The Visual System Flashcards
The andromeda galaxy
- It is 2.5 trillion kilometers away but you can actually see it with your naked eye.
- We don’t know if it is tilted or shaped as an elipse or a circle.
- This illustrates a fundamental problem of your visual system.
- Your visual system has evolved to use electromagnetic radiation (light to understand the world) and the world is 3D.
- But the light gets reflected on the back of your retina and so the 3D of the physical world are compressed onto the 2D dimensions of the retina.
- So the retina and the rest of your visual system has to reconstruct a 3 dimensional world base on a 2 dimensional image that is projected onto the retina.
- So how can your visual system know if it is looking at a circle tilted on its side or whether it is looking at something actually shaped like an ellipse. It cannot solve this problem.
Assumptions made by the visual system
- How are we going to understand the 3D world from a 2D image projected onto the retina.
- The way the visual system has evolved to handle this is to make assumptions about the way that the world works.
- These assumptions are the basis of a multitude of visual illusions.
- Visual illusions are all based on the idea that the visual system has built into it certain kinds of assumptions about the way the world works.
Illusions
- See a triangle even tho the lines are not there - see a complete shape
- See the lines as curved even tho you know they are straight.
- All because your visual system has these built in assumptions that are hard wired into the system and you can’t override them.
- Illusions based on shadding and colour - see two different shades of grey even tho they are the exact same. Your visual system is assuming that light is coming from above and so that the bottom part is in the shade and thus must be a lighter colour (colour in the shade must be lighter than colour in the light).
- black dots keep appearing in different places and dissapearing - when you look at the world, most of what you see is through a narrow cone in the center. Your visual system is filling in the gaps, filling in the patterns in the background.
Visual scene
Your retina and cerebral cortex are constructing a visual sxene. They are actually building the visual scene inside your head. And its not what is literally out there.
How does the visual world reconstruct the 3D world with sufficient accuracy that you can actually interact with the world effectively.
The eye
The eye is like a camera
* Visual system is not like a camera but the eye is.
* eye has a pupil (like a little opening for light to go through)
* lens that focuses the light
* surface that absorbs the light that has been focused onto it.
* the focused light that is being projected onto this very thin sheet of tissue (its acutally neurons on the back of the eye) retina. So the starting point of visual processing is light that is focused onto the retina in the back of the eye.
* There is a gap in the retina, the output of the retina are axons that form the optic nerve that is going to go up to the brain. Optic disk is the place where the axons are exiting the retina and it is a blind spit in your retina. So in each of your eyes, you have a little part of the visual space that you can’t see at all and that is called your blind spot.
* Blind spot is slightly different for both eyes. An image is never projected onto the blind spot of both eyes at the same time.
* You do not see this blind spot as a black spot, your visual system just fills in the gap.
The Retina
- The retina consists of five cell types.
- Photoreceptors are the cells that actually absorb light and transform light into an electrical signal (rods and cones)
- Photoreceptors are the input cells. Photons of light are bring absorbed by these photoreceptors and then being transformed into a change in the electrical properties of the photoreceptors (NTs being released by photoreceptor cells).
- Retinal ganglion cells are the output cells of the reitna. These cells have axons that come together to form the optic nerve that is heading up to the brain (heading up to the LGN).
- Neurons connecting the ganglion cells to the photoreceptors are called bipolar cells.
- Flow of info: light is shinning onto the photoreceptors, they are changing their electrical properties, the photoreceptors are making synapses with the bipolar cells and then the bipolar cells are relaying that information to the ganglion cells and then the ganglion cells are the output that is going up to the cerebral cortex.
- All of this circuitry is the initial stage of complex processing of visual information that is taking place in the retina itself.
- The light is actually shinning through all these layers to get through the photoreceptors. The photoreceptors are the only cells that are actually sensitive to light, they are the only cells sensing light and yet they are at the back of the eye. The main reason they are at the back us because of the metabolic needs of the photoreceptors (very demanding cells that require a lot of care and maintenance). The photoreceptors have to be close to the cells that are maintaining there health (which are at back of the eye).
- This does not affect visual acuity because one the retina is very clear and these cells are not colored, they are clear. So light just shines right through to get to the photoreceptors.
The 2 kinds of cells making horizontal connections
- They are connecting together different photoreceptors or different ganglion cells. These are horizontally oriented cells called horizontal cells.
- Horizontal cells are connecting together many photoreceptors allowing them to communicate with each other.
- Amacrine cells are connecting many ganglion cells together.
