young researchers who grew up with the Internet don’t see using it as a tool for research as a problem, whereas the older set are still skeptical. Their skepticism isn’t completely unfounded, Germine said. Through her research career, she’s stumbled upon new ways to unintentionally introduce bias into data multiple times. She cited Amazon’s Mechanical Turk—an online venue for people to fulfill the requests of others, for a price. It turns out paying people is too good of an incentive for them to take an anonymous test. Germine cited an experiment by a colleague of hers that required participants to be native speakers of Japanese. Through the Mechanical Turk, he offered a small compensation for participants and got a huge response—by people who lied about their proficiency in Japanese. Without a face-to-face interaction, or at least over a phone call, it can be harder to screen participants. That bias largely goes away when you don’t pay people for their time, Germine said.

She also cited the ever changing nature of technology as being a constant thorn in the side of Internet research. Ten years ago, differences between browsers made running online tests very difficult, as the timing of certain cues wasn’t universal. These days, as phones and tablets become commonplace, the headache is with motion-tracking technology.

Researchers must anticipate and account for biases in their data set, as well. To participate in online cognitive tests, you must first find them and then decide to take them. Germine said she’s observed a lot of biases this way. Her largest demographic encompasses educated, wealthier-than-average people aged 18 to 35. Those people are the most tech savvy, have the most spare time, and have the requisite Internet connection, she says. In contrast, one of her biggest challenges is getting data on elderly people.

But accounting for bias in this work is no different than in any other form of research.

“There’s a running joke-slash-severe-concern in psychology that we should call it the psychology of undergraduates,” Germine said. Indeed, many psychological experiments are carried out on university undergraduates—not exactly a representative sample of the human race, either. She said interest in certain tests ebbs and flows as well. For example, if a celebrity comes out and says that he has face blindness, as Brad Pitt did in a 2013 Esquire interview, Germine and Duchaine can reliably see a spike in the number of people who take the Famous Faces Test.

“When Jay Leno one night made a joke about ‘face blindness’ in his monologue, I realized we had arrived.”

While popular media can bring the Internet’s attention to something like face blindness, the opposite is true as well. Choisser said he felt the Internet, through the online tests and community presence, made face blindness more visible.

“When Jay Leno one night made a joke about ‘face blindness’ in his monologue, I realized we had arrived,” Choisser wrote in an email. “We had made the transition from being obscure to having entered the mainstream.” It’s quite easy to find references to face blindness everywhere on the Internet today, from New York magazine to

Synesthesia—a condition where people mistake one sense for another, like “seeing sounds,” for example—is another neurological phenomenon that’s gained increased attention thanks to the Internet. About 80 percent of the research articles on synesthesia have been published in the last 10 years, according to a search on PubMed, the National Institutes of Health’s database for biomedical articles. Whether that uptick in interest is directly related to the Internet is hard to say. In 2005, researchers developed the Synesthesia Battery, an exhaustive online test that determines if the user may be having some sensory cross-talk in their brain. Synesthetes tend to think of themselves as normal, since their condition is usually not obstructive to their everyday lives. Finding out they are not quite normal on the Internet, be it through research websites or webcomics, may have helped research along.

•  •  •

Face blindness has enjoyed an upswing in research interest as well. But like synesthesia, it’s still poorly understood. Less than 100 papers are published on face blindness each year, according to a PubMed search.

The Internet may continue to change that. It makes collecting data cheap and easy, leading researchers to develop diverse tests, ranging from tests for racial bias or social intelligence to tests that screen for signs of dementia.

Duchaine said he’s hopeful that as technology improves, it will enable researchers to collect more types of data. Germine said she was excited about the possibilities of smartphones, and expressed interest in harnessing the power of smartphone GPS and motion-tracking capabilities to explore ways cognition and consciousness change between activities like sitting, standing, and walking.

As for Choisser, he just hopes there will be more research into the complexities of social functioning for people like him. He’s added an appendix to Face Blind! about how to find medical research on prosopagnosia. Though he expresses great joy and gratitude for all the research that has been done on the physiological aspects of the disease, he says research into the inner lives of face-blind people is limited. “I have heard little to date from anyone interested in the social or psychological effects of the condition, so I suspect papers on such topics to remain sparse in the immediate future,” the page reads.

Perhaps some researcher will develop a survey for that.

Illustration by Tiffany Pai