1. Lecture 3 Vision and Agnosias
Humans, like most diurnal species, depend on sense of vision. While other senses such as hearing and touch are essential, visual perception dominates our perceptions and frames the way we think. Our cortex reflects this dominance of vision. For example, over 50% of the cortex in macaque monkeys is devoted to visual perception. However, these estimates were based on stimulation studies. Perhaps auditory information can activate 50% of the cortex as well.
One reason why vision is so important is that it enables us to perceive information at a distance , to engage in what is called remote sensing or exteroceptive perception. Advantages of remote sensing are numerous. An organism can avoid a predator better when it can detect the predator at distance. It is too late to flee once bear teeth pierce your skin.

2. Lecture Outline:
Visual perception from the eye to the primary visual cortex (pp )
Vision beyond primary visual cortex: analyses of movement and colour
Higher perceptual abilities: recognition of objects and disorders of visual recognition – agnosias
Are faces special objects?
Are there two ways of processing visual information?

3. Vision ~50% of cortex devoted to visual perception
Vision is important for primates
~50% of cortex devoted to visual perception
Stimulus in the visual system is light (electromagnetic energy)
We only see a small band of electromagnetic waves

4. The Eye Retina Photoreceptors Bipolar cells Ganglion cells
Rods – low levels of light
Cones – color
Bipolar cells
Ganglion cells

5. Cones

6. Form Eye to the CNS Each eye is divided into two identical halves
Within each eye, one half received stimulation from the left visual field and the other from the right visual field
Optic nerve
Lateral or temporal branch stays on the same side
Medial or nasal branch crosses over
Optic chiasm
Extensive signal processing of visual information is performed within the retina. The output from the photoreceptors is first processed in the bipolar cells and from there to ganglion cells. Humans have estimated 260 million photoreceptors but only about 2 million ganglion cells – the eye’s sole output source. Axons of the ganglion cells form a bundle, the optic nerve. By way of this nerve, visual information is transmitted to the central nervous system.
Before entering the brain, each optic nerve splits into two parts. The temporal, or lateral, branch continues to traverse along the same side. The nasal, or medial, branch crosses over to project to the contralateral side of the brain. This crossover takes place at optic chiasm.
Once inside the brain, each optic nerve divides into pathways that differ with respect to where they terminate within the subcortex.
In the figure you can see the retino-geniculate pathway – this is the projection from retina to the lateral geniculate nucleus of the thalamus. This pathway contains more than 90% of the axons in the optic nerve and provides input to the cortex via the geniculo-cortical projections. The remaining 10% of the fibers innervate other subcortical structures including the pulvinar nucleus of he thalamus and the superior colliculus of the midbrain. However, the fact that these other receiving nuclei are innervated by only 10% of the fibers does not mean these pathways are unimportant. The human optic nerve is so large that 10% of the the optic nerve constitutes more fibers than are found in the entire auditory pathway. The superior colliculus and pulvinar nucleus play a big role in visual attention, and the retino-collicular pathway is sometimes viewed as a more primitive visual system.
Geniculo-cortical pathway consists of axons that exits from the LGN and ascends to the cortex, with almost all of the fibers terminating in the primary visual area of the occipital lobe. Thus, visual information in the cortex has been processed by at least five distinct neurons: photoreceptors, bipolar cells, ganglion cells, LGN cells, and cortical cells.

7. Form Eye to the CNS Lateral geniculate nucleus (LGN) of the thalamus
Primary visual cortex (90%)
Geniculostriate pathway
Superior colliculus (10% - projects back to thalamus and then to cortex) – tectopulvinar pathway

8. Form Eye to the CNS

9. Visual Cortex – Primary Visual Cortex
Different names for primary visual cortex:
Brodmann’s area 17 V1
primary visual cortex
striate cortex (“striped” cortex)
The first cortical synapse for neurons carrying visual information are in the medial portion of the occipital lobe, area 17 in Brodmann’s map, also referred to as V1 area and primary visual cortex. This are is located medially and buried bellow the superficial surface of the cortex along the calcarine sulcus. Cytoachtectually this area is stippled or spotted which is why it is sometimes referred to as the striate cortex.

