Two pre-tests were conducted before scanning to verify the efficacy of our visual masking

1. The role of the left posterior fusiform gyrus in reading J. T. Devlin1,H. L. Jamison1, L. M. Gonnerman2, and P. M. Matthews1 1Centre for Functional Magnetic Resonance Imaging of the Brain, University of Oxford, Oxford, U. K. 2Department of Experimental Psychology, Lehigh University, PA, U. S. A. E79 Predictions Imaging results: Reading Words and Pseudowords Discussion Background • These results suggest that neither lexical nor pre-lexical word forms are stored in the left posterior fusiform gyrus: 1. Repeated presentation of words, but not pseudo-words, leads to a reduction in activation in the left posterior fusiform region. In fact, pseudo-word repetition led to a small increase in activation, exactly opposite the prediction of the pre-lexical hypothesis. 2. Orthographically overlapping words (e.g. ponder-pond) also led to a decrease in activation. The lexical hypothesis, however, claims that competition visual word forms should lead to an increase in activation, which is necessary to explain activation for pseudo-word reading (6). 3. Orthographically related words which also share meaning (e.g. government-govern) show a reduced neural priming effect, which cannot be explained by either the lexical or pre-lexical word form hypothesis. In contrast, this is predicted by the visual identification hypothesis as the shared visual form reduces processing demands on the area but this is partially offset by the increased demands necessary to differentiate between semantically similar words. • The visual identification hypothesis is not specific to visual word forms but also explains why precisely the same area is engaged by picture naming (5,7) • We argue that the “visual word form area” is a misleading label, suggesting a specificity for words which simply isn’t present. In addition, it suggests a neural correlate of the cognitive “visual word form system” (8) that is not correct or appropriate. • Future work is necessary to determine the cause of the apparent laterlity of this region with the left posterior fusiform engaged by written word forms while the right is engaged in the recognition of faces and other structurally similar objects (9-10). Functional neuroimaging studies have identified a left lateralized region of the posterior fusiform gyrus consistently engaged by word reading (1), although its precise functional contribution remains unclear. Currently there are two main hypotheses: 1) Pre-lexical visual word form hypothesis: The region stores abstract, orthographically legal letter combinations, which is why it is activated for both word and pseudo-word reading (2-5). 2) Lexical visual word form hypothesis: The region stores abstract lexical representations, which is why word frequency affects levels of activation (6). By these accounts, the function of the region is to store abstract orthographic representations. Here we consider an alternate explanation in terms of the information processing performed by the region. Visual word recognition relies on the reader’s ability to recognize subtle differences in visual form quickly and accurately in order to distinguish between different words (e.g. distinguishing “acne” from “acre”). This property of orthographic systems is not limited to English, or even Roman script, but is equally true for all written languages, be they alphabetic, syllabic, or logographic (see Fig. 1). Consequently, we propose a third explanation for the function of the left posterior fusiform gyrus in reading: 3) Visual identification hypothesis: The region is engaged when uniquely identifying a visual stimulus requires fine grained form processing. L R The three hypotheses generate different predictions regarding activation of the left posterior fusiform region in visual masked priming: Pseudo-word repetition: If pre-lexical orthographic patterns are stored in this region, then repetition of words and pseudo-words should show a similar reduction in signal. In contrast, the lexical hypothesis predicts no neural priming effect because pseudo-words do not have lexical representations. Similarly, the visual identification hypothesis also predicts no neural priming because the benefit of shared form is offset by the additional processing necessary to determine a unique referent (which doesn’t exist). Form and meaning relations: If lexical representation are stored in the area, then the competition between orthographically overlapping word pairs (e.g. corner-corn) should increase activation in the area. In contrast, the pre-lexical hypothesis predicts activation should decrease due to overlapping letter combinations. Similarly, the visual identification hypothesis predicts a neural priming effect because