CNS 2014 Blog: Q&A with Marlene Behrmann
The idea of being left-brained or right-brained is pervasive in society. But scientists now know that it is not so simple. While some skills may over time develop to depend more on one side of the brain, the two hemispheres work interdependently from birth. Case in point is visual recognition of words and faces: According to new research, while the left hemisphere becomes increasingly specialized for word recognition after children learn to read, the left hemisphere is not exclusively dedicated to words versus faces.
Marlene Behrmann of Carnegie Mellon University has conducted several studies to look at how the hemispheres develop to visually recognize different objects. In her latest work, she and colleagues showed both adults and children a face or a word in the middle of a computer screen. The object then disappears and immediately a face or a word pops up on either the right side or left side of the screen. The participant has to decide whether the two consecutively presented faces or words are the same or not and has to press a button to indicate the response.
“We examine whether the viewer is more accurate at making the perceptual decision when the face appears on the left – i.e. is projected to the right hemisphere of the brain – or on the right, and we ask the same question when the words are tested,” Behrmann says. “The more accurate performance on one or the other side tells us which side of the brain is better at processing faces and which side is better at processing words, and we can do experiments like this even with children as young as 5 or 6 years old.”
With these experiments, the researchers have been able to determine when children first develop “hemispheric superiority” for recognizing certain objects. The work involves multiple methods, including the use of electrical measurements, as well as computational models, coupled with the behavioral experiments. Behrmann, who will be presenting this body of work in her keynote address for the CNS annual meeting in Boston next month, talked with CNS about these experiments, popular myths of the brain hemispheres, and why the brain develops specialized regions for visual recognition.
CNS: Your work suggests overlapping and interdependent systems for visually processing words and faces. Does it overturn the previous left hem v. right hem model?
Behrmann: Yes, this in fact is the thrust of most of this work. The argument is that the brain is not as dichotomous and binary as previously thought, but, rather, that complex behavior emerges from the interaction of graded systems and that these graded systems come to be organized over the course of development.
CNS: Are there any myths about how the brain works that you’d like to clear up?
Behrmann: There are many myths that surround the division of labor between the left and the right hemisphere. The media often portrays the idea that the left hemisphere is more analytical and detail-oriented, whereas the right hemisphere is more creative and holistically-oriented. Although there may be some kernel of truth to this, the oft-cited left versus right hemisphere distinction is not valid. There are other aspects of hemisphere function and organization that also require closer scrutiny, including whether there are hemispheric differences from birth or only as a result of experience-dependent tuning. Also, we do not yet have a clear understanding of the relationship between handedness and hemispheric organization. There are also some interesting and controversial issues concerning the extent to which the hemispheres are different in other species who do not have language. And, finally, one could even ask the most basic of questions – why do we even have two hemispheres?
CNS: Why is it important to understand how we process words vs. faces in the brain?
Behrmann: This issue is relevant from both a basic science and a clinical perspective. In the former instance, the longstanding wisdom is that the psychological and neural mechanisms that support word recognition (subserved by the left hemisphere) are independent of those that support face recognition (subserved by the right hemisphere). Indeed, there has been substantial consensus on his point. One of the goals of our studies has been to examine the extent to which these systems are truly independent.
The motivation for the investigations grew out of our observation that in most, if not all, functional imaging studies, word-selective and face-selective brain activation is usually present in both hemispheres (is bilateral) rather than just in a single hemisphere (unilateral) and that, interestingly, the sites of activation are the same across the two hemispheres. It is also the case that both words and faces constitute a class of visual objects in which there are many, similar examples comprised of the same set of local components – for example, letters in the case of words and eyes/nose and mouth in the case of faces. Our findings suggest that word and face representations are not as independent as previously thought.
From a clinical perspective, this is also an interesting avenue for research. For example, damage to each hemisphere affects the perception of both words and faces and difficulty in acquiring mastery over say word recognition is associated with reduced use of just one hemisphere for face representations. Taken together, the mechanisms used for face and word recognition form part of a more complex, dynamic system, which, over the course of development, comes to be optimized for rapid and efficient recognition of complex visual inputs.
CNS: Can you briefly explain how use of the hemispheres to recognize words versus faces differs in children?
Behrmann: We know that in adults, the left hemisphere is better than the right hemisphere of the brain in recognizing words. This is probably because in the majority of the population, language systems are more left- than right-sided. We were particularly interested in exploring how this mature hemispheric profile emerges over development or whether it is present from a very young age.
