Understanding Human Vision

• Visual perception is a miracle that involves solving problems to see separate solid objects in surrounding space from tiny, distorted, upside-down images in the eyes.

Visual Processing Overview

• Vision enables not only object recognition but also guides movements through at least two parallel and interacting pathways.
• The binding problem is a central question of cognition that arises because of parallel pathways in the visual system.
• The question is: How are different types of information carried by discrete pathways brought together into a coherent visual image?
• Visual system interprets the scene and separates foreground from background unlike a camera that simply reproduces point-by-point the light intensities.
• Evolutionary perspective reconciles the visual system's efficiency at grasping the bigger picture with its detail inaccuracy
• Vision extracts useful information, identifies objects independent of lighting, and uses previously learned world structure rules, some hardwired, others plastic and based on experience.
• Complex processing occurs at all levels, starting at the retina, which picks out object boundaries rather than creating point-by-point representations of uniform surfaces.

History of Visual Perception Understanding

• Earlier ideas saw perception as an atomistic process assembling elementary sensations like color and shape additively.
• Gestalt psychology, founded by Max Wertheimer, Kurt Koffka, and Wolfgang Köhler, views visual perception as constructive, with the perceptual interpretation of a stimulus depending on its context and other visual field features.
• Gestalt psychologists state the visual system processes sensory information about shape, color, distance, and movement based on inherent computational rules and expectations that derive from built-in neural wiring and experience.
• Max Wertheimer: The behavior of the whole cannot be derived from its individual elements nor from how these fit together; rather, the properties of any of the parts are determined by the intrinsic structural laws of the whole.
• Gestalt psychologists worked out perception laws to group visual scene elements, including similarity, proximity, and good continuation, in the early 20th century.
• The visual system tends to impose a pattern
• The principle of good continuation is an important basis for linking line elements into unified shapes and is seen in contour saliency.
• The Gestalt features picked out also characterize objects in natural scenes and may have created evolutionary pressure to develop neural circuits sensitive to those features.

Visual Scene Analysis

• Separating figure and background is important for object recognition aided by geometric principles and cognitive influences like attention and expectation.
• Internal representation or priming stimulus can aid association of visual elements into a unified percept that reflects time scales and mechanisms of neural encoding.
• The brain analyzes a visual scene at low, intermediate and high levels.
• Local contrast, orientation, color, and movement are discriminated at the lowest level.
• The intermediate level looks at the layout of scenes and surface properties to parse the visual image into surfaces and global contours.
• Object recognition occurs at the highest level.
• After parsing and recognition, objects are matched with associated memories and meanings.
• Vision is important in guiding body movement and hand movement.
• Various features in vision form a unified percept, coming together in multiple brain areas fed by parallel but interacting neural pathways.

Neural Circuitry and Organization

• Understanding neural circuitry and visual pathway organization is required to grasp the neurobiology of vision.
• The principles are relevant for multiple brain areas and sensory information processing types.

Geniculostriate Pathway

• The brain's analysis of visual scenes starts in the two retinas, which use parallel processing to transform visual input through pixel-like bits of visual input falling on photoreceptors.
• Analysis by retinal circuits extracts about 20 local features, such as contrasts of dark versus light, red versus green, and blue versus yellow, which are computed by specialized neural circuits.
• Each visual field point is processed in multiple parallel channels simultaneously.
• Parallel streams are sent through retinal ganglion cell axons, the project neurons of the retina, that form the optic nerves.
• Optic nerve extends to a midline crossing point or the optic chiasm, beyond which fibers from each temporal hemiretina proceed ipsilaterally, and nasal hemiretina fibers cross contralaterally.
• Hemifield information is processed in the contralateral hemisphere's visual cortex.
• The pathway layout forms the foundation for useful diagnostic information relating lesion locations to visual deficits.
• Lesions create visual deficits with types dependent on location
• These shapes can vary if: entirely monocular, affect noncorresponding or corresponding parts of the visual field in both eyes, restricted to either the upper or lower visual field, extend into both fields, etc.
• Lesions can: cause total vision loss in one eye; losses in the temporal half of each visual hemifield; loss in the opposite half of the visual hemifield; loss of vision in the upper quadrant of the contralateral visual hemifield in both eyes; and cortical lesions can lead to deficits in portions of the contralateral visual hemifield.

