Steps of The Visual Pathway

1. The Optic Nerve & Optic Chiasm

• Information from the eyes converge for processing in the brain
• Optic nerve meet at optic chiasm –> nasal fibres cross over, temporal fibres stay on same side
• allows partial crossover of visual information so each hemisphere receives input about the same visual field from both eyes

Optic nerve:

• Made up of bundled axons of retinal ganglion cells
• Carries visual signals to the brain
• Blind spot where optic nerve leaves eyes

Optic Chiasm:

• Nasal nerves cross over
• Crossing over = decussation
• Important for binocular vision
• Allows integration of vision from both eyes for depth & perception

2. Lateral Geniculate Nucleus

• Visual signals travel from optic chiasm to LGN in the thalamus
• LGN organises & refines visual information before it reaches the cortex

Structure of the LGN:

• One in both brain hemispheres, deep in the thalamus
• 6 layers (each has map of visual field)
• LGN receives info from optic tract –> processes it –> sends to occipital love
• Gatekeeper for sorting visual info before conscious awareness
• LGN cells have centre-surround antagonism
• LGN cells are monocular (receive input from only one eye)

The 6 layers of the LGN:

There are two LGNs: one on the left and one on the right side of the brain. Both eyes send information to both LGNs because each eye sees both sides of the visual field. Each LGN processes the opposite visual field:

• Left LGN → right visual field
• Right LGN → left visual field

Within each LGN, information from the two eyes is kept separate in 6 layers:

• Layers 1, 4, 6 → contralateral (opposite-side) eye
• Layers 2, 3, 5 → ipsilateral (same-side) eye

So, an LGN receives information from both eyes, but only about one side of the visual world.

Why does each layer of the LGN organise input into only one eye per later?

• Keeps signals separate for precise mapping

Retinotopic Mapping

• The LGN keeps a map of where things are in visual space.
• Neighbouring areas of the retina are represented by neighbouring cells in the LGN.
• This preserves the spatial layout of what we see.

Relay & Modulation:

• LGN receives feedback from cortex to adjust information based on context

3. Optic Radiations & The Primary Visual Cortex

1. Signals leave the LGN in optic radiations, travelling through brain white matter
2. Optic radiations carry visual information to occipital lobe
3. Signals reach primary visual cortex (V1)
4. V1 maintains retinotopic organisation during processing

Inside The LGN

3 Cell Types in the LGN

Feature	Magnocellular (M) cells	Parvocellular (P) cells	Koniocellular (K) cells
Size	Large	Small	Intermediate
Axons	Thick	Thin	Intermediate
Receptive fields	Large	Small	Intermediate/varied
Main job	Motion & rapid changes	Fine detail & colour	Colour processing
Colour	Little/no colour detail	🔴🟢 Red–green	🔵🟡 Blue–yellow
Speed	⚡ Fast	Slower	Intermediate/varied
Spatial detail	Low/coarse	High/fine	Less specialised for fine detail
Good for	Detecting movement and sudden events	Reading, recognising fine details	Blue–yellow colour information
LGN location	Layers 1–2	Layers 3–6	Thin layers between M & P layers
Easy memory	M = Motion	P = Precise	K = Colour pathway (especially blue–yellow)

Parallel Processing

          VISUAL INPUT
               ↓
      ┌────────┼────────┐
      ↓        ↓        ↓
      M        P        K
    motion   detail   colour
      ↓        ↓        ↓
    processed AT THE SAME TIME

Retinotopic Mapping & Cortical Magnification in V1

Remember: Neighbouring regions on the retina correspsond to neighbouring regions in V1

Importance of Retinotopic Mapping

Spatial organisation in V1:

• Preserves spatial relationships on. the retina
• Keeps the positioning of things in your visual field accurate
• Helps with perception of shapes & patterns

Cortical Magnification:

• V1 has more cortical area for fovea
• More emphasis on fine detail & central vision
• Important for reading & facial recognition

Functional Significance:

• Fine details
• Higher level visual processing

V1 Cells as Visual Filters

What do V1 Cells do?

• Neural filters of visual information
• Specialised cells that detect specific visual information to break down visual information into smaller parts of higher level processing

How do they filter?