10. Retinal Topography

11. Retinal Topography and Cortical Blindness
Damage of the primary visual cortex causes blindness
By now you should know that primary visual cortex is the only area of cortex that receives visual info from the eyes through the lateral geniculate nucleus. While lesions of LGN are rare, they do happen and the result is total blindness, just as if the both eyes have been damaged. However, it is not uncommon to come across patients that had a stroke that involve primary visual cortex. These lesions are devastating for visual processing. The patients become blind to stimuli falling within the receptive field of the affected area.

12. Retinal Topography and Cortical Blindness
Hemianopia – loss of pattern vision in either the left or right visual field
Quadrantanopia –blindness in one quadrant of the visual field – damage to the optic tract, LGN or V1

13. Disorders of Visual Pathway

14. Cortical Blindness and Consciousness
Case D.B.
Area around the right calcarine fissure was removed for treatment of angioma
Reported not seeing anything in the left visual field
Able of pointing out where the light was in the left visual field
Blindsight – residual visual abilities within a field defect in the absence of acknowledged awareness
Importance of subcortical visual pathways

15. Visual Areas of the Cortex Outside the Primary Visual Cortex
~30 cortical visual areas with distinct functions
Each visual area has a topographic representation of external space in the contralateral hemifield (however, these get ‘less’ topographic as we get further up in the system)
In this figure we can see a map of the visual areas of the cortex of macaque monkey. Each box in the figure stands for a region of cortex that is purported to be a distinct region of visual processing. More than 30 distinct cortical visual areas have been identified in the monkey. While V1 areas is the initial projection region of lateral geniculate nucleus, you should know that V2, V3 and V4 does not mean that the synapses proceed sequentially from one area to the next. The lines connection these extrastriate visual areas demonstrate extensive convergence and divergence across visual areas. Also, connections between many areas are reciprocal, areas frequently receive input from an area to which they project.
How do we define a visual area? This depends on the criteria that we use. One criterion is that cells within one area respond to visual stimuli;

16. Two General Projections From the Primary Visual Cortex
Dorsal stream – Occipito-parietal stream spatial perception – action- “where” or “how to”
Ventral stream – Occipito-temporal stream object perception – identification – “what” stream
Now that you’ve learned quite a bit of detail about all these areas involved in visual information analysis let’s try to simplify things a bit. We have to do this in order to have a somewhat coherent theory of how things work in the visual system. What became obvious to many researchers early on and what is probably evident to you is that projections from primary visual cortex, area V1, are contained in two fiber bundles. One of them takes a dorsal path towards posterior regions of the parietal lobe. The other projection takes a ventral path towards the temporal lobes. Now the fact that we have two projection paths is meaningless unless there is a something fundamentally different about the way these two paths analyze visual information. It turns out that these are functionally distinct.
The ventral pathway also know as the occipito-temporal pathway is specialized for object perception and recognition, basically for determining what it is we are looking at.
The dorsal or occipito-parietal pathway is specialized for spatial perception, for determining where an object is, and for analyzing the spatial configuration between different objects in a scene.
WHAT AND WHERE ARE THE TWO BASIC QUESTIONS TO BE ANSWERED IN VISUAL PERCEPTION. SO IN ORDER FOR US RESPOND APPROPRIATELLY WE MUST KNOW WHAT SOMETHING IS AND WHERE IT IS.

