Cognition Chapter 3

Perception

The ability to recognize and interpret information from the senses

Ex: A self-driving car must perceive the road, obstacles and signals

Visual cognition

The body of research on the cognitive processes that contribute to perception

Ex: Completion, constancies, and figure-ground processes

Blind spot

A gap in the photoreceptors where the optic nerve passes through the retina; the mind fills in the missing information

Ex: We don’t notice the gap in our visual field

Modal completion

Perceiving in illusory object when your view is not obstructed and there is no objective boundary with the background

Ex: The light square over four black circles

Amodal completion

Perceiving an object despite an apparently obstructed view; the brain infers the missing information

Ex: A square seen through four circular holes; the partially hidden fire truck; “amodal nudity”

Bottom-up information

Sensory input, such as an image falling on the retina

Ex: light reflected by objects stimulating photoreceptors (feedforward process)

Top-down information

The knowledge and expectations that influence and enhance interpretation of sensory input

Ex: The central character reads as “B” in a row of letters but “13” in a column of numbers

Retina

The light-sensitive part of the eye

Ex: Holds more than 100 million photoreceptors

Photoreceptors

The light-sensitive cells on the retina (rods and cones)

Ex: They convert light into neural signals

Fovea

The center of the retina, with densely packed cones

Ex: Explains why detail decreases toward the visual periphery

Cones

Photoreceptors differentially sensitive to wavelengths corresponding to different colors

Ex: Color vision, concentrated in the fovea

Rods

Extremely light-sensitive photoreceptors that help in very dim light but don’t distinguish colors

Ex: seeing in a dim room

Sensation

The input for perception: the stimulation of the sensory receptors

Ex: Light hitting the retina

Transduction

Translation of physical signals from the environment into neural signals the brain can use

Ex: Light → neural signals sent to primary visual cortex

Primary visual cortex

The area at the back of the brain specialized for rudimentary visual processing, such as orientations and spatial frequencies of light and dark patches

Ex: Hubel and Wiesal’s work on its responses

Feedforward

The hierarchical account of the flow of visual information, moving forward through increasingly sophisticated processing

Ex: Primary visual cortex → later pathways

Reentrant (feedback) connections

Reciprocal connections that let sophisticated brain areas modulate earlier stages of visual processing

Ex: Resolving ambiguity in the coarse feedforward signal; (C_T)

Unconscious inference

Helmholtz’s term for combing bottom-up and top-down information to make the interpretation most likely to reflect reality; rapid, unconscious, fairly automatic

Ex: Assuming light comes from above to see bumps vs. dimples (Fig. 3.4); the blue and black/white and gold dress and shadows

Predictive coding

The brain constantly generates expectations about the world and compares them with bottom-up input

Ex: A blurry object looks like a hair dryer in a bathroom but a drill in a workshop (include pic)

Cognitively penetrable

Perception is changeable by beliefs, knowledge, or motivation

Ex: Backpack wearers overestimate distances; people want-chocolate seems closer

Cognitively impenetrable

Perceptual processing proceeds without influence from high-level cognition such as beliefs, knowledge, or motivation

Ex: The rotating snakes illusion persists despite knowing it’s static

New look

A historical movement, sparked by Bruner & Goodman (1947), claiming that value and need shape perception

Ex: Children estimated coins as larger than identical cardboard discs, especially poor children. Critics cite flawed methods and confusing perception with judgment

Object Segmentation

Visually assigning the elements of a scene to separate objects and backgrounds

Ex: Separating the dog, laptop, mug, and grill; camouflage makes it harder

Figure-ground organization

Determining which side of a boundary is the object (figure) and which is the background

Ex: The Rubin vase. Rules: enclosure, symmetry, convexity, meaningfulness

Occlusion

Our views of objects are often partially blocked by other objects

Ex: The mostly hidden fire truck in Fig. 3.13

Boundary extension

People tend to remember pictures as extending beyond their edges

Ex: Emerges within 1/20th of a second, suggesting it stems from perception

Inverse projection

The 3-D world reaches the as a 2-D image, and the same 2-D image could be produced by many 3-D objects

Ex: Two tables with identical surface dimensions look different (Fig. 3.15B)

Binocular disparity

The difference between the views of the two eyes; larger for close objects

Ex: alternately closing each eye shifts your finger’s position against the background

Binocular depth cues

Cues that require both eyes to be effective

Ex: Binocular disparity

Monocular depth cues

Cues used to construct 3-D understanding even when both eyes get roughly the same view

Ex: Linear perspective, texture gradient, relative size; pavement artists

Linear perspective

Parallel lines appear to converge on a vanishing point as they recede

Ex: The building at the upper left of Paris Street; Rainy Day (the painting)

Texture gradient

Similar-sized textural elements appear smaller and more densely packed with distance

Ex: The street cobblestone in the painting

Object constancy

Recognizing objects despite how different they look on the retina due to orientation

Ex: Rotating a book in your hand

Size constancy

Perceived size stays stable despite radical differences in retinal image size

Ex: Background people in the painting are smaller on the canvas but perceived as the same size; the ants in Fig. 3.18

Color constancy

Factoring illumination differences into colors perception

Ex: Identical paint chips in sun and shade look the same color; Fig. 3.19A.