the process of identifying the stimulus will benefit from overlapping visual forms. This hypothesis makes a further prediction, namely that if word pairs share both form and meaning, then the neural priming effect will be reduced because of the extra processing necessary to distinguish two words with similar meaning (e.g. boldly-bold). X = –42 Z = –18 Figure 6: Both words and pseudowords activated the left posterior fusiform gyrus relative to consonant letter strings. Activation for words is shown in red, for pseudowords is shown in orange, and the overlap is shown in yellow. Activations are displayed on the participants’ mean structural scan. • Words relative to consonant letter strings activated a region of left posterior fusiform cortex with a peak at (–42, –60, –18) and a Z-score of 3.6 (p<0.01 corrected) • Pseudowords activated an overlapping region with a peak at (–44, –54, –16) and a Z-score of 3.7 (p<0.01 corrected)  Demonstrates that our lexical decision task engages the specific region observed in previous studies of the “visual word form area” (2-5) • Within the posterior fusiform region commonly activated by words and pseudowords, the effects of lexicality and repetition were assessed using a 2 × 2 ANOVA • There was no significant main effects but there was a significant interaction (F1,21=5.2, p<0.05). • Case-independent repetition priming for real words lead to a regional reduction in BOLD signal similar to those reported previously (refs). In contrast, repeated pseudowords led to a small increase in BOLD signal, inconsistent with the prediction of the pre-lexical hypothesis. Behavioural pre-testing Alphabetic Arabic Logographic Mandarin Syllabic Hindi • Two pre-tests were conducted before scanning to verify the efficacy of our visual masking • Performance was at ceiling in both tasks for words presented for 200msec. • In contrast, words presented for 33msec were nearly impossible to perceive. Only 4/220 words were read aloud and performance on the matching task was not different from chance (binomial test, p=0.98). • These results demonstrate the efficacy of the visual masking used here. رمانزمان Neural priming effects Task: 1. Read the word aloud, if possible. 2. Select the word between the masks. If uncertain, guess. sky man today come pomegranate time/era نعمنغم WAVE MUST yes melody field letter enter human #$%%#$ #$%%#$ 500ms 500ms Effects of Lexicality cast wave 33 or 200ms مغلوبمقلوب 33 or 200ms $%##$% $%##$% 500ms 500ms defeated turned over time nail me find Time * Time 1000ms * 1000ms Peak at (–44, –62, –16) Z=2.9, p<0.08 corrected Unrelated Repeated Figure 1: Subtle visual form differences in written words often are critical for correctly identifying the word, independent of the writing system. Here we illustrated this with examples each from alphabetic (Arabic), syllabic (Hindi), and logographic (Mandarin) orthographies. Literacy in a language makes recognising these visual differences quick and effortless. In contrast, identifying these differences in an unfamiliar script typically involves conscious serial comparisons. 0.50 0.25 0.00 100 50 0 100 50 0 * Accuracy (% correct) References 1. Price, C. J. (2000). The anatomy of language: Contributions from functional neuroimaging. Journal of Anatomy, 197, 335-359. 2. Cohen, L., Dehaene, S., Naccache, L., Lehericy, S., Dehaene-Lambertz, G., Henaff, M., & Michel, F. (2000). The visual word form area: spatial and temporal characterization of an initial stage of reading in normal subjects and posterior split-brain patients. Brain, 123, 291-307. 3. Cohen, L., Lehericy, S., Chochon, F., Lemer, C., Rivard, S., & Dehaene, S. (2002). Language-specific tuning of visual cortex? Functional properties of the visual word form area. Brain, 125, 1054-1069. 4. Dehaene, S., Le Clec'H, G., Poline, J. B., Bihan, D. L., & Cohen, L. (2002). The visual word form area: a prelexical representation of visual words in the left fusiform gyrus. Neuroreport, 13(3), 321-325. 5. McCandliss, B. D., Cohen, L., & Dehaene, S. (2003). The visual word form area: expertise for reading in the fusiform gyrus. Trends Cogn Sci, 7(7), 293-299. 6. Kronbichler, M., Hutzler, F., Wimmer, H., Mair, A., Staffen, W., & Ladurner, G. (2004). The visual word form area and the frequency with which words are encountered: evidence from a parametric fMRI study. Neuroimage, 21(3), 946-953. 7. Price, C. J., & Devlin, J. T. (2003). The myth of the visual word form area. Neuroimage, 19(3), 473-481. 8. 9. Gauthier, I., Tarr, M. J., Anderson, A. W., Skudlarski, P., & Gore, J. C. (1999). Activation of the middle fusiform 'face area' increases with expertise in recognizing novel objects. Nat Neurosci, 2(6), 568-573. 