Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Jun 1.
Published in final edited form as: Neuropsychologia. 2017 Apr 7;100:35–43. doi: 10.1016/j.neuropsychologia.2017.04.003

An fMRI study of visual hemifield integration and cerebral lateralization

Lars Strother 1, Zhiheng Zhou 1, Alexandra K Coros 2, Tutis Vilis 2
PMCID: PMC5484576  NIHMSID: NIHMS868362  PMID: 28396097

Abstract

The human brain integrates hemifield-split visual information via interhemispheric transfer. The degree to which neural circuits involved in this process behave differently during word recognition as compared to object recognition is not known. Evidence from neuroimaging (fMRI) suggests that interhemispheric transfer during word viewing converges in the left hemisphere, in two distinct brain areas, an “occipital word form area” (OWFA) and a more anterior occipitotemporal “visual word form area” (VWFA). We used a novel fMRI half-field repetition technique to test whether or not these areas also integrate nonverbal hemifield-split string stimuli of similar visual complexity. We found that the fMRI responses of the both the OWFA and VWFA while viewing nonverbal stimuli were strikingly different than those measured during word viewing, especially with respect to half-stimulus changes restricted to a single hemifield. We conclude that normal reading relies on left-lateralized neural mechanisms, which integrate hemifield-split visual information for words but not for nonverbal stimuli.

Keywords: cerebral lateralization, hemispheric transfer, word recognition, visual cortex, fMRI

1. Introduction

Due to the division of neural fibers from the nasal hemiretinae in the optic chiasm and uncrossed fibers from the temporal hemiretinae, sensory information available in each visual hemifield is initially projected to the contralateral occipital lobe and subsequently combined via interhemispheric transfer. In addition to its role in binocular vision (Mitchell & Blakemore, 1970), interhemispheric transfer plays an integral role in word recognition (Brysbaert, 2004; Hunter, Brysbaert, & Knecht, 2007; Monaghan & Shillcock, 2008), face perception (Bridgman et al., 2014; Hsiao, Shieh, & Cottrell, 2008), the detection of mirror symmetry (Herbert & Humphrey, 1996; Saarinen & Levi, 2000) and other forms of perceptual organization (Genç, Bergmann, Singer, & Kohler, 2011; Pillow & Rubin, 2002). The precise nature of mechanisms involved in the hemispheric integration of foveal input is controversial (Ellis & Brysbaert, 2010), but hemispheric transfer is necessary for the neural integration of all hemifield-split visual input, even within a degree of visual angle from the vertical midline of the visual field (Berlucchi, 2014; Reinhard & Trauzettel-Klosinski, 2003). Given the reliance of reading upon visual processing within this range of the visual field, interhemispheric transfer clearly plays an integral role in the neural integration of hemifield-split words during reading (Dougherty, Ben-Shachar, Bammer, Brewer, & Wandell, 2005; Lavidor & Walsh, 2004), and its disruption is associated with dyslexia (Henderson, Barca, & Ellis, 2007).

Recently, Strother, Coros, & Vilis (2016) used functional MRI (fMRI) to reveal a hemispheric asymmetry in the visual integration of letters comprising a word split at fixation. Specifically, they reported an “occipital word form area” (OWFA), which contains neurons that bind hemifield-split word parts into a unitary bilateral hemifield word form representation. Their finding suggests that this process occurs earlier in visual cortex than proposed by models of hemifield integration in an anatomically anterior occipitotemporal “visual word form area”(VWFA; Cohen et al., 2003; Molko et al., 2002). Based on the Talairach coordinates reported by Strother et al., the OFA/OWFA is located either within or near the inferior occipital gyrus, intermediate to V4v (hV4) in the transverse collateral sulcus (Witthoft et al., 2014), and object-selective neurons on or near the lateral occipital gyrus (Larsson & Heeger, 2006; Strother, Aldcroft, Lavell, & Vilis, 2010), possibly corresponding to a coarsely retinotopic putative human posterior inferior temporal area (phPIT) or a putative V4 transitional (pV4t) zone (Kolster, Peeters, & Orban, 2010). In contrast, the VWFA is considerably more anterior, in fusiform cortex (lateral to the middle section of the fusiform gyrus), typically lateral and/or anterior to visual field maps VO-1 and VO-2, and inferior and medial to visual field maps TO-1 and TO-2 (Yeatman, Rauschecker, & Wandell, 2013). Strother et al. showed that the OWFA in left occipital cortex was precisely symmetric in anatomical location relative to an “occipital face area” (OFA; Gauthier et al., 2000) in the right hemisphere. The OFA is typically larger and more frequently found in the right hemisphere (Pitcher, Walsh, & Duchaine, 2011), and it represents visual features of faces and spatial relations between them during the early stages of processing (Liu, Harris, & Kanwisher, 2010; Pitcher, Walsh, Yovel, & Duchaine, 2007; Rhodes, Michie, Hughes, & Byatt, 2009; Strother et al., 2011). The right OFA is distinct from its more elusive left counterpart in its sensitivity to mirror and its role in interhemispheric integration during face recognition (Bona, Cattaneo, & Silvanto, 2015; Frässle et al., 2016; Kietzmann et al., 2015).

In addition to a growing interest in the parallels between word recognition and the visual processing of faces (Behrmann & Plaut, 2013; Dehaene, Cohen, Morais, & Kolinsky, 2015; Dehaene et al., 2010; Nestor, Behrmann, & Plaut, 2013), there is also considerable longstanding interest in the degree to which potentially word-specific mechanisms are engaged during the visual processing of nonverbal stimuli (Seghier & Price, 2011; Vogel, Petersen, & Schlaggar, 2012, 2014). Given the antipodal anatomical relationship between the right OFA and the OWFA reported by Strother et al. (2016), it is reasonable to expect that the role of the OWFA in word recognition exhibits some parallels with the role of the OFA in face perception, such as the representation of constituent “parts” of a configuration. A limitation of the study by Strother et al. is that the authors did not test whether or not the OWFA integrates hemifields-split parts of non-word configurations. Here we report results from two experiments that employed the same fMRI method used by Strother et al., and an original reanalysis of a subset of their data, to show that both the OWFA and the VWFA selectively bind hemifield-split letters of a word but not non-letter parts of nonverbal configurations. Specifically, we used non-verbal stimuli, which either repeated or changed in full, or repeated/changed on one half or the other. Our logic in using these conditions was the same as Strother et al. Suppression of fMRI responses should only occur for neural populations in the hemisphere contralateral to the hemifield location of repetition unless these neurons receive ipsilateral input (e.g. via the corpus callosum); we were specifically interested in asymmetries of contralateral and ipsilateral hemifield-specific repetition suppression or release from suppression. Our results are consistent with results of some previous studies of the VWFA, but emphasize the role of more posterior regions of visual cortex, the OWFA in particular. Our results also emphasize the utility of half-field manipulations of hemifield-split stimuli in conjunction with fMRI, which can improve our understanding of word recognition and also the process of inter-hemifield integration in object recognition more generally.

2. Material and methods

2.1. Participants

Twelve right-handed volunteers (21 – 27 years of age; 8 female) participated in Experiment 1; the subjects were a subset of those who participated in the study by Strother et al. (2016). Twelve different right-handed observers (20 – 33 years of age; 9 female) participated in Experiment 2. All participants were literate native English speakers and were literate in English only. All participants were recruited from the University of Western Ontario (London, Ontario, Canada), and all consent forms and experimental procedures described in these forms were approved by the University of Western Ontario’s research ethics board.

2.2. fMRI data acquisition and analysis

Imaging was conducted at the Robarts Research Institute (London, Ontario, Canada) using a 3-T Siemens Tim MAGNETOM Trio imaging system. BOLD data were collected using T2*-weighted interleaved, single segment, EPI, PAT = 2, and a 32-channel head coil (Siemens, Erlangen, Germany). Foam padding was used to reduce head motion. Functional data were aligned to high-resolution anatomical images obtained using a 3-D T1 MPRAGE sequence (echo time [TE] = 2.98 ms; repetition time [TR] = 2300 ms; inversion time = 900 ms; flip angle = 9°; 192 contiguous 1 mm slices; field of view = 240 × 256 mm2). Each functional volume included 33 contiguous slices. Scanning parameters for obtaining functional data with full coverage of OT: TE = 30; TR = 2 sec (single shot); flip angle = 90°; field of view = 148 × 148 mm2; 2 × 2 × 2 mm3 voxel size. Each run of the main experiment included 204 volumes.