Visual Information Transmission

• Axons from nasal and temporal hemiretinas with input from one hemifield join in the optic tract, to the thalamus' lateral geniculate nucleus (LGN) extending beyond the optic chiasm.
• Primate LGN has six primary layers: Four parvocellular and two mangocellular, paired with intercalated koniocellular layers where the name reflects small cell bodies.
• Channels established in the retinas remain segregated through the LGN. Layers process different types of information
• Parvocellular layers get input from midget retinal ganglion cells (~70%), carrying red-green opponent information.
• Magnocellular layers get achromatic contrast information (10%) from parasol ganglion cells.
• Koniocellular layers get input from small/large bistratified ganglion cells, carrying blue-yellow information, and other smaller classes of retinal ganglion cells
• Each geniculate layer either get ipsilateral or contralateral eye input, however aligned from the matching regions in the contralateral hemifield (concordant maps stacked atop one another).
• Thalamic neurons relay this information to the primary visual cortex
• The LGN modulates incoming retinal information via attention and inhibitatory/arousal connections and feedback from the visual cortex.
• The primary visual pathway or geniculostriate pathway passes through the LGN on the way to the primary visual cortex (V1) (striate cortex).
• There is a second pathway from the retina to the midbrain's pretectal area for pupillary reflexes controlling light entering the eyes.
• There is a third pathway from the retina to the superior colliculus to control eye movements to the pontine formation in the brain stem to the extraocular motor nuclei.
• Each LGN projects to the primary visual cortex through the optic radiation.
• Afferent fibers generate a complete neural map of the contralateral visual field in the primary visual cortex. Beyond the striate cortex are extrastriate areas, organizing visual field neural maps.
• Preservation of spatial layout of retinal inputs is retinotopy, describing neural maps of the visual field as retinotopic with retinotopic frame of reference.

Processing in the Visual Cortex

• The primary visual cortex is the first level of cortical processing for visual information that proceeds through two major pathways: dorsal and ventral
• The ventral pathway carries "what" information to the temporal lobe.
• The dorsal pathway carries "where" information (guiding movement) to the parietal lobe.
• The corpus callosum connects the two hemispheres to transmit information and enables unifying perception of objects spanning the vertical median.

Discrete Areas of the Cerebral Cortex

• The cerebral cortex was differentiated into discrete regions by Korbinian Brodmann and others.
• Anatomical criteria include neuron size, shape, packing density in the cortical layers, and myelin thickness and density.
• Functionally distinct cortical area's correspond loosely to Brodmann's classification

Functional Differentiation

• Macaque monkey visual cortex consists of over 30 areas.
• Functional magnetic resonance imaging establishes homologies between visual areas of macaque and human brains.
• Pathway tracing has revealed functional stream organization.
• Visual cortex areas can be distinguished by the functional properties of neurons.
• The eye sends information to thalamic nuclei (lateral geniculate nucleus and pulvinar) and then to cortical areas, progressing from primary visual cortex to parietal and temporal lobe areas supplemented by pulvinar which serves as a relay.
• Ventral pathway = object recognition
• Dorsal pathway = visually guided movement
• Midbrain pretectum relays light signals to preganglionic parasympathetic neurons
• The parasympathetic outflow innervates the pupillary sphincter and the muscles controlling the lens.
• Information moves from retina to superior colliculus to cortical areas and back to the superior colliculus for eye movement control.
• The superior colliculus projects to the pons. The pons send controls signals to oculomotor nuclei including the abducens, which controls the lateral movement of the eyes.
• Two hierarchical pathways are organized in visual areas: ventral (object recognition) and dorsal (movement guidance).
• Ventral pathway extends from the primary cortex to the temporal lobe.
• Dorsal pathway connects the primary visual cortex with the parietal lobe and then the frontal lobes.
• Connections between pathways enable information sharing for kinematic cues.
• All connections between areas are reciprocal to provide cognitive information to earlier levels of visual processing.
• The pulvinar in the thalamus = relay between cortical areas.

Dorsal Pathways in Detail

• The dorsal pathway uses information to direct movement via visuomotor integration.
• The lateral intraparietal area represents eye and reach targets in space where lesions to the parietal area mean patients fail to pay attention to one side of their body.

Ventral Pathways in Detail

• The ventral pathway extends into the temporal lobe, storing object shape and identity information.
• Inferior temporal cortex is also responsible for the ability to recognise faces.

