Memory can be quite fickle. Sometimes we remember the things we desire most to forget and forget the things we diligently seek to remember. This had led to the popularity of such phrases as, I never forget a face (if you are like Willy Wonka) or how could I ever forget (if you are like Jennifer Lawrence who tripped while receiving an Oscar for Best Actress).

Another example of information that is easily stored and retrieved in memory is visual images. It has been demonstrated that people have near perfect memory for thousands of images and retain that information over time.

Although memory for certain information is impressive, memory for other information is not nearly as good, for example memory for random words. Despite the privilege of visual images, even memory for images has its limits. For instance, visual images can fall prey to “recognition-induced” forgetting, which occurs when recognizing an object from memory (e.g., a purple jacket) leads to forgetting of other objects that are semantically related (a white jacket).

But what about faces; a domain for which people clearly have expertise? People claim to be so good at remembering faces (ahem, Mr. Wonka) that there is an entire game dedicated to the subject: eeBoo’s memory game, “I Never Forget a Face”.

In this matching game, faces of individuals from all around the world are printed on little blocks and placed face down  on a table. Participants have to recall the locations of matching faces and the individual with the most matches at the end of the game wins. So, if you are an expert at this game, you might be inclined to agree with Mr. Wonka.

So are faces cognitively privileged in some way as to not be lost in memory? Or, like images (let alone random words!), can they, too, fall prey to recognition-induced forgetting?

Kelsi Rugo and colleagues posed this question in a recent article published in the Psychonomic Society journal, Attention, Perception, and Psychophysics, where they assessed memory for Black and White faces.

The experiment was carried out in three phases: a study phase, a practice phase, and a test phase. During the study phase, participants were shown a series of Black and White faces and were told to remember them for a memory test.

During the practice phase, participants were presented with a subset of the studied faces (either half of the Black faces or half of the White faces) along with a set of practice lures from the same racial category (i.e., faces they did not study), and were asked whether each face was old (they remembered studying it) or new (they did not remember studying it). Each studied face was tested twice. This experimental design resulted in three types of faces: (1) practiced faces (faces presented during both study and practice phases), (2) related faces (practice lures only presented during the practice phase), and (3) baseline faces (faces from the other racial category only presented during the study phase).

During the final testing phase, participants were shown all of the faces from the study phase, along with an equal number of novel faces (from both racial categories) and were asked whether they had previously studied each face in the same fashion as the practice phase (sample stimuli are shown in the figure below).

Stimuli of the experiment. Dashed colored boxes were not presented to the subjects but are used here to illustrate the different face classifications. Faces practiced during the recognition practice phase became the practiced faces and are marked here in green. Novel faces (i.e., practice lures, marked here in yellow) warranted a “new” response and were never again presented in the experiment. Test phase was an old–new recognition judgment task that presented all the study phase faces and an equal number of novel Black and White faces (marked here in pink). New faces were entirely novel test lures, marked here in pink.

If participants demonstrate recognition-induced forgetting, they will have better memory for baseline faces compared to related faces because the practice faces (which were of only one race), will lead to forgetting of the same-race studied faces. In other words, practicing one black face (i.e. practiced item) interferes with the recognition of another black face (i.e. related item), but not a white face (i.e. baseline item) because the white face is unrelated.

Another element that enhances the applicability of this study is that participants were recruited from both a Historically Black University and a predominately White institution, resulting in a large sample of Black and White participants. This allowed the authors to also explore the contribution of the participants’ race on memory for faces from the two distinct races.

As expected, there were a number of interesting and noteworthy results.

First, participants demonstrated high accuracy in the recognition of faces, with better memory for the practiced faces relative to the related faces. Surprisingly, recognition-induced forgetting persistent for faces, despite the fact that this is a domain for which people have expertise. People demonstrated worse memory for the related faces relative to the baseline faces.

Rugo and colleagues also found that people who practiced White faces demonstrated recognition-induced forgetting with better memory for baseline faces relative to related faces. Conversely, people who practiced Black faces did not demonstrate recognition-induced forgetting regardless of the race of the participants.

So what is the verdict regarding memory for faces you ask? Although faces are highly memorability (the overall hit rate was 87%), unfortunately, they too can fall prey to recognition-induced forgetting.

So friends, I hate to be the bearer of bad news, but you can forget a face!

Featured Psychonomic Society article:

Rugo, K., Tamler, K.N., Woodman, G.F., & Maxcey, A.M. (2017). Recognition-induced forgetting of faces in visual long-term memory. Attention, Perception, and Psychophysics DOI: 10.3758/s13414-017-1419-1.