Visual Perception: Brain Processing

• Perception involves actively constructing a visual representation of the world, not just a literal reproduction of light patterns on the eyes
• The brain tends to perceive objects even when they are not necessarily present
• Inferences are made during visual perception, going beyond raw information
• Perception is the way sensory information is interpreted, organized, and consciously experienced
• Bottom-up processing is where perceptions are built from sensory input
• Top-down processing is where interpretation of sensations is influenced by existing knowledge, experiences, and thoughts

Factors Affecting Perception

• Sensory adaptation occurs when one stop perceiving stimuli that remain relatively constant over time
• Motivation, beliefs, values, prejudices, and expectations can all affect perception
• The process of concentrating on certain features in the environment, thoughts, or activities is attention, often excluding other information
• Inattentional blindness is when one fails to notice something completely visible due to lack of attention
• Change blindness is where one fails to notice changes in their environment
• The Sapir-Whorf Hypothesis is where the structure of a language influences how its speakers perceive and think about the world

From Eye to Brain

• Visual perception begins with processing light energy entering the eye
• Light energy enters the eye through the cornea
• Light passes through the pupil and is refracted by the lens
• It is then focused on the retina, at the back of the eye
• Photoreceptors in the retina, especially in the fovea, capture the light
• The light is converted into a neural signal
• The neural signal is transmitted to the brain via the optic nerve

The Retina

• The Retina is the internal surface of the eyes and contains photoreceptors that transduce light into neural signals
• Rods are specialized for low levels of light intensity
• Cones are specialized for high levels of light intensity and detecting different wavelengths
• Cones are packed more densely in the fovea, while rods are more in the periphery
• Neural signals are relayed to the brain through the optic nerves
• The blind spot is the point where the optic nerve leaves the eye and has no rods or cones

Primary Visual Cortex (V1)

• The optic nerve merges at the optic chiasm, just below the cerebral cortex
• The main visual pathway travels to the primary visual cortex (V1/striate cortex) via the lateral geniculate nucleus (LGN)
• The primary visual cortex is in the back of the brain and is responsible for initial processing of visual signals and visual consciousness
• V1 processes light/dark, color, edges, movement, and depth
• Primary Visual Cortex cells respond to light in particular orientations and lengths
• The primary visual cortex is retinotopic, mapping visual input from the retina to neurons

Beyond V1

• There are two distinct visual processing pathways, one is the Dorsal ("where") pathway, the other is the Ventral ("what") pathway
• Connections between visual areas are bi-directional
• Feedforward processing involves visual information moving from the primary visual cortex to the "what" and "where" pathways
• Feedback processing involves regions at the ends of the ventral and dorsal pathways sending information back to the primary visual cortex
• Jumbled Word effect is where the ability to read words even when some of the letters are mixed up
• Word superiority effect is where it is easier to identify a letter if it appears in a word than if it appears alone or in a non-word
• Visual processing is highly modular, meaning it uses the "Divide and conquer" methodology
• Some visual deficits are caused by damage to specific areas of the brain
• V4 is a region of the extrastriate cortex associated with color perception
• Achromatopsia is the failure to perceive color, causing the world to appear in grayscale
• V5 (or MT) is a region of the extrastriate cortex associated with motion perception
• Akinetopsia is the failure to perceive visual motion

"What” vs. “Where"/"How"

• Goodale et al. (1991) reported on patient DF with visual agnosia
• DF was unable to identify objects visually due to damage to the ventral stream
• DF could still act on objects because of an intact dorsal stream
• Jakobson et al. (1991) reported on patient VK with optic ataxia
• VK was unable to properly change grip aperture due to damage to the dorsal visual pathway
• VK could still visually identify objects because of her intact ventral stream
• Optic ataxia means spared “what,” but impaired “where”/“how”
• Parietal lobe damage and dorsal stream cause poor visual guidance of reaching
• Different types of visual information are processed in more specialized brain regions
• It is not understood how these streams of information merge, but it requires attention

