The Design of the Neuronal Correlate of Consciousness (NCC) for Visual Perception
Solving the inverse optics problem of “seeing”A Theory of Visual Consciousness

Introduction

The present day view is that visual perception that generates recognition of the visual experience is transformational. The transformations are generalizations of the Fourier analysis of local scale-based spatial and temporal characteristics of the retinal image. Vision is thought to be a creative process mediated by 3 or more parallel pathways that process information for motion, depth, form, and color. The unity is achieved not by one hierarchical neural system concerned with vision but by at least 3 (possibly more) parallel pathways in the brain. This view, which has become a major tenet of today’s neurobiological study of vision, had its origin in visual psychophysics (Cambell and Robeson, 1968), and a large number of experimental measurements the most prominent ones performed by Hubel and Wiesel (1959, 1962, 1968, 1998). It is now presented in almost all major textbooks in cognitive neuroscience (Kandel et al 1991; Gazzaniga et al 2002; Bear et al, 2001; Haines, 2002). Up to date historical reviews of the “standard model” have also been presented by Robert F. Hess (2004), M. Reisenhuber and Tomaso Poggio (2004).
In an article published in 1992 Francis Crick and Christof Koch selected the “study of mammalian vision” as the approach to the problem of consciousness. [Crick and Koch, 1992] The full title of the article was “The problem of Consciousness: It can now be approached by scientific investigation of the visual system. The solution will require a close collaboration among psychologists, neuroscientists and theorists”. Unfortunately, most of present day experimental research is not aimed at the problem of consciousness (it is aimed at processing associated with recognition and identification), and at the turn of the century Crick and Koch (1995, 2000, 2003, 2004), and a large number of neuroscientists and philosophers (Metzinger, Ed. 2000), recognizing that processing per se would not lead to the subjective experience of consciousness, were publishing extensively seeking “Neural Correlates of Consciousness” (Metzinger, 2000).
This paper presents a proposed NCC for visual perception. The design of the NCC for visual perception is an expansion of the design of the NCC for the tactile modalities presented by A. Rosen and David B. Rosen (Rosen et al, 2003a). The sensation of “seeing” or the experience of “vision” is mediated by the modalities of the receptors in the retinas of the eyes. The law of specific nerve energy, applied to the visual Consciousness Mechanism (CM), explains the subjective sensation of pinpoints of light, and visual colors (Guyton. Chapter 46 p.495). The law of collective receptor energy, applied to group-patterns of receptors that are activated simultaneously, explains the subjective experience of “seeing” shape or form generated by the pattern. Rather than processing information for motion, depth, form and color, the NCC for visual perception generates an experience of “seeing” which is based on the local feature representation (images) in the retinas of both eyes. No additional processing is required to combine the independently processed parallel pathways for form, color, depth, and motion.

Generalizing the Law of specific Nerve Energy: application to collective modalities

