What is Motion Perception?
Motion perception is the process by which the visual system detects and interprets movement.
It allows us to work out:
- whether something is moving
- its direction
- its speed
- where it is likely to move next
- whether the object is moving or we are moving
Motion perception is important for everyday tasks such as crossing roads, driving, catching objects and navigating crowded environments.
Why motion perception is difficult
The brain does not simply calculate:
distance travelled ÷ time
Instead, it uses specialised neural mechanisms that respond directly to motion.
Several problems have to be solved:
| Problem | What the brain needs to work out |
|---|---|
| Motion ambiguity | Is the object moving, or am I moving? |
| Speed | How quickly is it moving? |
| Direction | Which way is it moving? |
| Illusions | Is the apparent movement actually real? |
| Multiple cues | How do vision, vestibular information and proprioception fit together? |
For example, when sitting on a stationary train beside another moving train, it can briefly feel as though your train is moving. This happens because the visual motion signal is ambiguous.
Akinetopsia: When Motion Perception Fails
Akinetopsia = motion blindness.
A person can still see objects, but has difficulty experiencing their continuous movement.
Instead, movement may appear like a sequence of static snapshots.
Patient LM
Patient LM developed severe motion-perception problems after a stroke.
This made tasks such as:
- pouring tea
- crossing the street
- following moving objects
extremely difficult.
This is important evidence that motion perception is a specialised visual function, rather than simply ordinary object perception plus time.
Motion Illusions
Two basic examples show that motion can be constructed by the visual system.
Stroboscopic Motion
Stroboscopic motion is the perception of smooth movement from a rapid sequence of separate still images.
Example:
movies and animation
Nothing is physically moving continuously between frames, but the visual system interprets the sequence as continuous movement.
Wagon Wheel Effect
A rotating wheel may appear to:
- rotate backwards
- slow down
- remain stationary
This can occur when motion is sampled intermittently, such as in film.
The visual system has to infer how the wheel moved between samples, and that inference can be wrong.
Reichardt Motion Detector
The Reichardt detector is a model explaining how the visual system could detect motion directly.
The central idea is:
Motion can be detected by comparing signals from two nearby locations across time.
Rather than separately calculating position and then calculating movement, the system contains mechanisms specifically tuned to movement.
Basic Structure of a Reichardt Detector
A Reichardt detector contains:
- Two neighbouring receptors
- A temporal delay
- An AND/comparator unit
Step 1: Spatial Separation
Two receptors respond to different locations in the visual field.
For example:
Receptor A → Receptor B
If an object moves from A toward B:
A is activated first.
Then B is activated shortly afterwards.
Step 2: Temporal Delay
The signal from one receptor is deliberately delayed.
So:
A activated → signal delayed
Meanwhile:
object moves → B activated
If the movement occurs at the correct speed, the delayed A signal and immediate B signal reach the next unit at the same time.
Step 3: AND / Comparator Unit
The AND unit fires strongly when it receives both signals simultaneously.
So:
A → delay ─┐
→ AND → MOTION detected
B ─────────┘
If the two signals arrive together:
→ detector responds
If they do not:
→ little/no response
Direction Selectivity
The delay makes the detector sensitive to one direction of motion.
Imagine the detector is designed for:
A → B
If the object moves:
A → B
the delayed A signal and B signal arrive together.
→ strong response
But if the object moves:
B → A
the timing no longer matches.
→ weak/no response
Therefore:
A Reichardt detector responds preferentially to movement in one direction.
Speed Selectivity
The detector is also sensitive to speed.
To produce the strongest response, the stimulus must take approximately the right amount of time to travel between the two receptors.
Correct speed
A activates
↓
A signal delayed
↓
stimulus reaches B
↓
signals arrive together
↓
strong response
Too fast
B’s signal arrives before delayed A.
→ poor match
Too slow
Delayed A arrives before B.
→ poor match
So speed sensitivity depends on:
- distance between the receptors
- length of the temporal delay
Detecting Different Directions and Speeds
One Reichardt detector only detects a limited range of:
- direction
- speed
Therefore the visual system would need many detectors with different:
- spatial arrangements
- delay lengths
- preferred directions
For example:
Detector 1 → leftward motion
Detector 2 → rightward motion
Detector 3 → upward motion
Detector 4 → downward motion
and others would be tuned to different speeds.
