Interestingly, although Braille imagery activated the VWFA significantly less than vOICe SSD letters, it did generate widespread activations learn more as compared to passively hearing the letter names (which controls for both auditory stimulation and semantic content; see Figure S1B). One area of activation is of particular interest given theories on mental imagery originally framed in the context of vision (Kosslyn et al., 1999): we found robust activation to Braille imagery as compared to the semantic control in the hand area of S1 (Figure S1C; t = 6.5, p < 0.000001). This mental imagery reactivation was specific to the relevant part of the somatosensory homunculus,

as we found no such effect in the S1 foot area (p < 0.36). Moreover, we also found Braille imagery activation in the left vOT (Figure S1B; t = 4.6, p < 0.000005; see also Figure 3C, showing a similar selleck compound effect for vOICe imagery). Thus, our results demonstrate that imagery in the blind generates a pattern of activation similar to that seen when comparing visual perception and visual mental imagery in the sighted. Ventral visual cortex activation for imagery in the blind, as in the sighted, (1) is specific to the stimulus-selective cortical location (O’Craven and Kanwisher, 2000), in our case in the VWFA, and (2) is significantly less intense

than bottom-up perception of the same stimuli (Amedi et al., 2005; O’Craven and Kanwisher, 2000). Moreover, as in sighted subjects, imagery in the blind can generate activation in the primary sensory cortex related to the stimulus modality and location—in our case in the hand area of S1 (Kosslyn et al., 1999). Finally, we investigated a rather unique case of a single congenitally blind subject, T.B., who was highly literate in Braille but was completely unfamiliar with the shapes of the sighted alphabet in her native language (Hebrew). This allowed us to test

whether the VWFA could be recruited for reading using Mephenoxalone an SSD (i.e., in a novel modality) in a new script in the adult brain after a brief 2 hr training period (e.g., without enabling long-term plasticity). We taught T.B. to identify complex geometric shapes by using the vOICe SSD (see details in the Supplemental Experimental Procedures) but refrained from teaching her the shapes of letters. We then scanned subject T.B. twice in a single day, before and after a single 2 hr session of learning to read several letters of the regular alphabet using vOICe. We compared the activation for reading in the tactile and auditory modalities with modality-matched nonreading controls to look for reading-selective activations both in an ROI located at the VWFA and across the entire brain. Braille reading (BR; contrasted with its modality-matched control, Braille control, BC, homogenous Braille dots) activated a left-vOT/VWFA peak identically in both scanning sessions (Figure 4A).