The reflexive 3D cyclopean eye, and the NCC circuit associated with it is a Darwinian adaptation that has been developed by natural selection to obtain a high degree of correspondence between the subjective experience of the object and the 3 dimensional shape-depth color and motion of the object. The Darwinian adaptation must generate a high fidelity representation of the external environment under conditions of variable day-night illumination, variable speed motion of a predator or target, and variable speed-motion of the mounting of the eye sensors.
The design of the visual NCC is based on a generalization of the individually activated tactile modalities into collective visual modalities wherein groups of receptors generally arranged in patterns, are activated simultaneously. The sentient characteristic associated with a tactile modality is similar to the sentient characteristic associated with the activation of a single visual receptor in the retina (Guyton,1991). This visual modality may give rise to the subjective experience of “seeing” a pinpoint of light (by the law of specific nerve energy) (Brown, 2002)3 . But what if a pattern of pinpoints of light were to activate the same pattern of receptors in the retina of the eye? It is postulated herein that the subjective experience generated by the pattern is perceived as an outline of a shape or form with no additional (transformational) processing required to perceive (not “identify” or “recognize”) that shape or form. The law of specific nerve energy may therefore be generalized to patterns, arrays, or groups of nerves that are activated simultaneously. That is, the visual sensation of a form or shape is mediated by an array of receptors wherein simultaneous firing of a pattern of receptors within the array generates the subjective experience of the form or shape. The law of specific nerve energy is thus generalized to the law of “collective” receptor energy. The “collective” receptors in this case, called the form-shape collective, is an array of receptors in the retina of the eye that maintain a retinotopic organization, and that sub-serves the form-shape modality. The conscious experience of resolution and visibility of the perceived form-shape depends on the number of receptors per unit area in the array, the area of the form-shape, the signal strength generated by each individual receptor relative to the signals generated by inactive receptors (contrast sensitivity).
The law of collective receptor energy may also be applied to the superposition of colors that generate the sensation of the spectral hues resulting from the addition or subtraction of individual color receptors. Without superposition, the law of specific nerve energy would apply to individual L, M or S-cone receptors in the retina and would sub-serve red, green or blue color modalities. That is, the subjective experience generated by each modality is the experience of “seeing” a red green or blue pinpoint of light. If, on the other hand a collective is formed that combines color opponent cells, as for example the retinal collectives of a red cone receptive field center, aLc, with an opposing green-cone receptive field surround, bMs, the collective may then generate an output which is the sum or difference of 2 colors (aLc ± bMs). The circuit (collective) in this case is a red, green color-hue nerve-circuit that may be identified as the red, green color hue modality. The subjective experience generated by this modality is the experience of seeing one of thousands of different color hues depending on the magnitude of the sum or difference aLc ± bMs (The region covered by the color-hue depends on the size of the receptive field.). The coding of the different color-hues is therefore a collective (color addition or subtraction) modality of the L, M or S-cone receptors in the retina (De Valois 2004, p. 1002; Knobblouch and Shevell 2004, p. 892).

Visual modalities of the human retina

The visual modalities of the retinal layers are a Darwinian adaptation represented by neuronal coding of collections of photoreceptors. These retinal collections have been developed by natural selection to obtain a high degree of correspondence between the subjective experience of an object and the 3 dimensional shape-color and motion of the object in the external environment (while the organism performs survival tasks) (Rodiek 1998, p.8). It is generally assumed that this Darwinian adaptation gave rise to 2 retinal collectives, or a dual retina; a photopic retina for daylight viewing, consisting predominantly of cones, and a scotopic retina for night-time viewing consisting predominantly of rods (Rodiek, 1998). A review of the neuroscience literature, however, reveals that the human retinal layers may be viewed as giving rise to many more collective modalities. Five collective sub-modalities are identified in table 1.2. Table 1.2 shows 3 modalities that may tentatively be assigned to the photopic retina and 2 to the scotopic retina. References to the scientific literature that identify these 5 collectives are presented in the table. The identification of the various collective modalities generated by the pre-processing taking place in the retinal layers is at the heart of the issue of neuronal pathways from the retina to the striate cortex and thence to the self circuit.

Pre-processing in the retinal layers: Small-scale collective sub-modalities.

Before the signals reach the output ganglion layer, a significant amount of pre-processing is performed in the retinal layers that effects the modality (visual sensation) generated by light striking the retina. Light passing through the lens first passes through the (transparent) ganglion layer, thence through (transparent) bipolar cells, thence to the rod-cone photoreceptors that are activated by the light photons. The synoptically activated signals originating in the photoreceptors then travel in a direction opposite to the light through the bipolar cells and thence to the ganglion cells. The bipolar cells serve as intermediate processors by grouping rods, L-cones, M-cones, and S-cones by means of horizontal cells, and bipolar cells by means of amacrine cells. These groupings determine the receptive fields associated with each ganglion cell located in the output layer of the retinal laminar organization. These groupings give rise to small-scale collective sub-modalities.

Pre-processing in the retinal layers: The large-scale retinotopic collective.

