PSYC3015.

OBJECT (DISTAL STIMULUS) > IMAGE PROJECTION > RETINAL IMAGE (PROXIMAL STIMULUS) > SENSORY REPRESENTATION (OF THE PROXIMAL STIMULUS) > PERCEIVED OBJECT (THIS IS A PERCEPTUAL REPRESENTATION OF THE DISTAL STIMULUS).

S = FORMED BY OUR SENSE ORGANS E.G. OUR RESPONSE TO LIGHT VIA THE RETINA AND COLOUR VIA THE L, M AND S CONES + how the LGN and V1 cells encode information. P = HOW WE INTERPRET THE SENSORY INPUT E.G. WHAT WE ACTUALLY SEE.

PSYCHO (OF THE MIND) + PHYSICS (MEASURABLE). THE LINK BETWEEN THE PHYSICAL STIMULUS AND OUR PERCEPTION OF IT. IT'S MODERN AIM IS TO UNDERSTAND HOW THE VISUAL SYSTEM WORKS AND HELPS US TO PERCEIVE THE OUTSIDE WORLD.

absent/present e.g. how much intensity do you need there to be to see the stimulus. ABSOLUTE = DETECTING THE PRESENCE OF A STIMULUS. IT'S NORMALLY THE MINIMUM INTENSITY AT WHICH THE STIMULUS IS JUST VISIBLE E.G. HOW BRIGHT DOES A LIGHT HAVE TO BE FOR YOU TO BE ABLE TO SEE IT. AND RELATIVE = TELLING TWO STIMULI APART.

PLOTS DETECTABILITY VS. STIMULUS INTENSITY.

1. EXTERNAL NOISE = PROBLEM WITH THE STIMULUS E.G. THERE'S A CHANGE IN THE LIGHT SOURCE OVER TIME AND; 2. INTERNAL NOISE = PROBLEM WITH THE OBSERVER E.G. ATTENTION/% saying that they can see it. noise = variation.

1. RANDOM NOISE = THE ERROR IS RANDOM E.G. THE LIGHT LEVEL IS MEASURED TO BE TOO HIGH AND THEN TOO LOW AND; 2. SYSTEMATIC NOISE = THE ERROR IS CONSTANT (e.g. always in the same direction). E.G. THE LIGHT LEVEL IS ALWAYS TOO LOW and e.g. using hard-drives with different thresholds.

R = THE MEAN IS UNAFFECTED BUT THERE IS MORE VARIANCE. THE ^ STANDARD DEVIATION < SIGNIFICANT DIFFERENCES. SHOULD INCREASE THE NO. OF TIMES THE THRESHOLD IS MEASURED TO REDUCE THIS VARIANCE. S = THE THRESHOLD WILL BE WRONG AND THE RESULT WILL BE WRONG. makes false claims.

S = calibrate and check. Randomize controls and anything that you think will effect e.g. Fatigue. This doesn't decrease their effect but increasing random noise will fix systematic noise. Use a random order, breaks, explanation and test.

VALID = E.G. IT INCORPORATES THE THRESHOLD LEVEL AND RELIABLE = YOU WOULD GET THE SAME DATA IF YOU DID IT AGAIN. THE MID-POINT IS THE BEST POINT TO USE ON THE PSYCHOMETRIC CURVE.

A = bad. Not seeing the stimulus = No threshold and not valid. C = bad. Can see 100% but bad because the threshold is earlier. Not reliable. B = good. It has the smallest change in stimulus intensity and the biggest change in performance. Most reliable and valid. A + C were too extreme.

ADJUST THE STIMULUS INTENSITY E.G. THE DOT BRIGHTNESS UNTIL IT IS JUST VISIBLE. THE THRESHOLD IS THE SPOT WHERE THEIR RESPONSE CHANGES. NEED TO TO IT IN ASCENDING AND DESCENDING ORDER TO AVOID SYSTEMATIC ERROR. A= IT'S QUICK AND THERE'S NO PRE-TESTING. D = NOT PRECISE.