Structure of Pathways

• Pathways include hierarchical series of areas delineated by several criteria.
• First, inputs form a map of the visual hemifield at many relays.
• Second, an area in the dorsal pathway, the middle temporal area contains neurons uniquely suited to direction of movement.
• The classical view: visual cortical areas are a hierarchical system, where areas represent primitives of orientation, direction of movement, depth, and color; higher areas represent whole objects or intermediate views.
• More nuanced views state: Primary visual cortex plays a role in intermediate-level vision and higher level neurons process object components, a view inclusive of reverse information flow or feedback. Reverse information flow includes high order influences and the order is cognitive.
• Top-down influences play a role in segmentation.

Receptive Fields

• 1906, Charles Sherrington: Receptive field is a whole collection of points of skin surface from which the scratch-reflex can be elicited
• H. Keffer Hartline applied the concept to the retina of the horseshoe crab: Receptive field is the retina region which must be illuminated in order to obtain a response in any given fiber.
• A neuron's receptive field is "window" on visual field
• Kuffler demonstrated "not only areas from which responses can actually be set up by retinal illumination...but also all areas which show a functional connection, by an inhibitory or excitatory effect on a ganglion cell" Retinal ganglion cells have functional subareas that are center-surround organized, one of two categories, and they are on-center and off-center.
• On-center cells fire if a light is turned on within a circular central region.
• Off-center cells fire if a light is turned off.
• Later work revealed cells in the LGN have similar receptive fields.
• The surrounding annular region has the opposite sign of sensitivity to light. Center and surround areas are mutually inhibitory (on-center, off-surround) and borders and contours lead these to encode contrast information. According to a field's eccentricity (position relative to the fovea) as well as the position of neurons along the visual pathway, a receptive field's size on the retina varies. Small receptive fields are found closer to the fovea, and larger ones at later levels. The visual field covers nearly 180 degree. Receptive fields near fovea are the smallest. The amount of cortex changes with eccentricity. The smallest ones are in the centre of the field which gives it the highest spatial resolution Properties change from relay to relay along visual pathway by the properties show each relay assay and information becomes progressively analysed. Change and field structure indicate the LGN and cerebral cortex mechanism in brain's analysis. Hubel and Wiesel, 1958: Primary visual cortex shows selectivity for orientation of contours in visual field. Primary visual cortex shows a property of signal processing as it is a generated property within itself.

Organization of Visual Cortex

• The dominant feature includes: the systemic representation of the surface across the cortex.
• Similar properties are close together in columns which reflect that area in vision

Ocular Dominance

• Ocular dominance integrates thalamocortical inputs which come to one point from another area.
• Columns reflect a thalamocortical arriving segregation.

Orientation Preferences

• The organization's cells group into columns. Across the cortical surface: regular clockwise and counterclockwise cycling, with 180 degree cycle repeating in every 750 uM. A pair is called a "hypercolumn."
• The orientation and dominance are orthogonal to each other. Each consists for combinations and orientation for each.
• Each type was made from electrode penetration and in the cortex by cells in the lateral geniculate nucleus of thalamus.
• Researchers visual images of surface using living animals. For by with and in optical imaging the image obtained means that eye can be easily detected and imaged.
• Experiment can visualize the distribution using two ocular by subtracting the image where stimulation is one eyes. Orientation columns formed from stripes to pinwheels.

Interneurons

• Are neurons in the specialization that are contained in ocular-dominance columns but have selectivities.
• They are distributed with patterns and help with the identification with superlayers can and may

Visual Pathways and Strips

• In area V2 the stripes thin and are with labeling. The stripes contain that are directed by the and the cells that are the illusion one. The hold color in cells and are organized respectively.
• Projections are responsible for inputs which are in the hemisphere
• Inputs from the project and terminate, they and each have a termination. However these maintain the anatomical integrity of the different layers in the eye of cells.
• Neurons that the layers form an image, the and this will. This will help project.

Connections

• For every to be at a value it must be as possible
• To function in processing it must go through these layers in the brain
• The are the source for two of in the and
• This shows that the and

Benefits

• minimises that share a and to from. this connectivity saves which which the and there is the brain, that the
• area

Ciruits

• that is by local

Feedbacks

• which is to of it be the is
• These the the the the the the the the the the the the the in the in the the the the the the, such VIP

Horizontal Communication

• horizontal with the or the