17. PET studies of “What” and “Where” pathways
Where task: did the objects remain in the same position?
What task: did objects change?
Neuroimaging studies with human subjects have provided evidence that the dorsal and ventral streams are activated differentially by “where” and “what” tasks. In one elegant PET study trials consisted of pairs of displays containing three objects each. In a position task, the subjects had to determine if the objects were presented at the same locations in the two displays. In the object task, they had to determine if the objects remained the same across the two displays. The irrelevant factor could remain the same or change. The objects might change on the position task, even though the locations remained the same; similarly, the objects might be presented at new locations in the object task. Thus, the stimulus displays were identical for the two conditions, with the only difference being the task instructions.
The PET data were compared and indeed, the where task activated portions of the parietal lobe while the what task activated the portions of the temporal lobes.

18. Area MT or V5 MOTION
Cells in area MT respond to movement but not color
For example, this particular neuron in this monkey’s V5 area responds best when stimulus moved down and to the left
Why would it be useful for the primate brain to have evolved so many areas?
One possibility is that the areas form hierarchy in which each area successively elaborates on the representation derived by processing in earlier areas, representing the stimulus in a specific way. The simple cells of the primary visual cortex calculate edges used by more complex cells to detect corners and edge terminations used by higher-order neurons to represent shapes. Successive elaboration culminates in formatting the representation of the stimulus so it matches (or not) information in memory. But as you saw in previous slides there is no simple hierarchy; extensive patterns of convergence and divergence result in multiple pathways.
An alternative hypothesis relates to the idea of visual perception as an analytic process. Although each visual area provides a map of external space, the maps differ with regard to the type of information they represent. For example, neurons in some areas are highly sensitive to color variation. In other areas, the neurons may be movement sensitive but color insensitive. By this hypothesis, neurons within an area not only code where an object is located in visual space but also provide information about object’s attributes. Visual perception is divide-and-conquer strategy. Rather than each visual area representing all attributes of an object, each provides its own analysis. Processing is distributed and specialized. As we advance through the visual system, different areas elaborate on the initial information in V1 and begin to integrate this information across dimensions to form recognizable percepts. There is extensive evidence supporting this hypothesis. If you look at this figure you see that if visual stimulus, a white rectangular bar, is moved down and to the left the cells in area MT will be firing a lot. If the stimulus is moved up and to the right these same cells are virtually silent. Hence, neurons within this area are selective with respect to the visual field and direction of the moving stimuli. In addition, activity of MT cells is correlated with the speed of motion.

19. Imaging Visual Areas in Humans
Semir Zeki – What part of the brain processes movement in the visual field in humans?
PET scans
Experimental condition: black-and-white collage set in motion
Control condition?
Motion – area V5 (MT)
The same logic was used to design the motion experiment. For this experiment, the control stimulus consisted of a complex black-and-white collage of squares. The same stimulus was used in the experimental condition, except that the squares were set in motion. When the subtractions were completed the researchers found that area V5 also know as MT, was activated only in the motion part of the experiment. SO WHAT WE HAVE HERE? WE HAVE A DOUBLE DISSOCIATION, WE ARE SEEING THAT AREA V4 IS ACTIVATED BY COLOR STIMULI AND AREA MT IS NOT ACTIVATED BY COLOR STIMULI. ON THE OTHER HAND WE SEE THAT AREA MT IS ACTIVAED BY STIMULI THAT IS SET IN MOTION WHILE AREA V4 IS NOT ACTIVAED BY THIS SAME STIMULI.

20. Area V4 COLOUR
Semir Zeki – What part of the brain processes colour in the visual field in humans?
PET scans
Experimental condition: multicolor rectangles
Control condition?
Colour – area V4
Single cell recording studies have provided physiologist with a powerful tool to map out the visual areas in the monkey brain and characterize the functional properties of the neurons within these areas. This evidence provided strong evidence that different visual areas are specialized to represent distinct attributes of the visual scenes. Inspired by these earlier findings a number of scientists wanted to use new imaging techniques to learn about human visual areas.
One of the most prominent people in this field is Semir Zeki who used PET to verify that different visual areas are activated when subjects are processing color or motion information. Do you remember what technique did we say is used with PET studies?
Subtractive technique was used. In the COLOR experiment control condition consisted of subjects passively looking at a collage of achromatic rectangles. Various shades of gray were shown to the subjects. The experimenters here expected to activate neural regions with cells that are contrast sensitive. For the experimental condition, the gray patches were replaced by variety of colors. Each color patch was matched in luminance to its corresponding gray patch. Therefore, luminance sensitive neurons should be equally sensitive to both type of stimuli. However, the colored stimulus produced more activity in neural regions sensitive to chromatic information. The experimenters found that area V4 was activated when subtractions were completed.

