Senses 3: Visual perception Flashcards
how do we see?
object/person seen against background
perceived 3-D large upright colourful image
Size of projections onto the retina:
complex structure of vertebrate retina
functional classes of cells in retina
first stages of visual processing
functional classes of cells in the retina
4 classes of photoreceptors (3 cone types and rods)
50-70 classes of horizontal, bipolar and amacrine cells
20-30 classes of ganglion cells
first stages of visual processing
Edge detection in visual scenes
Edge enhancement in patterns
Filtering of spatial, wavelength, movement and directional information
fundamental visual tasks of imp for many behavs
Seeing and recognising objects, mates, predators or prey
major task of visual system
segregate objects and backgrounds, automatically and quickly
lateral inhibition
makes edges stand out
H.K. Hartline model of lateral inhibition in the retina (1956)
Neighboring neurons in the same layer of the retina inhibit each other mutually
The photoreceptors stimulated by the right edge of each grey bar are inhibited by the neighbouring photoreceptors stimulated by the lighter bar next door.
Thus, photoreceptors on the right edge report receiving less light than they actually do (i.e., that edge looks darker to us).
neighbouring neurons in same layer of retina inhibit each other mutually
If the light falling on the group of retinal neurons is uniform, then their reciprocal inhibitions cancel each other out without further effects.
When an edge (dark and light illumination) is created, the cells on both sides of the edge will influence each other strongly.
This changes their signals such that a much stronger contrast is coded than physically exists.
More distant cells are not affected.
As a result the perception of the edge is enhanced.
simultaneous contrast effect
Koffka-ring illusion
A grey ring on a dark and bright background
A white bar separate the two halfs of the ring. Do the two halves of the ring appear to be identical in brightness?
The right half has been moved in a vertical direction. Do the halves look identical?
identifying spatial r’ships and properties of objects
Edges and shadows provide context information about the spatial structure of objects (in the picture a three-dimensional object with inclined surfaces) or spatial relationships between objects (identical objects laying sidewise behind the central one)
Without context cues, we perceive the physical reflectance of the surfaces which carries little information
segregated rod- and cone-connected pathways in the retina
horizontal connections
vertical connections
Cones (or rods) converging on a bipolar cell form its receptive field.
Similarly, the receptive field of a ganglion cells is formed by all converging bipolar cells.
Receptive fields are large in the periphery (low acuity) and small in the fovea (high acuity)
horizontal connections
horizontal cells
amacrine cells
vertical connections
Fovea: 1 cone to 1 bipolar
Periphery: Many cones to 1 bipolar, many bipolars to 1 ganglion cell. Same for rods, but connect to rod bipolar cells other classes of ganglion cells.
centre-surround (CS) receptive fields
Bipolar and ganglion cells have centre-surround receptive fields (sometimes also classical receptive fields)
Horizontal cells influence bipolar cells either directly or by feeding back information to the cones (probably both)
The bipolar cell integrates EPSPs and IPSPs (spatial and temporal summation)
Signals from several bipolar cells define the activity in a ganglion cell (with additional modulation by amacrine cells)
on/off-bipolar cells
In the vertebrate retina, photoreceptors release neurotransmitter (glutamate) when not stimulated.
Exposure to light hyperpolarises the photoreceptor and decreases the release of glutamate.
The bipolar cells invert the receptor signal to the standard: depolarisation when light intensity increases and hyperpolarisation when light intensity decreases
the sombrero-shaped response of a cell with a CS receptive field
When a light spot is moved slowly through the receptive field of a bipolar cell, the summation of EPSPs and IPSPs varies
When plotted, the response curve looks like a Mexican sombrero seen from the side
retinal ganglion cells also have a sombrero-shaped response curve
Ganglion cells in the retina have long axons that project (through the blind spot of the eye) as optic nerve to the LGN in the midbrain. They generate action potentials (spikes) when transmitting a signal.
Most ganglion cells have centre-surround receptive fields and will show a sombrero-shaped response similar to the bipolar cells
Similar to the bipolar cells, they are specialised to respond to contrast changes in the receptive field which correspond to edges or small bright and dark spots in the visual scene
ganglion cells respond to ratios of light/dark
At rest, a ganglion cell fires action potentials (spikes) at a spontaneous rate
Whilst the ON-centre bipolar cell depolarises, the ON-centre ganglion cell responds by increasing its spike rate
Whilst the OFF-centre bipolar cell hyperpolarises, the OFF-centre ganglion cell responds by decreasing its spike rate
ganglion cells do not respond to uniform illumination
When the light spot covers the whole ON-centre, the ganglion cell responds with its highest spike rate
When a ring of light covers all of the surround but not the ON-centre, the ganglion cell responds with the lowest spike rate or even no spikes at all
When the whole receptive field is equally stimulated, the ganglion cell is at rest and fires with a spontaneous frequency
inverted responses in off-centre ganglion cells
Also respond to light/dark ratios but not to uniform illumination
Responses are inverted
If a spot illuminates the centre of the OFF-centre ganglion cell, the spike rate is reduced
If a spot illuminates the ON-surround of the OFF-centre ganglion cell, the spike is increased
If the whole receptive field is illuminated, the ganglion cell is at rest and fires at its spontaneous rate
P- and M- ganglion cells project to diff layers in the LGN
P-ganglion cells (project to Parvocellular layer in LGN) – small RFs, slower conduction speed, high acuity, poor response to transient stimuli, colour sensitive.
