Rethinking the Process of Vision

Reading Hazel Rossotti’s “Color, Why the World Isn’t Grey” (Princeton University Press).

From Chapter 14, “The Eye and the Brain”, p.130

She writes, referring to an earlier collection of selected writings from Newton, Goethe, DaVinci, etc. (”Sensations of Colour” pp104-107):

“…..how can we reconcile the phenomena they describe with our earlier interpretation of colour vision in terms of the absorption of photons by the three cone pigments?”

It is obvious from my explanation for light interaction with the retina that the eye evolved to detect the wave nature of light. It is the absorbing mass of the cone and rod retinal receptors that act as quantized electron (’nanowire’) sites thus preserving the quantum nature of the interaction. This interacton can (and should) be properly termed a ‘quantum interaction’ and avoiding the mental construction that ‘photons interact with pigment molecules’. The retina is actually a precise, geometrically spaced, array of receptor nanowires with the sub-optical wavelength spacings between them acting as tuned optical ‘antennas’. A great deal follows from, and is explained by, this realization..

A very prescient comment by Rossotti on p.130:

“Some of the psychological responses to colour may have a simple geometrical origin.”

The emphasis is mine. The author refers here to a number of visual responses that indeed support my explanation but, I am afraid that her basic premise is just more of the mistaken model

For example, an exaggerated longitudinal chromatic aberration (LCA) is used (Fig. 51) to explain why

“A red splodge often seems to advance from the page and to concentrate the viewers attention towards its centre, a blue splodge seems to recede and lead the eye outwards”.

The figure seems to say that a different focusing condition brings only one of the three RGB wavelengths to focus on the foveal region of the retina. Does this imply that one could see only one of these wavelengths at a time? How extraordinary! But she does for some reason introduce chromatic aberration – that in my explanation is not an ‘aberration’ at all but central to the vision process.

Again to review, I propose that the eye basically functions to continually bring the long (R) wavelength to focus on the fovea. The ‘aberrated’ (I would say simply ‘refracted’) GB wavelengths simultaneously (or almost simultaneously) impinge on the parafoveal region of the retina with the exact center of the RGB band incident at 7-8 degrees of retinal angle. The hues of color that the author wants to see is determined by comparing intensities on either side of this point (I tire of pointing out that this is what Edwin Land deduced)

Again, the unspoken but crucially important implication of the Figure is that the retina (or fovea) is the location or plane where the image formation process of the eye takes place. One imagines, as those espousing the existence of ‘RGB cone mosaics’ seem to believe, that this region is the analogue of the silicon receiver in a digital camera or film in a camera. formng the (intensity only) image. This is not the case at all as I explain.

In Figure 52 the author demonstrates how blue tends to be perceived in peripheral, and red in central, vision. I have written about this (see previous Comments) and it is inherent in my explanation.

The author on p.141 does provide a clear explanation of Land’s initial color vision experiments. For those who have forgotten I will transcribe directly from her text:

“A similar effect has been exploited by Land, who showed that red and white light could be mixed in such a way as to give sensations over a wide range of colours. He made black and white transparencies, photographing each scene twice, once through a red filter and once through a green one. He projected them both on to a screen, the first with a red light, and the second with a white one, and when the images were superimposed, obtained, in addition to red and black, a range of colours from turquoise ,through green and yellow to orange.”

GCH

:

The all-cone central foveal region (that contains ~ 99% of the cone receptors of the retina) occupying the region within one degree of retina angle, is solely responsive to the single, long wavelength (red) limit of visual response. The diffractive nature of retinal response that follows from this explanation must mean that the fovea is located at the Fourier (or focal) plane of the eye. The fovea performs a 2-D Fourier transform of image infrmation acquired by the eye producing an ‘outline sketch’ of the perceived image. I believe that this is the ‘primal sketch’ introduced by Marr. The completion of the visual imaging process requires that a second inverse Fourier transform be performed. This must occur in the visual centers of the brain using the input from the initial wavelength-diffractive FT information from the eye. The idea that the fovea is comprised of ‘classes of cones’ that are responsive to three separate wavelengths is just plain wrong.

The eye is a Fourier transforming device.

GCH

Perhaps the most fundamental aspect of this explanation of the process of vision is that pure, rigorously geometrical or dimensional, factors effect the exact determination of wavelength in the certainly non-rigorously geometrical realm of biology. Geometrical nanostructure is at the heart of (I want to use the term ‘overlays’) a biological process.I refer here to my proposal that it is retinal receptor dimensionality and interreceptor dimensionality that is the specific wavelength determining factor of the retina. The maximum intensity of the ‘green’ (or 550 nm) wavelength, for example, is detected at 7-8 degrees of retinal angle in accordance with light refraction within the eye by the maximum density of cone/rod appositions that geometrically occurs at this point. And further, this wavelength-determining apposition geometrically forms the location where the exact midpoint of the visible band interacts on the retinal surface. In addition to validating Land who deduced that such a fixed point must be present, it provides the basis for an understanding of the historically mysterious color constancy of the vision process.