Strengths and Limitations of the Reichardt Model
Strengths
The model:
- is relatively simple
- explains basic motion detection
- explains direction selectivity
- explains speed selectivity
- provides a useful model of early/local motion processing
Limitations
Each detector is only sensitive to a particular combination of speed and direction.
Therefore:
- many detectors would be required
- complex motion cannot be explained by one detector
- the model struggles with ambiguous stimuli
- it does not explain every motion illusion
- uniform visual fields are difficult because there may be no distinct changes for neighbouring receptors to compare
Why the Model Needs Extensions
A simple Reichardt detector can sometimes create conflicting signals.
For example, two motion detectors might simultaneously indicate:
LEFT
and
RIGHT
The brain obviously should not simply perceive an object moving in both directions.
Therefore later models add mechanisms that compare the outputs of multiple detectors.
Comparator Units
A comparator unit compares the responses of motion detectors tuned to opposite directions.
For example:
Left detector → 8
Right detector → 2
Comparator:
→ stronger left signal
→ perceive leftward motion
But:
Left detector → 5
Right detector → 5
Comparator:
→ equal signals
→ no net motion
Main idea
Motion is determined not just by whether one detector responds, but by the relative activity of opposing detectors.
Ratio Models
Ratio models compare the relative strengths of motion signals.
For example:
| Left signal | Right signal | Perception |
|---|---|---|
| Strong | Weak | Left |
| Weak | Strong | Right |
| Equal | Equal | No movement |
This makes motion perception more flexible than relying on a single detector.
Population Coding
Real motion perception relies on populations of motion-sensitive neurons.
Instead of:
one neuron = one movement
the brain examines the overall pattern of activity across many neurons tuned to different:
- directions
- speeds
This allows the brain to represent complicated motion patterns involving multiple objects or changing directions.
Double-Flash Problem
Imagine two locations flashing at the same time.
A simple system could activate motion detectors for:
A → B
and
B → A
at the same time.
That would incorrectly suggest motion in both directions.
However, we normally perceive:
two flashes, not movement.
Comparator solution
Equal activity from opposite motion detectors:
left signal = right signal
↓
signals cancel/balance
↓
no motion perceived
This illustrates why comparing motion signals is important.
Apparent Motion
Motion does not require an object to physically travel continuously through space.
If stimuli appear at different locations with the correct timing, the brain may interpret them as one object moving.
This explains things such as:
- animations
- movies
- flashing lights that appear to travel
So motion perception depends heavily on the relationship between:
SPACE + TIME
Wagon Wheel Effect
The wagon wheel effect occurs when a rotating wheel appears to:
- slow down
- stop
- rotate backwards
In film, the wheel is only sampled at particular moments.
Imagine one spoke appears here:
Frame 1: ↑
Then the next frame looks like:
Frame 2: ↖
The brain has several possible explanations for how the wheel moved.
It tends to select the smallest/most likely displacement between frames.
Sometimes that shortest movement is in the opposite direction from the wheel’s real movement.
Therefore:
physically forward motion can be perceived as backwards motion.
Motion Aftereffects
A motion aftereffect (MAE) occurs when prolonged exposure to motion in one direction causes a stationary stimulus to appear to move in the opposite direction afterwards.
Classic example:
Waterfall Illusion
Look at downward-moving water for a long period.
Then look at stationary rocks.
The rocks can appear to move upward.
Why Motion Aftereffects Happen
Motion aftereffects occur because of neural adaptation.
Suppose there are populations tuned to:
UPWARD motion
and
DOWNWARD motion
Normally, when viewing something stationary:
up activity ≈ down activity
The signals are balanced.
→ no movement perceived
Adaptation Phase
You stare at downward motion.
Downward-sensitive neurons fire repeatedly.
Over time they adapt and become less responsive.
So:
down detectors ↓ responsiveness
while
up detectors remain relatively normal
Test Phase
Now you look at something stationary.
Normally:
up = down
But the downward detectors are still adapted.
