Striate cortex

Question 1

what is the basis for all higher level vision?

Answer 1

mapping from the retina to the area of the neocortex
- particularly striate cortex in V1

Question 2

why are blood vessels in the eye important? Why are they a problem?

Answer 2

The blood vessels are really important because the cellular population of the retina (particularly photoreceptors) are very metabolically demanding and need oxygenated blood; The blood vessels are a problem because they block information and create gaps

Question 3

Sulci vs. Gyri

Answer 3

sulci are the folds in the brain while gyri are the portion of the brain between the folds
- we get sulci and gyri because the neocortex is a sheet of cells 2.5 mm thick that when you unfold it it can occupy the size of a desktop (for human)
- This maximizes the surface area of the neocortex into a volume ⇒ as you go up in the brains there are more sulci and gyri, more convolution, and a larger ratio of surface area to volume

Question 4

what does the metabolic profile of the cortex show compared to the retina/optic disk?

Answer 4

There is a clear pattern of inactivity in the cortex that reflects the pattern of occlusion of visual information created by the optic disk in the blood vessels
- This 1:1 mapping of inactivity in the cortex and occlusion factors like the optic disk and retina is shown at the cortex
- This is a problem because the structure that is supposed to precisely and accurately represent the outside world has gaps in it

Question 5

why don’t we segregate an object due an occlusion?

Answer 5

because we need to infer its presence even when at the particular location it is blocked

Question 6

how do we solve the problem of gaps in the retinal map to the cortex?

Answer 6

even if particular neurons are inactive, you use a population of activity to infer what is out in the world via the population encoding

Question 7

what do diagrams of activity over the cortical surface in V1 show w/ retinal stimulus?

Answer 7

the diagram shows images of activity over the cortical surface in V1 as a function of time after the onset of a small stimulus
- Within 200 ms there is a nice peak of activity corresponding with the retintotopic location of the stimulus

Question 8

why even when there is an absence of center stimulation is there activity around the square side in the diagrams?

Answer 8

This means that the area around the middle has bled in and we are inferring that there is actually a square sitting in front of a full field of grating in the back
- The square is an occlusion that should somewhat be ignored
- if we averaged the activity and asked what the center is, we would find the same point in the top and bottom activity graphs

Question 9

how does a neuron put in its vote?

Answer 9

To some extent a neuron with a specific receptive field is voting that a stimulus like that is in the world
- This is the representation that it is instantiated int he neurons receptive field properties

Question 10

how did Hubel and Wiessel figure out how the center surround connect to the cortex?

Answer 10

they had a microelectrode inserted in the striate cortex to listen to action potentials of individual cells (very few action potentials in the cell)
- When they switched a slide they were using to project the imaging and in the act of removing and replacing the slide the cell went crazy and fired
- They concluded it was the edge of the slide that evoked activity
- From here they constructed stimuli based on edges to characterize V1 receptive fields

Question 11

position selective receptive field

Answer 11

Slight changes from the X location has silence

Question 12

orientation selectivity

Answer 12

at different orientations the neuron is completely silent

Question 13

Simple cortical cell

Answer 13

an on orientation selective cell that is also positionally selective ⇒ if you line up appropriate receptor fields you can create this
- the 4 neurons are in the lateral genicular nucleus (relay structure) with a receptive field organization that largely mimics what is seen in the retinal ganglion cell with center surround organization
- The 4 neurons converge on a cell in the primary visual cortex (striate cortex) ⇒ when there is a light bar at the particular orientation then all the neurons will be active and their convergent input will create activity for the simple cortical cell to fire
- If we slightly change the angle then activity will be less ⇒ if we are orthogonal such as only activating one center surround with little contrast (deactivating the off region) we can predict silence

Question 14

what is the only way to create a simple cortical cell?

Answer 14

the only way to create a cell like a simple cell is to sample form retinal ganglion cells that have the same polarity and are colinearly aligned in visual space ⇒ any other selectivity will not create position selective orientation selective responses

Question 15

Complex cortical cell

Answer 15

there is not a level of on/off specific locations but they are orientation selective ⇒ as long as there is appropriate orientation the complex cell will discharge
- The hypothesized that if you add up appropriate simple cells that have a lot of different polarities and positions but are all orientation selective
- You get orientation selectivity by the simple cell and you get an invariance to brightness or darkness by summing up the on and off center simple cells

Question 16

what % of striate cortex cells are orientation selective?

Answer 16

90% (almost every cell)

Question 17

what type of image can we see with only V1?

Answer 17

if we think about V1 as edge or line detectors that are position and orientation selective that would look like an image with outlines
- That is a pretty understandable image ⇒ text, objects, important things with just an understanding of where edges are and how they are oriented

Question 18

when doing the blindspot demonstration with the vertical bars that have a gap in the middle of them what does this show when we see one line that doesn’t have a missing gap?

Answer 18

This means we need to infer orientation selectivity and infer an edge to make the straight line

Question 19

what did Hubel and Wissel find when they had a microelectrode that passed through at a shallow angle through the neocortex and at each regular point they recorded the orientation selectivity?

Answer 19

There is a smooth progression of the line angle and you can go on to make a complete cycle
- There is a smooth variation of orientation selectivity across the surface of the cortex and nearby neurons share a common orientation selectivity in the cortex

Question 20

what does optical imaging show us?

Answer 20

the functional selectivity across the surface of the cortex
- The results of the experiment when looking at the cortical surface can have color distribution demonstrating the angle of response of orientation columns

Question 21

Orientation columns

Answer 21

there are patches of common orientation selectivity on the neocortex

Question 22

what do we see with the projections (mm long) in the optical imaging?

Answer 22

If we inject in a green region that prefers 45 deg, then the predominant label is also in green and it avoids other colored areas
- the boutons indicate there is long range connectivity between similar orientation preferences

Question 23

what do long range horizontal connections provide a solution to?

Answer 23

filling in higher order representations
- if any one location is inactive its going to receive collateral input from long range of similarly orientation preferenced neurons ⇒ we can then infer the continuous presence of that orientation
- Within a few ms of information leaving the retina there is a mature pattern of orientation selectivity across the visual cortex ⇒ due to feed forward convergence between appropriate retinal ganglion cells (after 50 ms)
- All that needs to happen is the cells of an appropriate polarity and collinearity are activated and if these inputs are summed appropriately there is an orientation selective response

Question 24

Functional organization

Answer 24

how the receptive field properties systematically vary over the surface of the cortex
- the precision of functional organization is remarkable ⇒ in many animals if we look in an orientation column we will find segregation with the direction of motion

Question 25

for cell clusters that respond to opposite degree angles what do we see?

Answer 25

if you look at the orientation of the green and red columns that respond to a 45 degree angle there are some cells that respond when you move the bar upward and some that respond when you move the bar downward but those cells are not intermixed (complete direction selectivity) - There is a precise border where you go from one direction preference to the other - In the intermediate region (4) there are cells that are not direction selective or that respond to both but the intermediate region is very small and just a border between directions of 1-2 cells

Question 26

do we see clustering of feature selectivity in rodents?

Answer 26

No, the receptive field properties (orientation and direction selectivity) can also exist in the absence of of functional organization - When we look at individual neurons we can find strong direction selectivity and strong orientation selectivity - the patterning is all mixed together and there is no tendency for cells with nearby direction or orientation preference to be next to one another in rodents