Data were preprocessed and analyzed using BrainVoyager QX 2.1 (BVQX; Brain Innovation, Maastricht, The Netherlands). We performed corrections for slice scan time, head motion (always <2 mm), and low-frequency artefactual drift (linear trend removal and high pass filter of 3 cycles/run); each functional volume for a given participant was aligned to the functional volume collected closest in time to the anatomical volume. Functional data were superimposed on anatomical brain images, aligned on the AC–PC line, and transformed into Talairach (Talairach, Rayport, & Tournoux, 1988) space and co-registered with the anatomical image for each participant. Talairach transformation was performed using standard BVQX procedures (Goebel, 1996). The hemispheres were segmented at the gray/white matter boundary, and the resultant cortical sheet was then reconstructed, inflated, and flattened for functional data analyses and visualization. Functional data were spatially smoothed using a Gaussian kernel of 8 mm (FWHM). Predictors were generated using rectangular wave functions (with a value of 1 for l volume = 2 sec when the action was initiated at the onset of the inter-trial interval and a value of 0 for the remainder of the trial) that were convolved with a hemodynamic response function (Boynton, Engel, Glover, & Heeger, 1996).

2.3. Stimuli and Procedure

Both Experiment 1 and 2 employed the same general procedure as Strother et al. (2016), but with different stimuli. Figure 1 shows stimuli and conditions from the Strother et al. experiment and the experiments reported here. All word (Figure 1a), silhouette (Figure 1b), and Japanese character (Figure 1c) stimuli extended to a visual angle of ~ 5° × 1.5° (viewed via mirror at 15 cm distance). In all cases, observers fixated a small (~ 0.05°) dot centered on the screen. Silhouette string stimuli in Experiment 1 were comprised of four animal shapes. The Japanese character stimuli used in Experiment 2 were comprised of four different characters (including Kanji and Kana), and were roughly equated in spatial frequency to the word stimuli in the previous study (but not the silhouettes in Experiment 1). All stimuli were split in half between the left (LVF) and right (RVF) visual hemifields.

Figure 1.

Figure 1

Open in a new tab

Stimuli and conditions used in Strother et al. (2016) and Experiments 1 and 2: (a) words; (b) silhouettes; and (c) Japanese character strings. The top row shows paired examples of Same (repeated) stimuli. The second row shows examples of Different (non-repeating) stimuli. The third and bottom rows show RVF and LVF change stimuli, respectively, for which half the string repeats and half changes between successive presentations within a block. Note that although pairs are shown here, blocks contained sequences of twelve stimuli (indicated by dots).

As in Strother et al., four experimental conditions were used in each experiment: Same, Different, RVF change, and LVF change. A 12-s block design was adopted for both experiments identical to the previous study, and within each block, 12 stimuli were presented at a rate of 1 Hz, with words/strings displayed cyclically (until the end of a 12 s block) for 500 ms followed by 500 ms blank screen. For the Same condition, each block contained the same four string components repeated for 12 times. For the Different condition, each block contained 12 different stimuli changing in both LVF and RVF. For the RVF change condition, each block contained 12 stimuli with the same stimulus repeated in the LVF, but changing stimuli in the RVF, and vice versa for the LVF change condition. There were 34 blocks in each run, and 8 blocks per condition with 2 fixation blocks (1 block in the beginning and 1 block at the end). Ninety-six words sharing the left two letters and ninety-six words sharing the right two letters were used as stimuli in the Strother et al. experiment (stimuli for Same and Different conditions were subsets of these); twelve different words sharing the same first or last two letters were used in each block (12 × 8 blocks per run = 96 total). Silhouette and Japanese character stimuli were created by substituting single silhouettes or characters with each letter in the word stimuli. In all experiments, block order was counterbalanced across runs and each volunteer participated in a minimum of four runs.

3. Results

3.1. ROI analyses

We were primarily interested in the OWFA and VWFA reported by Strother et al. (2016). We therefore present region of interest (ROI) analyses of the OWFA and VWFA, and their anatomical counterparts in the right hemisphere, for each of the two new experiments and reanalyzed results from the Strother et al. study. Some of our results are reported in Supplementary Materials. We also report results from group-level whole brain analyses, both here and in Supplementary Materials.

We defined the OWFA and its right hemisphere counterpart based on anatomically mirror symmetric Talairach coordinates of the OFA reported by Strother et al. (2016). The OFA coordinates reported in their study were obtained in an independent experiment from their main study, the results of which we compare to those obtained in the new experiments reported here (i.e., our ROI definitions are not based on any of the fMRI results reported here, thus avoiding circularity and lack of independence). The OWFA ROI was centered at Talairach coordinates x = −39, y = −80, z = −10, anatomically mirror symmetric to the OFA (x = 39, y = −80, z = −10) reported by Strother et al. Rather than use the VWFA coordinates reported by Strother et al. (which was not defined using an independent localizer in their study), we defined the coordinates of our VWFA ROI based on coordinates reported by Glezer, Kim, Rule, Jiang, & Riesenhuber (2015), transformed from MNI (x = −42, y = −58, z = −17) to Talairach coordinates (x = −40, y = −59, z = −10); we also defined an anatomically mirror symmetric right hemisphere VWFA (rVWFA) centered at 40, y = −59, z = −10. All ROIs consisted of a ~800-voxel cube centered on the reported Talairach coordinates from which fMRI beta values were extracted for analysis.

3.1.1. ROI Analyses for Experiment 1 (silhouettes)

Figure 3a shows average beta values for all conditions, for the OWFA and OFA from Experiment 1 (silhouettes). A two-way repeated measures ANOVA was conducted on betas for each condition (Same, Different, RVF change, and LVF change) and ROI (OWFA and OFA), followed by paired tests with Bonferroni correction for multiple comparisons. The ANOVA results showed a main effect of condition, F(3, 33) = 46.39, p < .001, and an interaction between condition and ROI, F(3, 33) = 33.08, p < .001. The main effect of ROI was not significant, F(1, 33) = 1.13, p = .31.

Figure 3.

Figure 3

Open in a new tab

Results of whole brain analysis from Experiment 1, Strother et al. (2016), and Experiment 3. The smaller scale brain shows the activation yielded by the Different > Same contrast for each experiment, at q(FDR) < .05 (a) and (b), p < .005 uncorrected, (c). We observed repetition suppression in both the left and right vOTC for all three experiments. On the larger scale brain, the activation yielded by the RVF change > LVF change contrast is shown in blue and the LVF change > RVF Change contrast is shown in red. The thresholds are set at p < .01, uncorrected, for all three experiments. We observed there is no left hemisphere activation for RVF change > LVF change for Strother et al. (b), but the same contrast yield left vOTC activation for experiment 1 (a) and 2 (c). However, we observed that LVF change > RVF Change activated the right vOTC in all three experiments. White lines indicate regions showed repetition suppression observed from the smaller scale brain.

Results from all paired comparisons are reported in Supplemental Materials (Table S.1). We were mainly interested in comparisons of betas between the Same and Different conditions, and the RVF change and LVF change conditions. In the OWFA, Different betas (M = 0.804, SE = 0.072) were significantly larger than Same betas (M = 0.430, SE = 0.041), p < .001; and in the OFA, Different betas (M = 0.910, SE = 0.092) were also significantly larger than Same betas (M = 0.457, SE = 0.047), p < .001. These results suggest significant repetition suppression in both the OWFA and OFA for silhouettes.

We were especially interested in whether or not the OWFA and OFA would show different magnitudes of fMRI response to the RVF change and LVF change conditions. Bonferroni-corrected paired comparisons confirmed significant differences in both ROIs: in the OWFA, RVF change betas (M = 0.668, SE = 0.066) were significantly larger than LVF change betas (M = 0.500, SE = 0.040), p < .01; and in the OFA, LVF change betas (M = 0.840, SE = 0.086) were significantly larger than RVF change betas (M = 0.539, SE = 0.076), p < .001. In short, the results showed that the OWFA and OFA responses to the two change conditions were not the same, but rather, showed a contralateral bias (contralateral > ipsilateral fMRI responses). Results of all possible paired comparisons are summarized in Table S.1.