We conducted a half-field study in which we presented subjects with words or faces in each visual field, on the right or left. We showed that adults evinced the complementary superiority of faces in the left visual field (right hemisphere) and words in the right visual field (left hemisphere).
Although adolescents were overall as accurate as adults, they only showed a left hemisphere advantage for words and no hemisphere difference for faces. The same was true for younger children, although their overall accuracy was lower than either of the other groups. These findings suggest that the division between the hemispheres in visual recognition for words precedes that for faces. We also showed that the extent to which a participant was a proficient reader was predictive of specialization of face perception on the right hemisphere.
Taken together, these findings are consistent with the account that prior to the onset of reading, there is no pressure for face recognition to occur on one or the other hemisphere. It is only once a child starts reading, and there is a need to have the visual information be in close proximity to the language information in the brain, that the left hemisphere becomes increasingly – but not exclusively – tuned for word recognition.
Because words and faces are so different from each other and it is unlikely that the same area of the brain can represent both classes efficiently, the right hemisphere becomes increasingly (again, but not exclusively) tuned for face recognition.
These results are surprising: Even young infants are exposed to faces and so one might expect the brain to be tuned to faces early on. Instead, we find that there is no hemispheric preference for face perception (nor a region uniquely dedicated to face perception) and although young children can certainly recognize faces, it is the acquisition of reading skills that triggers the cascade that results in the adult hemispheric pattern.
CNS: What is the take-home message of your research?
Behrmann: The bottom-line is that the brain is rather complicated! Although one might think that having special parts each devoted to processing one type of information (like one area of the brain that recognizes faces and one area that recognizes words) might work well, brain organization is not so cut and dried. As we learn and become better at recognizing various images, our brain needs to develop skills to decode all stimuli and the brain changes dynamically over development and experience.
CNS: Why do you think our brains develop to recognize words and faces as they do?
Behrmann: The brain needs to be able to differentiate and individually identify very similar visual inputs (for example, consider how similar faces are compared with two different objects, say, an eraser versus a bulb) and has to be able to represent the inputs precisely and rapidly. The nerve cells or neurons of the visual system have to become tuned to these fine visual differences. There is not enough space in the brain that there can be different regions for all the different classes of objects we can recognize; so, there must be some clever organization that allows us to perceive so accurately and efficiently. We have put forward a theory that explains how the brain might have developed a system that allows for good recognition of faces and words. This theory goes further to explain how this system emerges over development.
CNS: How did you personally become interested in the topic?
Behrmann: My research focuses on the mechanisms by which the brain transforms the impoverished input from the eye into coherent and meaningful representations. In past research, I have explored how it is that we are able to recognize words so quickly and accurately notwithstanding the fact that they can be shown in different font and different size. I have also explored how it is that we are able to recognize faces so accurately and efficiently; it occurred to me, that there are many similarities in the systems that underlie these two classes. We had also conducted some fine-grained imaging studies, showing that the same brain data had enough information to discriminate between a set of four different faces and to discriminate between a set of four letter strings. The challenge was now to understand the system, its neural underpinnings, and computational implementation. And my enthusiasm was charged!
CNS: What is next for this line of work?
Behrmann: We have begun to examine how the standard pattern of hemispheric superiorities might go wrong in individuals with different brain organization. For example, right- handed individuals typically have language represented in the right hemisphere, but the hemispheric dominance for language in left-handers is more mixed (some have right hemisphere language, some left, and some bilateral). If, as we argue, the pressure for visual areas to be coupled with language areas is key for word recognition and has a knock-on effect driving face recognition into the other hemisphere, then we might see hemispheric organization in left-handers potentially being directly opposite to that of right-handers.
We are also interested in examining more detailed patterns of how a single word or a single face is processed temporally and how this might differ from adulthood to childhood. We are planning some studies using magnetoenecephalography to do so. There are many outstanding questions and many superb methods that are now available to help shed light on brain organization and the emergence of high level vision from the underlying neural correlate.
-Lisa M.P. Munoz
Behrmann will give a keynote address at the CNS annual meeting in Boston on Monday, April 7, 2014.
Media contact: Lisa M.P. Munoz, CNS Public Information Officer, firstname.lastname@example.org