*
Retinal detachment
Retinal detachment is when the retina pulll away from its supporting cells, the cells invovled in maintaining the sort of metabolic health of the photoreceptors and this causes the photoreceptors to die.
The layers in the retina.
The retina has 10 layers
* Retinal pigment epithelium are the supporting cells for the photoreceptors.
* do not need to memorize the 10 layers.
The important info about the layers in the retina
This image is the practical summary - need to know these.
* Where the photoreceptors that absorb light are found.
* where the synaptic layers are. The layer where the photoreceptors are making synapses with the bipolar cells and the horizontal cells.
* Then you have a layer with the cell bodies of the bipolar cells and horizontal cells.
* Followed by a layer of synapses between the bipolar cells and the ganglion cells and amacrine cells.
* Then you have the ganglion cells followed by the axons of the ganglion cells.
Ratio of photoreceptors to ganglion cells
There are more photoreceptors than ganglion cells. This means that information is being sort of funneled into the ganglion cells.
* Ganglion cells are a narrow point, a sort of bottleneck in the output of the retina. This is an example of convergence.
* Information being gathered by a large number of retinal ganglion cells is converging on a smaller number of retinal ganglion cells and those are the outputs of the retina.
The fovea
The fovea is the high-acuity center of the visual field.
* The thing you are looking at is being focused on the fovea (this is the part of your visual space that has the highest resolution).
* What you see is mostly right in front of you and that is because the part of the visual field right infront of you is the part that your eyes are focused on (part focused onto your fovea). Fovea us specialized, it has high resolution.
* In the fovea, all those other cells are pushed out of the way so that the photoreceptors in the fovea have direct access to light.
The fovea
At the fovea ganglion cells and bipolar cells are pushed to the side, so the photoreceptors have direct access to light.
- they still have the same connections
- fovea is indented.
Retina has 5 different types of cells but for each category there are multiple subtypes:
- Two kinds of photoreceptors (rods and cones) –> the input cells
- There are at least 17 distinct types of retinal ganglion cells (output cells from the retina)
- 10 types of bipolar cells
- and more than 30 types of amacrine cells.
- Within each of the 5 categories there are specialized different kinds of cells and so the retina is very complex.
- Retina is a good system for studying how a network of neurons can process information because it is comprised of a comparatively small number of cells and the information flow is always in the same direction.
Morphology of photoreceptors
- These are the cells that absorb light and transform it into an electrical signal.
- Two kinds of photorecpetors: rods and cones.
- There is a particular region of the photoreceptors that is involved in absorbing light (the upper part with rods and cones on it). The reason for that is that the disks anc the convolutions in the cone create a huge amount of surface area for absorbing photons of light. So these regions are specialized for capturing photons of light.
- Another part of the photoreceptor is involved in the metabolism, just basically keeping the cell alive and the protein synthesis.
- The bottom is the presynaptic terminal.
- So rods and cones are specialized neurons because they have the specal function of absorbing light and transforming it into an electrical signal.
- Rods and cones do not fire action potentials (they change their membrane potential but they do not fire APs)
Location of Cones vs Rods
- Cones are concentrated in the fovea, whereas rods are concentrated outside the fovea.
- Cones are responsible for high resolution visual processing and color processing.
- No rods in the fovea but the rod concentration goes way up outside the fovea. Lots of rods in the peripheral.
- Rods are for night vision so they are super sensitive to light, whereas cones are much less sensitive to light. Rods enable you to see in the dark (a single rod can absorb a single photon of light and change its electrical properties - cannot get any more sensitive than rods).
- Rods are completely saturated by daylight. Do not use rods in lighting.
- Cones have low sensitivity to light so they have a hard time seeing very faint light but your rods are very sensitive to light so if you can move an object away from the center of vision, out to the periphery, you can actually see it.
- Sensitivity of light is greater in the periphery but your acuity/precision is much higher in the central part of your vision (when using cones).
Transduction
- Transduction is the process of transforming some kind of information about the external world into an electrical signal in the nervous system.
- Light hyperpolarizes photoreceptor by closing a cGMP-gated cation channel.
- The disks that are inside the specialized part of the rod, those disks are made out of plasma membrane, the same thing that the outer membrane of the cell us made out of. They are stacked on top of each other.
- These disks are specialized for absorbing photons of light and the reason they can do this is that they are covered with a protein molecule called rhodopsin (a G-protein coupled receptor).
- Rhodopsin absorbs a photon of light and that absorption of light causes rhodopsin to change its confirmation.The change of conformation causes it to activate a g-protein called transducin and that transducing activates a protein called Cyclic GMP phosphodiesterase. This protein breaks down a substance that is floating around inside the photoreceptor, which is called cGMP.
