The Neurobiology of Reading
The phonological model incorporates a modular scheme of cognitive processing in which each of the component processes used in word identification is carried out by a specific network of brain cells. Until recently, however, researchers have had no firm indication of how that scheme maps onto the actual functional organization of the human brain.
Unlike many other functions, reading cannot be studied in animals; indeed, for many years the cerebral localization of all higher cognitive processes could be inferred only from the effects of brain injuries on the people who survived them. Such an approach offered little to illuminate the phenomena my colleagues and I were interested in. What we needed was a way to identify the regions of the brain that are engaged when healthy subjects are reading or trying to read.
Our group became quite excited, then, with the advent in the late 1980s of functional magnetic resonance imaging (fMRI). Using the same scanning machine that has revolutionized clinical imaging, fMRI can measure changes in the metabolic activity of the brain while an individual performs a cognitive task. Hence, it is ideally suited to mapping the brain’s response to stimuli such as reading. Because it is noninvasive and uses no radioisotopes, fMRI is also excellent for work involving children.
Since 1994, I have worked with several Yale colleagues to use fMRI in studying the neurobiology of reading. Bennett A. Shaywitz, Kenneth R. Pugh, R. Todd Constable, Robert K. Fulbright, John C. Gore and I have used the technique with more than 200 dyslexic and nondyslexic children and adults. As a result of this program, we can now suggest a tentative neural architecture for reading a printed word. In particular, the identification of letters activates sites in the extrastriate cortex within the occipital lobe; phonological processing takes place within the inferior frontal gyrus; and access to meaning calls on areas within the middle and superior temporal gyri of the brain.
Our investigation has already revealed a surprising difference between men and women in the locus of phonological representation for reading. It turns out that in men phonological processing engages the left inferior frontal gyrus, whereas in women it activates not only the left but the right inferior frontal gyrus as well. These differences in lateralization had been suggested by behavioral studies, but they had never before been demonstrated unequivocally. Indeed, our findings constitute the first concrete proof of gender differences in brain organization for any cognitive function. The fact that women’s brains tend to have bilateral representation for phonological processing explains several formerly puzzling observations: why, for example, after a stroke involving the left side of the brain, women are less likely than men to have significant decrements in their language skills, and why women tend more often than men to compensate for dyslexia.
As investigators who have spent our entire professional lives trying to understand dyslexia, we find the identification of brain sites dedicated to phonological processing in reading very exciting—it means that we now have a possible neurobiological “signature” for readng. The isolation of such a signature brings with it the future promise of more precise diagnosis of dyslexia. It is possible, for example, that the neural signature for phonological processing may provide the most sensitive measure of the disorder. Furthermore, the discovery of a biological signature for reading offers an unprecedented opportunity to assess the effects of interventions on the neuroanatomic systems serving the reading process itself.