COGSCI 1B Final Test Questions with Correct Answers 100% Pass
brain matter dedicated to vision - Answers 30-40%
early models of visual perception - Answers -only used top-down analysis
-big picture (upsidedown triangle)
-no neural implementation (came later in the 1980's)
2 visual systems hypothesis (2 pathways) - Answers dorsal stream and ventral stream
-experiments on monkeys where diff parts of brain were selectively removed
dorsal stream - Answers regions in visual cortex involved in perception of spatial location
-striate cortex to posterior parietal cortex
-where pathway
ventral stream - Answers regions in visual cortex involved in perception of form
-striate cortex to inferior temporal corted
-what pathway
tri-level hypothesis of information processing - Answers -David Marr
-3 levels
Levels of tri-level hypothesis - Answers computational level: highest most abstract level. what does
problem entail? what output do we want? purpose?
algorithmic level: programming level, steps? procedure?
implementation: lowest, where is hardware being used? physical realization?
aim of visual processing - Answers object recognition
Marr's model of visual processing - Answers - based on tri-level hypothesis
- system has to take a complex pattern of unstructrued stimuli and interpret them into representations
that will help object recognition
Steps of Marr's model of visual processing - Answers 1. image projected onto retina analyzed in terms of
light/dark, edges, and contours -- PRIMAL SKETCH OF IMAGE
2. features in raw primal sketch that are similar in size and orientation get grouped -- 2D sketch, viewer
centered (problem is we need to establish object constancy)
,3. 3D sketch -- symmetry axis, elongation axis: line that shows direction along which main bulk/mass of
shape is distributed. OBJECT CENTERED and solves object constancy problem
information in visual cortex is processed... - Answers heirarchically (like neural networks)
-visual system is also similar to neural networks because it is retinotopically organized like convolutional
layers in convolutional NN
human visual process step-by-step - Answers input --> retina --> optic nerve --> optic chiasm --> superior
colliculus of brianstem --> lateral geniculate nucleus of thalamus --> projects to area v1 in visual cortex --
> maps to striate cortex
Area V1 - Answers - retinotopically organized (neighboring regions of visual field are represented
accordingly in V1)
-neurons in V1 are sensitive to low levels of features (orientation and direction of movement)
-edges and contours
-simple cells that respond only to presence of line segments
neurons in V1 - Answers feature detectors
V1 projects to V2 - Answers -process same as V1 and more complex
-compelx edges, shape, depth
-extrastriate cortex: process movement, spatial frequency, color, retinal disparity
Depth is determined by... - Answers Retinal disparity: points on object located at different distances
from observer will fall on different locations of the 2 retinas (thumb moving)
-closer object = larger disparity
--> stereopsis or depth preception
V2 to ventral pathway - Answers V2 --> V4
-inferior temporal cortex: facial recognition area (fusiform face area)
-fusiform body areas
V2 to dorsal pathway - Answers V2 --> MT --> posterior parietal cortex
trichromatic (3 color) theory - Answers retina has 3 types of color receptors (cones) for red, blue, and
green
color blind - Answers lack of functioning red/green sensitive cones (see yellow and blue and red and
green look yellowish to them)
, opponent process theory - Answers -opponent process cells
-color sensitive receptor cells respond in an opposing center-surround manner to primary colors
-excited by light in the center, inhibited by light in the surroind
-green on red off, red on green off, yellow on blue off, blue on yellow off
blindsight - Answers blind in one or both visual fields but can guess identity or location of objects,
emotions of face in picture
-blindsight patient navigated cluttered hallway after being told that it was empty
explanation for blindsight - Answers instincts
- there is a second pathway for visual perception that doesn't go through the visual cortex instead it
does a short loop through the limbic system to the superior colliculus (emotional/instinctual parts of the
brain)
-faster response time but slows down cognitive reactions because other parts of the brain are not being
used
mental imagery - Answers represented either in analog code (pictures, visualizers, MOST SUPPORTED) or
propositional code (descriptive, verbalizers)
imagery and rotation studies - Answers rotate each object to the left and see if it matches with the one
on the right
-length of time to process --> depends on extent of rotation and quantity of information
-supports that we use pictorial form instead of description (unlike computers)
participants asked to focus on point of an image and then asked questions about other parts - Answers
length of time = longer based on distance from focal point
evidence supporting analog code - Answers -people make faster judgements about larger mental images
-longer to travel larger mental distance
-visual imagery may interfere with visual perception
-visual imagery activates 70-90% of same areas of brain as visual perception
evidence supporting propositional code - Answers -ambigious figures are hard to make a mental picture
of
-difficulty understanding that a part belongs to a whole if didn't include part in their original mental
description
brain matter dedicated to vision - Answers 30-40%
early models of visual perception - Answers -only used top-down analysis
-big picture (upsidedown triangle)
-no neural implementation (came later in the 1980's)
2 visual systems hypothesis (2 pathways) - Answers dorsal stream and ventral stream
-experiments on monkeys where diff parts of brain were selectively removed
dorsal stream - Answers regions in visual cortex involved in perception of spatial location
-striate cortex to posterior parietal cortex
-where pathway
ventral stream - Answers regions in visual cortex involved in perception of form
-striate cortex to inferior temporal corted
-what pathway
tri-level hypothesis of information processing - Answers -David Marr
-3 levels
Levels of tri-level hypothesis - Answers computational level: highest most abstract level. what does
problem entail? what output do we want? purpose?
