The Tell-Tale Brain: A Neuroscientist's Quest for What Makes Us Human

3. In lower, “projection,” synesthetes there are several lines of evidence (in addition to segregation) supporting the low-level perceptual cross-activation model as opposed to the notion that synesthesia is based entirely on high-level associative learning and memories:

     (a) In some synesthetes, different parts of a single number or letter are seen as colored differently. (For example, the V part of an M might be colored red, whereas the vertical lines might be green.)

     Soon after the popout/segregation experiment had been done, I noticed something strange in one of the many synesthetes we had been recruiting. He saw numbers as being colored—nothing unusual so far—but what surprised me was his claim that some of the numbers (for example, 8) had different portions colored differently. To make sure he wasn’t making this up, we showed him the same numbers a few months later—without letting him know ahead of time that he would be retested. The new drawing he produced was virtually identical to the first, making it unlikely that he was fibbing.

     This observation provides further evidence that, at least in some synesthetes, the colors should be seen as emerging from (to use a computer metaphor) a glitch in neural hardware rather than from an exaggeration of memories or metaphors (a software glitch) Associative learning cannot explain this observation; for example, we don’t play with multicolored magnets. On the other hand, there may be “form primitives” such as line orientation, angles, and curves that get linked to color neurons that execute an earlier stage of form processing within the fusiform than the one at which full-fledged graphemes are assembled.

     (b) As previously noted, in some synesthetes the evoked color becomes less vivid when the number is viewed off-axis (in peripheral vision). This probably reflects the greater emphasis on color in central vision (Ramachandran & Hubbard, 2001a; Brang & Ramachandran, 2010). In some of these synesthetes the color is also more saturated in one visual field (left or right) relative to the other. Neither of these observations supports the high-level associative learning model for synesthesia.

     (c) An actual increase in anatomical connectivity within the fusiform area of lower synesthetes has been observed by Rouw and Scholte (2007) using diffusion tensor imaging.

     (d) The synesthetically evoked color can provide an input to apparent motion perception (Ramachandran & Hubbard, 2002; Kim, Blake, Palmeri, 2006; Ramachandran & Azoulai, 2006).

     (e) If you have one type of synesthesia, then you are more likely to have a second unrelated one as well. This supports my “increased cross activation model” of synesthesia; with the mutated gene being more prominently expressed in certain brain regions (in addition to making some synesthetes more creative).

     (f) The existence of color-blind (strictly speaking, color anomalous) synesthetes who can see colors in numbers that they can’t see in the real world. The subject couldn’t have learned such associations.

     (g) Ed Hubbard and I showed in 2004 that letters that are similar in shape (e.g., curvy rather than angular) tend to evoke similar colors in “lower” synesthetes. This shows that certain figural primitives that define the letters cross-activate colors even before they are fully processed. We suggested that the technique might be used to map an abstract color-space in a systematic manner onto form-space. More recently David Brang and I confirmed this using brain imaging (MEG or magnetoencephalography) in collaboration with Ming Xiong Huang, Roland Lee, and Tao Song.

     Taken collectively these observations strongly support the sensory cross-activation model. This is not to deny that learned associations and high-level rules of cross-domain mapping are not also involved (see Notes 8 and 9 for this chapter). Indeed, synesthesia may help us discover such rules.

4. The model of cross-activation—either through disinhibition (a loss or lessening of inhibition) of back projections, or through sprouting—can also explain many forms of “acquired” synesthesia that we have discovered. One blind patient with retinitis pigmentosa whom we studied (Armel and Ramachandran, 1999) vividly experienced visual phosphenes (including visual graphemes) when his fingers were touched with a pencil or when he was reading Braille. (We ruled out confabulation by measuring thresholds and demonstrating their stability across several weeks; there is no way he could have memorized the thresholds.) A second blind patient, whom I tested with my student Shai Azoulai, could quite literally see his hand when he waved it in front of his eyes, even in complete darkness. We suggest that this is caused either by hyperactive back projections or by disinhibition caused by visual loss, so that the moving hand is not merely felt but is also seen. Cells with multimodal receptive fields in the parietal lobes may also be involved in mediating this phenomenon (Ramachandran and Azoulai, 2004).