7 Vision

The objectives for this lecture were to understand the key anatomy of the eye, to describe the two visual pathways of perception, and to explain the perceptual outcomes associated with neurological damage.

Light, and what different animals see

A beam of light can be separated into its wavelengths. Visible light is a narrow band of the electromagnetic spectrum, which runs from gamma rays (wavelengths around nm) through X-rays, ultraviolet, visible light and infrared to radar, FM, TV, AM and AC circuits (around nm). The light visible to the human eye runs from violet (the slide’s scale starts at 350 nm) through green (500 nm) and yellow (600 nm) to red (700 nm). Other animals see beyond this band: many insects and birds can see ultraviolet, and some species of snakes can see infrared.

Dogs see the spectrum differently from people. To a person the spectrum runs from red (700 nm) through orange, yellow, green (500 nm) and blue to violet (400 nm). To a dog the long wavelengths look dark brown shading into pale yellow, the middle of the spectrum (around 500 nm) looks greyish white, and the short wavelengths look light to dark blue. The slide’s photographs show a puppy among red flowers and some dog toys, each as a person and as a dog would see them: in the dog’s version the reds and greens turn into muddy yellows and browns, while blues and yellows stay distinct.

Cats have:

• poor distance vision, about 20/100 to 20/200 (a cat must be 20 feet away to see what a person with normal vision sees from 100 to 200 feet);
• a wider visual field, 200° as opposed to 180° in humans;
• great night vision;
• less colour vision than humans.

The slide shows street and city scenes in pairs, as a person sees them and as a cat would: the cat’s versions are wider, blurrier at a distance, much less colourful, and much brighter at night.

The eye

Light from an object (a pen, in the slide’s drawing) enters the eye through the cornea and the pupil, the opening in the iris. The lens focuses it; the ciliary muscle controls the lens. The light crosses the vitreous humor that fills the eyeball and lands on the retina, the layer at the back of the eye that contains the rods and cones. The rays cross inside the eye, so the image on the retina is upside down. The fovea is the point of central focus; it lies within the macula. The optic nerve leaves the back of the eye at the blind spot.

The retina up close

The retina is built in layers. The rod and cone receptors (each with an outer and an inner segment) sit at the back. In front of them are the horizontal cells, the bipolar cells, the amacrine cells, and the ganglion cells, whose axons run across the retina’s surface and leave the eye as the optic nerve. Light enters from the ganglion-cell side, so it passes through the other layers before it reaches the receptors. A stained micrograph on the slide shows the same layers in real tissue: receptors, bipolar cells, ganglion cells, and the axons from the ganglion cells.

Rods and cones

There are two types of visual receptors.

	Rods	Cones
What they perceive	black and white (they process light, not colours)	colour
Detail	poor at resolving details	detect details on objects
Where they are	mainly in the periphery of the retina	mainly in the fovea

The slide’s electron micrograph shows rods as tall cylinders and cones as shorter, tapered cells.

Counting the cone types gives four kinds of photoreceptors: rods (black and white) and three types of cones, sensitive to

• long-wavelength light (red cones),
• middle-wavelength light (green cones),
• short-wavelength light (blue cones).

Each receptor responds to a broad range of wavelengths, with peak sensitivity at 437 nm for blue cones, 498 nm for rods, 533 nm for green cones and 564 nm for red cones.

The slide shows a column of balloons three ways. With normal vision they are red, orange, yellow, green and pink against a blue sky; with red–green colour blindness the reds and greens all look yellow, olive or grey; with blue–yellow colour blindness the scene turns to pinks, reds and teals.

Past the retina

From the retina, the optic nerve carries signals to the optic chiasm, where some fibres from each eye cross to the other side of the brain. They continue to the lateral geniculate nucleus (LGN), a region of the thalamus, and from there to the primary visual cortex (V1). The slide’s diagram also shows a branch to the superior colliculus.

V1 is at the back of the brain, in the occipital lobe (the slide’s drawing labels the frontal, parietal, occipital and temporal lobes, with the primary visual cortex at the back of the occipital lobe). V1 does two things:

• it gives simple responses to visual stimuli: lines, orientations and movements;
• it routes information to different brain regions for further processing of colour, shape and recognition.