Lightness constancy

Factoring illumination conditions into the perception of brightness

Ex: Squares 1 and 2 in Fig 3.19B are the same gray; the chess pieces in Fig 3.20

Agnosia

The inability to recognize objects, caused by brain damage; patients aren’t blind and don’t seem to lack representations

Ex: Can’t recognize a cup by sight but can name it by touch

Apperceptive agnosia

Agnosia with impaired early vision; can’t perform simple visual feature tasks or copy images

Ex: Failed copying (Fig. 3.21A)

Associative agnosia

Agnosia with intact early vision; the naming failure arises at a later stage of recognition and categorization

Ex: Can copy drawings but can’t name the object (Fig 3.21B)

View-based approach

Matching images to representations like 2-D pictures or templates

Ex: Bank-check numerals read by template matching

Strengths/Limits: Simple and efficient for standardized stimuli; struggles with natural, noisy images. Predicts slower recognition the more an object is rotate from the learned view.

Template

A representation that fully describes the shape of an object.

Ex: A toy with shaped holes that only matching blocks pass through

Multiple-trace memory model

the view-based idea that we store multiple representations of multiple views of the same object

Ex: Quickly matching a new view of an object to a stored view

Structural descriptions

Models that represent objects as sets of 3-D parts in spatial relationships to each other

Ex: One “recipe” model per object, allowing a match regardless of orientation

Strengths/limits: More tolerant to rotation; the line between the two approaches is hazy and both seem to be used

Recognition by components

Biederman’s model in which an alphabet of about 36 or fewer basic shapes (geons), arranged by a recipe, represents objects

Ex: Visual input is matched against structural descriptions

Geons

The basic shapes in Biederman’s alphabet

Ex: Up to three geons yield about 154 million distinct structures (Fig. 3.23)

Holistic perception

Processing a whole object at once, including the relations of its parts

Ex: Gauthier et al. (2003): faces, and cars for car experts, are processed as wholes

Deep learning

A form of AI that uses deep neural nets to process, categorize, and label natural images

Ex: Photo software that groups or searches images by word

Neural nets

Brain like algorithms that analyze images in multiple steps

Ex: First layer processes the raw image; later layers extract abstract features; the output is a label such as “cat.”

Perception pathway

Allows us to determine what is located where (Goodale & Milner)

Ex: Reporting a slot’s orientation

Action pathway

Uses perceptual information to guide ongoing actions

Ex: Posting a card through a slot despite being unable to report its orientation; frogs lashing as fly like movement

What/where pathways

Separate but integrated pathways: ventral cortex for “what,” dorsal/parietal cortex for “where.”

Ex: Monkeys with inferior temporal damage failed object discrimination; those with posterior parietal damage failed landmark discrimination

Mental imagery

The act of forming a percept in mind without sensory input

Ex: Einstein imagining traveling on a beam of light; visualizing an athletic performance

Aphantasia

An inability to engage in mental imagery

Ex: People report nor being able to generate images in their minds

Mental rotation

Using mental imagery to compare and match rotated images; time is directly proportional to the amount of rotation

Ex: Shepard & Metzler (1971); assembling Ikea shelves

Mental scanning

Mentally moving from one point in an image to another; scan time is proportional to distance

Ex: Kosslyn’s island map: hut-lake vs. hut-tree

Depictive

Kosslyn’s view that the brain represents mental images like real images coming through the eyes

Ex: A color-highlighted route on a map

Propositional

Pylyshyn’s view that mental images are held post-perceptually in an abstract form, like a linguistic description

Ex: verbal directions to a destination

Epiphenomenon

Something that occurs together with a process of interest but isn’t central to its function

Ex: The Wi-Fi icon isn’t needed for the signal itself; Pylyshyn says imagery is similar

Topographic

Neighboring locations in visual space are represented by neighboring neurons in the cortex

Ex: Occipital cortex as a weather map; damage causes location-specific imagery errors

Spatial neglect

Inability to visually attend to objects on one side of the visual field

Ex: Bisiach & Luzzatti patients described only one side of the Piazza del Duomo, but described the other side when imagining standing at the opposite end.

Method of loci

A memory strategy using mental imagery, placing items at locations in an imagined room

Ex: walking through an imagery room, encountering each object at a different spot