Object Recognition

• The initial stage involves detection of basic visual elements like edges and bars
• The intermediate stage Groups basic elements into higher-order units that code depth cues and segregate surfaces into figure and ground
• The advanced/final stage involves recognizing the object and assigning meaning to it
• The detection of patterns occurs based on their features or properties
• Selfridge's (1959) Pandemonium Model is a pattern recognition model with three levels
• The first level are the Features, consisting of properties such as size, color, shape, etc. of a stimuli
• The second level are Cognitive Demons, who decide whether the stimulus matches its particular pattern
• Decision Demon is the top-level, and decides which pattern is being recognized based on input from the Cognitive Demons
• It is easier to make a list of basic features for letters or words, and more complicated for 3D objects.
• Recognition By Components is decomposing objects into their fundamental 3D geometric shapes (geons) and comparing these shapes to existing memory representations
• Recognition by components can account for how we recognize objects from different viewpoints
• One can make predictions about which viewpoints are difficult to recognize objects from
• Perception is holistic rather than atomistic and that the grouping of visual features to form a whole follows certain organizational principles
• Template Matching is where one compares the configuration of the current sensory input with a memory (prototype or template)

Visual Agnosia

• Apperceptive agnosia is the failure to understand the meaning of objects at the level of object perception
• Associative agnosia is the failure to understand the meaning of objects at the level of semantic memory
• Integrative agnosia is failure to integrate parts into wholes in visual perception

Face Recognition

• Face recognition is a within-category discrimination, unlike other object recognition
• Faces require mechanisms all their own because of the importance from a social/evolutionary perspective
• Cognitive Model (Bruce and Young) is:
  • Face recognition units (FRUs) are stored knowledge of the 3D structure of familiar faces
  • Person identity nodes (PINs) are abstract descriptions of people that links together perceptual knowledge (e.g., faces) with semantic knowledge
• The fusiform face area (FFA) is in the inferior temporal lobes and responds more to faces than other visual objects, also implicating processing facial identity
• Prosopagnosia is an impairment of face processing where an individual does not recognize faces

Speech Perception

• Sound can be composed of waves
• Frequency of the wave is perceived as pitch
• Amplitude of the wave is perceived as loudness
• In written text, each letter is a discrete symbol, and each word is separated from its neighbors by spaces
• The speech stream doesn't consist of discrete phonemes and clear word boundaries: a continuous flow of ever-changing frequencies and amplitudes that smears into each other
• The speech sounds smear into each other, as shown by coarticulation, that is the process of overlapping phonemes in the speech stream
• It takes the tongue, jaws, and lips time to move from one position to another, and that is why there is coarticulation
• Coarticulation also increases the rate of transmission for speech
• There is variability between and within speakers
• Speech sounds smear into each other due to a lack of invariance
  • no reliable relationship between a phoneme and the acoustic signal
• Speech perception infers intended phonemes and word boundaries based of multiple cues within the speech stream
• Speech perception requires the auditory system to:
  • Record sound vibrations of someone talking
  • Translate vibrations into a sequence of sounds you perceive to be speech
  • Distinguish the sound pattern of one word from irrelevant words
  • Separate voice of speaker from background noise and other conversations
• Context and speech perception involve top-down processing
• Phonemic restoration is the process of filling in missing segments of the speech stream with contextually appropriate material

Multimodal Perception

• The modality appropriateness hypothesis posits that depending on the circumstances, different sensory modalities will dominate at different times given how the different senses are better at processing some stimuli than others.
• Visual perception involves the visual system dominates the other senses when it comes to perceptual processing, as shown by the visual prepotency effect and Conflict Resolution
  • During Conflict Resolution, individuals rely more on visual information to resolve the conflict (in situations where visual and auditory information conflict)
• The McGurk Effect involves:
  • Audio recording a person speaking, and video recording the same person speaking but mouthing another sounds
    • In this example, audio recorded ‘/ba/’ while video recorded ‘/ga/’
  • When one only listens to the audio they will perceive ‘/ba/'
  • If one only watches the video, one will perceive ‘/ga/'
  • When one listens to the audio and watches the video at the same time, one will perceive ‘/da/’
    • ‘/da/’ is a speech sound halfway between ‘/ba/’ and ‘/ga/’