Question 1

Which anatomical pathway sends information through the following regions:
V1 -> V2 -> V4 -> IT?

Accepted Answer

The ventral/"what" pathway

Question 2

Which is NOT characteristic of invariances that are typically observed in neurons in IT cortex?

Accepted Answer

A) Invariance to an object's location in the visual field 
B) Invariance to the size of the object as projected on the retina 
C) Invariance to the identity of the object 
D) Invariance to changes in illumination of the visual scene C

Question 3

Name two properties of neurons in IT that make them particularly well-suited for involvement in object recognition.

Accepted Answer

In the context of Information Technology (IT), when we refer to "neurons," we are typically talking about artificial neurons, which are the basic units of a neural network in the field of artificial intelligence (AI) and machine learning. These artificial neurons are inspired by biological neurons found in the human brain. In the domain of object recognition, two properties of neurons (or more accurately, artificial neurons) that make them particularly well-suited for this task are:

1. **Learning Capability**: Artificial neurons have the ability to learn from data through a process called training. During training, the weights of the connections between neurons are adjusted based on the input they receive and the error in their output. This learning capability allows neural networks to improve their performance in recognizing objects as they process more data. They can identify patterns, features, and characteristics in the input data that are relevant for distinguishing different objects.

2. **Non-linear Transformation**: Neurons in a neural network can apply non-linear transformations to their inputs. This is crucial because the real world is inherently non-linear and complex, and the ability to capture non-linear relationships between features is essential for tasks like object recognition. Non-linear activation functions, such as the sigmoid, hyperbolic tangent, or rectified linear unit (ReLU), allow neural networks to create complex mappings from input data to the desired output. This enables them to recognize objects even under variations in position, scale, lighting, or other forms of distortion.

Together, these properties enable neural networks to perform object recognition with high accuracy, making them a cornerstone of modern AI applications in image and video analysis, autonomous vehicles, security systems, and many other areas of IT.

Question 4

Why are large receptive fields beneficial to object recognition?

Accepted Answer

The answer of Why are large receptive fields beneficial to...

Question 5

Describe the theory of a hierarchy of stimulus representation in the visual system.

Accepted Answer

The answer of Describe the theory of a hierarchy of...

Question 6

What is the binding problem?

Accepted Answer

The answer of What is the binding problem?...

Question 7

Between the inferotemporal cortex and posterior parietal cortex, which is more likely to process information about what an object is?

Accepted Answer

The answer of Between the inferotemporal cortex and posterior parietal...

Question 8

In terms of object categorization, your pet cat is a token.

Accepted Answer

A) True 
 B)False

Question 9

What is the "grandmother cell theory," and why is it problematic?

Accepted Answer

The answer of What is the "grandmother cell theory," and...

Question 10

If an individual had a difficult time following directions on a map, but could point to the locations on his own body corresponding to the analogous part of a mannequin, his impairment would be with an _____ frame of reference.

Accepted Answer

The answer of If an individual had a difficult time...

Question 11

What is the difference between an egocentric and an allocentric frame of reference?

Accepted Answer

The answer of What is the difference between an egocentric...

Question 12

Visual illusions provide evidence that the visual system must include what during processing?

Accepted Answer

The answer of Visual illusions provide evidence that the visual...

Question 13

Name two side effects of lesioning the magnocellular pathway.

Accepted Answer

The answer of Name two side effects of lesioning the...

Question 14

What is a double dissociation of function? How has it been important in cognitive neuroscience?

Accepted Answer

The answer of What is a double dissociation of function?...

Question 15

Describe the difference between a type and a token. Give an example of each.

Accepted Answer

The answer of Describe the difference between a type and...

Question 16

Describe an alternative, distributed view to higher-level object recognition that does not invoke specific feature detectors for individual people, such as your grandmother.

Accepted Answer

The answer of Describe an alternative, distributed view to higher-level...

Question 17

What is the STP and why is it given that name?

Accepted Answer

The answer of What is the STP and why is...

Question 18

In vision, the _____ pathway is also referred to as the "where" pathway and the _____ is also referred to as the "what" pathway.

Accepted Answer

A) Dorsal-stream; ventral-stream 
B) Ventral-stream; dorsal-stream

Question 19

What is the "aperture problem" in early visual areas? How might it be solved by feedback?

Accepted Answer

The answer of What is the "aperture problem" in early...

Exam 5: The Visual System

Which anatomical pathway sends information through the following regions: V1 -> V2 -> V4 -> IT?