21. Deficits in Motion Perception: Akinetopsia
Case M.P.
Bilateral damage to teporolateral corticies (MT?).
“When I’m looking at the car first, it seems far away. But then when I want to cross the road, suddenly the car is very near.”
Color discrimination OK
Object recognition OK
Researchers at the Max Planck Institute in Munich reported in 1983 a striking case of a woman who had incurred a selective loss of motion perception. For this woman world was a series of snapshots. When pouring a cup of tea, M.P would see liquid frozen in air, like a glacier.

22. Deficits in Color Perception - Achromatopsia
Congenital colorblindness (dichromats) vs. acquired colorblindness
Usually associated with damage to V4
Colorblind painter – case J.I.
Object recognition OK
Improved acuity
People were “rat-colored”
Dreams?
When we speak of someone who is color blind, we are usually describing a person who has inherited a gene that produces an abnormality in the photoreceptor system. These people have congenital colorblindness. Dichromats, or people with only two photopigments, can be classified as red-green colorblind or blue-yellow colorblind. These occur in about 8% of male population and less than 1% of female population.
A more rare condition is achromatopsia which is colorblindness associated with problems in the central nervous system. Oliver Sacks describes one case of total colorblindness that was a result of a car accident. This rather successful artist contacted Oliver Sacks and told about his condition. One day he was driving back home and he got into an accident that involved another small truck. In the hospital he appeared fine and he was quickly released. As soon as he came home he fell into a deep and long sleep. When he woke up he couldn’t remember anything about the accident, because he’s wife asked him about his car and he said that somebody probably backup into it. This is not so unusual since a lot of accidents that result in some kind of head trauma result in temporary retrograde amnesia. So he got up and went to his studio, but before he got there he was pulled over by a police officer for going through two red lights. He got a ticked and the police officer told him to seek medical help. When he got to his studio he was terrified to see all this paintings that were full of color, gray and dirty. In addition to not being able to see colors he temporarly could not read either. He said that letter’s appeared Greek to him. Over the next little while he grew more and more depressed. Ironically he was a painter and now he was colorblind. But not only could he not paint the way he use to, other things in his life were severely impacted as well. For example, he saw reds color as black, and hence to him tomatoes and tomato juice appeared black. People appeared gray and dirty to him, in fact he described most of the things as gray and dirty. He said that people appeared “rat-colored” to him. Interestingly he lost all his imagery of color too. This is not the case of people who go blind do to eye damage. So even when he closed his eyes if he imagined food or people, they appeared disgusiting to him as they were when he had his eyes open. Two years after the accident it appears that Jonathan I.’s despair has declined because he could not remember colores that well.
Most of the cases of acquired color blindness usually involve other deficits as well. Strokes and brain damage do not respect borders between different visual areas. This is why this case is so interesting. Jonathan I. did appear to have any other visual deficits, in fact his visual acuity actually improved after the accident he said : “Within days ….my vision was that of an eagle – I can see a worm wiggling a block away. The sharpness of focus is incredible”.
Achromatopsia has consistently been associated with lesions that encompass V4 and the anterior region of V4, but the lesions usually extend to neighboring regions of visual cortex.