FIXED SET E.G. 5-9 OF STIMULUS INTENSITIES ARE PRESENTED FOR A FIXED NO. OF TIMES IN A RANDOM ORDER.  THEN PLOT THE DATA. THE EXPECTED THRESHOLD VALUE SHOULD LIE AT THE MIDPOINT. A = GET A FULL PSYCHOMETRIC CURVE. THE RANDOM ORDER AVOIDS PROBLEMS WITH TMOA. D = INEFFICIENT E.G. TRIALS THAT ARE WAY ABOVE AND BELOW THE THRESHOLD. NEEDS A PILOT TEST.

START <> THE THRESHOLD VALUE AND CHANGE THE INTENSITY DEPENDING ON THE PARTICIPANT'S RESPONSE. PR = DI AND; NR = II. REVERSAL POINT = WHERE THE DIRECTION CHANGES. STOPS AT A CERTAIN NO. OF REVERSAL POINTS. A = QUICK AND ACCURATE. D = SPEND A LOT OF TIME AROUND THE THRESHOLD AND SENSITIVE TO CHANCE. 1/1 = 50%; 2/1 = 71%; 3/1 = 79% (N/1 = NTH ROOT OF 0.5).

1. CRITERION VARIES; 2. SOME PARTICIPANTS SAY YES WHEN THEY CAN SEE EASILY AND SOME PARTICIPANTS SAY YES WHEN THEY CAN BARELY SEE IT. CRITERION VARIES. 3. Confidence and 4. There's no way to know whether they can actually see it or not.

NO. OF CHOICES AND FORCE THE PARTICIPANT TO PICK AN ANSWER. E.G. 2AFC. THIS MEANS THAT THEY HAVE TO PICK AND THERE'S A CORRECT ANSWER.

USE A CHANCE-LEVEL PERFORMANCE. AT 2AFC THEY WILL GET THE RESPONSE CORRECT 50% OF THE TIME. AT 4AFC AND 6AFC IT'S 25% AND 16.7% WHICH MAKES IT HARDER TO GUESS. Getting it wrong 0% of the time means that they are lying.

PARTICIPANT'S ABILITY TO DISCRIMINATE BETWEEN TWO THRESHOLD STIMULI E.G. WHICH OF THE TWO LIGHTS IS BRIGHTER. THIS LEADS TO THE JND.

JND = THE DIFFERENCE IN INTENSITIES THAT IS REQUIRED TO TELL THEM APART. A DISCRIMINATION THRESHOLD. JND = KXI0 (K = THE CONSTANT E.G. WEIGHTS = 3% AND I0 = THE REFERENCE STIMULI E.G. 10KG JND = 0.3KG; 20KG JND = 0.6KG. THEREFORE == THE PERCEIVED MAGNITUDE OF THE STIMULUS IS PROPORTIONAL TO THE NO. OF JND'S IT IS ABOVE A THRESHOLD.

CONTRAST = LUMINANCE/LUMINANCE (B/G) e.g. luminance of the bar compared to the background / the luminance of the background. WHY DO WE USE CONTRAST? The receptor cells don't look for a difference in local contrast == They look for variation e.g. they look for a bright bar on a background. Therefore, the cell's response depends on the contrast.~~~

THE VISUAL SYSTEM TRIES TO RECONSTRUCT THE OUTSIDE WORLD e.g. It tries to break information down into little mosaic pieces and reconstruct it. IT USES SINEWAVE VARIATIONS IN LUMINANCE (e.g. change from light to dark). AT DIFFERENT SPATIAL SCALES TO DO THIS E.G. SPATIAL FREQUENCIES. EACH SINEWAVE SHOWS YOU HOW MUCH ENERGY IS IN THE STIMULUS AT EACH SPATIAL FREQUENCY E.G. HOW MUCH OF EACH SINEWAVE YOU NEED TO MAKE THE STIMULUS. V1 cells are little spatial frequencies analyzers. They break the sinewaves down into different luminances across space and time.

1. IT CAN TELL US WHAT THE VISUAL SYSTEM ACTUALLY DOES AND; 2. CAN DETERMINE WHAT ANY PERSON CAN SEE WHEN PRESENTED WITH A STIMULUS E.G. A CAT, CHILD OR VISUALLY IMPAIRED PERSON.

ANY PATTERN CAN BE REPRESENTED BY A SERIES OF SINEWAVE GRATINGS. THE RF'S ALSO LOOK LIKE SPATIALLY-LOCALIZED SINEWAVE PROFILES (E.G. A GABOR FUNCTION). sinewaves are the building blocks of vision. V1 cells can detect and break down the sinewave components.

