Study Guides (247,954)
United States (123,256)
Psychology (557)
01:830:301 (10)
Singh (7)

01:830:301 Final: UNIT THREE STUDY GUIDE

42 Pages
Unlock Document


UNIT THREE STUDY GUIDE Lecture 19:Attention (U3:1) -the video of the gorilla approx 50% of the subject never noticed the gorilla. -you are trying to pay attention to white and the other team you’re trying to ignore is black shirts  the gorilla is black.  SELECTIVEATTENTION Selective Attention: when you’re focused one type of stimulus, it’s harder to focus on other stimulus. -same thing happens when you’re talking on the phone and driving  doesn’t matter if you’re using hand-held device or microphone  your attention is on the conversation -subjects are HALF as likely to recognize relevant signs and objects in the “dual-task” condition -dual task condition: talking on phone and driving at the same time. -listening to audiobook doesn’t disturb but if you are engaged in something, it affects -Inattentional blindness: when you are inattentive, you are blind to the gorilla as well as objects on the streets. -attention is split with driving and phone (it has nothing to do with what you’re hands -are doing) -people have half of attention  ability to perceive relative objects, signs and how quickly you notice. (if you’re listening doesn’t affect but when you’re engaged in the conversation) Attention (famous quote by William James) “Everyone knows what attention is. It is the taking possession by the mind, in clear and vivid form, of one out of what seem several simultaneously possible objects or trains of thought. Focalization, concentration, and consciousness are of its essence. It implies withdrawal from some things in order to deal effectively with others..” -your ability to attend other things are compromised DividedAttention: divided attention between 2 things (ex: dual task) OvertAttention: directing the eyes and attention to a stimulus (obvious, explicit) -if I want to attend to you, I’m going to direct my eyes to you. Covert Attention: directing attention to a stimulus, while the eyes are visually fixated somewhere else -if I want to attend to something, but not actually look at that direction (peripheral) -you’re in a subway and there’s a crazy guy, you don’t want to look directly but still look at LIMA1 that area/direction -you can train your attention to move independently from the fixation. SelectiveAttention:when youareselectingonestimulustoattendtooutofmany(mayinvolve looking directly at it or covertly  where you’re attending to) Some metaphors: 1. Spotlight attention: putting a “spotlight” on things we are attentive to (dark room, direct spotlight in the room  directly that spotlight) -it misses some part: it can’t get narrow or detailed 2. zoom-lens: zooming in on things (you can zoom narrowly, spotlight might not capture that but a zoom light can) -it is possible to have your attention split in two areas that are not connecting Why have visual attention at all?  we can consider the evolution of visual organism (species evolving vision) and why not just have their brains process the entire image on the retina of the entire scene all at once?  Why have this extra mechanism of having allocation of specific attention? 1. Limited capacity in information processing (focusing on a specific) Strategy 1: all information all at once all the time (always) Strategy 2: spotlight of attention is so narrow and short so that you are completely blind (no consciousness) to anything that doesn’t fall within your focus of attention How has evolution discovered a balanced strategy that falls somewhere in between these two extremes? 2.Attentional “Bottleneck” -always processing the entire scene, all at once OR narrow spotlight and sharp, so anything that doesn’t fall in the spotlight is “dark” -how has evolution evolved so that we can find a balance? -initial stage: left of bottleneck, the entire image is processed by the ganglion cells (edges) -moresophisticatedprocessingthroughtheattentionalbottleneck(whateveryou’re paying attention to  only that part of information will then get further more sophisticated processing)  “SELECTIVE” 3.Asubset “selected” for further processing. LIMA2 VISUALSEARCH -how can we experimentally distinguish between those two types of situations (something that gets processed before the bottleneck –all at once OR something that only gets processed after the bottleneck – only if you’re attending to it)?  visual search Visual Search: well-known technique that distinguishes these two Two category: 1. some information is processed AUTOMATICALLY or “Pre-attentively” (before you chose to attend to it; left of bottleneck) 2. other information requires attention (meaning that other information will only get processed if you are attending to it and will not get processed otherwise) Q: How can we tell whether some information processing will fall on 1 or 2 categories? -Subjects look for a target (what they are supposed to be looking for) among distractors (not what they’re supposed to be looking for) -Measure Response Time (RT): how long does it take people to respond to tell you yes/no  how long it takes for them to make that judgement; how long does it take as you systematically increase the number of distractors. Key question we are asking: How does the RT vary with number of distractors? (Depending on that pattern of increase or not, we can make some discriminations) Based on the pattern of research, we can divide visual search into 2 type of categories: 1. Parallel Search: the entire image (all parts) are being processed as parallel -example: the tilted bar example -example: Is there a red bar -it’s not going to take longer  for some properties, RT are almost flat (doesn’t go up)  those properties are called “Basic Features” (color, the orientation of a line are both basicfeaturebecausethebrainapparentlyconsidersthesetobeimportant,fundamental properties that they are processed as parallel EVEN before attention) LIMA3 x-axi: # of distractors y: reaction time Twotypes oftrial: trials wheretarget was present or absent.  