The principle is that an admixture of two dimensions (as the cones and rods are intermixed on the retina according to Osterberg) results in the specification of an exact, geometrically defined midpoint. Receptor size itself determines wavelength sensitivity and, importantly, the ratio of the difference in size of two dimensions determines where this midpoint occurs.

I have noted that the specific size ratio of cone and rod receptors comprising the retina is 1.8:1 (cone/rod) corresponds to the visible band (700-400 nm). And herein, in addition to defining a symmetrical midband point on the retina, enters the greater idea of geometrical (nanostructural) symmetry. Back in the recesses of this work (search the term ‘epitrochoid’) I noted that this size ratio is unique to defining a spatially symmetrical epitrochoidal shape (the same necessarily symmetrical epitrochoidal shape that Felix Wankel used in his internal combustion engine). Who knows?

But the idea of symmetry…and following the Olive Sacks comment (see previous) that ‘Landian’ correlations seem to take place in the vision centers of the brain…..might the concept of a spatial ‘nanostructural symmetry’ follow onward from the retina to perhaps new insights into the functional mechanisms of the brain? The unexpected occurrence of a ‘geometrical spatial symmetry’ in the biological realm…..??????

I am reminded here of a quote that I have used from Roger Penrose:

“There is, in fact, at least one clear place where action at the single quantum level can have importance in neural activity, and this is the retina (recall that the retina is a part of the brain).”

GCH

A correspondent has pointed me to a beautiful book by Oliver Sacks “An Anthropologist on Mars” that describes the case history of an artist who, after an auto accident, and although retaining vision (with perhaps increased acuity), lost all sense of color. It is a fascinating tale, of which I will write about shortly, but it goes a long way, I believe, toward substantiating Land’s work and my subsequent corroborative explanation of the vision process. The author recognizes Land’s genius and understands well his color vision concept even terming certain brain functions ‘Landian’ (p.39, “…another part of the brain to perform the requisite Landian correlations”). According to the author, this individual’s injury was to one of the locations in the brain that processes color visual information, but it was the character of this information from the eye that led to this understanding. Land was certainly cognizant of the linkage between retina and brain in coining his term “Retinex” (i.e., retina + cortex).Unfortunately, the only theory that he knew of to explain his experimental results was the traditional one (namely, that ‘classes of cones’ exist). He had to resort to devising an external (and complicated) algorithm to simulate the visual response that he measured. I am fascinated that the new explanation presented herein seems to me to lead to new insights regarding the retina-brain linkage in the vision process. We now know for the first time the correct content of the visual input information to the brain. I will write more about this.

It occurs to me to try again to attempt to summarize the light interactive properties of the retina. To begin, I assert that there are no ‘classes of cones’ and that three (and only three) wavelengths are detected by the three receptor appositions (cone-cone, cone-rod, and rod-rod) of the retina. As shown in the following drawing that overlays these wavelength response regions onto the classic Osterberg 1935 diagram of the density of cones and rods on the retinal surface. PLEASE NOTE THAT WAVELENGTH RESPONSE IS ACTUALLY DERIVED FROM, AND BASED ON, THE OSETERBERG DATA.

Please note that these are not the usual spectral plots but rather show (or attempt to show) variations of intensity at the same wavelength across each band. Each band is, in antenna engineering terms, ‘tuned’ to a single wavelength. It is light intensity at this single wavelength (proportional to the apposition-defined density of antennas) that varies across each band. The retina is sensitive to light intensity at each of the three wavelengths exactly as Land deduced must be the case. These three bands correspond to what have become to be termed the ‘primary colors’. The red intensity band detected by foveal cone-cone appositions is extends to a degree or so of retinal angle. This red wavelength corresponds to, and defines, the precise long wavelength end of the visible spectrum. The highest density of cone-rod appositions, and thus ‘green’ response, occurs at 7-8 degrees of retinal angle as shown.This then is a geometrically fixed point /on the retinal surface that precisely defines ‘midband’. This is the fundamentally important point that Land brilliantly deduced from external measurements must be present (his ‘fulcrum’) and which explains the color constancy of vision. By 10 degrees the blue response of rod-rod appositions becomes dominant and this response continues to the peripheral retina. Importantly again, this precisely defined blue wavelength defines the short wavelength end of the visible band. The many hued sensation that we term ‘color’ is determined by comparing light intensity on either side of the fixed midband point at 7-8 degrees. There is always a geometrically-defined wavelength reference point!

I must say again that the above defines a surface that is sensitive to the /diffraction pattern /of the perceived image. The retina must be the Fourier (or focal) plane of the eye and the image formation lies in that domain with processing occurring somewhere between the retina and the brain in a yet to be determined manner. It is conclusively shown, however, that the eye detects the 2-D Fourier transform of the perceived image (as Hagan presciently proposed). Additionally, I have described how the receptor apposition-defined ‘antenna’ detectors of the retina are able to detect both light intensity and phase that is a requirement of the Fourier equation. This realization opens many new areas of inquiry, for example, bringing time into the imaging process (time is manipulated and essentially ‘brought into coincidence’ at this plane).. I have previously noted that the eye may not be the ‘passive’ receiver that has been historically assumed but rather, part of an active (based on antenna behavior) system interrogating external reality (acting as a ‘phase conjugate mirror’?).

Enough for now.

GCH