- So in the dark, there is a high concentration of cGMP in the cell and it binds to an ion channel.
- This ion channel is permeable to Na ions (it is not a Na channel). Therefore, in the dark, it is bound to cGMP so these channels are open so sodium is flowing in. So photoreceptor is depolarized in the dark.
- When light shines on the photoreceptor, it activates cGMP phosphodiesterase which degrades the cyclic GMP and so that reduces the concentrations of cGMP in the cell and these ion channels close. Thus, the photoreceptor hyperpolarizes.
Photoreceptors are depolarized in the dark and light causes them to hyperpolarize.
Bipolar cells
Bipolar cells connect photoreceptors to retinal ganglion cells
* out of the 5 cell types in the retina, none of them fire action potentials except for retinal ganglion cells (output neurons that will relay the signal up to the brain)
* In the dark, the photoreceptors are depolarized and continuously releasing neurotransmitter onto bipolar cells. When light shines onto the photoreceptor, it hyperpolarizes which closes those calcium channels and so the photoreceptor releases less neurotransmitter.
* The photoreceptors are releasing glutamate at their presynaptic terminal.
* The glutamate is actually hyperpolarizing the bipolar cell (it is acting as an inhibitory). So, in the dark, photoreceptor is releasing glutamate and so the bipolar cell is hyperpolarized. Therefore, the bipolar cell is not releasing neurotransmitters (bipolar cells are also releasing glutamate but in this case it is excitatory) onto the retinal ganglion cell. So, retinal ganglion cell is maybe firing a few APs in the dark, but it is basically quiet.
* When the cone stops releasing glutamate onto the bipolar cell (bipolar cell is being held at a negative potential by this inhibitory glutamate), this relieves the inhibition and the bipolar cell depolarizes and releases glutamate onto the ganglion cell. Therefore, ganglion cell will fire AP.
Why is glutamate inhibitory at the synapse between the photoreceptor and bipolar cell?
The neurotransmitter receptor at this synapse are metabotropic G-protein coupled receptors. Glutamate is not binding to an ion channel (which is the typical way for glutamate) instead when the receptor is activated, it activates a second messenger cascade.
Intensity of light effect on cascade
- Remember: the photoreceptors and the bipolar cells do not fire APs.
- The significance of this is that the AP is an all or nothing thing, if a neuron fires an AP, nothing is going to happen until you reach the threshold. However, the membrane potential of the photoreceptors and bipolar cells can change continously over a range, it is graded.
- The amount that the cone hyperpolarizes is going to be a function of the intensity of the light and the amount of neurotransmitter that the cone releases is going to be a function of its membrane potential. The intensity of light is transformed into the frequency of action-potentials in retinal ganglion cells.
- In the dark, cone will be releasing a lot of glutamate. In bright light, amount of glutamate is significantly reduced. At an intermediate level, the cone is only going to be somewhat hyperpolarized and so it is going to release sort of less glutamate. Therefore, **the amount of glutamate released is going to be a continuous function of how much light is absorbed by the cone. **
- same is true for the bipolar cells - membrane potential of bipolar cell is going to depend on how much glutamate gets released. No light = a lot of glutamate released and bipolar cell is going to be very hyperpolarized. Bright light = not a lot of glutamate, releaving inhibition on the bipolar cell causing it to depolarize.
- OVERALL, bright light causes the ganglion cell to fire a ton of APs and dim light causes the ganglion cell to fire fewer APs at a lower frequency.
Horizontal cells and amacrine cells
- What the amacrine cells do is very similar to what the horizontal cells do.
- The photoreceptors near the photoreceptor are connected to the same pathway indirectly through the horizontal cells.
- Horizontal cells are inhibitory. The horizontal cells are inhibiting transmitter release from the middle cone in the dark but when light shines on the horizontal cell that inhibition is partially relieved.
- The horizontal cells make these inhibitory connections through this pathway so that means when light shines on one of these cones that is connected indirectly to this pathway through horizontal cell, the end result is that the ganglion cell that is firing is going to be inhibited. So, if light shines on the central cone, the firing of the ganglion cell increases and if light shines on the cones adjacent, the firing of the ganglion cell decreases.
- The receptive field of the ganglion cell is made up of the cones that are directly and indirectly connected through these horizontal cells. The central cone is connected directly to the bipolar cell and then that is surrounded by a ring of other cones
- Example on lateral inhibition!