24. “What” Pathway
Early in the stream, the cells are responsive to simple stimuli
Further up the stream, cells respond to more complex stimuli
Receptive field of a cell is the area of visual space to which the cell is sensitive
Cells in the primary visual cortex have small receptive fields
Cells further down the “What” stream have large receptive fields

25. Deficits Following Damage to the WHAT Pathway
Visual agnosia – partial or total inability to recognize visual stimuli, unexplainable by a defect in elementary sensation or reduced level of alertness or memory
NO GENERAL LOSS OF KNOWLEDGE
Different from other neurological conditions such as Alzheimer’s disease
Tactile agnosia (Asterognosis)
Auditory agnosia

26. Dissociating Deficits in WHAT and WHAT/HOW Pathways
Patient D.F. severe disorder of object recognition following carbon monoxide poisoning that produced damage in the lateral occipital cortex
Dissociating the “what” and “where” system
D.F. had no problem with the “where” pathway

27. Deficits Following Damage to the WHERE/HOW Pathway
Patient V.K. – bilateral hemorrhages in the occipitoparietal regions
Deficits with “how/where” stream but not “what” stream (can recognize objects
Optic ataxia – difficulty in using visual information to guide actions that cannot be ascribed to motor, somatosensory, or visual-field or –acuity deficits.

28. Subtypes of Visual Agnosia
Apperceptive agnosia
Associative agnosia
I should also let you know that visual agnosia is different from visual deficits caused by an impairment in sensory abilities. A patient who is completely blind will be unable to recognize a visually presented object. But a blind patient does not have a problem linking visual information to stored knowledge about the world; rather a blind patient lack the perceptual input needed to activate this internal knowledge. The label visual agnosia is restricted to individuals who demonstrate object recognition problems despite the fact that visual information continues to be registered at the cortical level.
In the current literature, the general distinction is made between apperceptive agnosia and associative agnosia.

29. Apperceptive Agnosia
Apperceptive agnosia – is a fundamental difficulty in forming percept (a mental impression of something perceived by the senses)
cannot recognize, copy, or match objects, however elementary sensory functions appear relatively intact (i.e., patients are not blind).
usually bilateral damage to lateral portions of the occipital lobes (what stream – early deficits in visual perception)
Often associated with carbon monoxide poisoning
The inability to recognize objects can arise from a host of perceptual disorders. In patients with apperceptive agnosia, the perceptual deficits are subtle. In fact, a standard clinical evaluation may fail to reveal any visual problems, and the patient will have to insist that he or she is having difficulties in recognizing objects to receive a more detailed examination. The patient may perform normally on shape discrimination tasks yet make many mistakes when asked to recognize line drawings or photographs of objects. To demonstrate that the agnosia is truly of the apperceptive subtype and not an associative agnosia, it is necessary to devise refined tests of perceptual acuity. For example, tasks like the GOLLIN Picture task and Incomplete letters task examine whether patients can recognize objects in a degraded format.

30. Associative Agnosia
Associative agnosia – basic visual information can be integrated to forma meaningful perceptual whole, yet that particular perceptual whole cannot be linked to stored knowledge
What is affected? “higher cognitive” level of processing that is associated with stored information about objects – that is with memory.
Patients have either lost access to memories of what things should look like or actually lost these memories
Damage to regions in ventral stream that are further up the processing hierarchy, such as the anterior temporal lobe

32. Associative Agnosia - Example
Case F.R.A. – infarct of the left posterior cerebral artery
Copying of objects OK
Can describe objects when they are named
Can segment a complex drawing into parts (apperceptive patients cannot do this)
He could not name these objects
How would test if this is a language problem (finding words for objects)?

