The model

1
The model
good
2
(No Transcript)
3
Cortical Circuitry
Feedforward intracortical connections V1 (II/III)
to V2 (IV)
Feedback intracortical and subcortical
connections V2 (VI) to V1 (VI) V1 (V) to SC, V1
(VI) to LGN, V2 (V) to SC, Pons, Striatum
4
(No Transcript)
5
(No Transcript)
6
Histology of Cerebral Cortex 2

• Pyramidal neurons are large and complex
• Similar orientation
• Process input from many sources

7
Stuff and things

• Retina, lateral geniculate nucleus and primary
  visual cortex produce rich information about
  local points,
• but properties of more global objects are not
  represented (stuff versus things).

8
Color and motion

• Need to determine what goes together to represent
  a thing. How does it move? What is its true
  color? This takes place in extrastriate cortex.
  Farah begins consideration of global (rather than
  local) image perception with color and motion.

9
(No Transcript)
10
(No Transcript)
11
Color and motion areas
12
Color

• Color perception begins with wavelength detection
  in the retina. Color contrast is enhanced by
  center-surround receptive fields in retina and
  LGN. Double opponent effects occur in blob cells
  of Layers 2 and 3 in visual cortex and
  color-selective responses continue in the thin
  cytochrome oxidase stripes of V2 and are
  projected to V4. Up to V4, responses are
  wavelength-selective rather than color selective.
  To be color selective means context can influence
  color response.

13
Border of V1/V2 with blobs and stripes
14
Color constancy

• Our perception of the color of an object is based
  on the light reflected back to us from that
  object. That light depends on both the spectral
  reflectance (true color) AND the spectral
  composition of the incident light (light that
  bathes the object). We perceive color accurately
  because we can take the color of the incident
  light into account. This ability is called color
  constancy. The larger context helps us here.

15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
Mondrian colors

• Land (who invented the Land camera or Polaroid)
  showed that Mondrian colors are perceived
  accurately if the whole thing is bathed in the
  same light but if one light is used on a patch
  and another light on the next patch then we cant
  see color accurately.

19

• In the rosy light of dawn, for instance, a yellow
  lemon will reflect more long-wave light and
  therefore might appear orange but its
  surrounding leaves also reflect more long-wave
  light. The brain compares the two and cancels out
  the increases.

20
V4 had color constancy

• Zeki, recording from visual pathway neurons in
  monkeys, showed that while areas up to V2 only
  see light from a patch and cant compensate for
  the context light overall, neurons in V4 are able
  to compensate for the incident light and
  accurately report the color!

21
V4 receptive fields support color constancy

• These neurons have large receptive fields, even
  extending into the other visual field.
• The surrounds are inhibitory and sensitive to the
  same wave length as the center thus if the same
  wavelength is everywhere, it is discounted or
  dismissed (psychologically speaking) as ambient
  light.
• This is consistent with evidence that a split
  brain patient has trouble with color constancy
  for stimuli that cross the midline since he/she
  cant know what the overall light is on both
  sides at once.

22
Cerebral achromatopsia

• Lost of color vision or color blindness when
  acuity, motion, depth perception and object
  recognition are good is called achromatopsia.
• In some cases the other visual functions are
  transiently affected too or pattern recognition
  may also be a problem. Cases involving artists
  are particularly dramatic.
• Unilateral lesions can create loss in only one
  hemifield hemiachromotopsia.

23
(No Transcript)
24
(No Transcript)
25
"... as soon as he entered, he found his entire
studio, which was hung with brilliantly colored
paintings, now utterly grey and void of color.
His canvases, the abstract color paintings he was
known for, were now greyish or black and white.
His paintings--once rich with associations,
feelings, meanings--now looked unfamiliar and
meaningless to him. At this point the magnitude
of his loss overwhelmed him. "He had spent his
entire life as a painter now even his art was
without meaning, and he could no longer imagine
how to go on. Oliver Sacks, The Case of the
Colorblind Painter, 1995      In An
Anthropologist On Mars, p.6
26

• Achromatopsia is produced by lesions on the
  inferior surface of temporo-occipital regions,
  (lingual and fusiform gyri).

27
Other color related disorders

• color anomia problem in producing the names of
  colors
• color agnosia, loss of knowledge about colors
  (hard to define, cant learn paired associates
  where one word is a color and the other is a name
  or number)
• impaired color-object association, cant tell
  the typical colors of things.

28
Color anomia

• Color anomia, damage to the temporal segment of
  the left lingual gyrus, prevents you from
  naming colors even though you can match and
  discriminate between colors non-verbally. So,
  there seems to be a very specific locus where
  color perception gets coded as color language,
  but the perception can go on normally even when
  the ability to encode the perception as language
  is destroyed

29
Neuroimaging and color

• Positron emission tomography (PET) and
  event-related potentials (ERPs) are consistent
  with lesions. PET showed activation in
  lingual/fusiform region when colored Mondrians
  rather than gray scale Mondrians were presented.
• When attending to color, rather than multiple
  stimuli, activation was seen in collateral sulcus
  (between lingual and fusiform gyri) and also
  dorsolateral occipital cortex (new area).