When light shines on the central photoreceptors…
- When light shines on the central photoreceptors it activates the ganglion cell.
- Ganglion cell starts firing more APs.
When light shines on the surrounding
photoreceptors…
- When light shines on the surrounding
photoreceptors it inhibits the ganglion cell. - Ganglion cell fires less AP (inhibiting it).
When we shine light over all the photoreceptors (central and surrounding photoreceptors)…
- The central photoreceptor should increases the firing rate and the surround photoreceptors should inhibit its firing.
- If you put them together, nothing happens. There is hardly any change.
Summary shinning light on an on-center and off-surround receptive field
These retinal ganglion cells have on-center-off-surround receptive fields, reflecting convergent input from multiple photoreceptors.
Why does the visual system have on-center and off-surround receptive fields?
- It is important for the visual systems ability to detect edges and contrast.
- The visual system is not that interested in uniform lighting (not interested in areas of the visual world where lighting is all the same).
- The ganglion cells are tuned to be better at detecting contrast between light and dark regions then they are at detecting uniform illumination.
Summary shinning light on an off-center and on-surround receptive field
Photoreceptors also make excitatory connections with bipolar cells. In this case, the ganglion cells are off-centeron-surround.
Where does the difference lie in these on-center, off-surround and off-center, on-suroound?
- The difference is in these bipolar cells.
- If the cell is on-center, it has a bipolar cell that is hyperpolarized by glutamate.
- If the cell is off-center, it has a different kind of bipolar cell that is depolarized by glutamate.
In the fovea, do we have on or off center cells?
- Every single cone in your fovea is connected to an on-bipolar cell, and an off-bipolar cell. So each cone is then connected to these two streams: one on-center, off-surround and one that is off-center, on-surroud. In the fovea, each cone forms the center of an oncentered and an off-centered pathway.
- Why does it have this? this means that the output of this region of the retina is going to be very good at detecting increases in illumination and decreases in illumination. Sudden increase in illumination will cause an increase in firing in the on ganglion cell. Sudden decrease in illumination will cause an increase in firing in the off ganglion cell.
- Every single cone in the fovea is connected to an on bipolar cell, an off bipolar cell and horizontal cells. So it has its own central pathway. And every single cone contributes to the suround receptive field of other cones.
- The receptive field of each ganglion cell in the fovea is quite small. That makes sense because your fovea is where you have high precision vision. This precision is a result of the receptive fields of the ganglion cells being so small.
Receptive fields outside of fovea
- As you start to move away from the fovea, the receptive fields of the ganglion cells become larger. Outside the fovea, more photoreceptors feed into a single ganglion cell, so the receptive fields
are larger. - When you start moving outside the fovea, multiple photoreceptors are going to contribute to the center and then a whole bunch of photoreceptors will contribute to the surround.
- They are collecting light from more photoreceptors, meaning they will be more sensitive to light. So periphery of your visual field is more sensitive to light then the center of your visual field.
Retina and Retinal ganglion cells
- There more than a dozen distinct types of RGCs. Each type tiles the entire retina, creating multiple parallel labelled lines to the lateral geniculate nucleus. Each labelled line conveys a distinct type of visual information.
- Different types of ganglion cells tile the retina - each type of ganglion cell carpets the entire retina.
- somewhere between 15-20 different kinds of ganglion cells.
- The significance of this; each type is like a seperate labelled line or a seperate channel that is conveying a particular type of info from the retina to the visual system.
3 types of retinal ganglion cells we are focusing on.
The three best characterized channels from the retina originate with the midget ganglion cells (~ 70% of RGCs) parasol ganglion cells (~ 10% of RGCs) and bistratified ganglion cells (~ 8% of
RGCs).
* midget ganglion cells = P-pathway
* parasol ganglion cells = M-pathway
* bistratified cells = K-pathway
Each one of these types of ganglion cells are conveying different aspects of the visual scene.
Types of ganglion cells responsible for colour vision.
- The three types of ganglion cells are responsible for colour vision.
- In daylight, each of these channels gets its main input from cones. The human retina contains three types of cones, called S, M and L. Each type is tuned to respond best to a distinct range of wavelengths within the visible spectrum.
- You need all 3 cones to see different colours.
- S responds to short wavelength, blue light
- M responds to medium wavelength, green light
- L responds to long wavelength, red light.
- You need all 3 cones to see colour, colour blind = missing one of the cones (more common in men, usually missing red or green).
- If you only had green cones, you would not be able to tell blue from any other colour.
Combinatorial processing for colours
- Brain is assessing the relative activation of the 3 labelled lines and uses this to intepret colour.
