Object Recognition
Call Kai
Learn
Practice Test
Spaced Repetition
Match
Flashcards
Knowt Play1/80
There's no tags or description
Looks like no tags are added yet.
Name | Mastery | Learn | Test | Matching | Spaced | Call with Kai | Chat |
|---|
No analytics yet
Send a link to your students to track their progress
81 Terms
Visual information follows two main pathways after initial processing in V1. What are they?
Dorsal and Ventral pathway
Dorsal pathway
where pathway; occipital to parietal
Ventral pathway
“What” pathway; occipital to temporal
Dorsal “where” pathway neurons
Neurons can have eccentric (non-foveal) receptive fields
Ventral “What” pathway
Neurons with receptive fields that always encompass the fovea
Mishkin et al, Monkey lesion study
What were the monkeys doing
The object/landmark task: selecting a foodwell based on either the identity of an object or the location of a landmark
What are the independent variables?
Task, two conditions
Lesion location, two conditions
What are the two conditions of the task?
Landmark task: choose the foodwell next to the landmark
Object task: choose the object that hides the reward
What are the two conditions of the lesion?
One monkey has a dorsal lesion
Another has a ventral lesion
DV?
accuracy on the two tasks
If the dorsal pathway is necessary for perceiving where something is, how will the monkey with a dorsal lesion perform on each task?
Monkey with dorsal lesion will do well on the object task but do bad on the landmark task
If the ventral pathway is necessary for perceiving what something is, how will the monkey with a ventral lesion perform on each task?
Monkey with ventral lesion will do well on landmark and not on object task
Dorsal path damage results in…. which is….
Optic Ataxia
difficulty using visual information to guide movements accurately, such as reaching for an object.
How does optic ataxia affect reaching?
A person can see an object but has difficulty guiding their hand to its location.
What brain region is typically damaged in optic ataxia?
The posterior parietal cortex, part of the dorsal “where/how” pathway
Why could the patient in the video reach with the left hand but struggle with the right?
Their left parietal damage impaired visually guided reaching with the right hand, on the opposite side of the body.
How to test participants for ventral path damage? what was the first task?
A: Hold a card and match its orientation to the slot without inserting it, judging the slot’s angle.
Q: How did controls and the patient with ventral pathway damage perform on the orientation task?
A: Controls matched the angle accurately; the patient performed poorly, appearing to guess.
Q: What did the second task ask participants to do?
A: Put the card into the slot, using vision to guide the movement.
How did controls and the patient perform on the card-insertion task?
A: Controls performed well, and the patient was relatively accurate despite struggling with the first task.
Why could the patient insert the card but struggle to report its orientation?
A: Ventral pathway damage impaired conscious perception of orientation, while the relatively intact dorsal pathway could still guide the hand during an action.
Q: What does this experiment demonstrate about the two visual pathways?
A: The ventral pathway supports perception, while the dorsal pathway guides visually directed actions.
What are the main stages of object recognition in the “what” pathway?
Early vision → shape encoding → object matching → identifying the object, such as recognizing an apple.
Q: Where does shape encoding occur along the ventral pathway?
Across several visual areas, beginning in occipital visual areas such as V2 and becoming more complex toward the inferior temporal cortex.
Kanwisher et al, Human PET study
what are the participants doing
Participants are viewing stimuli
IV:
conditions?
type of stimulus, 3 conditions
Familiar stimuli
Novel stimuli (made up)
Scrambled stimuli (no object info)
DV?
A: Regional cerebral blood flow measured using PET, an indirect measure of brain activity.
Q: Which conditions should involve feature extraction?
A: All three, because each contains basic visual features, such as edges.
Q: Which conditions should involve shape encoding?
A: Familiar and novel objects, because their parts form coherent whole shapes.
Q: Which condition allows successful matching to a stored representation of a known object?
A: Familiar objects, because participants have encountered them before.
Q: How would you identify a brain region that supports shape encoding?
A: Look for greater activity for familiar and novel objects than for scrambled stimuli: conditions 1 and 2 > condition 3.
Q: Why compare both familiar and novel objects with scrambled stimuli?
A: Both have coherent shapes, but only familiar objects have known identities. This helps distinguish shape encoding from recognition based on memory.
Q: What did the red areas in the PET results indicate?
A: Greater activity for whole shapes than scrambled parts.
Q: Which region supports shape encoding?
A: The lateral occipital cortex (LOC).
Q: What is shape encoding?
A: A perceptual process that combines parts into a representation of a whole shape.
Q: What is object matching?
A: Linking the perceived shape to stored internal representations to identify the object.
Q: What happens when a novel shape is compared with stored representations?
then either it cannot be identified, or is compared to similar internal representations
Q: Can you tell that two shapes match without knowing what they are?
A: Yes. Perceiving matching shapes does not require recognizing their identity.
Q: What is object constancy?
A: Recognizing an object as the same object despite changes in lighting, position, viewing angle, or apparent size.