the slop of the line is almost 0 (flat) -slope is how steeply the line goes up as we increase the number of distracters. -the small dashed line is the RT slopes  (= ~0 ms/item)  why is ms/item the right unit?  on average, when you increase/add 1 more distractor, on average how much does the response time increase  it’s milliseconds -the two curves are BOTH close to having 0 slope -because these visual properties are often called basic features  it is also called feature search because you’re being asked to search basic features. -“Feature Search” :same as parallel search  way to find basic features (visual system considers them important enough to be processed right away) -basic features: simple building blocks for building larger, more complex shapes -feature computed over the entire image in parallel (opposite of parallel is serial  serial meaning one at a time) -property like color or line orienatation, the feature is computed over the entire image in parallel (meaning processed all at the same time)  you don’t have to wait – Assumption: the search does NOT require any attentional resources (it is a “pre- attention”  left of the bottleneck)  way to find basic features 1. Color 2. Line Orientation 3. Curvature -not goingto slowed downbythenumberofdistractorsor thelocationofthesefeatures  just as fast, because the visual system is processing the entire image all at once (you’re going to find it just as quickly) 2. Serial Search (opposite of parallel search) -RT increases with the number of distractors. LIMA4 Graph: 1. Slope of both lines 2. Slope of target present is much steeper than the target absent line. -target present: 20-30 ms/item (for every distracter, one average, it would take you an extra 20-30 to tell) -target absent: 40-60 ms/item Slope for target absent = 2 x slope for target present Ratio is 2:1 BasicAssumption: serial search  Each item needs to be scanned individually(cannot process all of the items of the display all at once) -another word for serial is SEQUENTIAL -This is called Serial, self-terminating, search  once you find the target, you can stop the search and give the response Slope for target absent = 2x slope for target present If you have 10 items and you’re looking for a target item and it’s random where the target item nd rd is and where you’re going start searching  you can get lucky (have it in the 2 -3 item) or unlucky (9 -10 )  and on average, you would have to search/go through half the number of items, before you come upon the distractor. -target absent: need to scan all items (10) -target present: on average, scan half of the items (5) CONJUNCTION SEARCH (conjunction means “and”  somethingAND something else) -simply means, I ask you to search for an item, but it’s not defined by one single feature -In Real-life searches are not defined by a single feature (if you’re picking up your friend from the airport, they’re not defined unless you tell them to wear red  it would then turn it into a parallel search but unless you do something like that, your friend’s properties (shape, height, etc)  so when we’re searching for objects in everyday life, they are complex objects and it’s not determined by one single property. -at the grocery store, you are trying to find red peppers but there are a lot of red fruits and veggies  the shape and colors are not going to help.  You are looking for a COMBINATION of Red and the shape of the pepper. LIMA5 -is there a green T? is there a red T?  People are slower to find it than parallel search and people are faster to find this than the serial search. -target defined by conjunction of two or more features -slopes between parallel & serial search -one feature is used to “guide” the search -find a red vertical bar: 3 distractors and 1 target 1. blue vertical 2. blue horizontal 3. red horizontal -only one of these four types is the target Set Size: the total number of items on display (the number of distractors + 1 )  almost the same thing (whether or not you’re including the target in your display) 1. slope of 5-15 ms/item (target present condition) 2. slope of 10-30 ms/item (target absent condition) -as we noticed, these slopes are somewhere between the two extremes that we saw with the parallel search where the slopes were close to 0, and the serial search where the slopes were between 20-30 ms/item  slopes are between the two extremes -why?  because there’s no way it can be as fast as parallel search (because you have to do more work) but at the same time it’s not going to be as slow as serial search (because once you’ve picked up a red item, you can pick out in the red vertical line)  that is why response time fall on the two extreme. -slower than parallel search, faster than serial search. BINDING PROBLEM -these basic features are processed/computed pre-attentively BUT they are not “bound” to any objects or locations  What does it mean to say that the feature is not bound to any oject or location? -two different image, but if there are the target, it would be the “same” even though the picture is different. -if all we ever had was basic features, they would all be the same  so you have to combine features (it has to make explicit the fact that the object and location needs to be binded to the LIMA6 basic features  the red bar needs to also be vertical) Lecture 20: Binding Problem Basic feature are computed pre-attentively BUT are not “bound” to any object or locations -binding requires attention -insufficient processing time  illusory conjunctions Cueing paradigm Red box: peripheral cue Valid cue: 80% of trials Invalid cue: 20% of trials Red arrow: symbolic cue -faster responses to a valid cue -cue benefit = RT (invalid) – RT (valid) -Effect of SOA? (= Time 3 – Time 2) LIMA7 Cueing Paradigm -benefit builds up more slowly for symbolic cue Object-Based Attention -what gets selected?  