• Orientation (angles)
• Size (Spatial frequency)
• Colour (wavelengths)
• Low pass = only respond to lower range
• Band pass = only respond within a certain range
• High pass = only respond when higher than threshold

Orientation Selectivity

• Many V1 neurons are tuned to a particular angle/orientation.
• E.g. one neuron might prefer vertical │, another horizontal —, another diagonal /.
• A neuron’s preferred orientation produces its strongest firing.
• As the angle moves away from the preferred orientation → firing decreases.

Orientation Tuning

• Tuning curve = graph showing how strongly a neuron responds to different orientations.
• Peak of curve = neuron’s preferred orientation.

Tuning	Meaning
Sharp tuning	Responds to a narrow range of angles → very precise/selective
Broad tuning	Responds to a wide range of angles → less selective
Bandwidth	How wide the range of orientations the neuron responds to is

Why does this matter?

• Orientation-selective cells help detect edges and boundaries.
• Edges are combined to help identify shapes and objects.
• This contributes to things like reading, recognising faces and navigating.

V1 neurons are feature detectors → some prefer certain orientations → preferred orientation causes strongest firing → many neurons’ responses combine to represent the edges/shapes in a scene.

Spatial Frequency

• Spatial Frequency = how often light & dark stripes alternate across an area of the visual field
• Measured in cycles per degree of visual angle
• High spatial frequency = thin, close lines
• Low spatial frequency = wide, further spaced lines
• High spatial frequency needed for sharp vision & fine details

V1 Modular Organization : The Ice Cube Model

V1 Column Organisation

Orientation Columns

• V1 neurons are organised into columns based on their preferred orientation.
• Cells in the same column prefer a similar angle/edge.
• Different columns respond to different angles: *│ / — *
• This allows V1 to detect different edge orientations.

Ocular Dominance Columns

• V1 is also organised according to which eye the input mainly comes from.
• Some columns respond more strongly to the left eye, others to the right eye.
• Helps compare information from both eyes → important for binocular vision and depth perception.

Hypercolumns / Ice Cube Model

• A hypercolumn = a little processing pack for one small area of the visual field.
• Each hypercolumn contains:
  • Cells for all different orientations
  • Cells representing both eyes
• Different areas of the visual field have their own hypercolumns.
• The ice cube model shows how orientation and ocular-dominance columns are organised together.

Concept	Simple meaning
Orientation column	Cells for a particular angle
Ocular dominance column	Cells organised by eye
Hypercolumn	One complete little pack containing cells for all angles + both eyes
Ice cube model	Model showing how these columns fit together

Simple, Complex, Hypercomplex Cells

What is hierarchical integration?

The process of building complex information by combining simple detectors

Earlier visual inputs
↓
Simple cells
detect basic oriented edges at particular locations
↓
Complex cells
combine simple-cell information across locations
↓
Hypercomplex cells
combine more information to detect endpoints/boundaries
↓
increasingly complex representation of shapes

Types of Cells

Simple Cells

• Detect bars/edges.
• Selective for a particular orientation AND location.
• Think: “This angle is HERE.”

Complex Cells

• Detect a particular orientation across different locations in their receptive field.
• Pool information from multiple simple cells.
• Show spatial invariance → exact location matters less.
• Useful for detecting moving edges and continuous contours.
• Think: “This angle is somewhere around HERE.”

Hypercomplex / End-Stopped Cells

• Sensitive to line length and endpoints.
• Help detect line endings, corners and boundaries.
• Response can decrease if a line extends too far.
• Think: “This edge ENDS here.”

Cell	What does it detect?	Easy thought
Simple	Orientation at a specific location	“│ is HERE.”
Complex	Orientation across different locations	“There’s a │ somewhere around HERE.”
Hypercomplex / end-stopped	Ends, lengths, corners and boundaries	“That │ ENDS here.”

Illusory Contours

• We can perceive edges that aren’t physically present.
• Edge/endpoint information can be integrated to fill in missing contours.
• Involves activity across multiple neurons (population coding) and feedback from higher visual areas such as V2.
• Helps recognise objects that are partially hidden or incomplete
• Example: Kanizsa triangle

Adaption, Aftereffects & Population Coding

key terms

Visual adaptation	Aftereffect	Example of aftereffect
A change in neural responsiveness after prolonged exposure to a stimulus, leading to altered perception.	Visual adaptation Aftereffect Example of aftereffect A change in neural responsiveness after prolonged exposure to a stimulus, leading to altered perception.	Seeing vertical lines as tilted after staring at slanted lines for a while.