Figure 3b shows results for the VWFA and rVWFA from Experiment 1 (silhouettes). A two-way repeated measures ANOVA was conducted prior to paired comparisons. The results showed a main effect of condition, F(3, 33) = 22.01, p < .001, and an interaction between condition and ROI, F(3, 33) = 16.63, p < .001. The main effect of ROI was marginally significant, F(1, 33) = 4.32, p = .062. Paired comparisons with Bonferroni correction (see Table S.1) indicated that, in the VWFA, Different betas (M = 0.330, SE = 0.052) were significantly larger than Same betas (M = 0.156, SE = 0.032), p < .005; and in the rVWFA, Different betas (M = 0.436, SE = 0.054) were also significantly larger than Same betas (M = 0.202, SE = 0.034), p < .001. This means repetition suppression occurred in both the VWFA and the rVWFA for nonverbal string stimuli.

As in the OWFA and OFA, we were mainly interested in whether or not the VWFA and rVWFA would show different fMRI responses to the RVF change and LVF change conditions. In the VWFA, paired comparison revealed that RVF change betas (M = 0.292, SE = 0.045) were not significantly differ from LVF change betas (M = 0.227, SE = 0.037), p = .11; in the rVWFA, LVF change betas (M = 0.385, SE = 0.042) were significantly larger than RVF change betas (M = 0.292, SE = 0.044), p < .001. In addition to our Bonferroni-corrected paired comparisons we conducted an additional considerably more liberal statistical test for non-significant paired comparison—a one-sample t-test (one-tailed) based on a contralateral sensitivity index (CSI), computed by subtracting the fMRI responses to LVF change from RVF change (for the left hemisphere ROIs, OWFA and VWFA) or subtracting the fMRI responses to RVF change from LVF change (for the right hemisphere ROIs, OFA and rVWFA), and dividing by their sum (i.e. CSILeft = (RRVF change - RLVF change)/(RRVF change + RLVF change); CSIRight = (RLVF change – RRVF change)/(RLVF change + RRVF change)). The result showed significant differences for the VWFA, t(11) = 2.09, p = .03. In short, the VWFA and rVWFA both showed greater fMRI responses to contralateral half-stimulus changes as compared to ipsilateral half-stimulus changes, although not to the same degree (statistical significance for the rVWFA test survived Bonferroni correction whereas that for the VWFA did not); this will be important in subsequent between-experiment comparisons of ROI results. To summarize the results of our analyses, we observed repetition suppression in all ROIs, and also a contralateral bias in all ROIs.

3.1.2. Reanalyzed results from Strother et al. (2016)

The 12 participants in Experiment 1 comprised a subset of the 18 participants who participated in the split-words study by Strother et al. (2016). Our main goal was to characterize the similarities and differences between the two experiments in the same set of participants, with a particular focus on beta values in the RVF change and LVF change conditions, which Strother et al. did not compare directly their published study. We also present results for overall repetition suppression.

Figure 3c shows results for the OWFA and the OFA of a reanalysis of data from Strother et al. (same 12 subjects as Experiment 1). We again performed a two-way repeated measures ANOVA, which showed similar results to those observed in Experiment 1 (a main effect of condition, F(3, 33) = 40.42, p < .001, and an interaction between condition and ROI, F(3, 33) = 8.68, p < .001). The main effect of ROI was not significant, F(1, 33) = 1.05, p = .33. Subsequent paired comparisons results are shown in Table S.2. In the OWFA, Different betas (M = 0.736, SE = 0.048) were significantly larger than Same betas (M = 0.420, SE = 0.028), p < .001; and in the OFA, Different betas (M = 0.834, SE = 0.105) were significantly larger than Same betas (M = 0.543, SE = 0.080), p < .001. Additional paired comparisons showed that RVF change betas (M = 0.667, SE = 0.047) were not significantly different from LVF change betas (M = 0.657, SE = 0.046), p = 1.00. Unlike the OWFA, LVF change betas (M = 0.835, SE = 0.101) were significantly larger than RVF change betas (M = 0.661, SE = 0.088) in the OFA, p < .005. We performed a more liberal statistical test on the non-significant results using the CSI. Unlike our corresponding result for Experiment 1, the result of this one-sample t-test on the CSI was still not significant: t(11) = 0.40, p = .35 (one-tailed). In Figure 3c, the “NS” refers to non-significance in this more liberal statistical test, and we interpret it to suggest a lack of contralateral bias in the OWFA for half-word changes.

Figure 3d shows results for VWFA and rVWFA. We performed a two-way repeated measures ANOVA on betas for condition and ROI, which showed a main effect of condition, F(3, 33) = 40.06, p < .001, and an interaction between condition and ROI regions, F(3, 33) = 3.80, p = .019. The main effect of ROI region was marginally significant, F(1, 33) = 4.62, p = .055. All of the following paired comparisons are reported in Table S.2. In the VWFA, Different betas (M = 0.736, SE = 0.048) were significantly larger than Same betas (M = 0.420, SE = 0.028), p < .001; and in the rVWFA, Different betas (M = 0.431, SE = 0.054) were significantly larger than Same betas (M = 0.260, SE = 0.033), p < .001.

In the VWFA, paired comparison revealed that RVF change betas (M = 0.298, SE = 0.042) did not differ from LVF change betas (M = 0.284, SE = 0.044), p = 1.00; and in the rVWFA, LVF change betas (M = 0.416, SE = 0.054) were larger than RVF change betas (M = 0.355, SE = 0.052), p = .015. When we used a more liberal one-sample t-test using the CSI (one-tailed), the result showed that the CIS for the VWFA was still not significant, t(11) = 0.89, p = .39 (Figure 3d, “NS”). Taken together, the VWFA (like the OWFA) showed equivalent fMRI responses to RVF change and LVF change for words, but not for silhouettes (Experiment 1), whereas the rVWFA showed a clear contralateral bias for LVF versus RVF half-word changes.

3.1.3. ROI Analyses for Experiment 2 (Japanese characters)

Figure 3e shows results for the OWFA and OFA from Experiment 2 (Japanese characters). We again performed a two-way repeated measures ANOVA, which showed a main effect of condition, F(3, 33) = 24.72, p < .001, a main effect of ROI, F(1, 33) = 7.75, p < .05, and an interaction between condition and ROI, F(3, 33) = 32.52, p < .001. The main effect of ROI indicates that fMRI responses in the OWFA were lower than those in the OFA. All of the following paired comparisons are reported in Table S.3. Paired comparisons with Bonferroni correction indicated that, in the OWFA, Different betas (M = 0.625, SE = 0.093) were larger than Same betas (M = 0.425, SE = 0.071), p < .001; and in the OFA, Different betas (M = 0.761, SE = 0.092) were significantly larger than Same betas (M = 0.540, SE = 0.071), p < .001. Additional paired comparisons showed that RVF change betas in the OWFA (M = 0.614, SE = 0.089) were larger than LVF change betas (M = 0.457, SE = 0.070), p = .014; and in the OFA, LVF change betas (M = 0.800, SE = 0.094) were significantly larger than RVF change betas (M = 0.650, SE = 0.071), p < .005. Thus, as in Experiment 1 (but not in the reanalyzed data from Strother et al. 2016), clear contralateral half-change biases were observed in the OWFA and the OFA.

Figure 3f shows results for the VWFA and rVWFA from Experiment 2 (Japanese characters). A two-way repeated measures ANOVA on betas for condition and ROI revealed: a main effect of condition, F(3, 33) = 16.38, p < .001, a main effect of ROI, F(1, 33) = 11.87, p < .005, and an interaction between condition and ROI, F(3, 33) = 15.55, p < .001. As before, we performed Bonferroni-corrected paired comparisons which are reported in Table S.3. In the VWFA, Different betas (M = 0.185, SE = 0.026) were larger than Same betas (M = 0.108, SE = 0.017), p = .011; and in the rVWFA, Different betas (M = 0.322, SE = 0.053) were significantly larger than Same betas (M = 0.209, SE = 0.040), p < .005. Again, we observed repetition suppression in both the VWFA and rVWFA for Japanese characters (as in the OWFA and OFA).