Midget cells
Midget cells are telling us how much green versus red is in the scene.
* Midget cells are concentrated in the fovea. They convey the amount of red versus green in their receptive fields.
* Their receptive fields are small, so they are good at resolving fine details, but not especially good at detecting rapid changes.
* They synapse with parvocellular layers in the LGN (p-pathway) which make a major contribution to cortical systems involved in processing of color and form.
* Start of a labelled line (p-pathway).
* They have a very small receptive field (high precision of fovea).
* Everything we have been talking about so far has been midget cells.
Red-On and Green-off –> when shining red and green light
- A red-on-green-off cell will increase firing to a** red** light covering the entire receptive field. This is because the center is activated by the red light and the surround will not be inhibited much as it is sensitive to green light.
- It will decrease firing in response to a green light covering its entire receptive field. There is strong inhibition because the center is not activated by green light and the surround is strongly sensitive to green light.
Red-On and Green-off –> when shining white light
- White light is the whole visible spectrum - has red and green component
- The same red-on-green-off cell will increase firing to white light in the center.
- It will decrease firing in response to a white light in its surround.
- So it also acts as an contrast detector.
Parasol cells
- Parasol cells contribute to the m pathway (another labelled line) that goes up all to way to the primary visual cortex.
- Parasol cells are found near the central part of the fovea, but they are not so concentrated right in the fovea.
- Parasol cells have larger receptive fields (because multiple cones contribute to center and multiple cones contribute to the surround). Not as precise.
- Convey an achromatic signal (i.e. no color discrimination).
- They have lower spatial resolution than midget cells, but better sensitivity to rapid changes in illumination. Detect rapid changes in illumination and movement.
- They project to the magnocellular layers in the LGN (m-pathway) which contribute to cortical systems involved in analysis of location and movement.
Why can parasol cells not detect colour?
- Parasol cells have larger receptive fields.
- They cannot detect colour because both their center and surround get input from both red and green cones. The center and surround of the receptive field is made up of a mixture of red and green cones.
Bistratified cells
- Bistratified cells detect the amount of blue versus
red/green (red + green = yellow). Relative amount of blue vs yellow. - They project to the koniocellular layers of the LGN (k pathway) and are involved in color perception.
- They do not have a center-surround. They are connected to on bipolar cells and off bipolar cells.
Bistratified cells - response to blue and yellow light
- Bistratified cells have larger receptive fields.
- Center and surround get input from both red and green cones.
- The receptive field has green, red and blue cones.
- If you shine blue light over the entire receptive field, it will activate the blue cones = excitation (firing of AP).
- If you shine yellow light over the entire receptive field, it will activate the green and red cones = inhibition of firing.
Can a single cone distinguish between different wavelenghts of light?
- A single type of cone cannot distinguish between different wavelengths of light, e.g. M cones respond best to green light, but will also respond to wavelengths covering most of the visible spectrum.
- The reason you can see colour is because you are getting ouput from the retina coming from all those different kinds of ganglion cells.
- Ie, if someone only has green cones. All that the cone can do is absorb photons of light and then chance how much NT is being released. So, the cells downstream from this cone cannot tell wether the cone is being exposed to a dim green lught or a very bright red light (bright red light will activate the cone, even though the cone is less sensitive to red light).
- This green cone is not very sensitive to red or blue light, but a bright red light or blue light will still activate the cone.
- The neurons downstream can only tell that there is a change in the amount of neurotransmitter that is being released. So, a single cone cannot descriminate between different colours.
How is colour encoded in the retina? How does this work - because you can see more than green,red, blue and yellow?
- Color is encoded in the retina by the relative output of the parasol, midget and bistratified cells.
- Parasol cells = give a measure of the overall luminance and brightness
- Midget cells = convey the amount of red vs green that is in the scene
- Bistratified cells = convery the amount of blue vs yellow that is in the scene. - The brain is interpreting the relative activation of these different pathways (labelled lines) and based on this is identiying the colour.
- If there is strong activation of red on center off suround ganglion cells and weak activation of these bistratified cells = red
- If there is strong activation of the blue on center bistratified cells and weak activation of the midget cells than that is probably going to be blue.
Projections from retina to the LGN
- The axons of retinal ganglion cells project as the optic nerves to the lateral geniculate nuclei. The fibers partially cross at the
optic chiasm and continue as the optic tracts.
Projections:
* Nasal side of retina crosses over (red)
* Temporal part of retina stays ipsilateral (blue)
* Left side of the retina of both eyes goes to the right side of the LGN
* Right side of the retina for both eyes goes to left side of the LGN.
