Chapter 4 - The Visual Cortex
G8 Ch 4 Outline
Following the Signals from Retina to Cortex
The Visual System
Processing in the Lateral Geniculate Nucleus
Receptive Fields of LGN Neurons
Information Flow in the Lateral Geniculate Nucleus
Organization by Left and Right Eyes
Organization as a Spatial Map
Receptive Fields of Neurons in the Striate Cortex
Do Feature Detectors Play a Role in Perception?
Selective Adaptation and Feature Detectors
Grating Stimuli and the Contrast Threshold
Selective Rearing and Feature Detectors
Maps and Columns in the Striate Cortex
Maps in the Striate Cortex
Columns in the Striate Cortex
Location Columns
Orientation Columns
Ocular Dominance Columns
Hypercolumns
How is an Object Represented in the Striate Cortex
Streams: pathways for What, Where, and How
Streams for Information About What and Where
Streams for Information about What and How
The Behavior of Patient D.F.
The Behavior of People Without Brain Damage
Modularity: Structures for Faces, Places, and Bodies
Face Neurons in the Monkey’s IT Cortex
Areas for Faces, Places, and Bodies in the Human Brain
Something to Consider: How Do Neurons Become Specialized
How Neurons Can Be Shaped by Experience
Striate cortex, area V1: About 250 million neurons (out of the 15-30 billion in the cortex).
Recall: The cortex is a sheet of neurons that covers the rest of the brain.
Side view of the cortex, with some of the possible sources of input
It used to be thought that the cortex was made up of 6 layers of cells. We now know that what used to be thought of as Layer 4 are actually 4A, 4B, 4Cα, and 4Cβ.
The Cortical Map of the Visual Field (From Wolfe, Kluender, & Levi)
Visual Image
Retinal Images
Cortical “Images”
Details of the representation
The cortex is organized as Hypercolumns
Hypercolumn: A 1 mm2 are of cortex receiving input from a small area on the retina. Stimulation of a small area of the retina leads to activity in the hypercolumn representing that area.
It’s called a column because it is collection of columns of cells, containing all 6 layers of the cortex.
It’s called a hypercolumn because it contains multiple individual columns, each one devoted to processing a the visual stimulus in a different way. Hypercolumns are analogous to states – each state has multiple sub-entities – counties, cities, towns – that perform different functions.
Adjacent small areas of the retina are represented by adjacent hypercolumns.
The hypercolumns form a distorted retinotopic map.
The cells within each hypercolumn have specific receptive fields in the retina – the area in the retina whose simulation leads to activity of those cells.
That is, each cell in each hypercolumn is “looking” at the activity in a unique part of the retina. Cells in different hypercolumns look at different parts. Those in adjacent hypercolumns look at adjacent parts of the retina.
Cortical magnificationG8 p. 82-84
Each small area of the retina is represented by a 1 mm2cortical hypercolumn.
Receptive fields in the fovea of the retina are smaller than those in the periphery.
Regardless of the size of the retinal area,each such area is represented by a same-sized 1 mm2cortical area.
This phenomenonis called cortical magnification, illustrated below.
In the figure, each circle in the retina represents a collection of ganglion cell receptive fields. Each circle in the cortex represents a collection of neurons that receive information from the retinal ganglion cells.
Recent evidence reported suggests that the foveal area is actually allocated more cortical neurons than the peripheral area.
Note that cortical magnification means that relatively more hypercolumns are involved in processing information from the foveal region than from other regions.
Types of cortical cells.
Layer 4
Layer 4 cells have circular receptive fields, similar to those of the LGN cells that drive them.
Other layers
In other layers, the neurons have receptive fields that are not simply circular.
Simple cells.
These cortical cells respond only to bars of light or slits of darknesslocated in a specific place in the visual field.
1) located in a particular place in the visual field and
2) have a particular orientation.
If the bar or slit is moved to a different location, the neuron quits firing.
If the orientation of the bar or slit is changed, ditto.
Consider the following. Each circle represents the visual field under a different stimulation.
Possible wiring input to simple cortical cells
Here’s how the simple cortical cell receptive fields might be created.
Schematic of “wiring diagram” of simple cortical cells.
Complex cells
These cells respond to bars of light or slits of darkness, as do simple cells.
But they respond best when the bar or slit moves within a certain area of the visual field.
Many respond best to a particular direction of movement.
The figure below attempts to illustrate the stimulus for a complex cell “looking” for movement of a particularly oriented bar moving from left to right across the visual field.
End-stopped cells (hypercomplex)
These cells respond to moving lines of a specific length (hence the term, end-stopped).
Some also respond to moving corners or angles.
Play VL 4.2 “Visual Cortex of the cat” here – about 20 min.
More on the Hypercolumns G8 p84
As stated earlier, it has been discovered that corresponding to each small area on the retina is an approximately 1 mm2 area in the striate cortex. These areas are called hypercolumns.
Within each hypercolumn are individual-neuroncolumns which receive specific characteristics of the visual stimulus.
Orientation column: A column of cells within a hypercolumn at the same location in each of the six layers, all of which respond to the same orientation of a visual stimulus. (Layer 4 excluded).
Adjacent columns respond to slightly different orientations.
Within each hypercolumn are 1000s of orientation columns covering virtually all possible orientations.
Ocular Dominance columns.
Ocular Dominance column: A column of cells at the same location across the six layers, all of which have the same eye preference.
Some columns respond only to stimulation of left eye
Some more to left eye stimulation than to right
Some equally to stimulation of either eye
Some more to right eye stimulation than to left
Some only to right eye
And all gradations between.
Blob columns.
Blob column: A collection of cells at the same location across the six layers, all of which respond to the same wavelength of light.
Hypercolumn Summary
Columns within each 1 mm2 area are categorized in 3 ways
Orientation preference
Eye preference
Wavelength preference
What’s it all mean?
Cells in the visual cortex respond to complex features – they’re feature detectors.
1) As we’ve said before, edges are probably more important for us than homogenous fields.
So immediate processing of the incoming stream of visual information for edges seems to be a smart thing to do.
2) Extracting features with edges may be the most efficient way of enabling the processing of the visual world as it continually changes around us.
Most of the things in the world that are important for us are defined by combinations of edges.
The Visual Cortex - 111/17/2018