Acuity and Visual Pathways PDF

Preview

Summary

Dr. Tracy Centanni Experimental Psych: Perception notes. These are comprehensive lecture notes and I attended every single class, getting an A in the class at the end of the semester. ...

Description

Acuity and Visual Pathways Friday, February 8, 2019

10:59 AM

• Visual acuity: the smallest spatial detail resolvable at 100% contrast • Snellen Chart: defines your visual acuity ○ A 20/20 letter is designed to subtend 5 arc minutes (0.083 degrees) designed to subtend 1 arc minute (0.017 degrees) at a distance of 20 ○ The first number is the distance you'd need to be to discern the lette ○ The second number is the distance at which someone with optimal a same letter ○ At 20/40, you would need 20 feet away to read a letter discernable th optimal vision could see at 40 feet (so someone with perfect vision c read that same letter that you would need to be at 20 feet away to see Minimum acuity (4 types) 1. Minimum visual acuity (0.00014 deg of visual angle): the smallest obj most sensitive you are in this specific domain, really small visual angle a. Detection of a feature-- can you tell something is there? b. Ex: thin black wire on a white background, you can see it if it is at th 2. Minimum resolvable acuity (0.017 deg): the smallest distance betwee detect a. Resolution of two features--are they different things? b. If you have two black bars on a white background that are getting clo between has to be at least .017 deg visual angle to see the gap c. Black and white bars 3. Minimum recognizable acuity (0.017 deg): the smallest size at which object a. Identification of a feature--what is it? b. Snellen Test--shows you all these different stimuli and asks you to id 4. Minimum discriminable acuity (0.00024 deg): the smallest change i objects that one can identify a. Discrimination of a change in a feature-b. If you have two black bars that are starting at even with each other, h to move one bar to see something has changed? When we talk about acuity we are talking about only CONES

with each stroke eet uity could discern that at somebody with uld be at 40 feet and ) ct one can detect. The s degree of visual angle n two objects one can ser together, gap in one can identify the ntify them a feature across two ow far away do we have

When we talk about acuity, we are talking about only CONES. • When you're measuring visual acuity, you're measuring it at the fovea The biological mechanisms of visual acuity--cone location • You need two cones solidly within each item • Need to have bars big enough that one cone is in each bar--single cone is • If they get so close together that no single cone can be within one bar, can one another. ○ If a single cone gets input from multiple bars, it cannot tell The biological mechanisms of visual acuity--ganglion cells • As we pull bars closer and closer together, changes input center vs. surrou • When the spatial frequency of the grating is too low, ganglion cell respond the fat bright bar of the grating lands in the inhibitory surround, which da • When the spatial frequency is too high the ganglion cell also responds we dark and bright stripes call within the receptive field center which washes ○ If you have a super fast wave, just as much input in center as you do other out • When spatial frequency is just right with bright bar filling the enter and d surround, the cell responds vigorously • Each cell responds best to a specific spatial frequency that matches its rec responds less to both higher and lower spatial frequencies Acuity: Center vs. Periphery • While you have really good motion perception in periphery, you don't hav • If something comes at you from the side, the first thing you do is look at i object to hit fovea and get the most detail • Less acuity in periphery than in fovea Massive data reduction • Consider that at the retina, there are 125 million receptors… ○ Plus all the activity in the millions of bipolar, horizontal, and amacri ○ These converge on mere hundreds of thousands of ganglion cells (a m ○ Ganglion cells summarize the data and output through their axons, w nerve ○ All cells of retina function as filters to be passed back towards the br • What type of data reduction can we do? E ti fi ld t i tf lti l h t t df