What layers of the cortex do the midget and parasol ganglion cells project to?
- Midget and parasol ganglion cells project to six distinct layers in the lateral geniculate nucleus.
- Retina is tilled with ganglion cells - each ganglion cell is telling your brain something about a small part of the visual field.
- Retinotopic map is preserved to cortex like the somatosensory map.
How do the midget and parasol cells segrate into distinct layers of the LGN?
- The output from the 2 eyes is going to stay seperate all the way to the cortex. The output from the midget and parasol cells also stay seperate all the way to the cortex.
- m pathway: layers 1, 2
- p pathway: layers 4-6
- k-pathway: project in between these layers
These 3 different labelled lines are stauing seperate from the thalamus all the way up to the cortex. - The outputs that are coming from the temporal part of the retina and will stay ipsilateral (red) will go to layers 2, 3 & 5.
- The contralateral lateral geniculate is projecting to layers 1, 4 & 6.
- Functionally what this means is that from the retina going through the lateral geniculate all the way up to the cerebral cortex, the output coming from the two eyes is going to stay seperate because each eye is projecting to different layers.
Where do the bistratified retinal ganglion cells project
- Bistratified retinal ganglion cells project to LGN layers in between the prominent magnocellular and parvocellular layers.
- They are very small and project to regions in between the prominent regions.
- These regions, which are referred to as koniocellular layers, are sparsely populated with small neurons.
- output from bistratified cells will also stay seperate all the way to cortex.
Relay from LGN to primary visual cortex
Lateral geniculate neurons faithfully relay inputs from retinal ganglion cells to primary visual cortex, without significant change in the receptive fields.
* The receptive fields of the neurons in the thalamus hardly change at all. This means that if a ganglion cell fires some AP and this ganglion cell is makinga synapse with the LGN, the LGN neuron is pretty much going to do the same thing.
* LGN neurons are basically relaying the input from the retina up to the cerabral cortex. They are not changing the receptive property.
* Most of the neurons in the LGN have some kind of center surround receptive field just like the retinal ganglion cells.
Retinotopic map in primary visual cortex
Primary visual cortex contains a retinotopic map of the
external visual field.
Notice:
1. The area corresponding to the fovea takes up practically half of the primary visual cortex –> larger representation of high precision vision
2. Left primary cortex represents the right side of the visual space
3. Top half of cortex represents bottom half of visual space and vice versa.
Layer IV
Layer IV is the main recipient of afferent inputs from the thalamus.
* Projections from the LGN terminate in layer IV of V1
* Layer IV in V1 is specifically large and has lots of sub layers.
What layer of V1 do the LGN neurons project to?
Lateral geniculate neurons project mainly to layer 4C of
primary visual cortex.
What do neurons in V1 respond best to?
V1 neurons respond best to bars of light at a specific orientation.
* Neurons in V1=more complex receptive fields.
* Respond to bars of light (the beggining of edges)
* The orientation of the bar of light matters as well as the direction of movement of the bar.
How are the receptive fields for bars created?
- The elongated receptive fields of simple cells are built from convergent input from many layer 4 cells with roughly circular receptive fields
- Receptive fields are getting more complex due to conversion.
V1 neurons and orientation-selective columns
- These orientation-selective neurons in V1 are organized into orientation-selective columns. All the neurons in the column will have the same response properties.
- Discovered by Hubel and Weisel.
- V1 is the brain region in which our understanding of cortical columns is the most developed.
- The columns are arranged in a complex pattern of swirls. A cycle around a swirl contains neurons responding to the full range of orientations for a particular location in space.
- Experiment: they used optical techniques to detect when neurons are active and they were able to colour code the activity.
- All the neurons in 1 column have the same response properties (white lines are the boundaries of the columns).
- We have a whole series of columns that make a pinwheel (circle) and if you go all the way around all the different columns in that circle, it corresponds to neurons that respond to bars of light of all the orientations going from this all the way to 180 degrees.
- Taken together, you have neurons that respond to a small region of visual space to bars of light of all the possible orientations going around 180 degrees.
- The whole pinwheel together is a functional unit - this single round structure is going to enable the little region of the retina to detect edges of all possible different orientations.
- The primary visual cortex, each little pinwheel is mapping out all the possible orientations for a bar of light in that particular region of visual space.
- Hypercolumn = entire pinwheel that includes all the possible orientations.
