The visual word form area (VWFA) is one of the most contested concepts of modern neuroscience. Its proponents claim that a dedicated slice of cortex in the occipitotemporal region of the brain - probably centered on the fusiform gyrus - underlie the ability to read. The most radical version of this hypothesis states that the VWFA is a specialized module - an idea that implicates some sort of genetical determinism. Since writing only has been around for some 6000 years many researchers doubt that there have been enough time for such a module to evolve. Instead, they expect reading to piggybackride on other, perhaps more general, visual processes.
The April 20 issue of Neuron contains a case study of man surgically treated for epilepsy that bears on these questions. As a consequence of excision of cortical tissue posterior to the putative VWFA this patient developed a case of pure alexia where, whereas before surgery he took ~600 ms to read a word, regardless of word lenght, after surgery his reading slowed to ~1000 ms for three-letter words and increased by ~100 ms per additional letter.
Crucially, postsurgical fMRI tests showed that the VWFA no longer responded more to words than to other objects, even when words were contrasted with viewing a simple fixation point. The authors suggest that this lack of activation is explained by the fact that postsurgically the VWFA is no longer receiving its normal input. Instead, a more widespread activation pattern, including frontal, parietal, and temporal sites, appear to "take over", resulting in a letter-by-letter reading strategy.
While the study clearly demonstrates that the VWFA plays some kind of (important) role in reading, it doesn't really illuminate what type of function the VWFA performs. And, at the same time, it cannot be said to settle the question of modularity, since the patient was a 46-old man who, of course, has had amble time to develop a reading skill from learning. As Alex Martin notes in a great accompanying resumé, the study do, however, leaves unresolved "the vexing problem of how to account for the intersubject consistency in the general location of the VWFA and other category-related regions in ventral occipitotemporal cortex. One possibility is that the VWFA performs a visual processing function that predisposed it to being co-opted for reading." But there may also be other reasons. In any case, it is a very nice study that will prove important for further debates on both reading and modularity.
Gaillard, R. et al. (2006): Direct intracranial, fMRI, and lesion evidence for the causal role of left inferotemporal cortex in reding. Neuron 50: 191-204.
Martin, A. (2006): Shades of Déjerine - forging a causal link between the visual word form area and reading. Neuron 50: 173-175.