Finally, with respect to contralateral bias for half-stimulus changes, RVF change betas in the VWFA (M = 0.200, SE = 0.026) were larger than LVF change betas (M = 0.145, SE = 0.022), p = .014. In the rVWFA, LVF change betas (M = 0.332, SE = 0.058) did not differ from RVF change betas (M = 0.298, SE = 0.053), p = .16. We again used one-sample t-tests (one-tailed) to examine the CSI for the rVWFA. The result showed a significant effect, t(11) = 3.04, p < .01. In short, the results of these analyses showed a contralateral bias, as observed in Experiment 1, but which was absent in the reanalyzed data of Strother et al. (2016). Comprehensive paired comparisons for this experiment are reported in Supplemental Materials (Table S.4.)

3.1.4. Direct comparison of ROIs results between experiments

To this point we have restricted our analyses to each experiment separately from the others. Here directly compare results between all three. Our first analysis investigated the differences in the effects of the contralateral bias to different stimulus types used across experiments for each ROI. A two-way mixed ANOVA on betas for ROI (OWFA and OFA) and experiment (silhouettes, word, and Japanese characters) revealed a main effect of ROI, F(1, 33) = 4.98, p = .033, a main effect of experiment, F(2, 33) = 11.19, p < .001, and a two-way interaction between ROI and experiment, F(2, 33) = 5.82, p < .01. This means the OWFA and the OFA showed different patterns of contralateral sensitivity depending on stimulus type (i.e. experiment), which we clarify next.

As suggested by the individual ROI × experiment analyses performed earlier, in the OWFA, a one-way between subjects ANOVA on the CIS for experiment suggested that there was a significant difference for silhouettes, word, and Japanese characters, F(2, 33) = 8.76, p < .001. Pairwise comparisons with Bonferroni correction indicated that the CSI for word (M = 0.008, SE = 0.020) was smaller than for silhouettes (M = 0.137, SE = 0.030) and Japanese characters (M = 0.155, SE = 0.033), both p values < .01, but there was no difference between results for silhouettes and Japanese characters (p = 1.00). In the OFA, a one-way between subjects ANOVA suggested that there was also a significant difference for silhouettes, word, and Japanese characters, F(2, 33) = 7.32, p < .005. Pairwise comparisons with Bonferroni correction indicated that CSI for silhouettes (M = 0.236, SE = 0.031) was larger than for word (M = 0.132, SE = 0.028) and Japanese characters (M = 0.093, SE = 0.022), both p values < .05, and there was no different between word and Japanese characters, p = 0.99. Taken together, these tests further confirm our earlier within-experiment results, which showed that whereas the OWFA showed equivalent sensitivity (CSI = 0.008) to both RVF change and LVF change for words, but a contralateral bias (CSI > .1) for silhouettes and Japanese character, the OFA showed a contralateral bias across all three experiments.

Next, we conducted a two-way mixed ANOVA on the CSI for ROI (VWFA and rVWFA) and experiment (silhouettes, word, and Japanese characters). Unlike the results for the OWFA and OFA across experiments, direct comparison of the VWFA and rVWFA across experiments failed to show any statistically significant differences—no significant effect of ROI (F(1, 33) = 0.21, p = .65), experiment, (F(2, 33) = 2.33, p = .11), or the interaction between ROI and experiment (F(2, 33) = 0.92, p = .41). This means that despite the lack of contralateral bias for half-word changes reported earlier for the VWFA, the VWFA and rVWFA behaved more similarly across experiments—with respect to contralateral bias (i.e., contralateral bias was always relatively weak as compared to the OWFA and OFA)—than did the OWFA as compared to the OFA. This suggests a distinction between the OWFA and VWFA in terms of contralateral sensitivity.

A second analysis investigated the lateralization of the repetition suppression between the OWFA and OFA, or the VWFA and rVWFA across all three experiments. We computed a suppression index (SI) for each ROI, which was defined by subtracting the fMRI responses to Same condition from Different condition, and dividing by their sum, SI = (RDifferent – RSame)/( RDifferent + RSame). A two-way mixed ANOVA on SI for ROI (OWFA and OFA) and experiment (silhouettes, word, and Japanese characters) revealed a main effect of experiment, F(2, 33) = 4.28, p = .022. There was no main effect of ROI, F(1, 33) = 0.41, p = .53, or interaction between ROI and experiment, F(2, 33) = 1.94, p = .16. This indicates that the suppression effect greatest for Experiment 1, but the degree repetition suppression in Experiment 2 and the reanalyzed data from Strother et al. (2016) were not different from each other. This observation required that we compare using the SI because of overall differences in the magnitude of the BOLD signal between the two experiments.

We repeated the previous analysis (two-way mixed ANOVA) on the SI for ROI (VWFA and rVWFA) and experiment (silhouettes, word, and Japanese characters). We observed a main effect of ROI, F(1, 33) = 4.51, p = .041, and an interaction between ROI and experiment, F(2, 33) = 4.00, p = .028. The main effect of experiment was not significant, F(2, 33) = 1.07, p = .36. Paired comparisons with Bonforroni correction indicated that, in Experiment 1, the SI for the VWFA (M = 0.316, SE = 0.084) was not significantly different from the SI for the rVWFA (M = 0.329, SE = 0.100), p = .84; in the reanalyzed data from Strother et al. (2016), the SI for the VWFA (M = 0.532, SE = 0.093) was significantly larger than the SI for the rVWFA (M = 0.256, SE = 0.029), p = .014; but in Experiment 2, the SI for the VWFA (M = 0.261, SE = 0.104) was not significantly different from the SI for the rVWFA (M = 0.236, SE = 0.045), p = .74.

Taken together, the SI analyses showed evidence of left lateralization of repetition suppression for words, and although right lateralization was not significant in Experiments 1 and 2, we found no evidence of left lateralization in these experiments. We revisit the possibility of right-lateralized repetition suppression for nonverbal stimuli in whole brain analyses in the next section. With respect to contralateral bias, which was of primary interest, our results confirm the lack of contralateral bias in the OWFA. While this was also observed in the VWFA, contralateral bias in the rVWFA was relatively weak as compared to the OFA, consistent with expected decreasing contralateral bias in general moving posterior-to-anterior in visual cortex (Hemond, Kanwisher, & Op de Beeck, 2007). Furthermore, in contrast to the VWFA, the OWFA showed greater fMRI responses during the Different condition as compared to the RVF change and LVF change conditions (Tables S.1 to S.4 in Supplemental Materials), as reported by Strother et al., although this effect was not limited to words and therefore suggests that OWFA representations may be pre-lexical.

3.2. Whole brain analysis

The goals of the whole brain analysis were to (1) determine the regions showed repetition suppression, (2) search for evidence of lateralization, and (3) investigate the contralateral sensitivity of each hemisphere to the half change stimulus in Experiment 1 and 2, and also in Strother et al. (2016). We reported the whole brain analysis based on three contrasts: Same > Different, RVF change > LVF change, and LVF change > RVF change. A percent signal change transformation and a correction for serial correlations were applied to each participant data, and a random effects GLM was applied to group data. Talairach coordinates of the voxels survived after correction for cluster size (> 10) were reported.

3.2.1 Repetition suppression

We first identified the voxels that showed significant repetition suppression by the different > same contrast based on a false discovery rate q(FDR) of .05 for Experiment 1 (which showed the greatest repetition suppression) and the previous study by Strother et al. (Figures 3a and 3b, smaller brains in upper left). For Experiment 2, no repetition suppression was found at q(FDR) < .05 (i.e., repetition suppression was relatively weaker overall as compared to the other experiments), thus we applied a statistical threshold of p < .005 uncorrected which yielded roughly equal numbers of voxels as compared to the previous study by Strother et al. (Figure 3c, smaller brain in upper left). We then performed voxel counts for each experiment to test for overall lateralization of repetition suppression (recall that in the previous section we focused only on our ROIs).