Two visual pathways: “what” and “where”

The initial theory (1982) proposed two pathways leaving the occipital lobe (the primary visual receiving area):

• a dorsal pathway up into the parietal lobe, the “where” pathway;
• a ventral pathway down into the temporal lobe, the “what” pathway.

The evidence came from monkeys with parts of the brain removed, tested on two tasks.

The slides show a side view of the monkey brain with the visual areas V1, V2, V4, MT and STP marked, along with the inferior temporal cortex (IT), which is the area removed for the object task, and the parietal area (P), removed for the landmark task. So the object task lined up with the ventral “what” pathway and the landmark task with the dorsal “where” pathway. But this theory was not fully correct.

“What” and “how”: Milner and Goodale (1995)

Milner and Goodale (1995) drew a more nuanced distinction between the pathways. They agreed that the ventral stream is the “what” stream. But they argued that the dorsal stream is a “how” stream, not a “where” stream: it is specialized in location and in how to direct actions. Their account was based on a patient, D.F., who had a brain lesion to the ventral pathway.

Dissociations

Behavioural functions may be present in some cases and absent in others. The same movement can come from different contexts. Raising your arm or hand, for example, can mean showing how many, showing that you know the answer or want to speak in a large audience, or waving at someone. These are products of different neural channels, so after a brain lesion you may be able to elicit the response in one context but not in others. (Another example from the slide: touching your nose.)

Case study: D.F.

D.F. had a visual form of agnosia (an inability to recognize what one perceives). She could describe the basic forms and colours of objects but was unable to identify or categorize them. She could identify objects by touch, but not by vision.

The slide’s figure (from Milner and Goodale, 1995) shows model drawings of an apple and an open book, D.F.’s attempts to copy them, and her drawings of the same objects from memory. Her copies are unrecognizable fragments of lines and dots, while her drawings from memory are recognizable as an apple and a book. Yet she could not recognize what her own drawings were.

Matching versus posting

D.F. was tested with a slot in a disc on the face of a box, in two tasks: in static orientation matching she had to judge the slot’s orientation and match it; in active “posting” she had to put a card through the slot, like mailing a letter.

Each line in the results shows the orientation of one attempt, with the slot’s true orientation marked at the top. In matching, D.F.’s attempts point in every direction, while a person without brain damage matches the slot almost exactly. In posting, D.F.’s attempts cluster around the correct orientation, nearly as tightly as the control’s.

So D.F. could not perceive and match “static” orientation models, but she could produce the right orientation when imitating an actual gesture such as mailing a letter: she did better when action was involved and the task was put in context. This suggests two mechanisms:

• one mechanism for judging orientation: the ventral pathway, “what” (damaged in D.F.);
• one mechanism for coordinating vision and action: the dorsal pathway, “how” (intact in D.F.).

The currently accepted theory

In the currently accepted theory, the ventral pathway into the temporal lobe is the “what” pathway (the object discrimination task), and the dorsal pathway into the parietal lobe is the “how” pathway (the posting task).

When recognition or action fails

Visual agnosia

The slide shows (a) an apperceptive patient’s copies of letters and numbers (A, B, X, O, 3, 4) and of filled shapes (a circle, a square, a diamond, a T-shape, a key), all of them unrecognizable scribbles; and (b) an associative patient’s copies of a tea bag, a ring and a pen, all accurate.

Visual ataxia

Visual ataxia impairs visuospatial skills, which obstructs visually guided movement, and it impairs attention. Together these make up Balint’s syndrome. The slide’s photograph shows a man reaching for a spoon held up in front of him.

Prosopagnosia

Prosopagnosia (“facial blindness”) is an inability to recognize faces. It comes from damage to the fusiform face area, which lies in the fusiform gyrus on the underside of the temporal lobe, along the ventral pathway. It commonly co-occurs with apperceptive visual agnosia. The slide’s photograph shows a family with their faces blurred to blank skin, suggesting what faces can look like to someone with prosopagnosia.

Check yourself