CONTRAST = the change from light to dark, SPATIAL FREQUENCY AND PHASE e.g. 0-360 degrees. It just changes the object's position. e.g. moving it by 180 degrees == the light bars are now where the dark bars are. Hasn't changed the spatial frequencies.

LESS = e.g. looks at broad outline of a face and; MORE = done when there is more spatial frequencies e.g. the fine details of the face.

The distance between the two peaks in the sinewave.

^the brighter the bright bits and the darker the dark bits. When the contrast decreases = the sinewaves get flatter. It tells us that there is less energy and contrast at that spatial frequency.

eg. has 1 peak and the next has 3 in that 1 and the next had 5 peaks inside the original 1 peak (dim) etc. 1/3 = less contrast and energy. An undetectable harmonic doesn’t impact the sinewave.

AT WHAT CONTRAST LEVEL CAN YOU TELL SQUARE-WAVE (when there is an instant drop in contrast) APART FROM A SINEWAVE? WHEN THE CONTRAST IS BIG ENOUGH TO MAKE ONE OF THE HARMONICS VISIBLE E.G. THE THIRD. = WE CAN BREAK A STIMULUS DOWN INTO ITS SINEWAVE COMPONENTS AND PREDICT IT'S VISIBILITY.

CAN PREDICT A PERSON'S RESPONSE TO ANY STIMULUS. IT PLOTS CONTRAST SENSITIVITY AGAINST SF. EG. AT EACH SF = TIMES THE CONTRAST (E.G. ENERGY) OF THE SINEWAVE BY HOW WELL THE SYSTEM TRANSMITS THE SINEWAVE. E.G. THE STEREO EXAMPLE.

THE MTF OF VISION. TELLS US THE WINDOW OF VISIBILITY. HIGH SF STOP BECAUSE OF THE OPTICS OF THE EYE. LOW SF STOP DUE TO OPPONENCY. How much contrast do you need to see the stimulus? 1. Find the sinewaves in the image. 2. times them together and it gives you the spectrum = spatial frequencies. Gives the image that they can see.

THE MAXIMUM SF THAT HUMANS ARE SENSITIVE TO = 30CPD AND CATS = 12CPD. WEDGETAIL EAGLE IS 140CPD BECAUSE THEY NEED TO SEE SMALL ANIMALS FROM A BIG DISTANCE. Babies e.g. tested with habituation and peripheral looking. Can see low-spatial frequencies and not high-spatial frequencies.

1. PREDICTS PERFORMANCE IN LOW LIGHT E.G. PILOTS. HAD TO AVOID HITTING AN OBJECT ON THE RUNWAY. IT IS MEASURED BY HOW CLOSE THEY GOT THE OBJECT BEFORE ABORTING. CSF WAS A GOOD PREDICTOR OF PERFORMANCE. 2. HOW DO VISUALLY IMPAIRED PEOPLE MOVE AROUND? 3. IT CAN ENHANCE DEGRADED IMAGES AND FIX GLARE.

THERE ARE 4-7 CHANNELS THAT ARE TUNED TO DIFFERENT SPATIAL FREQUENCIES. THE FIRING RATE TELLS YOU THE ENERGY AT EACH SF. COMPLEX STIMULI E.G. FACES CAN BE BROKEN DOWN INTO TO SMALLER SF'S.

HABITUATION e.g. looking at something for a long time makes the cell’s sensitivities go down (AKA. adaptation). If it is a single channel then sensitivity to all spatial frequencies will be decreased. However, the adaption happens to the cells that are sensitive to that stimulus. There will only be a reduction in the frequency you’re using.

HSF == rapid transition from light-dark. Gives you information about fine detail. LSF = blended e.g. can only see the mona lisa smile with this frequency.

TO PROCESS AN IMAGE THE VISUAL SYSTEM FIRST PROCESSES LOW SF'S AND THEN FILLS IN THE DETAILS WITH HIGH SF'S. LSF = MAGNOCELLULAR E.G. FAST AXON SPEED AND HSF = PARVOCELLULA E.G. SLOW AXON SPEED.

The firing rate of the cells tell you how much energy is at the spatial frequency.

MONOCULAR E.G. ONE EYE AND BINOCULAR E.G. TWO EYES.

