Visual perception and action pt2 Flashcards
Whats perceiving?
involves combining information from the senses with what we already know
whats the constructivist approach to perception?
constructivist as sensory information is incomplete. Perception must be constructed from this incomplete input.
•Top-down processes “fill the gaps”
whats gregory’s (1972) theory of indirect perception?
Top-Down: The image arriving on the retina is incomplete and needs to be “worked on” and supplemented by stored knowledge, past experience, etc.
•Recognition Driven: The end point (or purpose) of perception is to identify what we are looking at
key principles of this theory
The senses provide raw data, but stored knowledge enables us to construct a perception of the input
- The key principles:
- Perception is active
- Sense input is insufficient for perception – need interaction between sensory input and knowledge
- Much of the information that reaches the eye is lost by the time it reaches the brain
- Perception involves hypothesis testing to make sense of information from our sense organs
- Incorrect hypotheses will lead to perceptual errors
Depth cues - evidence supporting constructivism.
what are depth cues?
Depth Cues: Evidence in Support of Constructivism
- Depth cues: often taken as evidence for top-down processing in perception
- Critical because the perceptual system needs to transform the 2D retinal image into a 3D interpretation of the external world
3 main categories of depth cues?
Monocular cues (can be seen with one eye)
- Oculomotor cues (need both eyes: based on muscle movements)
- Binocular cues (need both eyes
evidence supporting constructivism.
whats monocular cues?
Give the perceptual system the impression of 3D
•Seen with one or two eyes
•AKA pictorial cues as artists use them to create the perception of depth in paintings etc
Monocular Cues: Linear Perspective- what is it?
The cue is so powerful that it can give rise to visual illusions (e.g., the Ponzo illusion)
.•Our perceptual system expects the upper line to be further away in 3D space due to the convergence of the line of pillar
Monocular Cues: Aerial Perspective - what is it?
Light is scattered as it travels through the atmosphere, making more distant objects look hazy
•Used in paintings to create the perception of depth
•Top-down knowledge about the world convince us the image is 3D
Monocular Cues: Interposition - what is it?
Where a nearer object hides a more distant one
•The fact that one object is in front of another gives a powerful sense of depth
•These objects are on the same plane
•Overlap = depth
Monocular Cues: Shading
Provides good evidence of the presence of 3D (2D images don’t cast shadows)
•The perceptual system makes inferences about the structure of the image based on top-down knowledge of where the light source (typically) is
Monocular Cues: Familiar/Relative Size
If we know the size of an object, we can use this to judge how far away it is
•Achieved by comparing the size of the image on the retina with long-term knowledge (top-down control) of how big objects are
Monocular Cues: Familiar/Relative Size
Ittleson 1951
used three sizes of playing cards: normal size, half size and double size
- Participants had to judge the distance of the card
- Distance judgments were based on familiar size
- The normal size card was judged to be 2.28 meters away.
The half size card was judged to be twice as far as this and the double size card half as far away
LimitationsL hypothesis testing
Hypothesis testing:
•How are hypotheses generated, and how to we know when to stop and accept one as correct?
•Why does knowledge sometimes help and at other times hinder perception?
•Put another way: how can we ‘know’ something is wrong, and yet still perceive it incorrectly?
Limitations - the hollow face illusuon
Some familiar stimuli cause a strong bias towards accepting an incorrect hypothesis
Although the face viewed from the back is hollow, it still appears perceptually as a normal face
Therefore, the perceptual hypothesis conflicts with what you know
The Perception-Action Model
- who proposed it?
Milner and Goodale (2008)
TPA model - whats vision for perception stream?
Vision-for-perception system: based on the ventral stream. Used to decide whether an object is a cat, or a buffalo or when admiring a landscape. Role – to identify objects. Akin to Gregory’s ideas
TPA model whats visual for action stream?
based on the dorsal stream and used for visually guided action (i.e., interactions with the environment). Akin to Gibson’s idea
what did Milner and Goodale say about the streams?
originally implied the two streams were independent of each other.
- They have now accepted the two streams often interact.
- Example: The dorsal stream “seems to carry visual and semantic complexity, thereby allowing us to bring meaning to our actions” (Milner, 2017, p 1305
The Perception-Action Model: Evidence from Neuropsychology – Optic Ataxia
- Results from damage to the posterior parietal cortex (forming part of the dorsal stream).
- Patients appear impaired at precise visually guided movements although their vision and ability to move their arms are okay.
- Perenin & Vighetto (1988) found patients with optic ataxia had trouble rotating their hands appropriately when reaching towards a largely oriented slot
Critique of optic ataxia evidence?
Not all optic ataxia patients conform to this pattern
•Oversimplification: e.g., Pisella et al. (2006) found evidence for much less impaired visually guided movements in central compared to peripheral visio
Evidence - visual agnosia
James, Cullham, Humphrey, Milner & Goodale (2003)
•Visual agnosia patient DF
•Damage to the ventral pathway
•Impaired at object recognition
•Unimpaired visually guided movements (grasping for objects)
critique of visual agnosia?
Himmelback et al. (1994) reanalyzed data with DF obtained by Goodale et al. (1994) and found that DF had more difficulties with visually guided action than first thought.
evidence from visual illusions
One explanation for this illusion is its reliance on depth cues
•“Angles-out” corresponds to an object that is further away
•“Angles-in” corresponds to an object that is closer
Milner & Goodale (1995; 2006) suggest most studies of visual illusions have tapped the vision for recognition system
- However, the vision for action system is the one that provides accurate information about our position in space:
- Predicts: the perception for action system should be unaffected by these illusions
evidence from visual illusions - Bruno et al 2008
Bruno et al. (2008):
conducted a meta analysis of Muller-Lyer studies where observers pointed rapidly at one figure (using the vision for action system)
.•The mean illusion effect was 5.5%.
•In contrast, the mean illusion effect was 22.4% when observers provided verbal estimations of length (using the vision for perception system
critique of bruno
The model seems to predict that there should have been no illusion effect at all with pointing.
development to TPA model
We now know more about interactions between the ventral and dorsal streams. •The ventral stream is often involved in visually guided action.
- The dorsal stream is involved in visual object recognition.
- There are more than two visual processing streams:
(e.g., the dorso-dorsal stream used for grasping; the ventro-dorsal stream, that makes use of memorized object knowledge
strengths of TPA
It’s been highly influential.
•The central assumption about visual system is partly correct
.•It is a neat way to reconcile previously competing theories
weaknesses of tpa?
•Evidence from brain damage patients is weak.“The double dissociation between optic ataxia and visual agnosia is no longer tenable” (Rossetti et al., 2017, p.130).
Findings based on illusions provide only partial support.
The model exaggerates the independence of the two streams (though has recently been improved in this regard