33. Associative Agnosias Category Specificity Patient J.B.R.
Herpes simplex encephalitis
Living objects (6% correct)
Inanimate objects (90% correct)
Other patients the other way around
Why is this the case?
Associative agnosia  loss of semantic knowledge
Semantic knowledge has categories

34. Why Is It More Likely That Animals Won’t be Recognized?
Manufactured objects are manipulated
Associated with kinesthetic and motoric representations
Manufactured objects are easier to recognize because they activate additional forms of representations
Individuals can “where” or “how to” pathway to derive knowledge of an object might be

35. Patient C.K.
Using hand movements in order to identify an object

36. Special Category - Faces
Prosopagnosia is the inability to visually recognize familiar faces including their own
Can recognize people by their voice or birthmark or characteristic hairdo
Prosopagnosia patients can be object-agnosia free
Are faces special?
Do the processes for faces and object involve physically distinct mechanism?
Are the systems (object and face recognition) functionally independent?
Double dissociation?
Do the two systems process information differently?

37. One Possibility- Hierarchical Model
If this model is correct, what would you expect regarding the dissociation between object and face recognition
FACE PERCEPTION
OBJECT RECOGNITION
EARLY VISUAL PROCESSING

38. Dissociations of Face and Object Perception
Patient C.K. was presented with Giuseppe Arcimbaldo’s paintings
severe object agnosia but no prosopagnosia
Hierarchical model probably not correct
2 parallel systems (object recognition and face recognition)
Perceives a face
No perception

39. More Evidence that Faces are Processed Separately
Yin, 1970
Inverted faces and inverted houses
More deficits with inverted faces
Patients with prosopagnosia  no inversion effect

41. More Evidence that Faces are Processed Separately
Tanaka & Farah 1993
Face perception not simply analysis of parts
For house perception it did not matter if the house was presented as a whole or in parts
Faces need to be perceived as whole

42. Are Faces Really Special or Are They Similar Examplars of the Same Category?
McNeil and Warrington, 1993
Farmer with prosopagnosia
Tested on human faces – failed
Tested on sheep faces – OK
Is prosopagnosia due to the fact that faces are similar members of the same category?

43. Are Faces Really Special or Are They Similar Examplars of the Same Category?
Bornstein, 1963
Following brain damage, avid bird watcher could no longer distinguish between different bird species
Sergent & Signoret, 1992
Toy car expert (5000 in his collection)
Following brain damage he became prosopagnosic
He was able to identify different cars

44. Expertise Inversion Effect for Other Objects
College students and judges of show dogs presented with:
Faces  upright, inverted
Dogs  upright, inverted
What would predict about the results?

45. Neural Mechanisms for Face Perception
Martha Farah (1990) – looked at 71 prosopagnosic patients
Bilateral lesions – 65%
29% right hemisphere lesions
6% left hemisphere lesions lesions
Conclusions?

46. Neural Mechanisms for Face Perception in Humans
Human fMRI studies
Face stimuli associated with activation of the right fusiform gyrus
Fusiform face area – FFA
FFA is also activated by other stimuli

47. Face Perception and Face Recognition
Overall the evidence suggests that the right occipital and ventral temporal regions are important for face perception
Posterior regions are important for perceptual representation necessary to create a configural representation of a face
Anterior regions are involved in linking a particular facial representation with biographical data

48. Activation of the FFA by Non-Facial Stimuli
Individuals were trained to be experts at recognizing “greebles”
In “greebles” experts, FFA is activated during identification

49. Two Systems for Object Recognition?
Alexia – reading problems as a result of brain damage
Alexia left angular gyrus
Prosopagnosia right FFA
Right hemisphere
Left hemisphere
High incidence of alexia and object agnosia
High incidence of prosopagnosia and object agnosia
No patient with prosopagnosia and alexia without object agnosia
Object recognition by two routes

50. Right Hemisphere  Holistic Left Hemisphere  Analytic
Extreme cases
Patients with either left- or right sided strokes were presented with stimuli and later asked to recall this stimuli
Patients with left-sided lesions show intact global features but no detail
Patients with right-sided lesions produce detail only

51. Implicit Recognition of Faces
In some instances individuals with prosopagnosia can recognize faces (implicitly……SCRs and priming)