Simple cells vs Complex cells
- The neurons in one single orientation-selective column were not all the same, they had 2 different types of neurons:
- Simple cells are orientation-selective and have centersurround receptive fields. They have an on-center and off-surround.
- Complex cells don’t have the inhibitory surround region in their receptive fields. Really good at detecting edges (all they need is an edge moving into their receptive field and they will start firing).
How are the complex cells receptive field made?
The receptive fields of complex cells result from integration of input from simple cells.
* You have a bunch of simple cells converging and that can give you a complex cell with more complex receptive fields
* Even in the primary visual field, we are buiding up this complexity, center-surround receptive fields are being assembled together to create simple cells, and then simple cells are being assembled together to create complex cells.
* The ability of the primary visual cortex to detect edges is getting more and more integrated.
Direction-selective field
- Some V1 cells respond best to bars of light that are moving in one direction (a subset of the orientation selective neurons).
- Within these orientation-selective neurons, you have a subset of direction selective neurons.
Neurons selective to length of bar
- Some V1 neurons are sensitive to bar length.
- Prefer a bar of light but only if the bar of light only goes part way through the receptive field.
- These neurons are really good for detecting the ends of edges.
Blobs
- Blobs are found just in layers 2 and 3 (superficial layers).
- Blobs are interspersed within the orientation columns.
- Neurons in the blobs tend to be more responsive to color and less sensitive to stimulus orientation.
- The functional significance of this is that the neurons in the blobs actually don’t care that much about orientation. They do not have the same kind of orientation preference that the regions outside the blobs do.
- the blobs are the starting point for colour processing in the cortex.
- the interblob regions are the regions where you have the orientation selective neurons. Where simple cells and complex cells are located.
- Blobs and interblobs are in essence columns: all the neurons in the blobs are going to prefer colours and all the neurons in the interblob are going to prefer certain orientations.
The two different types of colums that are superimposed on top of each other.
- Superimposed on the orientation columns and blobs are columns corresponding to alternating input from the ipsilateral and contralateral eyes. These are called ocular dominance columns.
- Inputs coming into layer 4, are seperated into 2 different columns (one group of inputs coming from the ipsilateral eye and one group of inputs coming from the contralateral eye) - columns of neurons responding to either the contralateral eye or ispilateral eye.
- Mixing together of the two eyes starts to occur in primary visual cortex, but at least all the way up to the primary visual cortex the inputs from the two eyes are segregated.
- There are multiple different columns in primary visual cortex. So if you think about the cortex as a flat sheet but there are a whole bunch of different properties that have to be mapped onto that sheet.
- we have to map on orientation selectivity
- we have to map colour
- we have to map motion (direction)
- we have to map depth
These are all organized as columns, so you really do not have one kind of column in the primary visual cortex. You have a whole bunch of different columns mashed together!! - Blobs and interblobs
- ocular dominance columns
- orientation selective columns
- columns for motion and depth perception
Dorsal and Ventral stream
Visual information flows from V1 through separate dorsal
(where) and ventral (what) streams.
* The dorsal stream encodes location and movement. From the occipital lobe to the parietal lobe.
* The ventral stream encodes form and color. From the occipital lope to the ventral part of temporal lobe
Dorsal and Ventral stream in the macaque monkey
dont need to know these
Area MT
Area MT in the dorsal stream is involved in detection of motion.
Lesions to MT
Lesions in MT neurons in monkeys produce deficits in detecting global motion.
* Using a screen that has dots on it and they are moving around.
Aperture problem
Because V1 receptive fields are small (they are only seing a small part of the object/edges of object), they may give misleading information on direction of movement. This is referred to as the aperture problem.
* Concerns the idea that receptive fields get larger as you move higher and higher up.
* deceived about the direction of movement.
How does Area MT aid with the aperture problem?
- Neurons in area MT integrate input from neurons with smaller receptive fields, so they can detect the overall direction of movement of an object.
- The little receptive fields are being combined together to from larger receptive fields. It is therefore seing a large part of the object!
- Importance of the receptive field getting larger, because as the receptive field gets larger, you can see more and more of the object and you can understand the global properties of the object like its overall direction of movement.
Dorsal stream
The dorsal stream contributes to multimodal sensory pathways from parietal cortex to premotor cortex that are involved in transforming perception into action.
The ventral stream
The ventral stream is involved in perception of color, form and object recognition.
* The ventral stream has specific regions that are involved in colour perception
* Ventral stream is involved in your ability to see.
* Lesions to specific stream that processes colour = no longer see colour.