In Experiment 1, we found that there were 33905 voxels activated in the left hemisphere with roughly equal numbers of voxels in the right (32686 voxels) suggesting no lateralization between left and right hemispheres. This is consistent with results from the ROI analysis which also did not find evidence of lateralization in all the ROIs. However, when we applied a more restricted statistical threshold (q(FDR) < .005), we observed more voxels were activated in the right hemisphere (numbers of voxels = 4138) than in the left hemisphere (numbers of voxels = 3555), see Figures S.1a and S.2a in Supplementary Materials. Taken together, the voxel counts show no evidence of left lateralization (consistent with the SI results), but rather, the possibility of right-lateralized repetition suppression for nonverbal silhouettes stimuli.

The whole brain reanalyzed results of our previous study showed that there were fewer voxels activated in Experiment 1 (Figure 3b), which suggests greater repetition suppression in Experiment 1 (recall that the same 12 participants were included in the both experiments, and the same statistical thresholds were applied (q(FDR) < .05)). More importantly, we observed more voxels in the left hemisphere (numbers of voxels = 19025) than in the right hemisphere (numbers of voxels = 16792), which indicates left-lateralized repetition suppression for words (Figure 3b, small brain in upper left). As before, a more restrict threshold at q(FDR) < .005 was applied to the Different > Same contrast (Figures S1b and S2b in Supplementary Materials), and we again observed more activation in the left hemisphere (numbers of voxels = 2164) than in the right hemisphere (numbers of voxels = 748). In short, whole-brain analysis showed clear left lateralization of repetition suppression for words, which was not observed in Experiment 1.

In Experiment 3, at the threshold of p < .005 uncorrected, we found that there were more voxels activated in the right hemisphere (numbers of voxels = 20064) than in the left hemisphere (numbers of voxels = 15018). This right lateralization (2690 voxels in the right hemisphere and 2073 voxels in the left hemisphere) was also evident when we increased the threshold to p < .0005 uncorrected (Figures S1c and S2c in Supplementary Materials). In short, the whole-brain analysis of the repetition suppression in Experiment 3 showed an overall right lateralization for Japanese character strings, more pronounced than that observed in Experiment 1 (silhouettes); both experiments showed an opposite pattern of lateralization for repetition suppression than the left lateralization observed for words.

3.2.2. Contralateral bias

Our final whole-brain analyses focused on contralateral bias and the prospective relative lateralization of the bias or its absence (i.e., given that the OWFA and VWFA did not show contralateral bias for words). We identified brain regions that showed contralateral bias by the RVF change > LVF change and LVF change > RVF change contrasts at a statistical threshold of q(FDR) < .05. The results showed that the left vOTC and right early visual cortex (in the vicinity of visual cortical area V4) were activated for RVF change > LVF change and LVF change > RVF change contrasts in Experiment 1, respectively. However, this later proved problematic for comparison with Experiment 2 (and the reanalyzed data) because of the relatively small magnitude of the difference between fMRI responses in the RVF change versus LVF change conditions in these data compared to those from Experiment 1 (as observed in the ROI analyses of CSI reported earlier). When we used a more liberal threshold (p < .01, uncorrected), equated across experiments (Figure 3a to 3c), we observed different patterns of contralateral bias depending on experiment. In Experiment 1 (Figure 3a), contralateral bias was observed in both hemispheres. In both hemispheres, this overlapped highly with the repetition suppression results (white outlines in Figure 3a), overlapped with the OWFA (but not the VWFA, at the chosen threshold), and was more pervasive in the right hemisphere than in the left.

In the reanalyzed data of Strother et al. (2016), contralateral bias was observed in the right hemisphere only, even at an extremely liberal threshold (p < .01, uncorrected; Figure 3b). In the right hemisphere, the voxels showing a contralateral bias overlapped considerably with the repetition suppression results (white outlines in Figure 3b). This indicates that contralateral bias throughout right vOTC, especially in areas showing repetition suppression, but a complete lack of contralateral bias in the left hemisphere (i.e., not restricted to the OWFA and VWFA), a result that was unique to the reanalyzed results of Strother et al.

In Experiment 2 (Figure 3c), contralateral bias was again observed in both hemispheres, as in Experiment 1. In both hemispheres, this overlapped highly with the repetition suppression results (white outlines in Figure 3c), and also overlapped with both the OWFA and VWFA, and was more pervasive in the right hemisphere than in the left. In short, the whole brain analysis of contralateral bias showed lack of contralateral bias in the left hemisphere only to word stimuli, which is consistent to the previous ROI analysis results.

To summarize, along with the results from the ROI analysis, we found different patterns of lateralization, both in terms of repetition suppression, and also contralateral bias for half-stimulus changes. In contrast to Experiments 1 and 2, repetition suppression in the reanalyzed results Strother et al.’s data showed left-lateralization of repetition suppression for word stimuli, and right lateralization with respect to contralateral bias (i.e., contralateral bias was uniformly weak in the left hemisphere, much more than in Experiments 1 and 2).

4. Discussion

Recent studies have reported left-lateralized involvement of occipital cortex in visual word form processing (e.g., Cohen, Dehaene, McCormick, Durant, & Zanker, 2016; Yu, Jiang, Legge, & He, 2015), and Strother et al. (2016), proposed a specific function of neurons in this “occipital word form area” (OWFA)—to combine hemifield-split letters into a unified pre-lexical representation of visual word form. We performed two experiments and a reanalysis of a subset of Strother et al.’s data to study inter-hemifield integration of words as compared to nonverbal visual stimuli. Both experiments employed an fMRI repetition suppression technique used by Strother et al. to identify an occipital word form area (OWFA) in the left hemisphere, which integrates letterform information split between the right (RVF) and left (LVF) visual hemifields. Our study focused on the OWFA and the visual word form area (VWFA) first reported by Cohen et al. (2000), as well as anatomically mirror-symmetric locations in the right hemisphere. The VWFA was initially characterized as a neural population that equivalently represents words presented to either the contralateral (RVF) or ipsilateral (LVF) visual hemifield (Cohen et al., 2002). In conjunction with other studies, this led to a model of visual word form processing in which hemifield-split letters are initially processed bilaterally in visual cortex, but eventually combined into a whole-word representation in the VWFA (Cohen et al., 2003). Our findings are consistent with this model but extend it to a more posterior anatomical locus in left occipital cortex, the OWFA. Crucially, we show that while both the OWFA and the VWFA participate in the inter-hemifield integration of visual input, this occurs to a lesser degree, if at all, for nonverbal stimuli. In addition to this main result, our experiments revealed interesting hemifield-hemisphere relationships in the visual processing of nonverbal visual stimuli.

In Experiment 1, in viewed strings of familiar animal silhouettes. Like words, these nonverbal stimuli were comprised of easily recognized two-dimensional shapes. However, unlike words, the animal silhouette strings formed no meaningful configuration. Our ROI analyses showed repetition suppression for repeated silhouette strings, and more importantly, consistently greater fMRI responses to contralateral versus ipsilateral half-string changes versus repetitions in all ROIs, including the OWFA and VWFA (i.e., both showed fMRI responses to RVF change > LVF change), which demonstrates greater contralateral versus ipsilateral visual hemifield sensitivity for nonverbal visual stimuli split between the RVF and LVF. While this result is consistent with heightened fMRI responses to whole objects viewed in the contralateral visual hemifield, even in anterior regions of visual cortex (Hemond et al., 2007; Sayres & Grill-Spector, 2008), it was not guaranteed given our stimuli and methods, which differed substantially from previous studies of contralateral bias in visual cortex. Our finding of contralateral bias for hemifield-split strings of silhouettes—which, like object images, elicit strong shape-related fMRI responses in visual cortex (Kourtzi & Kanwisher, 2001)—therefore extends previous findings of contralateral bias for objects viewed in isolation to those presented as part of a configuration viewed centrally. Additionally, a whole-brain analysis showed that overall repetition suppression was right-lateralized, consistent with other findings of right-lateralization during object recognition (Nakamura et al., 2005). A second whole brain analysis showed that the degree of contralateral bias was greater in the right hemisphere than in the left. While this may be related to an asymmetry of visual attention between the RVF and LVF, especially in the context of serially presented visual stimuli (Matthews & Welch, 2015), no hemispheric asymmetries of contralateral bias were observed in the single-object studies mentioned earlier (Hemond et al., 2007; Sayres & Grill-Spector, 2008).

