Last year we took the Famous Faces test to demonstrate how faceblindness works. Famous Faces is somewhat flawed because if you aren’t familiar with most of the celebrities in the test, it gives a less than accurate measure of how good you are recognizing faces.
This week’s test is a better gauge of faceblindness or prosopagnosia. The Matching Faces in Photographs test at Test My Brain is being used by researchers to understand the difference between how we recognize standardized versions of faces under ideal conditions versus how we recognize faces in changing conditions. For example, if you recognize your chemistry professor by his beard and glasses and the fact that you generally encounter him in the chem lab building, you may or may not recognize him at the beach in swim trunks and baseball cap, especially if he’s clean-shaven and not wearing his glasses.
Often, people with prosopagnosia rely on hairstyle, facial hair, glasses, voice, mannerisms, gait, or other “auxiliary” features to identify friends and acquaintances. Some people are mildly faceblind, meaning they recognize close friends and family quite easily but struggle with quickly identifying acquaintances if we meet them “out of context.” Others have difficulty recognizing everyone, even close family members, and may not recognize their own reflection right away.
Prosopagnosia can create embarrassing social moments and those of us who are more than a little faceblind have come up with all sorts of tricks to disguise our confusion. For example, when I’m introduced to someone, I always say “Good to see you,” never “Nice to meet you.” That way, if I’ve met the person before, they won’t immediately correct me.
TAKING THE TEST
The test takes 15 to 20 minutes. To start, go to testmybrain.org and click on the Go! button next to the “Matching Faces in Photographs” test.
You’ll be asked to make your browser window large. I kept mine the size I normally use and it worked fine. The next screen is a simplified informed consent form. You’ll be told what the research is being used for and asked to consent (agree). The next screen collects some anonymous demographic information.
Once you’ve entered your demographic info, you’ll begin the test. There are two parts. In the first, you’ll be shown two faces side by side and asked to determine if they are the same person or different people. This is the test of recognizing people in changing conditions. There are 100 pairs, but they flash by pretty quickly.
The second part is the Cambridge face memory test. You’ll be shown a computer generated human face to memorize from three different angles, then you’ll be asked to pick it out from a set of three faces. Next you’ll get to study all six faces and be asked to pick out one of them from a set of three faces. This section repeats twice, once with regular faces and once with faces that have an overlay of computer generated noise.
Once again, there are some spoilers in the scoring section. If you plan to take the test, do it before reading further.
SCORING THE TEST
Part One: My score was 55. The average score is 64.94. I scored higher than 1 out of every 10 people who took the test.
No surprise there. I typically recognize people based on things like how they walk or their voice, both of which change very little over time. If I have to remember a face, I rely on auxiliary features like hairstyle and color, facial hair, scars, glasses, etc. The changes in most of the comparison photos were so significant to me that I guessed at man of the answers. Also, some flashed by so quickly that I barely had time to take in one of the photos, let alone compare the two.
Part Two: My score was 46. The average score is 52.49. Again, I scored higher than 1 out of every 10 people who took the test.
I found this part even harder because the faces were so generic looking to me. In the single face comparisons I found myself picking out specific features related to the person’s facial expression to identify them (slightly raised lip, one eyebrow higher, etc.). Again I find myself relying on details over gestalt to identify visual images.
In the 6-person sets, after a few trials, I had mostly forgotten what the original 6 faces looked like and couldn’t bring up a mental visual representation of any of them. I was reduced to guessing based on the face that seemed most familiar, which may have meant I was picking whichever faces came up most often, even if they weren’t one of the 6 target faces.
I would love to know what strategies you used on this test, especially if they worked.
THE BOTTOM LINE
This felt like a fairly accurate test of the ability to remember and identify faces under a variety of conditions. Although it doesn’t take into account our ability to recognize the faces of people we know from memory, it does give an indication of how good we are at recognizing the same face under differing conditions or at remembering a new face based on facial features alone.