Acuity: Center vs. Periphery • While you have really good motion perception in periphery, you don't have good acuity • If something comes at you from the side, the first thing you do is look at it because you want object to hit fovea and get the most detail • Less acuity in periphery than in fovea Massive data reduction • Consider that at the retina, there are 125 million receptors… ○ Plus all the activity in the millions of bipolar, horizontal, and amacrine cells ○ These converge on mere hundreds of thousands of ganglion cells (a massive reduction) ○ Ganglion cells summarize the data and output through their axons, which make up the optic nerve ○ All cells of retina function as filters to be passed back towards the brain • What type of data reduction can we do? ○ Every receptive field gets input from multiple photoreceptors and from this creates a single output M and P cells • Two classes of ganglion cells that further process light signals in preparation for transmission into the head • Parasol cells (M cells) transmit to magnocellular (big) cells in the Lateral Geniculate Nucleus (LGN--a structure we will talk about later) ○ Input from diffuse bipolar cells that pass on to parasol cells (M type) so they will not have color vision, getting input from lots of different cones ○ Really good at motion, timing, depth perception • Midget cells (P cells) transmit to parvocellular (small) cells in the LGN ○ Only getting input from a single cone, can tell colors apart ○ Also have better acuity--can resolve better detail • Both cells interact with each other--when you're focused on an object d th f thi t th t d 't t t

and the move eyes from one thing to the next, you don't want a trace of everything you last saw. These shut off spatial vision • Focus on bottom chart for parvo vs. magno Leaving the eye… • Each eye has a left visual field and a right visual field • Everything in the middle is overlapping • As it enters your eye, it is backwards when it hits the retina • Your right visual field hits left side of the retina on each eye, and your left visual field hits the right side of the retina on each eye • Matters because we have to be able to track left side of world vs right side of world

•

Two paths to the middle of the head: • All input leaves the eye as the optic nerve • On the ventral surface of the brain, the optic nerves meet in the middle (left and right) on a place called the Optic Chiasm • Optic chiasm: left and right visual field input gets consolidated and inputs cross over • Once it leaves the chiasm, now the optic tract • Right goes to left, left goes to right, all crosses over • Eventually make it to primary visual field at the back of the brain • There are two pathways from the eye to the brain ○ Tectopulvinar pathway: evolutionary very old, share it with lots of reptiles, primary way that most animals see ○ Geniculostriate pathway: newer pathway ○ Both leave the eye and cross chiasm together, after they leave this is where they stray off

Tectopulvinar System • Leave the chiasm and go to the superior colliculus • Most of the projections in this system come from magnocellular cells • Goes through superior colliculus… • Superior colliculus: deep brain structure ○ Controls eyes as they move from location to location ○ Retinotopic map: organized based on where inputs came from in retina ○ Anterior portion of SC gets input from fovea (center of vision), posterior sides of SC gets input from periphery • Maintaining fixation, planning new targets, making that movement, fixating on that target • Never get to V1 The experimental task: • Monkey supposed to move eyes to cross in periphery when he sees it, will get sugar water • Most activation is in the front of SC when he is looking directly at cross in front of him, not activity in the back really because there is nothing in periphery • 100 ms before he moves his eyes, you get a reduction in activity on anterior side of SC--system releasing its hold on fixation, preparing to move eyes • Immediately before the movement, you have a reversal of activity-activation in posterior side now, away from center and towards periphery • Now when peripheral object appears, you have moved your eyes, had a release of activity from the frontal to allow eyes to move • Once eyes have moved and you are refocused on new target, go back to how it was in the beginning (anterior focus) Geniculotriate System • Our primary visual system, gets to primary visual pathway • Newer pathway • Most of the input from your eye goes through this route • Inputs go to the LGN (lateral geniculate nucleus) ○ Part of thalamus (relay station)

○ Part of thalamus (relay station) ○ Takes all sensory information coming in and routes it to other areas of cortex • To chiasm Lateral Geniculate Nucleus (LGN) • Gets its input from the eyes • 6 cortical layers in LGN--different layers for different eyes and for different cell types • Magnocellular cells go to layers 1 and 2 • Parvocellular cells go to layers 3, 4, 5, 6 • So this gets most of its input from parvocellular (p-type cells) • For every 10 inputs that come into LGN from eye, only about 4 outputs--reducing the amount of data that we are passing back to visual cortex From LGN to V1 • Long fiber tracks that go from LGN to V1--doesn't hit any more synapses one it leaves LGN until it hits V1 • V1 is the first stop in cortex • Hubel & Weisel: did a lot of visual experience regarding how V1 takes input from eye and uses it ○ They identified this striping pattern in the primary visual cortex--call it striate cortex because of striping ○ Can track same type of organization into V1 from LGN (layers from different eyes; left eye, right eye separation is maintained, haven't consolidated information from both eyes yet) ○ They also experimented on firing properties in V1: § Had a cat and put electrodes directly into V1--presenting complex scenes to cat and were not getting any response at all from the neurons § Got frustrated and one day were doing this experiment and the slide got stuck at a weird angle so there was a shadow of an angle at the screen; neurons started firing a lot § Built on this and started presenting different angles, lines, orientations, and found that there was an organization in V1 based on line orientation