The main purpose of Experiment 1 was to compare the results to those obtained from the same participants in a previously published study. We therefore reanalyzed a subset of data (12 of the original 18 participants) from Strother et al. (2016), using new analyses, and we used the results of these analyses to perform direct comparisons with the results of Experiment 1. In terms of repetition suppression for whole strings—silhouettes in Experiment 1 and whole words in the published study—we observed some noteworthy differences between the two experiments. First, greater suppression in the right hemisphere than in the left (in whole-brain analyses), and the opposite pattern for words (which was evident in both ROI and whole-brain analyses). Left-lateralized repetition suppression for words was expected, but nevertheless validated our results with respect to known patterns of lateralization for words as compared to nonverbal stimuli.

Our comparison of primary interest was the degree to which fMRI responses in the OWFA and VWFA differed, or not, for the half-word change conditions (LVF change and RVF change conditions, in which letters in one hemifield changed but repeated in the other). In contrast to the contralateral bias observed Experiment 1, such that left hemisphere ROIs (OWFA and VWFA) showed greater responses to RVF half-word changes, and vice versa for the right hemisphere ROIs (OFA and rVWFA), neither the OWFA nor the VWFA showed any difference in fMRI responses to contralateral versus ipsilateral half-word changes. This was not true of either the OFA or the rVWFA, both of which showed heightened fMRI responses to contralateral half-word changes. This result is consistent with the results Strother et al. (2016), and it is also a novel reanalysis of data from a subset of their participants, which was necessary for direct comparison with Experiment 1 (in which data were collected from the same participants). In short, the results of our reanalysis of published data and direct comparison with the results of Experiment 1 show that equivalent sensitivity of the OWFA and VWFA to half-word changes, irrespective of hemifield, is not necessarily observed for nonverbal stimuli, a prediction we tested further in Experiment 2.

The lack of contralateral bias for half-word changes in the OWFA and VWFA was further substantiated by a whole-brain comparison of fMRI responses in the LVF change and RVF change conditions for words as compared to silhouette strings—indeed, in contrast to the whole-brain analysis results for Experiment 1, no significant differences in fMRI responses between the two half-word change conditions were observed in the entire left hemisphere (in a group level RFX analysis). This means that, contralateral bias for half-word changes is relatively weak in the left hemisphere as compared to the right hemisphere, and that the OWFA and VWFA show concomitant maximal repetition suppression to words and minimal or non-existent contralateral bias for half-word changes, despite contralateral bias for half-stimulus changes in Experiment 1. While this result is consistent with the predictions of some existing models of word recognition (Cohen et al., 2003; Molko et al., 2002), these models predict our VWFA results only, not the word-specific lack of contralateral bias in the OWFA. Our results also highlight the relationship between our results and our method—in contrast to presenting whole words or whole objects to either the RVF or LVF (e.g., Cohen et al., 2002; Hemond et al., 2007; Niemeier, Goltz, Kuchinad, Tweed, & Vilis, 2005; Rauschecker, Bowen, Parvizi, & Wandell, 2012; Sayres & Grill-Spector, 2008), we used a novel half-field fMRI method applied to words and nonverbal strings. A possible criticism is that our word stimuli were artificially large, which could be worth studying in future experiments.

A limitation of our comparison of the Experiment 1 results is that our choice of nonverbal stimuli were definitively nonverbal. That is, none of our participants would have mistaken them for linguistic stimuli. In Experiment 2 we therefore used Japanese character string stimuli that were recognized as linguistic but processed as nonverbal stimuli because all of our participants were monolingual English readers. Experiment 2 served three main purposes. First, it involved a new set of participants. Second, the stimuli were more closely matched in spatial frequency to those used in our previous study with words, but nevertheless elicited right-lateralized repetition suppression (as in Experiment 1). This is extremely important given known hemispheric differences in sensitivity to high versus low spatial frequencies (e.g., Roberts et al., 2013; Woodhead, Wise, Sereno, & Leech, 2011). Third, and more importantly, the results of Experiment 2 replicated the results of Experiment 1 with respect to contralateral bias for half-stimulus changes, which was observed in all ROIs (both hemispheres) and also in a whole-brain analysis. That is, the only evidence of equivalent fMRI responses to contralateral and ipsilateral half-stimulus changes was observed for words, in the OWFA and the VWFA, further evidence that neurons in these areas are involved in visual processing that is unique to word recognition.

We interpret the results of our experiments, and the comparison of these results of a subset of previously published data analyzed using different methods, as evidence of word-specific hemifield integration in the OWFA and VWFA. While the VWFA result is consistent with existing models and experimental findings (Cohen et al., 2003; Molko et al., 2002), including finding from fMRI repetition suppression (Glezer, Jiang, & Riesenhuber, 2009), our observation of the same effect in the OWFA suggests that inter-hemifield integration of word form information occurs outside of the VWFA, and possibly during an earlier pre-lexical stage of shape processing than predicted by existing models of whole-word representation in the visual system. Interestingly, a similar proposal has been offered recently to explain right-lateralized face processing (Frässle et al., 2016), which exhibits striking parallels with word recognition (Behrmann & Plaut, 2013; Dehaene et al., 2015; Dundas, Plaut, & Behrmann, 2012), especially with respect to symmetry processing (Bona et al., 2015; Kietzmann et al., 2015), which may have a parallel in the OWFA for “symmetry breaking” in visual word form processing (possibly in addition to the VWFA; Pegado, Nakamura, Cohen, & Dehaene, 2011). Furthermore, given that defective callosal transfer and inter-hemispheric coordination is associated with dyslexia (Fabbro et al., 2001; Henderson et al., 2007), our findings highlight the possibility that at least some cases of dyslexia may be the result of impaired inter-hemifield integration (Kelly, Jones, McDonald, & Shillcock, 2004).

Supplementary Material

1
2
3
4

Figure 2.

Figure 2

Open in a new tab

Results of ROI analyses from Experiment 1 (top row), Strother et al. (2016; second row), and Experiment 2 (bottom row). Four ROIs are paired by OWFA and OFA (left), and by VWFA and rVWFA (right). All ROIs show repetition suppression (Different > Same) for Experiment 1, Strother et al. (2016), and Experiment 2. Importantly, in Experiment 1, (a) and (b), and 2, (e) and (f), all ROIs show contralateral bias for half change nonverbal stimulus (RVF change > LVF change in the OWFA and VWFA, LVF change > RVF change in the OFA and rVWFA). However, in Strother et al. (2016), (c) and (d), the OWFA and VWFA show equal fMRI responses to both left and right half-word changes (dashed circle; RVF change = LVF change) while the OFA and rVWFA still exhibit contralateral bias to left half-word change (LVF change > RVF change). “NS” refers to a lack of significant difference when using a liberal statistical test (see text for details).

Highlights.