a location in space or an entire object? Cued location Egly, Driver & Rafal (1994) 80% 10% 10% Valid Cue Invalid Cue Invalid Cue Same Object Different Object -benefit of valid vs. invalid cue -benefit of same vs. different object RT (same) < RT (different) -more boundaries to cross? -object benefit still obtained Duncan (1984) – Ss better at judging 2 properties of an object vs. 2 properties on separate objects LIMA8 Which of the two vertices is higher? (Baylis & Driver – 1993) FFAactivated  attend to faces PPAactivated  attend to houses Hemi-Neglect -lesion in right parietal lobe -neglect in contralateral visual field -line bisection -line cancellation -Tipper & Behemann (1996) LIMA9 Lecture 21- Scenes: Picture Memory -extremely good memory for pictures -98% accuracy with ~600 pictures -85% with 2500 – 10,000 pictures -visual memory with these pictures is extremely good -look at image as a whole and not remember -scenes and individual objects Scene Perception - we can see the world in its entirety in rich visual detail - the rich visual detail we experience may be an illusion  in our consciousness - good at quickly extracting the gist and spatial layout of a scene Gist – main idea Spatial layout – if you have a mountain scene, you have a very different sense of space vs sitting in a small classroom -how structures and objects in the scene lay out -memory for pictures are good -we fail at change detection tests  are not very good at details and individual objects -attention is object based -if you have to split your attention between 2 objects or multiple objects you are bad at attending to all, good at extracting the gist and space of the scene, but not all at the same time -reason why we’re not very good at change detection test - bad at attending to many objects at once - in everyday life, the world can serve as a memory -really good at remembering pictures but bad at saying what changed in a photo  for example, if I’m looking here and I don’t remember what was going on in this part of the scene, unconsciously I would move my eyes and see what was happening, we don’t need to store in memory in detail Scenes: Change Blindness Importance of intervening blank screen Scenes: Change blindness -the bush disappeared in this pic -the engine of the plane has disappeared LIMA10 Motion Perception The visual disorder complained of by the patient was a loss of movement vision in all three dimensions. She had difficulty, for example, in pouring tea or coffee into a cup because the fluid appeared to be frozen, like a glacier. In addition, she could not stop pouring at the right time since she was unable to perceive the movement in the cup (or a pot) when the fluid rose. Furthermore the patient complained of difficulties in following a dialog because she could not see the movements of a face, and, especially, the mouth of a speaker. In a room where more than two other people were walking she felt veryinsecure and unwell, and usuallyleft the room immediately, because “people were suddenly here or there but I have not seen them moving.” The patient experienced the same problem but to an even more marked extent in crowded streets or places, which she therefore avoided as much as possible. She could not cross the street because of her inability to judge the speed of a car, but she could identify the car itself without difficulty. “When I’m looking at the car at first, it seems far away. But then when I want to cross the road, suddenly the car is very near.” She graduallylearned to “estimate” the distance of moving vehicles bymeans of the sound becoming louder - patient with motion agnosia (akinetopsia) Motion Perception -motion helps to: 1. Draw attention 2. Segment objects from background -if you have a tiger hiding in the bushes, because their shape is broken into pieces by the bushes but as soon as the tiger starts to move you can see the outline of their form 3. Relative depth (motion parallax) - ones that are stationary are farthest away, vice versa (the stars that are moving more are closer to us, the ones that are stationary are farther from us) 4. 3D shape (kinetic depth effect) – as soon as you put the dots in motion, you see a rotating 3D structure, but if you stop the motion you only see dots -motion provides information to our brain about 3D shape of an object 5. Object recognition in impoverished displays -impoverished 0 if you’re in thick fog  impoverished condition -Especially useful for animate objects LIMA11 Biological Motion: to receive recognition of motion of inanimate objects -white dogs against the black screen, you don’t have any depth cues, etc -Johansson “Walker (1979) : minimum information for perceiving biological motion -based on point-light motion observer can tell: -sex of walker/dancer -action/kind of dance -identity of a friend -kind of animal Motion: Basics -motion has two components: -Direction -Speed (=distance/time) -how does directional selectivity emerge? -Recall: complex cells in V1 are direction-selective How does directional selectivity emerge? Reichardt Model Alpha and beta are two locations Photoreceptor at location alpha and location beta on the retina Retinal surface DS is a directionally selective neuron -both of these guys are sending their signals to the DS cell Time Delay -one thing, you have to do in sending the signal to the DS cell  you have to add Directionally- selective cell -one