1. ABSOLUTE E.G. THE DISTANCE BETWEEN THE PERSON AND THE OBJECT AND 2. RELATIVE E.G. THE DEPTH BETWEEN OBJECTS. Don’t know their distances but we know that one is closer than the other.

1. SIZE = can be an absolute depth cue if the size is known. Depends on other cues e.g. it’s not used in a cue-rich environment. Can be a relative depth cue when there are lots of objects that are the same size. Misleading e.g. The moon illusion 2. LIGHTING/SHADOW = predicts that light comes from above. Depth = more shadow so less depth is perceived when light is directed straight onto a surface. 3. CLARITY/ELEVATION = clear objects are perceived to be closer. High = far away. 4. PERSPECTIVE E.G. THE VANISHING POINT e.g. the football field. 5. Texture = the visual detail on a surface e.g. rough. Depth/FAR = HSF = more texture. 6. OVERLAP = The brain uses what a "normal shape" looks like to infer depth. E.G. Rectangle with a shape in front of it is still a rectangle.

Mountains look closer when there’s just been a storm == happens because there is less light scatter and refraction. Light scatter gets rid of the HSF'S.

MONOCULAR. VIEWED BALLOONS IN DIM LIGHT. VARIED THEIR SIZE E.G. BIGGER = SEEN AS CLOSER E.G. MISLEADING E.G. THE MOON ILLUSION.

WHEN THERE ARE 2 OPAQUE OBJECTS IN THE SAME LINE OF SIGHT THE CLOSER ONE WILL OCCLUDE THE OTHER. TELLS YOU DEPTH INFORMATION ALONG A LINE OF SIGHT.

THE LENS WILL CHANGE ITS SHAPE AND FOCUSING POWER TO MAINTAIN A CLEAR IMAGE AT DIFFERENT DISTANCES. IT IS A CUE FOR ABSOLUTE DEPTH. WE CAN'T SEE THIS CUE BUT CHAMELEONS CAN.

1. MOTION PARALLAX = THE PERSONS MOTION CAUSES RELATIVE OBJECT MOTION. DEPENDS ON THE DIRECTION AND MAGNITUDE OF RELATIVE DEPTHS. 2. THE KINETIC DEPTH EFFECT = USE DIFFERENT SPEEDS TO SHOW THE 3D OBJECT.

1. SIDE-FACING e.g. Bunnies which use monocular vision AND; 2. FORWARDS-FACING that use binocular vision. Prey sacrifice depth perception for a good field of view.

CONVERGENCE = RELIES ON ACCOMODATION CHANGES. THE EYES MOVE IN DIFFERENT DIRECTIONS TO FIXATE ON OBJECTS THAT ARE AT DIFFERENT DEPTHS. AND; 2. BINOCUALR DISPARITY = THE HORIZONTAL DISPLACEMENT OF EACH EYE MEANS THAT THEIR IMAGES ARE SLIGHTLY DIFFERENT E.G. FOR NON-FIXATED OBJECTS THE IMAGE WILL PROJECT TO DIFFERENT SPOTS IN EACH EYE. MAGNITUDE AND DIRECTION CUES ARE IMPORTANT.

THEY PRESENTED A DIFFERENT IMAGE TO EACH EYE AND VARIED THE AMOUNT OF LATERAL SEPERATION EBTWEEN THEM TO CHANGE CONVERGENCE. FOUND = ^ CONVERGENCE E.G. GIVING A SMALLER VIEWING DISTANCE = SMALLER SIZE EVEN THOUGH THE DISC SIZE REMAINED THE SAME. THEREFORE, IT IS A CUE TO ABSOLUTE DEPTH.

THE SIZE OF THE RETINAL AFTERIMAGE IS FIXED BUT THE SIZE DEPENDS ON THE DISTANCE BETWEEN THE PERSON AND THE SCREEN. IF THE TEST DISTANCE IS SMALLER THAN THE ADAPTATION DISTANCE THEN THE AFTERIMAGE WILL BE SMALLER THAN THE ADAPTION STIMULUS. LONGER = BIGGER.

