Lecture 5 - visual perception PDF

Preview

Summary

Detailed revision notes combining lecture notes and core reading. ...

Description

Lecture 5- Detecting Motion – part 1



Movement is vital for our perception of the world.

Where is motion processed within the brain? - Magnocellular cells are responsible for motion in the visual system. - The principle area sensitive to motion is V1 - V1, V2 & V3 feed into an area called V5 (this is early on in the dorsal/where pathway) - After area V5 it travels to MST

 

Principle area for motion processing is V5/MT & MST There is an express connection from the LGN straight into area V5/MT (this means that if you have a stroke which inactivates area V1, you become functionally blind but you can still respond to moving targets- unconscious part of vision)

There are two ways to see motion. - We can process motion differently depending on what we are doing with our eyes. - Are we sensitive to motion on the retina vs. motion of the retina? Example: red dot moving across on a screen over a still target in the centre  If you keep your eyes still on a target in the centre of the screen, the red dot moves across your retina – you see it move. The blue star stays still in the same place on your retina and you don’t see it move. If you track the red dot with yours eyes (achieved through smooth pursuit eye movements), it does not move on your retina (as it’s in the centre of your vision), but you still see it as moving (as you’re tracking it). The blue star moves on your retina but you see it as stationary How does this work?  Your brain knows when you move your eyes/head intentionally and codes all movement/motion relative to this. - Therefore every motion you perceive is coded relative to what your eyes/head are doing.

Therefore we have two routes to perceive movement; one that detects movement across the retina (retinal- movement system) and one that detects movements of the eyes in the hear (eye/head movement system).

Theories explaining this: Inflow vs. Outflow

The difference between the two theories is that; in the outflow model, the efference signal is taken from the ACTUAL muscle movement whereas in the inflow model the signal is taken from the intention to move the eye. These two theories can be tested by either moving the eye artificially or trying to make them move.

Outflow – Helmholtz

Inflow – Sherrington

Proposed that rather than comparing image motion of the retina with eye muscle movement, the comparision is made with the signal from the brain that tells the eye muscle to move.

Proposed that we monitor the movements of our eye muscles and compare these eye muscle movements to the retinal motion His theory works by: A signal generated in the muscle in eye tells the brain the eye is moving.

We take a copy (efferent copy) of the eye muscle signal to move our eyes and compare this with the retinal image motion.

Muscles have nerves which send feedback messages to the brain. At the same the movement on the retina is also going into the brain and so comparison takes place between them

Comparison is made with the signal from the brain that tells the eye muscle to move. We take a copy of the signal to move our eyes (efferent copy) and compare this with any retinal image.

Signals about eye movement Come from the eye movement come from the brain - what actually happens from the eye

Signals about eye - what are the intentions of the eye?

Inflow vs outflow model predictions; 1. Sherrington- Inflow / Muscle theory - If you poke your eye whilst looking at a dot (the dot should stay static/ still) – the poke moves the eye muscle so any movement on the retina should be down to the eye muscle movement. 2. Hemholtz – outflow theory

- If you poke your eye whilst looking at a dot you would see motion because you had no intention to move your eyes – it’s only an efferent signal from the brain. 3. Sherrington- Inflow/ Muscle theory - If you poke your eye whilst looking at an afterimage – you should see motion 4. Hemhotlz – outflow theory - If you poke your eye whilst looking at an afterimage - Because there is no intention to move the eye you see static 5. Sherrington- Inflow - If you fix the eye in a certain position (prevent the eyeball from moving) and look at a dot and try to make eye movement, Sherrington would predict that there would be no movement because there will be no retinal movement and no eye-muscle movement- static. 6. Hemholtz- outflow - If you fix the eye in a certain position (prevent the eye ball from moving) and look at a dot and try to make eye movement, Hemholtz would predict that the stationary dot would appear to move because the eye movement signal is not cancelled by the expected retinal movmenent – would see motion

Who is right? = Hemholtz! (all of hemholtz are correct but all of sherringtons are incorrect). - If you stare at a stationary object (red dot) and give your eyeball a poke the object appears to move. This is because you have moved the eye and therefore stretched the eye muscle, but no efferent signal was sent by the brain to tell the muscle to move. - According to Sherrington the eye muscle movement is the important thing (it doesn’t matter whether it is moved by pushing it with a finger). - However according to hemholz the eye muscle movement is important- as only an efferent signal from the brain will do - We detect motion across our retina via an efference copy of our intended action. THERFORE OBJECT MOTION IS PERCIEVED AS RELATIVE TO THE MOTION OF OUR BODY, HEAD AND EYES. - A copy of the self-motion signal (efferent signal) is sent to motion processing areas in the brain.

