perception-visual architecture 2 Flashcards
structure of v1 - and population coding summary
-due to v1s orientation and structure it can use something known as population coding to determine what it is seeing out there, despite neurons being noisy in themselves,
-use pop coding and do a form of code fitting to find out exactly what is out there
ocular dominance in v1
-stripes of neurons in the visual cortex of certain mammals (including humans) that respond preferentially to input from one eye or the other.
-left half of the world seen by right half of brain and vice versa
-in v1 the information from the left eye and right eye do not get fused (neurons that get input from right vs left eye stay separate)
-we have orientation columns then we have these ocular dominance columns so these neurons are dominated by one of the eyes
the ice cube model of v1 architecture
what is it composed of
-we have orientation columns (each column has slightly diff preference of orientation )
-there are also ocular dominance columns which means there are some neurons getting information only either from the right eye or left eye.
-also then have little columns of cells called blobs(they dont have a particular structure) which process colour
-taking all of it, orientation columns, a right and left eye and blobs constitutes of one hyper column
-in the slide example eg there are two hyper columns
difference between model and real world
-orientations are organised very systematically, in the model they are organised in a straight line but in the real world they are organised like a wheel (pinwheel structures ) (each colour representing different orientation preference)
-shows photo of actual v1
v1 vs v2 processing illusion
illusion in slides
v1 cannot tell you the triangle is there as there are no lines (cant process these illusory lines if they are not there)
so neurons in v1 will not fire
so v2 allows us to see these things
what do illusions tell us about our visual system
-the visual system doesn’t faithfully present things as they are (unlike a camera)
-it constructs the best and consistent interpretation, given the input
where is v2 - secondary visual cortex
-2 halves of v2 one above and one below v1
-the two halves : top one is dorsal (its closer to the back)
-the lower one is called ventral
v1 versus v2 receptive fields
-generally, directly looking at something the receptive fields are very small and they get bigger and bigger on the outside
-smaller perceptive field allow you to see in more detail (thats why things appear weaker in your periphery
-v2 the receptive field size is much bigger
-central ones still have larger than v1 and they continue to become bigger
-the reason they have bigger receptive fields is because each v2 cell gets information from many v1 cells and put them together
v2 retinotopy (mapping of how we see)
-characteristic of v2 is that it is reverse related to v1
-then v3 is reversed from v2 (so v1 and v3 representation is the same)
-retinotopy of v2 also split into two halves (remember dorsal and ventral) both are reverses of v1 on the section they are in
v2-does it also have cortical magnification?
yes
remember
-Cortical magnification describes how many neurons in an area of the visual cortex are ‘responsible’ for processing a stimulus of a given size, as a function of visual field location.[1] In the center of the visual field, corresponding to the center of the fovea of the retina, a very large number of neurons process information from a small region of the visual field. If the same stimulus is seen in the periphery of the visual field (i.e. away from the center), it would be processed by a much smaller number of neurons
v2 architecture
Look at notes
stripes : characteristic of v2. V2 is the only structure with a striped architecture.
what happens when you stain v2
staining it and putting it under a microscope :
-you can actually see the striped structure along it,
-some stripes are thicker and some stripes are thinner
-called them thick stripes , thin stripes and pale stripes (pale stripes in between the darker ones)
the stripes in v2 architecture -function
-of thin stripe
-pale stripe
-thick stripe
-all the different stripes are getting information from different parts of v1, therefore there function depends on what kind of input they are getting
-eg thin stripes get their info from blocks which process colour info, so thin stripes also process colour
-pale stripes get info from in between the blocks , therefore pale stripes process orientation and form etc
-thick stripes get info from the lower layers which is where you get info from left and right eye and therefore processes depth
v2 thick stripe
-where does it get info
-what does it process
-what cells does it contain
-gets info from ocular dominance columns (so left eye , right eye)
-this is the first region in the brain to represent 3d info, so has neurons which care about depth
-it has near cells and far cells (cells with neurons which fire when something is close to you or fire when objects are far away)
-only cells in the thick stripe have cells which prefer either close objects or far objects
v2 orientation in thick stripe
-has some amount of orientation information, what it does with that is it can also process motion
-they dont process orientation but if something is going in a certain direction the neurons fire
-so thick stripe processes depth and motion