Population Coding

• Population coding = the brain represents a visual feature using the combined activity of many neurons, rather than one neuron.
• Each neuron has a preferred value (e.g. a particular orientation).
• Neurons also respond less strongly to similar values because their tuning curves overlap.
• The brain uses the overall pattern of firing to determine what you are seeing.

Example: Orientation

If you see a 10° tilted line, you don’t need a neuron specifically tuned to exactly 10°.

Different neurons might respond:

Neuron preference	Response to 10°
0°	xxx
10°	xxxx
20°	xxx
30°	x

The brain combines this activity to estimate:

“The line is about 10°.”

Why Population Coding Is Useful

• Allows perception of intermediate values → we can perceive angles/sizes that don’t exactly match a neuron’s preferred value.
• Allows a continuous range of perception rather than needing a separate neuron for every possible value.
• Makes perception more robust → if some neurons respond less, other neurons can still contribute.

Population Coding + Adaptation

• Adaptation makes some feature-selective neurons less responsive.
• This changes the balance of activity across the population.
• The brain interprets the altered firing pattern differently.
• This can produce perceptual aftereffects.

Example:

Stare at tilted line
→ neurons tuned to that orientation adapt
→ their firing decreases
→ population activity becomes unbalanced
→ new line appears tilted away from the adapted orientation

Key Terms

Term	Simple meaning
Population coding	Many neurons work together to represent a feature
Preferred value	Feature that makes a neuron respond most strongly
Intermediate feature	A value between neurons’ preferred values
Population code shift	The balance of neural firing changes
Aftereffect	Perception changes because adaptation has shifted the population code

Adaption & Aftereffects

Population coding = perception of orientation from combined patterns of firing

Balanced population activity → accurate orientation

Tilt Aftereffect

Neural Adaptation

• Prolonged exposure to a particular orientation causes neurons tuned to that orientation to become less responsive.
• This is sometimes called neural fatigue, although neural adaptation is more accurate.
• These neurons temporarily contribute less to orientation perception.

Population Coding Shift

• Orientation is determined by the combined activity of many orientation-tuned neurons = population coding.
• Normally, their activity is relatively balanced.
• After adaptation, some neurons respond less → the balance of activity shifts.
• Other orientation neurons therefore have relatively more influence.

Tilt Aftereffect

• After staring at a tilted stimulus, a new stimulus can appear tilted in the opposite direction.
• This happens because adaptation has shifted the population response.
• The brain interprets this altered activity as a different orientation.

Example:
Stare at / → / neurons adapt → look at | → | appears tilted away from /.

Why is the effect strongest for similar orientations?

• Orientation neurons have overlapping tuning curves.
• A neuron responds to its preferred angle and somewhat to nearby angles.
• Therefore, adaptation has the greatest effect on similar but not identical orientations.

Size Aftereffect

Spatial Frequency Tuning

• Spatial frequency = the width/spacing of stripes in a pattern.
  • High spatial frequency → thin, closely spaced stripes
  • Low spatial frequency → thick, widely spaced stripes
• Different visual cortex neurons are tuned to different spatial frequencies.
• Some respond best to fine/thin patterns, while others prefer broad/thick patterns.

Adaptation

• Staring at stripes of a particular width causes neurons tuned to that spatial frequency to become less responsive.
• When you then look at medium-width stripes, the population response is unbalanced.
• This can make identical medium stripes appear thinner or thicker than they really are.

Size Aftereffect

• Size aftereffect = perceived stripe width changes after adapting to a different spatial frequency.
• Shows that recent visual experience can temporarily change our perception of size and spacing.

Tilt Vs Size Aftereffect

Tilt Aftereffect	Size Aftereffect
Neurons tuned to orientation adapt	Neurons tuned to spatial frequency adapt
Population coding shifts	Population coding shifts
Changes perceived angle	Changes perceived stripe width/spacing