Abstract
Neuroimaging and neuropsychological evidence suggests a specific module in the human ventral cortical stream, called the word form area (WFA), which responds selectively to visually presented word forms. This region is consistently localized adjacent and lateral to the left posterior FFA. However, since humans have used written language for only ∼7000 years, it is unlikely that a novel cortical module evolved during that time. Instead it has been suggested that the WFA may reflect top-down processes activating a pre-existing module, which is tuned to relevant lower level cues. To test this, we presented images of faces and places to human subjects and awake rhesus macaques, in a blocked-design fMRI experiment. The stimuli were either highpass-filtered, lowpass-filtered, or unfiltered for spatial frequency (SF). Results in humans showed selective activation in response to high SFs in the area reported as WFA. The functional signature of this area thus includes a specialization for high SF, which is crucial for perceiving (and presumably processing) letters. However, the high SF regions are bilateral, whereas the WFA is reported to be largely unilateral. Nevertheless, extensive fMRI tests in two awake monkeys showed no homologue of the WFA. This is notable because it is increasingly unusual to not find monkey homologues of known cortical areas in humans. We conclude that the WFA is coextensive with, or built up from, a preexisting cortical area that is activated selectively by high SFs, even using non-word stimuli. No such cortical area appears to exist in monkeys, when tested with the same visual stimuli that elicit WFA activation in humans. Therefore, it appears that a specific area/module has evolved in the human cortex in the past 25 million years, which is used for, but not exclusively activated by, word form processing. NEI Grant #R01 EY017081 to RBHT.