M-ganglion cells (project to Magnocellular layer in LGN) – large RFs, higher conduction speed, sensitive to motion, low acuity, no colour discrimination.
P- and M-pathways from retina to V1
P- and M- pathways
P – parvocellular (small soma)
M- magnocellular (large soma)
spatial layout of retinal ganglion cell projections is preserved
Retinal ganglion cells project retinotopically to each layer of the LGN
Right and left eye projections are also segregated in the LGN
LGN neurons project to V1
6 main layers (stripes) in striate cortex (V1)
neurons on M- and P-pathway project to diff layers in V1
M-cells project to layer 4cα
P-cells project to layer 4c and 2,3β (interblob)
columnar structure of V1
In addition to the six horizontal layers, the neurons in V1 are further segregated into functionally distinct hypercolumns (ca 1mm2)
Some cells in V1 tuned to orientation of bar stimulus
Neuron responds with maximal spike rate when a slowly rotating bar reaches the neuron’s preferred orientation
Hubel and Wiesel’s (1977) ice-cube model of the V1
All neurons in an orientation column share the same preference for a particular orientation of a bar stimulus in their receptive field
Retinotopic organisation: Signals from co-located ganglion cells in the retina are processed by respectively co-located cortical neurons within each orientation column
Hubel and Wiesel (Harvard, USA) were awarded the Nobel prize in 1981 for their work in the 60s and 70s
Responses of neurons in orientation columns of V1
When recording from neurons of a particular orientation column, some neurons respond to orientation columns only within a small part of the visual field which corresponds to their receptive field
Different to the retinal ganglion cells, these neurons fire at the maximal spike rate when a bar stimulus shows their preferred orientation
Other cortical cells in V1 respond with the maximal spike rate to a preferred direction of motion of bars or patterns
Microelectrode recordings reveal that cells differ greatly in their receptive fields.
(A) Visual cells in the lateral geniculate nucleus (LGN) have concentric receptive fields, like those of retinal bipolar cells and ganglion cells
(B) Visual cells in the cerebral cortex are more responsive to bars of light and may show orientation specificity or respond only to motion, or
(C) they may respond only to motion in a particular direction
the receptive field of a simple cell in V1 (striate cortex)
Hubel and Wiesel proposed that simple cells in the orientation columns receive input from several neighbouring retinal ganglion cells
Low- and high-bandpass filters for diff behavioural responses
Filtering that reveals very low spatial frequencies (left panel) or moderately low spatial frequencies (middle) makes it look as if she is smiling, but the high-spatial-frequency components (right) make her look almost sad.
Is this why when we see her from the corner of
our eye, she seems to be smiling, but when we look directly at her mouth, it conveys sadness?
complex cells respond to bar orientation across visual field
Simple cortical cells (aka bar detectors or edge detectors): respond best to an edge or a bar of particular width, orientation, and location in the visual field.
Complex cortical cells: respond best to a bad or particular size and orientation anywhere within a particular area of the visual field
perceptual shape constancy
We recognise the same shape from different view points and directions despite the distortions in the retinal projections
size constancy
When objects are at different distances, the size of their retinal projection varies
Context information allows the brain to judge whether the same object is seen from a different distance or a different object is seen
functions of simple and complex cells
Analysis of contours and boundaries analysis of objects
Shape and positional invariance
Contour enhancement for object identification
V1 is fundamentally important for perceptual processing
main visual streams in the cortex of primate/human brain
dorsal stream (where system) - interacting with the world
ventral stream (what system) - making sense of the world
seeing for action: Eye-hand coord
Guiding hand movements requires two processes
Deciding which objects to interact with
Interacting with objects skillfully
These processes require different type of information from both the dorsal and the ventral streams
size of projections onto the retina
1° visual angle = 0.288 mm (size of thumb nail when extending your arm = 1.5°), Fovea – 0.6mm, horizontal extension of retina – 32mm
where does lateral inhibition occur?
where the neurons in a region—in this case, retinal cells—are interconnected, either through their own axons or by means of interneurons, and each neuron tends to inhibit its neighbours.
one group of bipolar cells
contains on-center bipolar cells: turning on a light in the center of an on-center bipolar cell’s receptive field excites the cell because it receives less glutamate, which otherwise inhibits on-center bipolar cells.
second group of bipolar cells
contains off-center bipolar cells: turning off light in the center of an off-center bipolar cell’s receptive field excites the cell because it receives more glutamate, which depolarizes off-center bipolar cells
what do on-centre bipolar cells have?
an inhibitory synapse (i.e. IPSPs in presence of glutamate) with those cones (or rods) that form the centre of its receptive field (metabotropic glutamate receptors mGluR)
what do off-centre bipolar cells have?
an excitatory synapse (i.e. EPSPs in presence of glutamate), ionotropic glutamate receptor (iGluR)
the sombrero-shaped response of a cell with a CS receptive field - In an ON-centre bipolar cell:
When the spot is in the OFF-surround, there will be more IPSPs than EPSPs (because no or fewer cones are stimulated in the On-centre than in the OFF-surround)
When the spot reaches the ON-centre, there will be more EPSPs than IPSPs because more cones in the ON-centre are stimulated than in the OFF-surround
each hypercolumn in V1 composed of
Left and right eye ocular dominance columns (L, R)
Orientation columns (rainbow colours)
Blobs (drawn as cylinders)
Hypercolumns (ca. 1mm2)