D = where the nerve fibres cross over from one side of the brain to the other. H = partial cross over. The ratio of crossed to uncross fibres is proportional to the size of the binocular visual field e.g. humans = 50% and rabbits = 0%. Good because the cells in the left and the right eye that have RF'S that cover the same part of the binocular vision will project to the same part of the brain. Therefore, cortical cells will get information about the same region from both eyes == Have continuous mapping of the outside world.

: IMAGES PRODUCED BY OBJECTS ON THE HOROPTER WILL FALL ONTO CORRESPONDING POINTS = THEY CREATE 0 BINOCULAR DISPARITY. IMAGES THAT ARE NOT ON THE HOROPTER WILL FALL ONTO NON-CORRESPONDING POINTS WHICH = BINOCULAR DISPARITY.

WE WANT IT TO BE A FRONTAL PLACE BECAUSE THEN THE PATTERN OF BINOCULAR DISPARITY WOULD BE DIRECTLY RELATED TO THE DEPTH OF THE OBJECTS. ITS SHAPE VARIES WITH FIXATION DISTANCE E.G. GOES FROM CONCAVE TO CONVEX. THEREFORE, THE ABATHIC DISTANCE IS WHEN IT IS A FRONTAL-PARALLEL PLANE. Close == it’s behind the horopter and far == it’s in front of the horopter.

THE REGION AROUND THE HOROPTER WHERE IMAGES CAN BE FUSED. **OFFSET FROM THE HOROPTER ^ RETINAL DISPARITY. THIS ^ MEANS THAT WE REACH A POINT WHERE THE IMAGE CAN'T BE FUSED TO MAKE A SINGLE IMAGE.

THEY TELL DEPTH. DIFFERENT IMAGES ARE PRESENTED TO EACH EYE. E.G. THERE ARE TWO CAMERAS AND THE OFFSET CORRESPONDS TO THE INTERPUPILLARY DISTANCE (E.G. IPD). INCREASING THE IPD WILL INCREASE DEPTH.

• IMAGES PRESENTED AT DIFFERENT LOCATIONS E.G. MIRROR, PRISM E.G. They refract the light so you can see the image AND;
• IMAGES PRESENTED AT THE SAME LOCATION E.G. ANAGLYPHS e.g. overlap red and green images and the selectively present them to each eye; POLAROID e.g. 2 projectors with Polaroid filters; SHUTTER e.g. different images are presented really quick on a computer; AUTOSTEREOGRAM.

IMMEDIATELY SEE THE FUSED IMAGE. THEN IT SWITCHES BETWEEN THE TWO IMAGES E.G. EITHER ONE AND THEN THE OTHER OR A MIX. Can be tested via horizontal and vertical grating in a lab.

WE HAVE TO DETERMINE WHICH IMAGES IN EACH EYE CORRESPOND TO THE SAME OBJECT. NOTE THAT DIFFERENT OBJECTS HAVE DIFFERENT DISPARITIES. ***DIAGRAM. DOES THE VISUAL SYSTEM PROCESS INFORMATION IN EACH EYE AND THEN MATCH IT?

USES A STIMULUS THAT DOESN'T HAVE OBVIOUS MONOCULAR FEATURES. HAS LIGHT AND DARK DOTS. FOUND THAT THE STIMULUS CANNOT BE IDENTIFIED PRIOR TO BINOCULAR MATCHING.

RANDOMLY ASSIGN PIXELS TO BLACK OR WHITE. THE DEPTH IS SHIFTED IN OPPOSITE REGIONS IN THE TWO IMAGES E.G. CONVERGING = NEAR AND DIVERGING = FAR. THEN PUT THE UNCORRELATED PIXELS INTO THE BLANK REGIONS. WE CAN SEE DEPTH WITH AN RDS WHICH = WE DON'T DO A FORM ANALYSIS BEFORE JUDGING DEPTH.

FM = THE TWO IMAGES BEING MATCHED ARE NOT OF THE SAME OBJECT. MORE FALSE MATCHES ARE POSSIBLE THAN CORRECT ONES. TO AVOID THIS, OTHER CONSTRAINTS NEED TO BE USED BY THE VISUAL SYSTEM.

COOPERATIVE, COARSE-TO-FINE AND FEEDFORWARD.