  • Results are reported from an fMRI half-field repetition paradigm

  • Left and right hemispheres show distinct patterns of repetition suppression

  • Half-field suppression is different for words and non-verbal stimuli

  • An occipital word form area (OWFA) underlies split-word binding

Acknowledgments

Research funded by Canadian Institutes of Health grant 9335 to T. Vilis and the National Institute of General Medical Sciences of the National Institutes of Health under grant number P20 GM103650.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Behrmann M, Plaut DC. Distributed circuits, not circumscribed centers, mediate visual recognition. Trends Cogn Sci. 2013;17(5):210–219. doi: 10.1016/j.tics.2013.03.007. [DOI] [PubMed] [Google Scholar]
  2. Berlucchi G. Visual interhemispheric communication and callosal connections of the occipital lobes. Cortex. 2014;56:1–13. doi: 10.1016/j.cortex.2013.02.001. [DOI] [PubMed] [Google Scholar]
  3. Bona S, Cattaneo Z, Silvanto J. The causal role of the occipital face area (OFA) and lateral occipital (LO) cortex in symmetry perception. J Neurosci. 2015;35(2):731–738. doi: 10.1523/JNEUROSCI.3733-14.2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Boynton GM, Engel SA, Glover GH, Heeger DJ. Linear systems analysis of functional magnetic resonance imaging in human V1. J Neurosci. 1996;16(13):4207–4221. doi: 10.1523/JNEUROSCI.16-13-04207.1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bridgman MW, Brown WS, Spezio ML, Leonard MK, Adolphs R, Paul LK. Facial emotion recognition in agenesis of the corpus callosum. J Neurodev Disord. 2014;6(1):32. doi: 10.1186/1866-1955-6-32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Brysbaert M. The importance of interhemispheric transfer for foveal vision: a factor that has been overlooked in theories of visual word recognition and object perception. Brain Lang. 2004;88(3):259–267. doi: 10.1016/S0093-934X(03)00279-7. [DOI] [PubMed] [Google Scholar]
  7. Cohen L, Dehaene S, McCormick S, Durant S, Zanker JM. Brain mechanisms of recovery from pure alexia:a single case study with multiple longitudinal scans. Neuropsychologia. 2016 doi: 10.1016/j.neuropsychologia.2016.07.009. [DOI] [PubMed] [Google Scholar]
  8. Cohen L, Dehaene S, Naccache L, Lehericy S, Dehaene-Lambertz G, Henaff MA, et al. The visual word form area: spatial and temporal characterization of an initial stage of reading in normal subjects and posterior split-brain patients. Brain. 2000;123(Pt 2):291–307. doi: 10.1093/brain/123.2.291. [DOI] [PubMed] [Google Scholar]
  9. Cohen L, Lehericy S, Chochon F, Lemer C, Rivaud S, Dehaene S. Language-specific tuning of visual cortex functional properties of the Visual Word Form Area. Brain. 2002;125:1054–1069. doi: 10.1093/brain/awf094. [DOI] [PubMed] [Google Scholar]
  10. Cohen L, Martinaud O, Lemer C, Lehéricy S, Samson Y, Obadia M, et al. Visual word recognition in the left and right hemispheres: anatomical and functional correlates of peripheral alexias. Cereb Cortex. 2003;13(12):1313–1333. doi: 10.1093/cercor/bhg079. [DOI] [PubMed] [Google Scholar]
  11. Dehaene S, Cohen L, Morais J, Kolinsky R. Illiterate to literate: behavioural and cerebral changes induced by reading acquisition. Nat Rev Neurosci. 2015;16(4):234–244. doi: 10.1038/nrn3924. [DOI] [PubMed] [Google Scholar]
  12. Dehaene S, Pegado F, Braga LW, Ventura P, Nunes Filho G, Jobert A, et al. How learning to read changes the cortical networks for vision and language. Science. 2010;330(6009):1359–1364. doi: 10.1126/science.1194140. [DOI] [PubMed] [Google Scholar]
  13. Dougherty RF, Ben-Shachar M, Bammer R, Brewer AA, Wandell BA. Functional organization of human occipital-callosal fiber tracts. Proc Natl Acad Sci U S A. 2005;102(20):7350–7355. doi: 10.1073/pnas.0500003102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Dundas EM, Plaut DC, Behrmann M. The Joint Development of Hemispheric Lateralization for Words and Faces. J Exp Psychol Gen. 2012 doi: 10.1037/a0029503. [DOI] [PMC free article] [PubMed]
  15. Ellis AW, Brysbaert M. Split fovea theory and the role of the two cerebral hemispheres in reading: A review of the evidence. Neuropsychologia. 2010;48(2):353–365. doi: 10.1016/j.neuropsychologia.2009.08.021. [DOI] [PubMed] [Google Scholar]
  16. Fabbro F, Pesenti S, Facoetti A, Bonanomi M, Libera L, Lorusso ML. Callosal transfer in different subtypes of developmental dyslexia. Cortex. 2001;37(1):65–73. doi: 10.1016/s0010-9452(08)70558-6. [DOI] [PubMed] [Google Scholar]
  17. Frässle S, Paulus FM, Krach S, Schweinberger SR, Stephan KE, Jansen A. Mechanisms of hemispheric lateralization: Asymmetric interhemispheric recruitment in the face perception network. Neuroimage. 2016;124(Pt A):977–988. doi: 10.1016/j.neuroimage.2015.09.055. [DOI] [PubMed] [Google Scholar]
  18. Gauthier I, Tarr MJ, Moylan J, Skudlarski P, Gore JC, Anderson AW. The fusiform “face area” is part of a network that processes faces at the individual level. J Cogn Neurosci. 2000;12(3):495–504. doi: 10.1162/089892900562165. [DOI] [PubMed] [Google Scholar]
  19. Genç E, Bergmann J, Singer W, Kohler A. Interhemispheric connections shape subjective experience of bistable motion. Curr Biol. 2011;21(17):1494–1499. doi: 10.1016/j.cub.2011.08.003. [DOI] [PubMed] [Google Scholar]
  20. Glezer LS, Jiang X, Riesenhuber M. Evidence for Highly Selective Neuronal Tuning to Whole Words in the “Visual Word Form Area”. Neuron. 2009;62(2):199–204. doi: 10.1016/j.neuron.2009.03.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Glezer LS, Kim J, Rule J, Jiang X, Riesenhuber M. Adding Words to the Brain’s Visual Dictionary: Novel Word Learning Selectively Sharpens Orthographic Representations in the VWFA. J Neurosci. 2015;35(12):4965–4972. doi: 10.1523/JNEUROSCI.4031-14.2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Goebel R. Brainvoyager: A program for analyzing and visualizing functional and structural magnetic resonance data sets. Neuroimage. 1996;3(3):S604. [Google Scholar]
  23. Hemond CC, Kanwisher NG, Op de Beeck HP. A preference for contralateral stimuli in human object- and face-selective cortex. PLoS One. 2007;2(6):e574. doi: 10.1371/journal.pone.0000574. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Henderson L, Barca L, Ellis AW. Interhemispheric cooperation and non-cooperation during word recognition: evidence for callosal transfer dysfunction in dyslexic adults. Brain Lang. 2007;103(3):276–291. doi: 10.1016/j.bandl.2007.04.009. [DOI] [PubMed] [Google Scholar]
  25. Herbert AM, Humphrey GK. Bilateral symmetry detection: testing a ‘callosal’ hypothesis. Perception. 1996;25(4):463–480. doi: 10.1068/p250463. [DOI] [PubMed] [Google Scholar]
  26. Hsiao JH, Shieh DX, Cottrell GW. Convergence of the visual field split: hemispheric modeling of face and object recognition. J Cogn Neurosci. 2008;20(12):2298–2307. doi: 10.1162/jocn.2008.20162. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hunter ZR, Brysbaert M, Knecht S. Foveal word reading requires interhemispheric communication. J Cogn Neurosci. 2007;19(8):1373–1387. doi: 10.1162/jocn.2007.19.8.1373. [DOI] [PubMed] [Google Scholar]
  28. Kelly ML, Jones MW, McDonald SA, Shillcock RC. Dyslexics’ eye fixations may accommodate to hemispheric desynchronization. Neuroreport. 2004;15(17):2629–2632. doi: 10.1097/00001756-200412030-00014. [DOI] [PubMed] [Google Scholar]
  29. Kietzmann TC, Poltoratski S, König P, Blake R, Tong F, Ling S. The Occipital Face Area Is Causally Involved in Facial Viewpoint Perception. J Neurosci. 2015;35(50):16398–16403. doi: 10.1523/JNEUROSCI.2493-15.2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kolster H, Peeters R, Orban GA. The retinotopic organization of the human middle temporal area MT/V5 and its cortical neighbors. J Neurosci. 2010;30(29):9801–9820. doi: 10.1523/JNEUROSCI.2069-10.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kourtzi Z, Kanwisher N. Representation of perceived object shape by the human lateral occipital complex. Science. 2001;293(5534):1506–1509. doi: 10.1126/science.1061133. [DOI] [PubMed] [Google Scholar]
