Friday, August 7, 2015

Red Like You've Never Seen It Before

In his recent book, The Tell-Tale Brain, V. S. Ramachandran describes some cases of synesthesia.  A synesthete's  reaction to a stimulus can involve multiple senses.  If the stimulus is the sound of a trombone, for example, a synesthete hears the trombone, but might also see a colour, such as light blue.  While synesthetic cross-sensory effects are well known, p. 114 of the book reports a lesser-known phenomenon: some synesthetes see colours that are “unreal” or “Martian,” colours they have never seen previously.  This Hue Angles column hypothesizes a plausible cause and speculates about its implications.

A commonly suggested explanation for synesthesia is “cross-wiring” in the brain.  The idea is that certain stimuli activate not only the receptors they should activate, but also some that they shouldn't.  The sound of a trombone, for example, ordinarily affects only the aural apparatus and related brain regions.  In a synesthete, however, the signal pathway for a trombone's sound might overlap with the signal pathway for light blue, or the trombone circuit might cause activity in the light blue circuit, as if the two circuits were linked by a relay.

 Such a linkage might explain the perception of unreal colours.  The figure shows the Stockman-Sharpe cone fundamentals for the eye's red, green, and blue receptor cones.  Colorimetry is based on the colour-matching functions which, at least in theory, are linear transformations of these fundamentals, so ultimately these response curves determine human colour perceptions.  In particular, the overlaps in the response curves limit our perceptions.  For all light stimuli, except those very near the infra-red, the green cones will respond only if the red cones also respond.  Similarly, a physical stimulus cannot excite the blue cones appreciably without also  exciting the red and green cones slightly. 




 If, however, cone combinations were not limited by physical stimuli, colours could be produced that were outside our usual perceptual limits.  Synesthesia, it is suggested, can produce cone combinations that could not be produced by physical light sources, so the resulting colours are physically unreal.  Suppose, for example, your brain was wired so that only the red cone response was received.  Then the red signal would exist without the green signal, which, as we have seen, is an impossibility for physical stimuli.  You would see only a red, but since the red was unadulterated with other signals, it would be a red like you've never seen red before.  

 Any attempts to find such a red in the real world would be doomed to failure.  Whether the stimulus came directly from a light source, or indirectly, after reflection off an object colour, none of its wavelengths would be able to stimulate just the red cone any significant amount, without stimulating the other cones, too.  The new red would be “unreal,” because it could never have the same colour as any physical stimulus.

 A similar unreality occurs with so-called ''supersaturated'' colours, discussed by Glenn Fry, and also by Hurvich and Jameson.  A receptor can be temporarily adapted so that it is not responsive; this adaptation is often offered as an explanation for complementary afterimages.  If the green receptor were briefly inactive, then a stimulus between 550 and 700 nm would produce a red that was more chromatic than any real red.  
  
 An interesting question is the hue of a synesthetic or supersaturated colour.  The chromaticity of a pure red cone perception would be a point R outside the standard chromaticity diagram.  Suppose a neutral chromaticity N is chosen in the chromaticity diagram.  Join R and N by a straight line, that intersects the spectrum locus (which, along with the purple line, makes up the diagram's boundary)
at some point D.  By definition, D would give the dominant wavelength of the perception R.  Dominant wavelength is roughly equivalent to hue, but the two are not identical.  In fact, lines of constant hue curve noticeably as they radiate outward from N.  If R is far enough outside the diagram, then the curving might be significant, so D might not be a very good hue indicator at all.  It is possible, perhaps, that the perception resulting from a pure red cone might actually be an orange or a purple---which would be a whole new hue angle.

Paul Centore

Dr. Centore is a freelance colour scientist who is available for colour-related projects. He combines technical ability with an art/design/graphics background. For more information, see www.99main.com/~centore or send an email to centore@99main.com .