1. THE E AND I LINKS BETWEEN CELLS HELP TO; FIX THE FALSE-MATCHING PROBLEM AND EXTEND THE DISPARITY RANGE; 2. INVERSE-PROJECTION/THE DISPARITY MATRIX = LAYS OUT ALL THE POSSIBLE STEREO-MATCHES (e.g. right and wrong). E LINKS ALONG THE DEPTH PLACE AND I LINKS ALONG THE LINE OF SIGHT E.G. WINNER-TAKES-ALL = ONLY ONE CELL CAN BE ACTIVE. THIS IS FOR 2 REASONS; 1. most surfaces aren't opaque e.g. see-through. Doesn't mean that the nearest one is the right one. 2. SURFACE CONTINUITY.

3 STEPS; 1. MAKE ALL THE POSSIBLE MATCHES; 2. APPLY THE OPACTITY AND CONTINUITY RULES = E AND I INTERACTIONS E.G. when a cell is firing strongly it will ^ the firing rate of its neighboring cells and inhibit those tuned to different depths; 3. A STABLE SOLUTION IS FOUND THAT IS CONSISTENT WITH DISPARITY INFORMATION AND THE OPACITY AND CONTINUITY CONSTRAINTS. This is because there will be more active cells around the depth place with ^ excitation.

MORE LIKELY THAT WE MATCH EDGE/SF GIVEN THE RF'S RATHER THAN SINGLE PIXELS. E.G. BIG DISPARITIES ARE ENCODED BY LSF'S AND SMALL DISPARITIES ARE ENCODED BY HSF'S.

GOES FROM LSF'S TO HSF'S. THIS LETS THE SYSTEM FIND THE CORRESPONDING MATCHES AT A COARSE SCALE E.G. LARGE DISPARITY AND THEN FINE TUNE THEM TO MATCH THE HSF'S E.G. SMALL OFFSET.

The aliasing on wagonwheels happens in the movie not the eyes. 180-180 degrees would cause ambiguous movement (the 180 offset). Still would = 360 degrees. INSTEAD = cut out the HSF'S to see the image. Squinting also does this. Aliasing can only be applied to rotation etc. Therefore, LSF'S can give you better depth information.

1. OBJECT FORM (E.G. WHAT) AND OBJECT LOCALIZATION (E.G. WHERE).

TO PROVE THAT MOTION IS PROCESSED INDEPENDENTLY OF THE P/W (WE SEE MOTION INSTEAD OF JUST SEEING A CHANGE IN LOCATION AND INFERRING MOTION) WE NEED TO PROVE THAT WE CAN SEE MOTION WITHOUT SEEING A CHANGE IN LOCATION. EVIDENCE = E.G. THE MAE.

WHEN WE SEE MOTION IN ONE DIRECTION AND THEN VIEW A STATIC IMAGE, IT WILL LOOK LIKE IT'S MOVING IN THE OPPOSITE LOCATION EVEN WHEN THERE'S NO CHANGE IN LOCATION.

1. WE HAVE CELLS THAT ARE TUNED TO DIFFERENT DIRECTIONS OF MOTION; 2. WE HAVE AN I CONNECTION BETWEEN CELLS THAT ARE TUNED TO DIFFERENT MOTION; 3. MOTION IS CONTROLLED BY THE P/W.

• RC =M AND CC = B. 
• 1. ADAPT MOTION WITH ONE EYE; 2. TEST FOR MAE WITH THE OTHER. IF THERE IS A TRANSFER OF THE MAE THEN THE MOTION CELLS ARE IN THE CORTEX.

WHERE INFORMATION IS EXTRACTED AT DIFFERENT LEVELS. HIGH = MORE COMPLEX INFORMATION. LEVELS = 1. LOCAL-MOTION PROCESSING; 2. GLOBAL-MOTION PROCESSING AND; 3. THE OPTIC-FLOW LEVEL.

USED IN LOCAL-MOTION EXTRACTION. STAGES; 1. SPATIALLY OFFSET RECEPTIVE FIELDS; 2. TEMPORAL DELAY E.G. DELTA T ON THE OUTPUT OF THE SECOND RF AND; 3. SUMMATION CELLS THAT MULTIPLES THE 2 OUTPUTS. The 2 signals need to arrive at the same time otherwise e.g. signal x no-signal = 0. Moving in the OD means that the signal will arrive at different times.

1. TEMPORAL DELAY E.G. ^ = THE SLOWER THE OPTIMAL SPEED AND; 2. THE SPATIAL OFFSET OF THE FILTERS E.G. ^ = THE GREATER THE OPTIMAL SPEED.