The A and B ovals represent 2 receptive fields separated in space.

How do we detect motion?

(a)- does not detect motion, (b)detects motion left to right, (c)detects motion right to left, (d) – detects the balance of motion in both directions

o N.b. The comparative cell signals only when it receives a signal from A and B at the same time. o Imagine a spot of light that moves from spot A to B. o We have receptive fields in our brain that detect light at spot A and also receptive fields in our brain that detects light at spot B. o In order to detect the motion of the light moving from target A to B we need the signal from receptive field A and the signal from receptive field B to fire within a certain time period – simultaneously within the same time period so that the comparative cell can receive both signals. o However for the first diagram (A): The light is moving from target A to target B without any time delay. Therefore, receptive field A will be active first, followed later by receptive field B, but the two signals do not fire/ are not active simultaneously/ within the same time period and so the comparison cell does not see motion. o HOWEVER if you introduce a TIME DELAY between the time the receptive field detects the spot of light and the sending of this signal to the comparative cell - the signal from A will arrive at the comparison box the same time as the signal from B provided that the delay is as long as the time taken to travel from A > B. - This is also true if you introduce a time delay between the receptive field detecting the spot of light at B and the comparative cell if you were detecting light from B> A. – therefore a time delay will allow us to detect motion in the opposite direction too e.g. if the spot of light is moving in the opposite direction from B to A!

* Therefore diagram (b) and (c) show that with the use of a time delay system we can detect motion in different directions. - The last diagram demonstrates opponent motion because it is trading off leftward motion and rightward motion (see AV blog)

Where do we detect motion?

- We have no motion sensitive cells in human retina or LGN. Motion processing in humans starts in area V1 and extends into dorsal regions (V5/ MT, MST, V6, STS). - Each successive area processes more complex motion patterns.

V3 (complex of sub regions)

If see pic of Usain Bolt, will excite + imply motion without actually moving

MCQ: Which of the following are true? - When you poke your eye the image appears to move - When you fix your eye but try to make it move the images appear to move - When you poke your eye a previously established after image appears to move.

Motion selectivity in area MT (V5) - Figure shows the response of a directionally selective cell (in MT/V5) to different directions of motion. - This cell is particularly sensitive to motion at the top and motion to the left. - In humans - Cells in area V5 are direction-selective cells. - HOWEVER, in frogs & rabbits – direction- selective cells are at a retinal level

When the cells like movement (fire) When the cells don’t like movement (they don’t fire) Motion contrast in

V5/ MT

Smaller receptive field in central area

Surround OFF

Centre ON

- We have a collection of motion detectors for all different directions (have a set of cells each tuned to a different direction of motion and the one that is the most active signals the direction of motion of our object). - Opponent motion detectors respond to the balance between motion cells preferring opposite directions. - If you present motion only in the central region they will respond most strongly. - If you present motion only in the surrounding region the cells response turns off - We have centre surround cells for motion. INHIBITION OVER TIME: Motion after-effect – waterfall illusion > motion has direction. - results from inhibition over time - Aristotle (350 BC) - Rober Addams (1834) – looking at Falls of Foyers, Scotland.  Observe waterfall, looked at rocks, they appear to move upwards. - Main idea of the MAE: After you have been looking at something moving in one direction for a while, stationary things will appear to move in the opposite direction. Explanation

-

Perceived motion is encoded by cells (neural code) For a stationary object, both ‘up’ and ‘down’ cells fire a little. They are compared and there is no difference, therefore it is perceived as stationary. For a waterfall (downward) down cells excite and fire more. (up don’t fire much) After prolonged adaption to a given direction cells responsive to that direction will reduce their output (become suppressed as a result of inhibition) – adaption. The neural code for a subsequent static stimulus will be biased in favour of cells that normally respond to the opposite direction of movement.