The large scale collective is a retinotopic array made up of a multiplicity of small-scale sub-modalities. For example, the form-shape collective is a large scale array collective made up of small-scale individual modalities of rod and cone receptors. In order to experience the sensation of color-hue, form, and shape, the receptive fields of all the color opponent small-scale sub-modalities must be combined into a retinotopic organization. The conscious experience of resolution, color-hue, and visibility of the form-shape depends on the size of the receptive fields of the color-hue sub-modalities, the number of receptive fields per unit area, the area of the colored shape-form, and the signal strength of each receptive field relative to the signals generated by inactive receptive fields. (Note that the center surround organization may enhance sensitivities to boundaries between low contrast shaded regions.)

Post-processing.

The collective modalities of the various receptive fields of the retinal laminar organizations are collected by the ganglion layer and then project via the optic nerve to a number of different nuclei in the brain. The collective modality specificity is maintained in the central connections of sensory axons, so that retinal stimulus modalities are represented by various retinotopic organizations generated in the retinal layers, afferent groups of axons (optic nerve-ganglion cells) associated with each retinotopic organization type, and central pathways that maintain the retinotopic organizations that are activated by each modality. One of the most prominent projections is to the dorsal part of the lateral geniculate nucleus (LGNd), a multilayered structure in the thalamus (Goodale 2004, p.873; Kandel et al 1991; Gazzaniga et al 2002; Bear et al, 2001; Haines, 2002). The total visual CM is dependent on a superposition of the collective modalities of the visual receptors and the biological solution of the inverse optics problem of obtaining a correspondence between the image and objects giving rise to that image.
The human retina forms at least 5 large-scale retinotopic collectives, one large-scale retinotopic collective for each of the small-scale sub-modalities listed in table 1. And although the 5 sub-modalities are interwoven into a single retinal ganglion surface layer containing a very high density of ganglion cells, the 5 retinotopic collectives may be overlaid one on top of the other so that the 5 retinotopic organizations correspond one to one, to the single retinal ganglion layer but with a smaller density of ganglion cells per layer.
Note that the CM applied to the overlay of the 5-retinotopic collectives shown in table 1, yields a possible solution to the binding problem. That is, the subjective experience of the image formed by the 5 overlaid retinotopic collectives is that of an object in the external environment that gave rise to the overlaid images on the 5 overlaid retinotopic collectives. It is assumed that the CM combines the modalities that are sensitive to boundary, shape, form, color and motion into an experience of a high resolution image with shapes having sharp boundaries, color-hues within each boundary, and internal motion associated with each shape, just as we “see” it.
The Darwinian visual CM has been developed by natural selection to obtain a high degree of correspondence between the subjective experience of the configuration-color-hue-shape and motion and the images that gave rise to that subjective experience. No additional connectivity or brain processing is required in order to “bind” parts of the connectivity into a whole visual experience.