LM UNITS HAVE SPATIALLY LOCALIZED RF’S E.G. APPETURES. WHEN THE OBJECT EXTENDS BEYOND THE RD = THE CELL CAN ONLY TELL MOTION THAT IS ORTHOGONAL (R.G. AT A RIGHT-ANGLE) TO THE CELL’S PREFERRED DIRECTION. TO TELL WHAT THE ACTUAL MOTION IS WE NEED TO POOL THE OUTPUT OF LOTS OF LM CELLS.

WE POOL THE LM SIGNALS FOR OBJECTS THAT EXTEND THE RF. ONE MODEL = IOC.

WE CAN FIGURE OUT ALL THE POSSIBLE MOTIONS (E.G. A LINE OF CONSTRAINT) BECAUSE WE KNOW THE ORTHOGONAL COMPONENTS OF THE OBJECT MOTION. WE THEN MAKE ANOTHER LINE OF CONSTRAINT FOR ANOTHER ONE OF THE OBJECTS SURFACES. WHERE THEY INTERSECT TELLS YOU THE OBJECT/S MOTION E.G. BECAUSE IT SATISFIED BOTH OF THE CONSTRAINTS.

USES 1. SIGNAL DOTS = MOVE IN THE SAME DIRECTION AND; 2. NOISE DOTS = MOVE IN RANDOM DIRECTIONS.THRESHOLD = THE NO. OF SIGNAL DOTS REQUIRED TO FIGURE OUT THE GM DIRECTION. NEED TO POOL LOCAL-MOTION SIGNALS OVER S + T. THE MOTION SYSTEM IS PRETTY GOOD AT THIS E.G. 6%T.

NEEDED BECAUSE THE V1 CELLS HAVE SMALL RF SIZES. NEED TO POOL THESE SIGNALS. EXAMPLES = 1. CONTOUR AND; 2. GLASS PATTERNS. GOOD IF YOU WANT TO TEST GM IN THE DORSAL PATHWAY AND THE EQUIVILANT IN THE VENTRAL E.G. V5 VS. V4.

LM EXTRACTION = V1 AND GM EXTRACTION = V5 AND MT.***DIAGRAM.

LESION AND CLINICAL STUDIES; 2. TMS; 3. FMRI; 4. SINGLE-CELL RECORDING E.G.SEE WHETHER THE V5 CELLS ARE TUNED TO THE GM SIGNAL AND; 5. CORTICAL MICRO STIMULATION E.G. STIMULATING THE V5 CELLS COLUMN THAT IS TUNED TO A SPECIFIC MOTION WILL BIAS THE MOTION SEEN.

WE HAVE FIRST-ORDER AND SECOND-ORDER PATHWAYS. 1. = TUNED TO SPEED. ARE INDEPENDENT UP UNTIL THE GM LEVEL? EVIDENCE = EXPERIMENTAL + CLINICAL.

SELF-MOTION. PROVIDES INFORMATION ABOUT THE ENVIRONMENT AND OUR MOTION THROUGH THE ENVIRONMENT. THREE TYPES; 1. TRANSLATIONAL E.G. FRONTOPARALLEL MOTION; 2. RADIAL E.G. DEPTH AND; 3. ROTATIONAL E.G. ROTATION.

SEEING A OPTIC-FLOW MAKES YOU THINK THAT YOU’RE MOVING. E.G. YOU’RE IN A PARKED CAR AND THE CAR NEXT TO YOU MOVES = THINKS THE BIG OBJECT (E.G. THE CAR) IS STATIONARY AND THE SMALLER OBJECT (E.G. YOU) IS MOVING. WORKS BEST WHEN THE CAR IS MOVING SLOWLY. OTHERWISE THERE’S A CUE CONFLICT BETWEEN THE VISUAL AND VESTIBULAR SYSTEM.

DETECTS CHANGES IN MOTION E.G. ACCELERATION. NO RESPONSE WHEN YOU'RE DRIVING AT A CONSTANT SPEED IN THE SAME DIRECTION. THIS CAN CAUSE MOTION SICKNESS.