algorithmic level: programming level, steps? procedure?
implementation: lowest, where is hardware being used? physical realization?
aim of visual processing - Answers object recognition
Marr's model of visual processing - Answers - based on tri-level hypothesis
- system has to take a complex pattern of unstructrued stimuli and interpret them into representations
that will help object recognition
Steps of Marr's model of visual processing - Answers 1. image projected onto retina analyzed in terms of
light/dark, edges, and contours -- PRIMAL SKETCH OF IMAGE
2. features in raw primal sketch that are similar in size and orientation get grouped -- 2D sketch, viewer
centered (problem is we need to establish object constancy)
,3. 3D sketch -- symmetry axis, elongation axis: line that shows direction along which main bulk/mass of
shape is distributed. OBJECT CENTERED and solves object constancy problem
information in visual cortex is processed... - Answers heirarchically (like neural networks)
-visual system is also similar to neural networks because it is retinotopically organized like convolutional
layers in convolutional NN
human visual process step-by-step - Answers input --> retina --> optic nerve --> optic chiasm --> superior
colliculus of brianstem --> lateral geniculate nucleus of thalamus --> projects to area v1 in visual cortex --
> maps to striate cortex
Area V1 - Answers - retinotopically organized (neighboring regions of visual field are represented
accordingly in V1)
-neurons in V1 are sensitive to low levels of features (orientation and direction of movement)
-edges and contours
-simple cells that respond only to presence of line segments
neurons in V1 - Answers feature detectors
V1 projects to V2 - Answers -process same as V1 and more complex
-compelx edges, shape, depth
-extrastriate cortex: process movement, spatial frequency, color, retinal disparity
Depth is determined by... - Answers Retinal disparity: points on object located at different distances
from observer will fall on different locations of the 2 retinas (thumb moving)
-closer object = larger disparity
--> stereopsis or depth preception
V2 to ventral pathway - Answers V2 --> V4
-inferior temporal cortex: facial recognition area (fusiform face area)
-fusiform body areas
V2 to dorsal pathway - Answers V2 --> MT --> posterior parietal cortex
trichromatic (3 color) theory - Answers retina has 3 types of color receptors (cones) for red, blue, and
green
color blind - Answers lack of functioning red/green sensitive cones (see yellow and blue and red and
green look yellowish to them)
, opponent process theory - Answers -opponent process cells
-color sensitive receptor cells respond in an opposing center-surround manner to primary colors
-excited by light in the center, inhibited by light in the surroind
-green on red off, red on green off, yellow on blue off, blue on yellow off
blindsight - Answers blind in one or both visual fields but can guess identity or location of objects,
emotions of face in picture
-blindsight patient navigated cluttered hallway after being told that it was empty
explanation for blindsight - Answers instincts
- there is a second pathway for visual perception that doesn't go through the visual cortex instead it
does a short loop through the limbic system to the superior colliculus (emotional/instinctual parts of the
brain)
-faster response time but slows down cognitive reactions because other parts of the brain are not being
used
mental imagery - Answers represented either in analog code (pictures, visualizers, MOST SUPPORTED) or
propositional code (descriptive, verbalizers)
imagery and rotation studies - Answers rotate each object to the left and see if it matches with the one
on the right
-length of time to process --> depends on extent of rotation and quantity of information
-supports that we use pictorial form instead of description (unlike computers)
participants asked to focus on point of an image and then asked questions about other parts - Answers
length of time = longer based on distance from focal point
evidence supporting analog code - Answers -people make faster judgements about larger mental images
-longer to travel larger mental distance
-visual imagery may interfere with visual perception
-visual imagery activates 70-90% of same areas of brain as visual perception
evidence supporting propositional code - Answers -ambigious figures are hard to make a mental picture
of
-difficulty understanding that a part belongs to a whole if didn't include part in their original mental
description
