Pre-lexical visual word form hypothesis Put forward by Cohen and colleagues (2000). The basics of this theory state that the neurons in the ventral occipital-temporal cortex (vOT) – which the posterior
fusiform gyrus is a part of – have receptive fields that are sensitive to bigrams, or two letter combinations that commonly occur in words. The neurons sense and process the bigrams, to detect their legality. Here the posterior left fusiform gyrus (part of the vOT), is thought to be one station in a long line of processing areas. The processing starts with visual feature detectors in extrastriate cortex, proceeding through letter detectors and letter-cluster detectors in the posterior fusiform, and then activating lexical representations stored in more anterior multimodal fusiform area. The theory states the function of the VWFA is pre-lexical as it occurs before the word is understood to have meaning.
Lexical visual word form hypothesis Put forward by Kronbilcher et al. (2004), was based on functional imaging data that showed, in a parametric
fMRI study, that a decrease in activation in the left fusiform gyrus was seen in response to an increase in the frequency of the word - where the frequency is how common the word is. This data refutes the previous pre-lexical theory as if the VWFA was pre-lexical one would expect equal activation throughout all frequencies. Instead a lexical theory was proposed where the left fusiform gyrus neurons are thought to detect words by attempting to match them to stored representations of known words. This would explain the data as more common words would take less time to detect than the less common words, reducing the energy needed for computation and therefore potentially reducing the magnitude of the haemodynamic response that is detected by BOLD fMRI. A recent intracranial
electrocorticography study shows that the activity in the VWFA goes through multiple stages of processing. Using
classification with direct neural recordings from the VWFA, Hirshorn et al. showed that early VWFA activity, from approximately 100-250 milliseconds after reading a word, is consistent with a pre-lexical representation and later activity, from approximately 300-500 milliseconds is consistent with a lexical representation. These results potentially mediate between the pre-lexical and lexical hypotheses by showing that both levels of representation may be seen in the VWFA, but at different latencies after reading a word. Previous studies using fMRI did not have the temporal resolution to differentiate between these two stages. == Alternative functions for the cortical area ascribed to the VWFA ==