Cognitive Neuroscience Lecture 5: Visual Systems/Pathways

L5: Visual systems/pathways

Visual system

-         Optics/eyeballs

-         Retina

-         Pathways

-         Cortex

Overview of the visual pathway

-         There is energy in the world, in the form of light

o Light is an electromagnetic wave

o Light travels in a straight line

o We see objects because they reflect light

§  But light reflects in all directions from each point of objects

§  And it arrives from all points

§  So the pattern of light and dark is destroyed

o Solution: pinhole camera: select one direction per point

§  The problem with this is that when the pinhole is too small, not enough light

§  When the pinhole is too large, it is too blurry

o Speed of light is slower in some materials than others

o Refraction:

§  Because it travels at a diff speed in diff materials, light bends when crossing the boundary between materials

§  The refraction is proportional to the ratio of speeds

o Prisms: divergent > <, convergent <>

o Lenses

§  Light from objects scatters in many directions

§  Lenses gather light from a single point in world and focuses it on a single point on the screen (at a certain distance beyond the lens, say on a screen, or your retina)

o Human eye

§  Ciliary body: muscles that control the shape of the lens

§  Cornea: most of the focusing power

§  Lens: mainly for fine tuning

§  Retina: no blood vessels on cornea or lens, but plenty of blood vessels right on retina

·       We don’t see them because visual system interested in change

·       If you were to shine a bright light, you’d see the shadow of the retina

§  Fovea: blood vessels and nerves pushed away, densely packed receptors, cones only

o Accommodation (focusing)

§  Emmetropia: normal

§  Myopia: nearsightedness

§  Hyperopia: farsightedness

-         Light enters the eye, an image is formed on the retina

o The retina has receptors that react to light and convert into neural signal to use in the brain

o Photoreceptors: rods and cones

o Cones = fovea

§  Daylight, packed in fovea, good for high-res vision and color

§  Short wavelength, medium wavelength, and long wavelength

§  S:M:L = 1:5:10

§  Dearth of blue receptors in fovea (middle of image)

§ Retinal color blindness: lose L cones (protanopia) or M cones (deuteranopia)

o Rods != fovea

§  Night time, not in fovea, everywhere but blind spot, low-res vision, no color, good for low-light conditions

o NO RECEPTORS AT ALL in blind spot

-         Retina has receptors that react to light and transmit signals to the brain

o Convergence of cones and rods

§  In fovea: low convergence (one to one mapping; high spatial res, high acuity)

§  In periphery: high convergence (many to one mapping; low spatial res, low acuity)

·       1 RGC à many bipolar à many more rods

o Optic nerve

§  No photoreceptors: blind spot

§  All axons from RGC come together

§  Form optic nerve bundle and exit eye

-         Signals get transmitted down the visual pathway from eye to cortex

o Retina à optic tract à thalamus LGN à V1/striate cortex

o More depth: retina àoptic nerve àoptic chiasmà optic tractàLGNàV1

o Your visual system is “cross-wired”

§  Signals from nasal half of each retina cross at the optic chiasm

§  Left VF àright v1

§  Right VFàleft v1

§  NOT ‘left eye to right cortex, right eye to left cortex’

§  “what happens if”….do it the way you did, but then switch left to left after

o Subcortical waystations

§  LGN receives 90% of retinal input and transmits info to cortex

§  Superior colliculus receives remaining retinal input, important for eye movements

o Parallel pathways

§  Two types of RGC: magnocellular (rods), parvocellular (cones)

§  Temporal frequency higher for magnocellular

-         The brain tries to determine what is where in the world

o Receptive fields: the area over which stimulation affects the response of a cell

§  \mapping a receptive field: allowed recording from otherwise normal animals

§  Use micro-electrode

§  Screen, amplifier, loudspeaker, oscilloscope, micro-electrode, anesthetize animal

o What makes a V1 neuron fire? '''Hubel & Wiesel'''

§  Take a cell in v1

§  Stimulate photoreceptors on retina

§  Determine which pattern of light the neuron fires for

§  Talk about a needle in a haystack!

§  They found out it was straight lines/bars of light

§  Simple v1 cells: bar of light

o Receptive fields in RGC: center-surround (‘Mexican Hat’)

§  Not just single spot, small region with concentric ON/OFF responses

§  On/off = whether middle region “likes” light or dark

§  Excitation and inhibition cancel out if diffuse light, and NEITHER turns on if 0 light

o What does the retina care about?

§  Sensitive to changes in local contrast

§  Only when it doesn’t cancel (light only hits concentric circle or middle )

§  Retina signals edges, not uniform luminance

§  Sensitive to changes in local contrast : exaggerates it (actual luminance distribution is more sudden than  perceived luminance distribution)

o Not only stronger responses

§  ON: neuron fires more when light shown

§  OFF: neuron stops firing when light shown

o V1: calcarine fissure

§  Adjacent regions of space are coded for by adjacent neurons

§  Retinotopic mapping in humans with fMRI  (polar and eccentricity)

o Cortical retinotopic map: cortical magnification (more for fovea/center)

§  Convergence of rods and cones: why is there cortical magnification?

§  Density of RGC encoding visual info decreases as you go from center to periphery

§  Ganglion cells representing central vision provide finer grained info (high res) from central vision

o Receptive fields of simple v1 cells: edge detectors

§  Receptive fields become larger and more complex in concentric cells

§  Orientation selectivity and tuning (matters \ | - /)

§  Detect edges

§  Sensitive to specific orientation in specific position

§  Orientation “mapped” along cortex: preferred orientation changes as you move

o Orientation columns

§  V1 neurons tuned to same orientation are stacked in vertical columns

§  Adjacent columns hav similar orientations, resulting in “pinwheels”

o Complex v1 cells

§  Receptive fields larger and more complex

§  Oriented bar in general location (not as specific as in simple v1 cells)

o Hierarchy of processing up to v1

§  ‘building’ progressively complex representations as you move up the visual hierarchy

§  Photoreceptors à LGN à V1

§  Size of receptive field increases as you go from V1 to v4

o ‘Preferred’ stimuli

§  We say a neuron codes for whatever makes it respond most

§  Simple v1 cells: Line orientation, in one spatial location      /

§  RGC: a small point in space .

§  Complex v1 cells: line orientation, in a general location /////////

-         Further processing in dorsal/ventral pathways

o Dorsal: “where”

o Ventral: “what”

-         Pigmented cell à Rods/cones à horizontal cell (integrates and adjusts for light/dark) àbipolar cellà amacrine cell (intercept RGC/bipolar; directional motion, light adaption) àRGC