  32. Larsson J, Heeger DJ. Two retinotopic visual areas in human lateral occipital cortex. J Neurosci. 2006;26(51):13128–13142. doi: 10.1523/JNEUROSCI.1657-06.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lavidor M, Walsh V. Opinion - The nature of foveal representation. Nature Reviews Neuroscience. 2004;5(9):729–735. doi: 10.1038/nrn1498. [DOI] [PubMed] [Google Scholar]
  34. Liu J, Harris A, Kanwisher N. Perception of face parts and face configurations: an FMRI study. J Cogn Neurosci. 2010;22(1):203–211. doi: 10.1162/jocn.2009.21203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Matthews N, Welch L. Left visual field attentional advantage in judging simultaneity and temporal order. J Vis. 2015;15(2) doi: 10.1167/15.2.7. [DOI] [PubMed] [Google Scholar]
  36. Mitchell DE, Blakemore C. Binocular depth perception and the corpus callosum. Vision Res. 1970;10(1):49–54. doi: 10.1016/0042-6989(70)90061-1. [DOI] [PubMed] [Google Scholar]
  37. Molko N, Cohen L, Mangin JF, Chochon F, Lehericy S, Le Bihan D, et al. Visualizing the neural bases of a disconnection syndrome with diffusion tensor imaging. J Cogn Neurosci. 2002;14(4):629–636. doi: 10.1162/08989290260045864. [DOI] [PubMed] [Google Scholar]
  38. Monaghan P, Shillcock R. Hemispheric dissociation and dyslexia in a computational model of reading. Brain Lang. 2008;107(3):185–193. doi: 10.1016/j.bandl.2007.12.005. [DOI] [PubMed] [Google Scholar]
  39. Nakamura K, Oga T, Okada T, Sadato N, Takayama Y, Wydell T, et al. Hemispheric asymmetry emerges at distinct parts of the occipitotemporal cortex for objects, logograms and phonograms: a functional MRI study. Neuroimage. 2005;28(3):521–528. doi: 10.1016/j.neuroimage.2004.11.055. [DOI] [PubMed] [Google Scholar]
  40. Nestor A, Behrmann M, Plaut DC. The neural basis of visual word form processing: a multivariate investigation. Cereb Cortex. 2013;23(7):1673–1684. doi: 10.1093/cercor/bhs158. [DOI] [PubMed] [Google Scholar]
  41. Niemeier M, Goltz HC, Kuchinad A, Tweed DB, Vilis T. A contralateral preference in the lateral occipital area: sensory and attentional mechanisms. Cereb Cortex. 2005;15(3):325–331. doi: 10.1093/cercor/bhh134. [DOI] [PubMed] [Google Scholar]
  42. Pegado F, Nakamura K, Cohen L, Dehaene S. Breaking the symmetry: mirror discrimination for single letters but not for pictures in the Visual Word Form Area. Neuroimage. 2011;55(2):742–749. doi: 10.1016/j.neuroimage.2010.11.043. [DOI] [PubMed] [Google Scholar]
  43. Pillow J, Rubin N. Perceptual completion across the vertical meridian and the role of early visual cortex. Neuron. 2002;33(5):805–813. doi: 10.1016/s0896-6273(02)00605-0. [DOI] [PubMed] [Google Scholar]
  44. Pitcher D, Walsh V, Duchaine B. The role of the occipital face area in the cortical face perception network. Experimental Brain Research. 2011;209(4):481–493. doi: 10.1007/s00221-011-2579-1. [DOI] [PubMed] [Google Scholar]
  45. Pitcher D, Walsh V, Yovel G, Duchaine B. TMS evidence for the involvement of the right occipital face area in early face processing. Curr Biol. 2007;17(18):1568–1573. doi: 10.1016/j.cub.2007.07.063. [DOI] [PubMed] [Google Scholar]
  46. Rauschecker AM, Bowen RF, Parvizi J, Wandell BA. From the Cover: Position sensitivity in the visual word form area. Proc Natl Acad Sci U S A. 2012;109(24):E1568–1577. doi: 10.1073/pnas.1121304109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Reinhard J, Trauzettel-Klosinski S. Nasotemporal overlap of retinal ganglion cells in humans: a functional study. Invest Ophthalmol Vis Sci. 2003;44(4):1568–1572. doi: 10.1167/iovs.02-0313. [DOI] [PubMed] [Google Scholar]
  48. Rhodes G, Michie PT, Hughes ME, Byatt G. The fusiform face area and occipital face area show sensitivity to spatial relations in faces. Eur J Neurosci. 2009;30(4):721–733. doi: 10.1111/j.1460-9568.2009.06861.x. [DOI] [PubMed] [Google Scholar]
  49. Roberts DJ, Woollams AM, Kim E, Beeson PM, Rapcsak SZ, Lambon Ralph MA. Efficient visual object and word recognition relies on high spatial frequency coding in the left posterior fusiform gyrus: evidence from a case-series of patients with ventral occipito-temporal cortex damage. Cereb Cortex. 2013;23(11):2568–2580. doi: 10.1093/cercor/bhs224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Saarinen J, Levi DM. Perception of mirror symmetry reveals long-range interactions between orientation-selective cortical filters. Neuroreport. 2000;11(10):2133–2138. doi: 10.1097/00001756-200007140-00015. [DOI] [PubMed] [Google Scholar]
  51. Sayres R, Grill-Spector K. Relating retinotopic and object-selective responses in human lateral occipital cortex. J Neurophysiol. 2008;100(1):249–267. doi: 10.1152/jn.01383.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Seghier ML, Price CJ. Explaining left lateralization for words in the ventral occipitotemporal cortex. J Neurosci. 2011;31(41):14745–14753. doi: 10.1523/JNEUROSCI.2238-11.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Strother L, Aldcroft A, Lavell C, Vilis T. Equal degrees of object selectivity for upper and lower visual field stimuli. J Neurophysiol. 2010;104(4):2075–2081. doi: 10.1152/jn.00462.2010. [DOI] [PubMed] [Google Scholar]
  54. Strother L, Coros AM, Vilis T. Visual Cortical Representation of Whole Words and Hemifield-split Word Parts. J Cogn Neurosci. 2016;28(2):252–260. doi: 10.1162/jocn_a_00900. [DOI] [PubMed] [Google Scholar]
  55. Strother L, Mathuranath PS, Aldcroft A, Lavell C, Goodale MA, Vilis T. Face inversion reduces the persistence of global form and its neural correlates. PLoS ONE. 2011;6(4):e18705. doi: 10.1371/journal.pone.0018705. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Talairach J, Rayport M, Tournoux P. Co-planar stereotaxic atlas of the human brain : 3-dimensional proportional system: an approach to cerebral imaging. Stuttgart; Thieme u.a: 1988. [Google Scholar]
  57. Vogel AC, Petersen SE, Schlaggar BL. The Left Occipitotemporal Cortex Does Not Show Preferential Activity for Words. Cereb Cortex. 2012 doi: 10.1093/cercor/bhr295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Vogel AC, Petersen SE, Schlaggar BL. The VWFA: it’s not just for words anymore. Front Hum Neurosci. 2014;8:88. doi: 10.3389/fnhum.2014.00088. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Witthoft N, Nguyen ML, Golarai G, LaRocque KF, Liberman A, Smith ME, et al. Where is human V4? Predicting the location of hV4 and VO1 from cortical folding. Cereb Cortex. 2014;24(9):2401–2408. doi: 10.1093/cercor/bht092. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Woodhead ZV, Wise RJ, Sereno M, Leech R. Dissociation of sensitivity to spatial frequency in word and face preferential areas of the fusiform gyrus. Cereb Cortex. 2011;21(10):2307–2312. doi: 10.1093/cercor/bhr008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Yeatman JD, Rauschecker AM, Wandell BA. Anatomy of the visual word form area: adjacent cortical circuits and long-range white matter connections. Brain Lang. 2013;125(2):146–155. doi: 10.1016/j.bandl.2012.04.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Yu D, Jiang Y, Legge GE, He S. Locating the cortical bottleneck for slow reading in peripheral vision. J Vis. 2015;15(11):3. doi: 10.1167/15.11.3. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS868362-supplement-1.TIF (23.3MB, TIF)
2
NIHMS868362-supplement-2.TIF (22MB, TIF)
3
NIHMS868362-supplement-3.docx (1.1MB, docx)
4
NIHMS868362-supplement-4.docx (1.1MB, docx)

RESOURCES