"You look familiar to me…"   Neural Mechanisms Involved in the Identification   of Human Faces

Research team:
Researcher : Elana Zion-Golumbic
Adviser: Prof. Shlomo Bentin
Collaborators: Dr. David Anaki, Tal Golan, Viki Aizenberg
The Department of Psychology and the Department of Cognitive Science, Hebrew University of Jerusalem
The research was conducted as part of Elana Zion-Golumbic's doctoral work.

The Research Question
How are we able to identify human faces so quickly and accurately? After all, we all share the same facial features- two eyes, a nose, a mouth, ears, and hair. Yet we're very good at distinguishing between the faces of different people.

In this research, we wanted to examine the neural mechanisms behind our special ability to identify faces.

The Research Tool
We know that human beings have a special ability to process faces, and scientists have pinpointed an area of the brain involved specifically in this task. We wanted to examine how it happens.

To do that, we used a method for measuring electrical activity in the brain - electroencephalography, or EEG for short. Electrodes attached to the head of the subject detect the electrical field produced by tens of thousands of neurons in the brain's cortex.

The highly complex electrical signal recorded by EEG provides us with data about different types of brain activities. This experiment focused on two measurements of brain activity:

1. The N170 Component
This is a well-known brain response that, as measured by EEG, occurs 100-200 milliseconds (ms) after the appearance of a stimulus. Its origin is the posterior-lateral region of the brain (mainly on the right side). Previous research has shown that the N170 is greater for faces than for other objects.

Conclusion A: The N170 indicates brain activity related to the processing of faces.
In the past, we discovered that the N170 for responses to human faces and responses to monkeys' faces is similar.

Conclusion B: Although the N170 is associated specifically with faces, it does not explain our special ability to identify faces (since we're not very good at identifying different monkeys' faces).

2. Activity in a High Frequency range called the "Gamma Band"
Previous research associated gamma-frequency brain activity with "high level" cognitive functions such as visual and auditory perception, memory, learning, and attention. Gamma activity also helps synchronize different areas of the brain that work together to perform complex tasks.

Since face identification is a complicated process that requires cooperation between our visual processing system and the retrieval of information from our memory, we hypothesized that gamma activity may be involved in this process.

The Experiment
To answer the research question, we compared the brain's activity in response to familiar faces and unfamiliar faces. The 20 subjects participating in the experiment were shown faces on a computer screen. Half of these faces were of famous people and half of unfamiliar people. The subjects were asked to look at the faces and describe them as "good-looking," "ugly," or "average." We had two reasons for choosing this particular task:

1. We wanted to be sure the subjects were really looking at the faces and processing them.
2. We didn't want the task to relate directly to the familiarity of the faces. This was an important point, since we were interested in exploring the processes in the brain that lead us to automatically identify familiar faces.

The Results of the Experiment
N170 in response to familiar faces and unfamiliar faces

For Picture 1, we see the average electrical brain response over time after presentation of the stimulus (the X axis). The N170 appears about 150 ms after presentation of the stimulus, and there is no difference in the N170 response for familiar faces and unfamiliar faces.

Gamma frequency activity in response to familiar faces and unfamiliar faces

For Picture 2, we see brain activity at frequencies of 40-100 hertz over time (the Y axis), beginning with presentation of the stimulus (the X axis, with the stimulus appearing at the zero mark). The different colors indicate the intensity of the activity (red = high, blue = low). Approximately 100 ms after the stimulus appears, there is increased activity at these frequencies that lasts about one second. This response is stronger for familiar faces than for unfamiliar faces.

Discussion and Conclusions
The results of the experiment indicate that the two measurements we examined reflect brain processes that are related to different aspects of face processing.

While the N170 is a specific response to faces (as opposed to objects), we saw that it is not sensitive to the identity of the face. (That is, the N170 does not differentiate between familiar and unfamiliar faces.) We suggest, therefore, that this response indicates identification of the overall form of the face, with no additional information. This idea is consistent with the finding that the N170 is similar for responses to human faces and monkeys' faces.

Unlike N170, gamma activity is affected by face identification, and it is higher for familiar faces than for unfamiliar faces. How can we interpret this finding? One possibility is that gamma activity reflects the retrieval of information from our memory, which helps us identify the face we see. Another possibility is that heightened gamma activity indicates that greater attention resources are being used to process familiar faces, or that secondary mechanisms (such as personal memories about the individual or emotional responses) are being activated when a face is identified - mechanisms that are irrelevant for unfamiliar faces.

It is important to note that while the N170 response begins and ends in a brief instant, the gamma response lasts for more than a second after the stimulus appears - a very long time in terms of the brain, perhaps indicating that gamma activity reflects several processes rather than only one.

The Proposed Model for Face Identification
These findings (along with others reached in our laboratory) led us to propose the following model for face recognition:

The Importance of the Research and Plans for the Future
Do you ever see people that seem familiar but you can't remember their names? Sometimes it seems that the name is right there "on the tip of your tongue", but you cannot retrieve it from your memory. Then, on the other hand, you may run into someone you saw only once before, a long time ago, but you remember every little detail about her.

The research we presented here was the first in a series of experiments we conducted to explore the neural mechanisms involved in learning new faces, learning details about individuals, and retrieving what we've learned from our memory when the need arises. We also began examining what occurs in the brain when we misidentify a face or fail to recognize a face that should be familiar to us.

The subject of this research is close to the heart of anyone who has experienced the frustration of not recognizing the face of a person who should be easily identified. We hope that by shedding light on the brain mechanism responsible for this important and special ability, we will better understand the process and, perhaps, help to develop methods for improving the "face memories" of healthy people and people with impaired memories or visual identification problems.

More about...

To raise new questions, new possibilities,
To regard old problems from a new angle
Requires creative imagination and marks real advances in science.
(Albert Einstein)

The Research Question - At the root of all research is a question. Contained within this question is the scientific curiosity that led the researchers to conduct the particular experiment they did.

The scientific process will continue as long as we keep asking new questions. Each question leads to a new experiment and yields results that supplement and expand the existing body of scientific knowledge.

The research question not only describes the subject of the research; it plays a central role in all its stages. The researchers need to formulate a focused and precise research question, one that allows them to design an empirical experiment that can answer it. In other words, the question must be testable.

Usually, researchers can make predictions about the research question based on scientific theory. Nevertheless, they must state the question as objectively as possible in order to allow for answers they did not necessarily expect.

The research question is also crucial for understanding the results of the experiment and their significance - these will be scrutinized in light of the original question asked.