Neurons in the fusiform face area
- Neurons in the fusiform face area respond
preferentially to faces (FFA = neurons that only respond to faces). - The fusiform gyrus is involved in the processing of faces
- Processing of face is built up from dots, bars … to complex image.
- Will respond to the face ANYWHERE in visual space.
FFA
Neurons in the fusiform face area respond
preferentially to faces.
Neurons localized in different regions
- Clusters of neurons in localized regions respond preferentially to different types of visual objects.
- Some areas in ventral stream are specialized for recognizing different things.
- Point is that there is not just a face area. In the ventral stream, there seems to me a lot of different areas that are specialized for a lot of different things.
Receptive fields as we move up the visual hierarchy
- Receptive fields get progressively larger at progressively higher levels in the visual hierarchy
- Visual information is getting all the way up to the frontal lobes of the brain.
Range of connections
There are extensive short-range and long-range reciprocal connectivity low and high levels in visual processing.
* Information is being fed forward to higher levels and at the same time there is feedback projections.
* 2 way arrows: every connection has feedforward and feedback projections.
Role of feedback connections
- Feedback connections may be involved in visual attention and in topdown anticipatory mechanisms that enable us to differentiate an object from the background.
- Feedback connections enable us to use prior knowledge to organize information (ie: organize the visual info flowing up)
- Decide what part of the visual field you attend to.
- The idea is tha the info is flowing up (bottom-up) and the control is being done via top-down feedback projections.
Predictive Processing Models
According to Predictive Processing models of sensory perception, higher levels in the sensory hierarchy feedback perceptual expectations to lower levels. This feedback is compared with sensory input. Differences between the feedback expectation and the feedforward signal create a prediction error signal that is fed forward to update the predictive model at the next level in the heirarchy.
- High levels build a simulation of what the world is like and then runing it –> Feed this back to lower levels.
- Lower levels will compare the sensory information coming in to the existing visual model from the higher levels. If there is a discrepency between the two, then the new information is fed forward to update the model. What is different has to be fed forward to update.
Main visual pathway responsible for conscious visual perception.
The main visual pathway responsible for conscious visual
perception goes through the LGN to primary visual cortex.
* ventral stream: conscious visual perception
* dorsal stream: important for you ability to use visual information, location and movement of the objects to guide your own movements.
Alternate route
- There is an alternate pathway by which visual information gets up to the cerebral cortex.
- An alternate route goes through the superior colliculus to regions of the parietal and frontal lobes involved in detecting motion and controlling eye movements. This alternate may be responsible for blindsight.
- 10% of from optic nerve project to the superior colliculus.
- Superior colliculus is involved in rapid eye movement. Superior colliculus is connected to regions in the parietal lobe and frontal lobe as well (involved in detecting the location of objects and detecting visual motion).
- These connections bypass the primary visual cortex and bypass the ventral stream of visual processing.
- In people with blindsight, who have lesions to the primary visual cortex (no conscious visual perception), but they are getting visual information that is reaching these higher order levels of the parietal lobe that are important for detecting location and motion - so they can still detect where things are and that things are moving.
This suggests:
1. the dorsal stream is really not involved in conscious visual perception (ventral stream and V1 are more critical for conscious visual process).
2. A lot of the things you use your visual field for don’t actually require a conscious visual perception at all. Suggests a lot of what your visual system is doing is below the level of concsiousness.
Another unconscious pathway
- Another unconscious pathway (bypasses the primary visual pathway and ends in amygdala) goes from the thalamus (LGN and pulvinar) to the amygdala.
- This pathway is thought to be involved in rapid, unconscious emotional responses to visual stimuli.
- Rapid direct visual pathways that go direct to the amygdala. Enables us to react emotionally.
- This pathway enables the brain to detect threatening situations (situations that elicit fear before you become consciously aware).
Projections to the prefrontal cortex are involved in…
Projections to the prefrontal cortex are involved in visual working memory.
The binding problem
The binding problem: How are the different components of vision (e.g. form, color, location, motion), spread out over disparate regions of cortex, bound together to form a unified, coherent percept.
* When you look at the world. The visual world is unified (you perceive the visual world as being unified - all put together in a single perception). BUT visual processing is not unified. There are areas of the ventral stream that are processing color, other areas processing form of object, dorsal stream processing location, movement. Different anatomically seperated regions of the brain that are processing different aspects of the visual scene.
* How does that happen? how does it get bound all together? We don’t really know.
one idea is that the neurons that represent the different aspects of the car (red, location, direction of movement…) , start firing AP in ways that synchronize their behaviour. Activity of these neurons becomes synchronized.
Lesions to different parts of visual pathway
