Lecture 2 - The Retina and LGN

Página 1

Chapter 1

Retina

Light has to travel to a good amount of neural gunk and blood vessels to reach the photoreceptors, but that is okay because it is usually diffused enough so that we don't see shadows in our field of vision.

Aside from bipolar and retinal ganglion cells, horizontal cells and amacrine cells connect across the retina.

Página 2

The optic nerve

After the optic chiasm, the optic nerve turns into the optic tract.

Página 3

The lateral geniculate nucleus (LGN)

Information from both eyes is still kept separate after the optic nerve enters into the LGN.

The LGN has layers: magnocellular layers for M-cells, one for each eye, and parvocellular layers for P-cells, two for each eye (6 in total).

Layers: contra, ipsi (magno), ipsi, contra, ipsi, contra (parvo).

Contralateral - right eye

Ipsilateral - left eye

Each layer forms retinotopic mapping which means that the signals form an orderly mapthe visual world.

Schiller et al, 1990 - trained animals to fixate on a point, then move gaze to the "odd one out" on the screen. Then damaged specific parts of P or M divisions of the LGM (resulting in loss in specific areas in field of vision) and recorded their performance. P division damage => difficulty to detect target based on colour, subtle shape change or fine-detailed texture. M division damahe => difficulty with movement or flicker. Big changes in shape or depth had no change in performance.

K cells - detect blue-yellow colour vision, as opposed to P cells (green-red)

It is believed that the LGN filters out unnecessary information before sending it to the cortex (because it also receives a large amount of info top-down, from the cortex itself)

Página 4

Chapter 2

Receptive fields - grouped together by the retinal ganglion cells

- useful for detecting changes in environment so we can filter out the rest pretty much (all unnecessary info)

- the area of the retina which a ganglion cell is influenced by

- these receptive fields are antagonstic, because the centre has the opposite effect of the surrounding area

If every detail was picked up by the retina, that would create a huge blind spot and would require way too much energy

A neuron's baseline activity level is it's frequency of firing when it isn't being especially stimulated.

Receptive fields have 2 regions:

the ON field - where light increases the frequency of firing

the OFF field - where light decreases the frequency of firing

Keeping this in mind, there are 2 types of cells:

ON centre (off surround) - light in the centre increases frequency (maximum excitation) and light in the surrounding area decreases it (maximum inhibition)

OFF centre (on surround) - light in the centre decreases frequency and light in the surrounding area increases it

Página 5

Receptive fields and image processing

On ON centre cells: the centre has an excitatory input and the surround an inhibitory one.

ON centre cell responses:

no input: baseline frequency

light only on centre: high frequency (maximum excitation)

light only on surrounding area: low frequency (maximum inhibition)

light everywhere: baseline frequency (little to no change)

The effect the light has on the areas is d.p. to the percentage of the area that is covered by light.

Many receptive fields overlap.

Therefore, retinal ganglion cells only signal changes within their receptive fields and not overall modifications. This antagonism of their centre and surround fields is called lateral inhibition.

Illusion caused by the centre surround receptive fields:

The Hermann/Hering grid - back squares on a white background - illusory grey dots appear in peripheral vision (because foveal receptive fields are too small to catch the whole of the intersection)

Explanation: receptive fields at intersection are exposed to more white than those between intersections. Therefore, the ganglion cells are more inhibited and fire slower (thus, look "darker") than those between intersections.

BUT: this can't be the whole story, because if we make the lines "squiggly", the whole illusion disappears. Also, if we make it smaller, the foveal receptive fields do not respond in the same way as the peripheral ones. Rotating the image 45 degrees weakens it too.

Another illusion, the scintillating grid (Schrauf et al, 1997), looks identical, except the intersections have white circles in them (which turn black in your peripheral vision), but it seems to follow completely different rules.

The simultaneous contrast effect might also be due to the receptive fields of the ganglion cells.

Página 6

Change over time

We have 2 types of cells: sustained and transient cells.

Sustained cells give off constant signals when stimulates, while transient cells only signal when the stimulus turns on or off.

However, even sustained cells get bored, so they become inhibited over time.

If the eye didn't have tremors (minuscule movements we aren't aware of), the whole retina would become inhibited and the visual field would turn grey.

Because of this, we encounter a trick called Troxler fading (Wolfe at al, 1984), in which tho blobs (one black, one white) disappear after staring at a fixed point for a period of time. If gaze is switched to another fixed point, we see their after-images.

Change over space?

Because of our cell's receptive fields, we encounter an effect called simultaneous contrast, which means our minds can see the exact colour in two different shades, simply because of it's surroundings.

Filling in

The Craik–O’Brien–Cornsweet illusion is a prime example of how our minds collect information from edges and use it to "fill in" the whole image, until they reach a new edge. The illusion shows two apparently different rectangles, one lighter and one darker grey. In reality, both of them are the exact shade of grey, but the edge between them has been made gradually lighter (on the light side), then gradually darker (on the darker side), therefore our brain assumed the whole rectangle was the same.

Página 7

Chapter 8

Children

Researchers think children have shitty vision. :( (maybe it's blurry??)

Studying their vision

Behavioural methods:

For infants, researchers use preferential looking .

Ex: children are shown a card with a blank and a pattern. If the child looks more towards one side (usually the pattern), it usually means they can distinguish the two (so they can see the pattern). This is how researchers measure acuity (clarity of vision) and contrast sensitivity (the contrasts they can perceive). Preferential looking is very commonly used with stripe patterns.

However, stripe patterns stop being interesting for babies at around the 1 year mark. The researchers then use optotype cards. Very commonly used in the Cardiff Acuity Test, they consist of carefully constructed patterns of black and white lines which, if you do not have the necessary acuity, cancel out into the same colour as the grey background and are no longer visible.

Another thing researchers measure, other than looking time, is sucking time. Infants tend to suck on their fingers, for example, when they are bored. If they interrupt that when a stimulus is presented, we could assume that they can distinguish it from a blank.

Non-behavioural methods: EEG

We show kids an image about 100 times and we assume that the visually evoked potentials (VEPs) that are consistent throughout all the trials are due to the stimuli that the infant is being presented with. Like this, we can see if a child can distinguish between different stimuli (their visually evoked potentials change).

Página 8

ALSO:

General use: V1, V2, V3

Ventral: colour, fine detail, outline - V4 & V8

Dorsal: movement, shape, depth - V5 (MT) , V6, V7