10. Kanwisher, N., McDermott, J., & Chun, M. M. (1997). The fusiform face area: a module in human extrastriate cortex specialized for face perception. J Neurosci, 17(11), 4302-4311. Mean % BOLD signal Current Study 200ms 33ms Duration 200ms 33ms Duration Words Pseudowords Z = –16 Figure 3: Results of the behavioural pre-tests. On the left is a schematic of the reading aloud task where participants read words aloud presented for either 33 or 200msec between visual masks. On the right is a schematic of the forced choice matching task. • The aim of the current study was to evaluate these three hypotheses of posterior fusiform function during reading • A visual masked priming paradigm with a lexical decision task (see Figure 2) was used to engage the region and fMRI was used to measure its activation Figure 2: A schematic diagram of the visual masked priming paradigm. The prime was forward masked by a visual noise pattern and backward masked by the target. • The design included eight conditions which could be divided into two 2 × 2 factorial designs and a low level baseline (Table 1) Abbreviations: lex = Lexical, orth = Orthographic overlap, rep = Repeated, and sem = Semantic overlap. • Lexical stimuli were matched for familiarity, frequency, imageability and syllable length, while all stimuli were matched for letter length. • Twelve native British English speakers (5F, 7M) participated in two scanning sessions (B0=3T, gradient echo EPI, TR = 3sec, TE = 30msec, FOV = 192 x 256mm, matrix = 64 x 64). • Median RTs for correct responses were calculated per condition per subject and used in the behavioural analyses. • The functional imaging data was realigned to correct for small head movements, spatially smoothed with a 5mm3 FWHM Gaussian filter, and registered into standard space with an affine transformation. Group results were analysed with a random effects analysis. • A spherical region-of-interest [center = (–42, –56, –16), radius = 1cm] limited results to the left occipito-temporal region. Figure 7: Neural repetition priming effects. The left part of the figure shows the mean percent BOLD signal change per condition relative to consonant strings for words and pseudowords within the left posterior fusiform region. Only words showed a signficant repetition priming effect which is shown in the right part of the figure. Relative to unrelated word pairs, repeated words reduced signal in the left posterior fusiform region, even though the prime and target were not in the same case. 1500ms PASS 200ms passive 33ms $%##&@#@ 500ms Time * 1000ms Behavioural results • One subject had reaction times (RTs) more than 200msec longer than the group mean was therefore removed from all subsequent analyses. • Accuracy across conditions was 96.6%, indicating that subjects had no difficulty performing the task. • Responses to words were faster than to pseudowords (F1,10=111.8, p<0.001) • Repetition priming was present for words but not pseudowords (F1,10=4.4, p=0.06) • There was a significant main effect of Form (F1,10=8.2, p<0.05) indicating that when prime-target pairs shared visual form, this facilitated responses Effects of Form and Meaning Table 1: Sample stimuli in the lexical decision task Condition Lexical/Repetition Form/Meaning Example 1. Unrelated [+lex, –rep] [–orth, –sem] event – RUG 2. Identity [+lex, +rep] plant – PLANT 3. Pseudowords [–lex, –rep] dollar – TAVE 4. Identical pseudo [–lex, +rep] purit – PURIT 5. Orthographic [+orth, –sem] winsome –WIN 6. Semantic [–orth, +sem] profit – GAIN 7. Morphological [+orth, +sem] hunter – HUNT 8. Consonant letter string baseline brutally – PLB Effects of Lexicality * 800 700 600 • Within the posterior fusiform region commonly activated by words and pseudowords, the effects of form and meaning relations were assessed using a 2 × 2 ANOVA • There was a significant main effect of Form (F1,21=5.3, p<0.05) and a significant interaction (F1,21=5.0, p<0.05) but no main effect of Meaning • In other words, words which shared visual form led to a significant reduction in BOLD signal relative to unrelated words, in contrast to the prediction of the lexical hypothesis • Moreover, this effect was reduced when the words also shared meaning, consistent with the visual identification hypothesis Unrelated Repeated (*) Reaction Times (msec) Unrelated Related in meaning Related in form Related in both form & meaning * 0.50 0.25 0.00 * Mean % BOLD signal Words Pseudowords Figure 4: Reaction times to words and pseudowords in the lexical decision task. Figure 8: Effects of form and meaning relations on BOLD signal in the left posterior fusiform cortex. Effects of form and meaning * 650 625 600 Unrelated Related in meaning Related in form Related in both form & meaning Reaction Times (msec) Unrel Sem Orth Morph Figure 5: Reaction times to orthographically and semantically related word pairs.