REMOVE VISUAL FEEDBACK E.G. STAND ON ONE FOOT AND CLOSE EYES AND; 2. GIVE THE WRONG VISUAL CUES E.G. PUT THEM IN A SWINGING ROOM. SWINGS FORWARDS = THEY FALL BACKWARDS AND SWINGS AWAY = THEY FALL FORWARDS.

GET INPUT FROM THE V5/MT. V5 RF’S ARE 10X BIGGER THAN V1 RF’S AND MSTD RF’S ARE BIGGER THAN V5. ITS CELLS ARE TUNED TO OPTIC-FLOW PATTERNS.

EXPANDING = F AND; 2. CONTRACTING = B. WE DON’T KNOW WHICH ONE WE ARE MORE SENSITIVE TO. EXPANSION = AN ENVIRONMENTAL BIAS AND; CONTRACTION = OUTCOME. GM THRESHOLDS SHOW THAT WE ARE MORE SENSITIVE TO CONTRACTING PATTERNS. BIG AND SMALL STIMULI HAVE THE SAME RESULTS.

GIVE INFORMATION THAT IS EITHER CONSISTENT OR INCONSISTENT AND WORK OUT THE THRESHOLDS FOR E PATTERNS WHEN THE PARTICIPANT IS MOVING FW OR BW (C VS. IC). THRESHOLDS ARE LOWER IN THE CONSISTANT CONDITION.

MOTION E.G. PROCESSING AND USAGE; COLOUR ~E.G. PERCPETION AND FORM AND; FORM E.G. AGNOSIA. CAN ALSO HAVE DECIFITS IN ACTON VS. PERCEPTION.

HAD A CORTICAL BLOOD-FLOW BLOCKAGE AND STARTED SEEING OBJECTS AS BEING “RESTLESS” OR “JUMPING AROUND”. FELT QUEASY.  WAS FIRST DIAGNOSED AS HAVING AGORAPHOBIA. AFTER THREE YEARS OF THERAPY SHE COULD COPE WITH DAY-TO-DAY LIFE, BUT HER MOTION PROCESSING NEVER IMPROVED.

WRITING = EASIER WITH HER EYES OPEN; COULD RECOGNIZE OBJECTS, PLACES AND FACES; COULD FIND OBJECT WITH HAND AND EYE MOVEMENTS; HAD A LOWER RANGE OF COLOUR PERCEPTION AND; HAD NORMAL STEREOPSIS BUT HER ACQUITY WAS A LITTLE HIGH.

LOCALIZED DAMAGE MEANS THAT YOU ARE BLIND TO THE CORRESPONDING VISUAL FIELD. DESTROYED = 1. NO CONSCIOUS AWARENESS AND A CAN’T REALLY SEE OBJECTS.

RT = was 4/5x more than normal. Didn’t improve. Couldn’t process motion because of a “perceptual slowing” and; 2. TEMPORAL RESOLUTION = her critical-flicker thresholds were normal e.g. could see the change between high and low luminance flickers. Didn’t just see a merge of the two. Could determine the order of stimuli.

COULD PROCESS NON-MOTION = V1 IS FUNCTIONING. IT WAS MOST LIKELY CAUSED BY V5 DAMAGE+.

HAD A SMALL IMPAIRMENT OF MOTION DETECTION. HER DISCRIMINATION OF MOTION WAS EXTREMELY IMPAIRED. GM THEsHOLDS WERE REALLY HIGH. E.G. THIS MEANS THAT HER LM WAS WORKING BUT SHE COULDN’T COMPARE THESE OUTPUTS.

YOU ATTACH SMALL LIGHTS TO A PERSON’S JOINTS AND GET THEM TO DO A TASK.

SHE KNEW WHAT ACTIVITY THE PERSON WAS DOING E.G. WALKING. COULDN’T TELL WHAT DIRECTION THEY WERE DOING IT IN.  SHE WENT TO CHANCE-LEVEL WHEN THEY PUT A SMALL NO. OF STATIC DOTS ON THEM.

D = V5 AND PROCESSES VIEWER-BASED MOTION E.G. THE OBJECTS MOTION RELATIVE TO THE VIEWER E.G. GM, DIRECTION AND SPEED. V = V4 AND PROCESSES OBJECT-BASED MOTION. E.G. HELPS TO IDENTIFY WHAT AN OBJECT IS E.G. BIOLOGICAL MOTION.