A major difficulty of the English language is that mappings between how words are spelled and how they sound are only partially consistent. (For example, 'cough', 'bough', 'dough' and 'tough' are all pronounced differently.) Readers of Italian have an easier task because a particular letter or letter combination is almost always associated with the same sound. Now an international team of researchers reports that these differences in language structure lead native speakers to adopt different strategies for pronouncing words, which are reflected in activation of different brain regions when subjects read words and pronounceable nonwords (like 'jat') aloud. In this study, English subjects took significantly longer to begin reading each word, and were even slower when the stimuli were nonwords. The faster reading times for Italians are consistent with the idea that Italian speakers can rely almost exclusively on a procedure that translates letters to sounds, whereas English speakers have to use additional information such as the meaning of a word to decide how it should be pronounced. The authors then used positron emission tomography to measure blood flow (a measure of neuronal activity) in the brains of Italian and English speakers during this task. Brain activation depended on subjects' native language in three areas of a widespread network of brain regions previously associated with reading. Italians had greater activation in a left superior temporal region during both word and nonword reading than English subjects, who showed greater activation in a left frontal and a posterior inferior temporal region during nonword reading.
The finding that these regions are sensitive to the consistency of a subject's native language suggests that they are critically involved in phonology, the derivation of phonemes from spelling, although the specific contributions of each region may be debated. This ability is critical for skilled reading, suggesting that these findings may have important implications for reading disorders such as dyslexia. Equally exciting are the implications for understanding how experience can shape the organization of our brains. Both groups of subjects can use spelling or meaning to decide how to pronounce a word, but experience has optimized their use of these procedures for the language they read. Reading procedures may become tuned based on experience, with much of the optimization occurring automatically and continually. An interesting theoretical issue with profound practical implications is how much this tuning may be influenced by instructional strategy and the initial structure of reading materials. For instance, an early instructional strategy (such as phonics) that emphasizes the consistent features of spelling-to-phoneme transformation in English may lead children to emphasize such a procedure, compared to an early instructional strategy (for example, natural language) that focuses on the relationships between entire word forms and their corresponding pronunciations and meanings.
This article is discussed in an accompanying News & Views by Dr. Julie Fiez.
It is generally believed that the brain does not come 'hard-wired' but must instead rely on experience to form the correct connections. There are plenty of animal experiments to support this view, and it makes sense that the brain should be able adapt itself to an unpredictable world. But testing this idea in humans is more difficult.
It is obviously not possible to do experimental manipulations on children, but David Whitaker and Paul McGraw (University of Bradford) have done the next best thing. They have cleverly taken advantage of a natural experiment to show that our experience determines what we perceive. We are all exposed to letters printed in italics, which have a characteristic tilt to the right (typically 5 to 10 degrees), as do many digital numerical displays. As the authors show, this familarity causes our brains to play tricks on us. They asked volunteer subjects to judge the orientation of letters, numbers or meaningless patterns of stripes. For the letters and numbers, subjects judged leftward-tilted numbers to be farther from vertical than rightward-tilted ones. Yet they showed no difference in their sensitivity for stripes. The familiarity of rightward-tilted letters and numbers has apparently reduced our sensitivity to their tilt; as the authors say, it seems that everyday perceptions can be markedly influenced by visual experience. The ideal way to prove their idea would be to repeat the experiment in a culture that provides little exposure to italic lettering or digital displays. But where in the world these days can you find people who have not seen a digital watch?
Other people's faces tell us not only who they are, but also what they are thinking and feeling. We can read their emotions, we recognize instinctively when they are making eye contact, and we also tend to follow their gaze when they look away; our attention is drawn to whatever they are looking at. Where in the brain do we decode these powerful social signals? It is known that stroke damage to certain brain regions can leave patients unable to recognize other individuals but still able to recognize their emotions or direction of gaze, or vice versa. But brain damage seldom affects just one area, and so while the existence of these patients tells us that there are separate pathways for recognizing identity and social cues, their exact location is not known.
Now Elizabeth Hoffman and James Haxby (National Institutes of Health) have identified brain areas that are active during the two different tasks. They used functional magnetic resonance imaging (fMRI) to scan the brains of volunteers as they observed pictures of computer-generated faces that were looking either straight at the viewer or away to one side. They found one brain area, called the lateral fusiform gyrus, that was active when subjects had to identify the face on the screen. A different area, the posterior superior temporal sulcus, was active when they had to determine which way the eyes were looking. Other studies have implicated the STS not only in gaze perception but also in recognizing emotions and even in lip-reading. The authors suggest that the STS may be a general-purpose area involved in recognizing all the changing aspects of faces (as opposed to identity, which remains constant no matter what the facial expression).
Recovering drug addicts are often advised to avoid revisiting people and places that are associated with drug use in their minds, for fear that these memories may trigger a relapse. Similarly, in rats, repeated administration of morphine in a particular setting can lead to tolerance (requiring more drug for the same pain relief) that is expressed only in the same environment, termed 'associative tolerance'. Alternatively, long-term morphine exposure in different settings leads to 'non-associative tolerance', which occurs in any environment. Howard Fields and colleagues (University of California, San Francisco) now show that associative tolerance requires a peptide neurotransmitter called cholecystokinin acting in the amygdala, a brain region known to be important for emotional learning. In rats showing associative tolerance, morphine increased the expression of the activity-related protein Fos in the lateral amygdala, but only when it was given in the associated environment. Rats with non-associative tolerance did not show this response. Blocking the receptor for this peptide in the amygdala reduced the expression of associative tolerance, making the rats more responsive to morphine. The amygdala has previously been proposed to participate in the process of learning to associate drug use with environmental cues, which contributes to drug addiction. Further understanding of this process may lead to better treatments for addiction.
