Primary Visual Cortex

Lecture 5. Tues, 9/15

Jonathan Pillow Perception (PSY 323), Fall 2009 The University of Texas at Austin

Figure 3.6 Sine wave gratings illustrating low (a), medium (b), and high (c) spatial frequencies

More Channels: Spatial Frequency Channels

Visual Acuity: Spatial frequency: The number of cycles of a grating per unit of visual angle (usually specified in degrees) Another way to think of spatial frequency: # of times a pattern repeats per unit length

Visual Acuity: Why sine gratings? Patterns of stripes with fuzzy boundaries are quite common The edge of any object produces a single stripe, often blurred by a shadow, in the retinal image The visual system breaks down images into a vast number of components; each is a sine wave grating with a particular spatial frequency

The contrast sensitivity function Human contrast sensitivity illustration of this sensitivity

Retinal Ganglion Cells and Stripes The response (right) of a ganglion cell to gratings of different frequencies

Retinal Ganglion Cells and Stripes Not only is the spatial frequency important, but so is the phase

Phase: The phase of a grating refers to its position

within a receptive field

Figure 3.11 The primate lateral geniculate nucleus

The Lateral Geniculate Nucleus We have two lateral geniculate nuclei (LGNs): This is where axons of retinal ganglion cells synapse Ipsilateral: Referring to the same side of the body (or brain) Contralateral: Referring to the opposite side of the body (or brain)

Primary Visual Cortex Striate cortex: Also known as primary visual cortex, or V1 Primary visual cortex = first place in cortex where visual information is processed
(Previous two stages: retina and LGN are pre-cortical)

Primary Visual Cortex Major transformation of visual information takes place in V1

Circular receptive fields

found in retina and LGN are replaced with elongated stripe receptive fields in cortex cells!

It has about 200 million (vs. 1 million Retinal

Ganglion Cells)

Primary Visual Cortex Two Erroneous Ideas about What V1 Does

(or more generously, incomplete)

1. Edge detection

(extracting just the edges)

2. Fourier Analysis (breaking things apart by spatial frequency)

Edge detection hypothesis for V1:

The role of V1 is to detect edges not obvious that this is a good way to process images led to many frustrated projects in computer vision

Fourier Decomposition theory of V1

The role of V1 is to do Fourier decomposition,

i.e., break images down into sine waves

Fourier Decomposition theory of V1

Summation of two

spatial sine waves idea that each spatial frequency comprises a channel

Fourier Decomposition of a 2D image (into high and low spatial frequencies).

Original

High Frequencies

Low Frequencies

Nice idea, elegant mathematical theory more correct than edge detection hypothesis but incorrect--neurons respond to broad range of frequencies
(more on this in Chap 4!)

Primary Visual Cortex Two important features of striate cortex:

Topographical mapping:
! Nearby parts of visual space are represented at nearby anatomical locations in V1

Cortical magnification:
! Dramatic scaling of representation from different parts of visual field ! The amount of cortex devoted to processing the fovea is proportionally much more than the amount of cortex devoted to processing the periphery

Figure 3.14 The mapping of objects in space onto the visual cortex

contralateral
representation of visual eld (visual space represented in opposite hemisphere)

cortical
magnication (a slight misnomer) highly unequal representation of fovea vs. periphery

Striate Cortex

Visual acuity declines in an orderly fashion with eccentricitydistance from the fovea

Receptive Fields in Striate Cortex

Orientation tuning: neurons in V1 respond more to bars of certain orientations and less to others Response rate falls off with angular difference of bar from preferred orientation

Figure 3.16 Orientation tuning function of a cortical cell

Receptive Fields in Striate Cortex Cells in striate cortex respond best to bars of light rather than to spots of light simple cells: prefer bars of light, or prefer bars of dark complex cells: respond to both bars of light and dark

[Hubel & Weisel movie]

Receptive Fields in V1 How are the circular receptive fields in the LGN transformed into the elongated receptive fields in striate cortex?

Figure 3.17 Hubel and Wiesels model of how cortical simple cells get their orientation tuning

Receptive Fields in V1: how do you build them?

Hubel and Wiesel: Very simple scheme to

accomplish this transformation ! A cortical neuron that responds to oriented bars of light might receive input from several retinal ganglion cells ! If you string several retinal ganglion cells together, they can form an oriented bar ! A cell that is tuned to any orientation you want could be created in cortex by connecting it up with the appropriate retinal ganglion cells

Receptive Fields in Striate Cortex Many cortical cells respond especially well to: Moving lines Bars Edges Gratings Direction of motion

Receptive Fields in Striate Cortex Each LGN cell responds to one eye or the other, never to both Each striate cortex cell can respond to input from both eyes By the time information gets to primary visual cortex, inputs from both eyes have been combined Cortical neurons tend to have a preferred eye, however. They tend respond more vigorously to input from one eye or the other

Figure 3.20 A simple cell and a complex cell might both be tuned to the same orientation and stripe width, but might respond differently

Receptive Fields in Striate Cortex

End stopping: Some cells prefer bars of light of a certain length

Summary

V1 receptive elds orientation tuning spatial frequency sensitivity & tuning Fourier analysis Hubel & Weisel experiments simple vs. complex cells cortical magnication contralateral representation of visual eld