30
(No Transcript)
31
PET Imaging
Upper row Control PET scans (resting while
looking at static fixation point is subtracted
from looking at a flickering checkerboard
stimulus positioned 5.5 from fixation
point). Middle row Subtraction produces a
somewhat different image for each of 5
subjects. Bottom row The 5 images are averaged
to eliminate noise, producing the image at the
bottom.
32
PET Identification of Inferior Occipital Region
Activated by Color
Activation produced by staring at colored
stimuli. Panel A shows the blood flow images
before subtraction. Panel B show activation after
subtracting responses to the gray stimuli. Panel
C depicts statistical significance of the
responses. White is highest significance. Panel D
shows the location of the most significant
responses in a sagittal, coronal, and axial view
(Courtesy of Frackowiak and Zeki).
Multicolor abstract display (top) and version of
the same display in shades of gray (bottom) used
as stimuli
33
(No Transcript)
34
Event related potentials
ERPs are a non-invasive method of measuring brain
activity. Weak electrical fields, representing
the activity of neural populations within the
brain, can be detected at the scalp using
electrodes connected to an amplifier. The
amplifier enhances the electrical signal so that
it can be reliably recorded. This signal is
time-locked to an event, such as the presentation
of a stimulus (like a word or picture) or
production of response (like a button press),
then averaged to reveal changes in brain activity
specifically associated with different aspects of
cognitive processing.  The temporal precision
of ERPs is superior to all other currently
available neuroimaging techniques.
35
(No Transcript)
36
(No Transcript)
37

• ERPs were examined using checkerboard patterns in
  color after adapting to same color or different
  color. There is a different ERP response to
  different colors in lingual and fusiform gyri and
  in dorsolateral occipital cortex. Also using
  electrical stimulation in those same places could
  alter color perceptions.

38
Control issues

• Another study found the classic area
  (lingual/fusiform area) and also widespread
  activation in other regions, but may not have
  been well controlled. They used color random
  noise patterns, judging if mostly red, versus
  black and white random noise, judging if mostly
  white. One versus 4 colors one versus 2 for
  black and white also stimuli were not matched
  for luminance.

39
A good fit for V4 as the color area, but some
cautions

• In monkey and man color areas are different in
  location, monkey V4 is high up on lateral
  surface. In man color center is on the inferior
  surface and more medial. But that may be ok.
• V4 also inputs to object or form areas of
  temporal lobe and cells in V4 have some response
  to form.
• Finally, monkeys with V4 lesions can relearn
  color discrimination tasks, but have permanent
  problems with form discrimination. Same form
  tasks ok in human achromatopsics. Need to do
  imaging studies with monkeys.

40
V4 Color Caveat

• V4 has been suggested to be the color center by
  Zeki. It is more involved with perception of
  colors than other areas, but it may not be both
  necessary and sufficient for color perception or
  have no other role than color perception

41
And please let Mom, Dad, Rex, Ginger, Tucker, me
and all the rest of the family see color.
42
Motion perception

• Local measurement of motion is also ambiguous.
  You need to look at global indices (multiple
  local pieces of information). In early areas,
  most neurons respond better to moving than
  stationary stimuli (habituation).
• M cells are optimized for movement and provide
  input to first direction selective cells (striate
  visual cortex, 4B). They project to the middle
  temporal area (MT) and thick cytochrome oxidase
  stripes of V2 that then project to MT.

43
(No Transcript)
44
From local to globalINTEGRATION
Small receptive fields (input) LOCAL information

• (output) GLOBAL perception of motion

45
Aperture problem

• With local view cant tell which way edge is
  really moving when it passes through the local
  field, as several patterns of motion look the
  same to a small window.
• Need to combine information. Plaid patterns have
  been used to test for cells that can extract
  global, rather than local, motion.

46
The motion center MT

• V1 cells respond to local or component motion of
  plaid while MT cells can also respond to pattern
  or global motion. MT cells have larger receptive
  fields. MT projects to the medial superior
  temporal area (MST), which has cells with even
  larger receptive fields and more complex motion
  detectors like flow field properties of shinking,
  enlarging, rotation and translation.

47
(No Transcript)
48
Compare psychophysics and cells

• Discriminate dot motion when varying proportions
  of dots are moving consistently and others
  randomly. More consistent dots, the easier the
  discrimination.
• Monkey performance and performance of direction
  selective cells in MT was more or less the same.
  If task is at threshold for making the correct
  discrimination, when the cells are correct, the
  monkey is correct.

49
(No Transcript)
50
(No Transcript)
51
Necessary and sufficient

• If you are working with random dots and stimulate
  a column of cells for a particular direction then
  the monkey will judge that that is the direction
  the dots are moving. Thus MT activity causes
  motion perception.
• Newsome lesioned the MT with ibotenic acid.
  Monkeys were impaired in motion discrimination in
  the contralateral hemifield, but not for color,
  acuity or depth. (Recovery issues?)

52
(No Transcript)