V1 based on line orientation § Tuning: neuron is tuned to a very specific orientation § There are different neurons that are tuned to every possible orientation Three classes of cells in V1 • All 3 receive inputs from all areas of visual field • Each has a specific type of receptive field • The three classes: ○ Simple ○ Complex ○ Hyper-complex (end-stopped) • Each of these classes has a different specialty • Now that we're in cortex, cell types will start doing code • CODERS, not detectors!! Simple cortical cells • Respond to orientations and have receptive fields like ganglion cell • Have a center and surround and fire most strongly when center is activated • Receptive fields here come from the LGN--depending on which LGN cells fire will determine whether or not these simple cells fire • When a vertical part is activated in LGN (there is a vertical bar), it will fire more • At level of LGN, if LGN cells fired at a specific angle, then those cells will tell simple cell in V1 that is tuned to that angle to fire as well

Complex cortical cells • Take movement into account • Doesn’t care where in the receptive field the light hits as long as it's

• Doesn t care where in the receptive field the light hits as long as it s in the receptive field and at the right angle • More resistant to movement--angle can be anywhere in field and will still get firing End-stopped (aka hypercomplex) Cortical Cells • Still have an orientation tuning • Really interested in relationships between stimuli • Fires the most when edges of angle line up with edges of the box-when the relationships are precise 3-D columns in V1 • V1 has multiple layers of organization • 1st layer of organization is those who like orientation ○ All these layers of cells in the same column like a specific orientation ○ Organized by orientation--from a flat line to a vertical line as you move from column to column • Then have blobs responsible for color ○ Groups of cells that respond to specific colors ○ Short wavelength, medium, and long wavelength cones --blue, green, red • Hypercolumns: a 1 mm block of striate cortex containing at least two sets of columns, each covering every possible orientation (0-180 degrees), with one set preferring input from the left eye and one preferring input from the right ○ group input from eyes together ○ Each hypercolumn has input from both eyes and orientation columns ○ Have blobs and orientation columns ○ Don't exist in isolation--talk to each other § If you have a really big stimulus; talk and say here is what I see, do you see something similar? Same thing or two different things next to each other? ○ A single hypercolumn is reflective of a certain point in space § Code for visual input at a specific part of visual field ○ Not quite like a grid, more like a gradual change in gradients

Activating a hypercolumn • There is a line in the far left side of vision in A • Activates from one component of hypercolumn and one orientation column because you have a specific horizontal bar that comes through from one eye • The larger the stimulus, the more hypercolumns you will activate-the larger the bar the more will activate ○ In this case you are only activating the left eye (either right is closed or blocked) ○ So only getting left half of hypercolumn activated ○ If you were to expand input into the right eye, you will get the other half of the hypercolumn involved Lateral connections in V1 • Hypercolumns don't exist in isolation--talk to each other ○ If you have a really big stimulus; talk and say here is what I see, do you see something similar? Same thing or two different things next to each other? Muller-Lyer illusion • Which one is the longest line? All the same • Because hypercolumns are talking to each other, the angle of the arrows tricks them into thinking that there is more information on the end than there actually is • Not just a primary visual cortex illusion but also helps us clarify point about lateral communication within V1 Cortical magnification • One way we magnify input: from retina to the brain • about .01% of axons that come out of eye come from fovea • 8% of V1 responds these outputs from fovea --magnifying ○ A lot more cortical space dedicated to axons ○ Need to spend more time to process foveal inputs because of the detail it detects • Dense packing of cells in fovea • More ganglion cells devoted to fovea...