So now:
up activity > down activity
The comparator interprets the imbalance as:
UPWARD MOTION
even though the stimulus is stationary.
Important
The upward neurons have not necessarily become unusually active.
Instead:
downward neurons are temporarily less active, creating a relative imbalance.
Waterfall Illusion
The waterfall illusion is therefore:
Adapt to downward motion
↓
downward neurons become less responsive
↓
look at stationary object
↓
upward activity is now relatively stronger
↓
stationary object appears to move upward
The lecture uses this as a classic demonstration of sensory adaptation in motion-sensitive mechanisms.
What Motion Aftereffects Tell Us
Motion aftereffects provide evidence that:
- the brain contains neurons sensitive to particular motion directions
- opposing motion directions are compared
- motion perception uses population activity
- prolonged stimulation changes neural responsiveness
- perception depends on the relative balance between neural populations
Experimental work also shows that direction-selective neurons can drop below baseline firing after prolonged stimulation, matching predictions from these models.
Neural Pathway for Motion
The main motion pathway covered in the lecture is:
Retina
↓
LGN
↓
V1
↓
MT/V5
Each stage adds more sophisticated processing.
Retina
The retina detects changes in incoming light.
In some animals, such as rabbits and frogs, motion/direction-selective cells can already be found in the retina.
In humans and other primates, however, much of the important direction-selective processing develops later in the cortical pathway.
LGN and the Magnocellular Pathway
The LGN relays information from the retina toward the cortex.
For motion perception, the particularly important system is the:
Magnocellular pathway
Magnocellular cells:
- respond rapidly
- have relatively large receptive fields
- are sensitive to changes in luminance
- are well suited to detecting movement and rapid changes
Therefore:
Magnocellular pathway = major pathway supporting motion information.
Direction Selectivity in V1
In humans and primates, direction selectivity first emerges clearly in V1.
Some V1 neurons prefer a particular direction of motion.
For example:
Neuron A:
→ → → = strong firing
← ← ← = weak firing
Another neuron may prefer the opposite direction.
So V1 provides the first cortical building blocks for determining which way something is moving.
MT/V5
Motion information is then sent to MT/V5, a cortical area strongly specialised for motion processing.
MT/V5 contains many neurons tuned to:
- motion direction
- motion speed
These neurons are organised into columns according to preferred direction.
Role of MT/V5
MT/V5 integrates motion signals so that we can perceive more complex movement.
It helps determine:
- direction
- speed
- coherent motion across larger areas of the visual field
Evidence That MT/V5 Causes Motion Perception
Researchers have electrically stimulated specific motion-direction columns in MT/V5 in non-human primates.
This can bias the animal’s judgement of motion direction.
Therefore:
changing activity in MT/V5 can change perceived motion.
This provides strong evidence that MT/V5 activity is functionally involved in motion perception rather than merely correlated with it.
MT/V5 and Akinetopsia
Damage to MT/V5 is associated with profound impairments in motion perception.
People with severe motion blindness may experience the world as a sequence of static states rather than continuous movement.
This provides another important link between:
MT/V5 → normal motion perception
The Whole Motion System
1. Visual stimulation changes over space and time
↓
2. Local motion detectors compare neighbouring locations
Reichardt model
- receptor A
- receptor B
- delay
- AND unit
↓
3. Timing gives speed + direction selectivity
Correct direction + correct speed
→ signals coincide
→ motion detector responds
↓
4. Multiple detectors are compared
Comparator / ratio models
Opposing direction signals compete.
↓
5. Populations represent overall motion
Many neurons with different preferred speeds and directions work together.
↓
6. Adaptation changes the balance
Prolonged movement reduces responsiveness of neurons tuned to that direction.
→ motion aftereffect
↓
7. Motion information travels through the visual pathway
Retina → LGN → V1 → MT/V5
↓
8. Main roles
LGN / magnocellular pathway
→ fast motion-related information
V1
→ direction selectivity first clearly emerges
MT/V5
→ integrates and interprets motion
↓
Conscious perception of movement
That is the core structure of the lecture: it moves from how motion is detected locally, to how those detector signals are compared and adapted, and finally to the neural pathway that turns those signals into motion perception.
