Figure 3. Spatial mapping of selectivity to hierarchical orthographic stimuli.
(a) MEMA activation map showing the regions of significant activation to the real word stimuli during passive viewing (100 – 400 ms, 27 patients). This activation map was used as a mask and a normalization factor for the activations of the non-word stimuli (b). Normalized amplitude maps showing regions with preferential activation to words (red) or non-words (blue). (c) Electrode selectivity profiles grouped in 20 mm intervals along the y (antero-posterior) axis in Talairach space from all word responsive electrodes (207 electrodes, 20 patients). Horizontal dashed lines represent word response. Individual data points are overlaid. (d) Contrasts of the lettered non-words against words for electrodes within mid-fusiform cortex (e; 39 electrodes, 13 patients), showing latency differences for when each non-word category is distinguished from words. Coloured bars under the plots represent regions of significant difference from words (one-tailed Wilcoxon signed rank, q<0.01). (f) Spatial map of the initial timing of significant word selectivity (207 electrodes, 20 patients). Electrodes that did not reach significance shown in black.