CONCLUSION

In the standard model it is theorized that a visual scene is simultaneously processed by the Hubel and Wiesel cortical modules, with each module “looking” at a portion of the scene (Bear et al, 2001 p. 337). Experimental data supports the hypothesis that the visual image is created by several relatively independent parallel processing channels. Each one appears to be specialized for the analysis of a different facet of the visual scene. In the inter-blobs of layers III and II (P1-interblob channel), it is theorized that analysis of stimulus orientation may be required to discriminate and identify objects on the basis of their shape. The blob regions (Konio and P1-blob channels) are specialized for the analysis of color, whereas the receptive fields of the M-channel are specialized for the analysis of motion. The assumption that is NOT supported by experimental data is that analysis is performed in those regions in order to “see”
The visual NCC model is supported by the same experimental data that supports the standard model.
A careful study of the experimental data, biological structure and the associated neuronal pathways supporting the standard model, reveals that this data also supports the reverse engineered NCC model. In the visual NCC model, it is assumed that the inter-layer communication within the striate cortex facilitates the formation of the overlaid retinotopic collectives that define the reverse engineered Cyclopean eye. Spiny stellate cells facilitate inter-layer communication within the striate cortex. Spiny stellate cells in Layer II and III send their axons, and project the retinotopic organizations to layers I, V, and VI.
In the visual NCC model it is theorized that the Cyclopean eye is made up of the same cortical modules observed by Hubel and Wiesel. The cortical modules are arranged in an overlay of retinotopic collectives that correspond to the retinal ganglion collective. The various collectives are shown in figure 8 in striate cortical layers IVCb and layers II and III that show the projection of 2 pairs of layers from the Parvo-cellular pathway. An image that falls on the retinal layers is assumed to be made up of regions that are projected to the cortical modules in the striate cortex. Each region corresponds to the area of a cortical module (2mm x 2mm), which images a point (or region) in space according to Hubel and Wiesel. A single 2mm x 2mm cortical module is illustrated in figure 8.
The image that falls on the retina is reproduced in terms of shape, color and boundary by an array of cortical modules. The overlaid layers in each cortical module (in a direction perpendicular to the array surface) includes a variety of receptive fields that are representative of the 3 parallel processing channels and representatives of each of the processing channels from the right and left eyes. According to the visual–NCC model each parallel processing channel is associated with a collective retinotopic modality. For example, orientation preference is a collective modality made up of an orientation selective retinal sub-modality (a ON-ON-ON alignment along one axis with a OFF surround on each). The variation or orientation preference across a cortical module in a direction tangential to the layer surface as measured by Hubel and Wiesel is illustrated in figure 8. In a direction perpendicular to the array each orientation shown in figure 8 forms an orientation column. The image generated by a retinotopic collective modality made up only of orientation sensitive receptive field cells (that generate orientation columns in layer III of the striate cortex) is illustrated by M. F.Bear et al (Bear et al, 2001 Figure 10.32 p.346). The generated image falls on an area slightly larger than a 3x3 mm array of cortical modules (6x6 mm). It shows the boundary only of the shape of the image. The boundary is poorly resolved since the orientation selective receptive fields are relatively large and are all are geometrically arranged with receptive center OFF and an opposing surround. This collective modality shows great sensitivity to low contrast-shaded boundaries. An image may also be generated by a collective made up of cells of the shape form modality in the Parvocellular collective channel, shown in table 1. These cells have small receptive fields, are represented by numerous receptive fields that cover the surface of each cortical module and generate high spatial resolution image in the region of the cortical module that is representative of the shape form modality. The inter-blob region of each cortical module responds to form, shape, boundary orientation and boundary contrast. The blob region of each cortical module (each cortical module is made up of 16 blobs) responds to color. The total image made up of an overlay of retinotopic collectives, is projected onto an area made up of large number of adjacent cortical modules.
In the visual NCC model, the various overlaid retinotopic collectives generally emphasize different aspects of the image. A modality is associated with each emphasized aspect of a collective. When the law of collective receptor energy is applied to a superposition of collectives (where each collective emphasizes a different aspect of the image), the sensation that is generated may be of a highly resolved 3 dimensional color-hue image with very accentuated sharp boundaries.

Extra striate communication in the striate cortex from the Cyclopean eye to the NCC- “self” circuit.

Extra striate communication from the striate cortex (Cyclopean eye) to other cortical areas is facilitated by pyramidal cells that synapse with the retinotopic organization in layers I, II, III, IV, V, and VI. Layer VI pyramidal cells give rise to the cortifugal (feedback) massive projection to the LGN (the size of this projection is factor of 4 greater than the retinal ganglion cell projection to the LGN). In a local feature representation theory, one might reasonably expect that this cortifugal feedback loop would be used to fill in gaps in each of the six layered retinotopic organizations of lower density collective modalities (while maintaining the retinotopic organization of the higher density ganglion layer). Pyramidal cells in layers I through V send their projections to “another 2 dozen distinct areas of the cortex, each of which contains a representation of the visual world” (Bear et al, 2001 p. 337).
The extra striate communication, pyramidal cells from layers I to V, to the 2 dozen distinct areas of the cortex, represent reverse engineered communication between the receiving neurons of the Cyclopean eye and the NCC-“self” circuit. No additional processing is required in order to experience (“see”) the retinal image, as long as the pathways of the collective modalities adhere to the reverse engineering connectivity constraints of the Cyclopean eye and the NCC circuit as shown in figure 7.