week 12 (vision + recognition) Flashcards
(43 cards)
1
Q
explain: retinotopic organization in LGN
A
- topographically mapped
- each layer = info from 1 eye
⤷ monocular - 1, 4, 6 get info from contralateral eye (nasal retina)
- 2, 3, 5 get info from ipsilateral (lateral retina)
**each layer = monocular but overall, each LGN = binocular
2
Q
explain: functional segregation in LGN
A
- M ganglions project to M layers
⤷ good for large fast moving objects - P projects to P layers
⤷ good for details and stationary
3
Q
question: where else do the RGC project to? (aside from LGN) (3)
A
- superior colliculus
- pretectum and edinger-westphal nucleus
- suprachiasmatic nucleus (in hypothalamus
4
Q
explain: other structures tha LGN projects to (aside from LGN) (3)
A
-
superior colliculus
- integrates vis and somatosensory
⤷ coordinates eye and head mvt. w/ stim.
⤷ moves focal point if needed (saccades) -
pretectum and edinger-westphal nucleus)
- controls lens and iris
- iris = in too bright/dim -> sends motor output to eye from cranial nerve II to adjust pupil
- lens = out of focus -> sneds motor output to ciliary musc. to change accommodation -
suprachiasmatic nucleus (SCN)
- in hypothalamus
- photosensitive (resp. to ambient light)
- circadian clock
⤷ in less light -> projects to pinneal gland -> prod. melatoning
5
Q
explain: striate cortex
A
- V1
- in occipital lobe
- 6 layers
⤷ 2 - 3 = coordinate w/ other visual areas
⤷ 4 = primary receiving area
⤷ 5 - 6 = comms w/ subcortical struc. (LGN, SCN)
6
Q
explain: layer 4 of the striate cortex
A
- sublayers A, B, C
- primary receiving area
⤷ thickest layer - most LGN neurons terminate on 4C leurons
⤷ M axons (layers 1 - 2 of LGN) on upper 4C
⤷ P axons (layers 3 - 6 of LGN) on lower 4C
7
Q
question: does V1 have topographical mapping or cortical magnification?
A
- both
- topo. map
⤷ R of V1 has map of L vis. field - cortical mag
⤷ fovea = magnified compared to peripheral vision
8
Q
explain: cortical magnification in V1
A
- center of vis field = closer to fovea -> more representation
- peripheral = further from fovea -> less cortical space
- 20 degrees of vis field has same cortical space as fovea
⤷ fovea in reality is only 2.5 degrees of vis field
9
Q
explain: consequences of cortical magnification (2)
A
-
acuity declines w/ eccentricity
- eccentricity = how far stim. is from center of gaze
- stim. further away from center -> less acuity bc most cortical space is for the fovea
- solution: move head to foveate -
visual crowding in periphery
- objects that can be easily identified in isolation get jumbled when crowded
- deleterious effect of clutter impairs ability to see objects in the clutter
10
Q
explain: binocularity and ocular dominance of V1
A
- LGN = first place where input from both eyes intermingles
**above and below 4C = binocular but 4C = monocular
**layers 1, 4, 6 = contralat (nasal), 2, 3, 5 = ipsilat (lat)
**LGN layers/cells are each monocular but each LGN is binocular - ocular dominance caused by diff. neurons having preference for L or R eye
⤷ layers 2 - 6 are binocular
⤷ 1 = contralateral
⤷ 7 = ipsilateral
11
Q
question: role of binocular neurons
A
- perceive depth
- increases sensitivity of vis. system
12
Q
explain: orientation selectivity in V1
A
-preference of a V1 neuron for a vis. stim of a specific orientation
⤷ along a certain axis
⤷ ex. bars of light
- firing more = more like preferred orientation
⤷ more light falling on the excitatory region of that neuron
13
Q
explain: directional motion sensitivity in V1
A
- neuron shows greater firing to stim. mvt. in a particular direction
- represented by circle drawn on a cartesian plane
- if all distances same = non-directional neuron bc all same sensitivity
- closer to center = fires less = does not prefer (vv)
14
Q
explain: colour contrast detection of V1
A
- parvo path = red green light
⤷ 3 - 6 - konio path = blue yellow light
⤷ koniocellular = in between each layer of LGN - causes blobs of colour sensitive cells in V1
15
Q
question: simple vs complex cells?
A
- simple = clearly defined excitatory and inhibitory regions
- complex = receptive fields don’t have defined regions
16
Q
question: ocular dominance columns vs orientation columns?
A
- ocular dominance = vertical columns in V1 where neurons arranged in alternating L-preferring or R-preferring manner
- orientation columns = vertical columns where orientation preference changes in a regular manner unal all oritations are represnted
17
Q
explain: columns in V1 (ice cube model)
A
- have ocular dominance vs orientation columns
- hypercolumn = one pair (both L and R eye) of ocular dominance columns including all orientation columns
⤷ not totally correct bc Hubel and Wiesel only measured one neuron at a time
18
Q
explain: new model exp. instead of ice cube columns
A
- shine light onto brain and see a pop. of neurons
⤷ not just a single one - shows active areas
⤷ absorb more light -> appear darker - use detector to see where light is absorbed
⤷ more absorbed = less light to camera = darker is more active
19
Q
question: what is the updated model of colouring in V1?