- The Adaption (fire less) diagram above has 2 non opponent motion Up fires more so appears to move up

cells and an opponent motion cell working together. - A stationary object will excite both ‘up’ and ‘down’ detectors equally – the opponent motion cell that compares the up and down detector responses would find no difference and thus the stimulus appears stationary. - If we look at a downward motion, we have a stronger signal from the downward cells (the down detectors fire more) and a weak signal from the upward cells (upward detectors fire less) and so the stimulus appears to be down – we see a downward motion. - If we look at this downward motion for a prolonged period of time, the down detectors/cells are adapted and suppressed as a result of inhibition (don’t fire as much). - Therefore when we look at a stationary stimulus again, the down cells go into an ‘off’ period – it’s activity has been suppressed by inhibition and we have a slight signal coming from the up cells and so the opponent cell perceives the stimulus to be up 

-prolonged stimulation of the ‘down’ detector mans that it gets adapted and can fire as much as it used to. Following the adaption, the stationary pattern that would normally excite both up and down detectors equally will now excite the up detector more than the suppressed down detector.

So far, we have discussed that motion has direction but it also has speed.

Our perception of speed is influenced by stimulus duration (over time), position on the retina (over space) and contrast.

Perception of speed over time

(The faster the speed, the faster the motion cell fires – speed is encoded by firing rate) The figure shows the firing rate of a directionally selective ganglion cell in the rabbit retina. (NOTE humans do not encode motion until the cortex – not the retina!!)

(Motion after effect)

Adaption- There is a drop as the cells get tired – This is a deliberate strategy to reduce energy used (save energy). - The figure shows the recorded activity from a single ganglion cell in the retina of a rabbit to motion: - At stationary, the cells fire at ‘resting level’ – spikes at 10 impulses per second. - When stimulated by motion in its preferred direction, the cell it becomes more active and starts firing vigorously (firing rate increases to 60 impulses per second) - However as the motion of the stimulus continues, the firing rate / response slows down over time (adapts) until a plateu at 25 impulses per second is reached. (Adaption) - When the motion stops altogether, the cells firing rate drops below the resting level of 10 impulses per second for a few seconds > this is the motion after effect. This occurs when we look at a moving pattern for a long time and it appears to slow down quite dramatically and in some cases can come to a complete stop = illusion of ‘slowness’ is often experienced after driving quickly for a while on a motorway you feel as though you are driving slowly. (Experiencing a motion after effect).

Perception of speed over space The speed of a rotating disc

is altered by both how close it is to the fixation point and for how long we look at it. If we look at the disk on the left, then the discs on the right (falling in out peripheral vision) appear to be rotating more slowly. The effect gets larger the further away they are from fixation. The discs in our periphery get slower and even stop. More Sensitive to motion in our fovea than we are to motion in our peripheral vision – especially speed.

Perception of speed – image contrast

Respond best to high contrast and faster motion.

Same result – appear to be moving at same speed

   

Motion cells (at least in V1) respond to contrast (more strongly to high contrast stimuli) as well as the speed of stimulus (more strongly to faster motion) However the Principle of Univariance says that an individual cell can only have one output (it cannot signal two things – they will be confused). Therefore a low contrast stimulus moving quickly is perceived as moving about the same speed as a high contrast stimulus moving more slowly. Therefore it is difficult for us to perceive the speed of objects in fog! If we reduce the contrast of stimulus, the perceived speed drops (perceived speed reduces with reduced contrast) This is what happens in fog- reasons why you should have lights on in fog! Fog – hard to see correct speed. Things appear slower as low contrast

MCQ: which of the following will appear slowest? - A car in the fovea with bright sunlight. - A car in the fovea in fog. - A car in the periphery in fog. - A car in the periphery in sunlight. (This is because we perceive things as slow in our periphery vision and when it has low contrast).

Development of motion perception: o Sensitivity to motion develops at around 1-0-12 weeks of age o However, sensitivity to key types of motion may develop sooner o One study suggests that a looming stimuli (things that may hit a child are detected very early or may be present at birth (child moves out of way) o However some abilities that children are born with are later lost, and then have to be re-learnt later.

Motion Blindness – damage to MT (v1) – Akinetopsia (inability to percieve motion)

-

-

-

Loss of are MT (V5) makes you motion blind The world is perceived as still photos.

Losing v1 makes you functionally blind so you can still respond to some stimuli (especially movement) Losing v3 would hinder your motion perception but not destroy it. Losing MST/V6 would inhibit navigation but not stop you seeing motion. Whereas if you loose area MT you would not be able to see motion at all.

Normal person – 10% of motion uses v1 - as bypass it to v5 Patient – 90% of motion uses v1 - as don’t have v5 (MT)...