Q: What is an example of object constancy?
A: Recognizing the same car from different angles or under different lighting.
Q: What are the two key processes in the object recognition pathway?
A: Shape encoding: forming wholes from parts.
Object matching: linking those shapes with stored representations.
Deficit in object recognition
Visual agnosia: a failure of visual recognition
Subtypes of visual agnosia
Apperceptive agnosia
Associative agnosia
Q: What is apperceptive agnosia?
A: Difficulty recognizing objects by sight because of impaired visual shape perception.
Q: What process is impaired in apperceptive agnosia?
A: Shape encoding—combining visual parts into a coherent whole shape.
Q: Why does impaired shape encoding interfere with object recognition?
A: Without an accurate representation of the shape, the brain cannot successfully match it to a stored object representation.
Q: What visual difficulties did patient DF show?
A: Difficulty recognizing shapes and objects by sight and accurately copying drawings of objects she saw.
Q: How could DF identify an object she could not recognize visually?
A: By holding and feeling it, using tactile information.
Q: What does DF’s ability to recognize objects through touch demonstrate?
A: Her deficit was specific to visual processing; knowledge of the objects was still available through another sense.
How do we test for apperceptive agnosia?
Unsual views or shadows test
Q: How does the unusual views test assess object perception?
A: It presents objects from unusual angles and asks whether they are the same object.
Q: How does the shadows test assess object perception?
A: It changes lighting or shadows and asks whether the images show the same object.
Q: What ability do both the unusual views and shadows tests require?
A: Object constancy—recognizing the same object despite changes in its appearance.
Q: How do patients with apperceptive agnosia perform on these tests?
A: They have difficulty recognizing that the images show the same object despite changes in viewpoint or lighting.
Q: Where was DF’s damage associated with impaired shape perception?
A: Bilateral damage to the lateral occipital cortex (LOC), near the occipital–temporal border.
Q: What is associative agnosia?
A: An inability to recognize objects by sight despite intact shape perception.
Q: What can patients with associative agnosia do despite not recognizing an object?
A: Copy its drawing and match its shape, because they can perceive its visual form.
Q: How might a patient respond to seeing an apple?
A: They can draw its shape but cannot identify it as an apple or access its meaning from sight.
Q: Can patients recognize an object through touch?
A: They may recognize and name it when holding it because tactile input accesses object knowledge through another sensory pathway.
How to test for associative agnosia?
Matching by functions task - choose the objects that have the same function (with a distractor shape)
Q: How does the matching-by-function task work?
A: Patients choose objects with the same function, while ignoring a distractor with a similar shape.
Matching a closed umbrella with an open umbrella, rather than a similarly shaped cane.
Q: Why do patients with associative agnosia struggle with matching by function?
A: They perceive the shapes but cannot access the objects’ meaning or function through vision.
Q: Does poor performance on matching by function uniquely identify associative agnosia?
A: No. Patients with apperceptive agnosia also struggle because their shape perception is impaired.
What is impaired?
Object matching ; the ability to associate percepts (visual input) with meaning, that connection is lost
What is damaged?
Anterior temporal cortex/lobe
How are 2 ways objects represented?
Ensemble-coding
grandmother cell coding
Q: What is ensemble coding?
A: An object is represented by a pattern of activity across a group of neurons/voxels , rather than one neuron alone.
Q: How could ensemble coding represent a car?
A: Neurons responding to features such as shape, color, and texture contribute to a combined activity pattern representing the car.
Q: Can the same neuron participate in representing multiple objects?
A: Yes. Different objects can activate overlapping groups of neurons, but with different overall patterns.
Q: What is grandmother cell coding?
A: The idea that one highly specific neuron represents a particular object or person.
Q: What are potential limitations of strict grandmother cell coding?
A: Losing that neuron could eliminate the representation, and representing every new object with a separate neuron would require many neurons.
Q: What was the hypothesis of Haxby et al.’s human fMRI study?
object categories are represented by patterns of activity in VTC
Q: What did participants do in the study?
A: Viewed pictures from different object categories, such as scissors, shoes, chairs, and cats.
Q: What was the IV?
A: Stimulus category—the category of object participants viewed.
Q: What was the DV?
A: The fMRI BOLD signal across voxels in the VTC.
analysis they used for this?
Multivariate pattern analysis - how information is representative in the cortex
Q: How were object categories represented across VTC voxels?
A: Different categories produced distinct, distributed activity patterns across overlapping voxels.
Q: What did the classifier attempt to do?
A: Identify the object category being viewed from its pattern of brain activity.
Q: What would successful classification suggest?
A: The activity patterns contain information that distinguishes object categories.
Q: What were the main results?
A: Object categories could be distinguished with high classification accuracy from VTC activity patterns.
Q: How did the results support ensemble coding?
A: They showed that category information was present in activity patterns across many voxels, supporting distributed representation.