A
- lines marking borders of orientation columns are in radial pinwheel fashion
20
Q
question: how is colour sensitivity represented in V1?
A
- blobs and interblobs
⤷ magno path = interblob
⤷ parvo path = blob - parvo gets fovea -> cones -> colour
⤷ so blobs mean colour - stained cortex w/ cytochrome oxidase to see blobs
21
Q
explain: new model for hypercolumn
A
- 1mm block of V1 that has all the machinery needed to see everything the V1 is responsible for
22
Q
question: how does orientation preference change after adaptation?
A
ORIGINAL = prefer 0 degrees
- adapt to 20 degree stripes
NEW = prefer -10 degrees
- 20 degrees cells = most tired
⤷ followed by 10 degrees and 0 degrees - so the -10 degrees cells fire the best -> bars look like they bend to the left
23
Q
define: interocular transfer
A
- transfer of adaptation from an adapted eye to a nonadapted one
**input doesn’t converge until V1 so adaptation must happen in cortical neurons (V1 or beyond)
24
Q
explain: spatial freq. channel
A
- pattern analyzer by cortical neurons
- each set of neurons is tuned to limited range of special freq.
⤷ high freq. = detail
⤷ low freq. = broad outlines (large changes in contrast)
25
name: ways to study vision in infants (2)
- preferential looking
⤷ but baby can get distracted + too many other cog. factors to consider
- visually-evoked potentials
⤷ like EEG
⤷ measures sig. from brain when 2 stim. presented
26
question: when does cone system dev.? acuity and contrast sensitivity?
- 2 - 3 mths
- infants can't discriminate colours until then
**acuity and contrast sensitivity dev. much later
27
question: when does contrast sensitivity function develop?
- sensitivity to low spatial freq. dev. earlier
- CSF decreases w/ age
⤷ esp. decreases sensitivity for higher spatial freq.
28
name: examples of abnormal visual experience in infant dev. (3)
1. cataracts
- opacity in lens
- light can't reach eye as much
2. strabismus
- one eye is turned so that the eyes see diff. views
3. anisometropia
- two eyes have very diff. refractive errors
29
question: when is the vision critical period in humans?
- 3 - 8 yrs
30
define: amblyopia
- reduced acuity and lack of binocular depth
- result of abnormal early vis. experience
31
define: extrastriate cortex
- visual areas that lie just outside V1
32
question: what are the what and where pathways?
- what = ventral path = P channels
- where = dorsal path = M channels
**segregation of the streams et more blurred at higher lvls
33
explain: IT cortex
- inferotemporal cortex
- in monkeys
- important for object recognition
- lower part of temporal lobe
- close connections to hippocampus (memory formation)
34
question: do humans have a face-sensitive area like the monkey IT cortex?
- fusiform face area = activated when faces are perceived
⤷ homologous to monkey IT cortex
35
explain: jennifer aniston quiroga 2005 study
- studied cells from humans when looking at diff photos
- some cells responded to specific stim. (jennifer aniston)
⤷ only to her or her name
⤷ also fired when her + other people but not when only other people
- suggested neurons in brain respond to specific stim.
⤷ one cell per stim.
**probably not actually one cell per stim.
⤷ more likely neuron is part of a larger network
36
compare: PPA, FFA, EBA (functional imaging)
PPA = parahippocampal place area
⤷ spaces in the world (places, scenes)
FFA = fusiform face area
⤷ faces
EBA = extrastriate body area
⤷ body structures other than faces
37
explain: feed-forward processes
- carry out computation one neural step after another
⤷ no need feedback from later stages
- gives crude info about objects
**comple recognition needs feedback connections
38
explain: reverse-hierarchy theory
- feed-forward processes gives overall image first
⤷ get details after, when re-entrant feedback does back down vis pathway
- get bigger picture first, before zooming in to details from feedback
39
question: low-lvl, mid-lvl, high-lvl vision?
LOW
- early
- extract basic features of an image
MID
- middle
- combining features into objects
HIGH
- object recognition and scene understanding
40
explain: kanizsa figure
- illustrates illusory contours
⤷ borders even if nothing changes from one side to the other in an image
- happens bc vis. system makes inferences about contours
⤷ best guess
- might be occlusion
⤷ object blocks another
41
explain: structuralism
- way to understand perceptions by looking at the components
- got disproved by illusory contours
- replaced by gestalt school of thought
42
explain: gestalt grouping rules (6)
1. **texture segmentation**
- carving an image into regions of common texture
2. **similarity**
- image chunks that are more similar are more likely to be grouped together
3. **good continuation**
- two elements group if they seem to lie on the same contour
4. **proximity**
- things near each other are more likely to be grouped
5. **parallelism and symmetry**
- parallel or symmetrical contours are likely to be grouped
6. **figure ground**
- object vs background
43
define: camouflage
- object's features group with envrt.
- persuades observer that object doesn't form its own group