connection from one photoreceptor from the alpha, you only want to introduce the delay from one of those two connections -if something is moving across the retina, at time T1, the image of that object falls on a certain location at location alpha -sometime later at time T2, it is going to project its image onto location beta LIMA12 -that object or image of that object takes some small amount of time in order for its image to move from location alpha to location beta -cells fire only when it receives inputs from both simultaneously -it’s an “and” operator  need BOTH to fire at the same time before it will fire Reichardt Detector time T1 time T2 T1: light at position α -signal delayed for delta t T2: light at position β -no delay -time that it takes light to move from alpha to beta should Cells fire only match the time delayed, which will ensure that the delayed inputs from both simultaneously and non-delayed signal from beta will BOTH arrive simultaneously -respond to motion from left to right -we have achieved directional selectivity -if stimulated at the same time  no response -if β is stimulated before α  no response -does NOT respond to motion from right to left Move delay to other connection  Reverse directional selectivity -ifwemovethetimedelayfromthealphaconnection to the beta connection that will reverse the preferred motion detection -logic is same as before. LIMA13 Lecture 22- Correspondence andAperture Problem Review Reichardt Detectors Responseprofile of the cell DS given what wesaid for theprevious time, what would change in the way that it responds? (for both signals to arrive at the same time)  the light would have to move faster (because the light has to travel greater distance in the same amount of time) -if you’re keeping the delta T the same, but increasing the distance, then the light/object that crosses the retina would have to move at a faster speed. -you would have by altering the distance, effectively create a DS cells that responds to faster motion than it did before  (DS cells are selective to particular direction of motion and also a particular speed because the time difference (T2-T1) that it takes to get from alpha to beta has to match the time delayed. -if you have a greater separation then the light has to travel a larger distance in the same amount of time  you altered the DS cell to respond to faster speed. -greater separation between α and β  respond to faster motion Remember: Speed= distance/Time  decrease deltat  responds tofastermotion -if you’re not changing the distance, but decreased the time delayed  it would respond to faster motion because light would have to cover the same distance in less time therefore the only way you could do that is if it’s moving at a faster speed than before. -There are d ifferent cells that are selective for: -Different d irections of motion -Different speeds -across the population of these different cells, between them, they cover all possible speeds and all possible directions of motion -if you’re looking at an image that is moving, there are going to be some cell that matches that preferred motion and speed and therefore, that certain cell is going to respond. APPARENT MOTION Apparent Motion : you don’t have continuous motion, but it’s a bunch of static frames that have a blank frames in between. -frame 1 the dot is on the left side, frame 2, the dot is on the right side and in between are blank frames  this is an example of animation/movies (number of LIMA14 frames running on the projector - 30)  Each frame is a static frame , but if you have a sufficiently large number and speed perameters are correct, we’re going to see a smooth motion even if though it’s just frames -the dot only occupies one of t wo positions (the left or the right  the dot never takes immediate positions ) -Inter -frame interval : how long you show the time delayed in frame 1 and frame 2  determine the interpretation that you get from the same frames and what motion you’re going to see -video ex: it looks like the perceptually it takes position somewhere in between  apparent motion  looks like it’s a con tinuous motion but it’s just a static frame Correspondence Problem -Ternus Effect -two interpretations: the same two frames can lead to different possible interpretation s based on what we call the correspondence (the way in which our brains sets up the correspondence to the dots  this is called the Ternus Effect ) -Which element in Frame 1 corresponds to which element in Frame 2? -depending on how the brain solves the correspondence problem of this mapping, we’re going to see one of two different possible motion interpretation 1. Group motion: see all three dots moving back and forth as a group. (what is the notion of motions) 2. Element motion: only one of the dots/elements moves  (how it affects) If you alter the correspondence of the two frames , it’s going to alter the motion interpretation that you’re going to see. -any time you have more than one element in the image in each frame, there’s goingtobe morethanonewaytoset upthecorrespondence inframe 1 and 2 this is what we call corres pondence problem (which element in frame 1 corresponds to which element in frame 2) -depending on how the brain solves this correspondence problem (how it sets up), we’re going to see one of two different possible motion or interpretations of these motions . Example of two balls that can be seen as going vertically or horizontally across -Default preferences for SHORTER motion paths  proximity (preference for LIMA15 grouping the shorter distances) -why default? Because we can bring other factors in the picture and violet the preference for shorter Role of Context -Picture of a boy with arms crossed 2 Possible paths: -3/4 turn; biologically possible (longer) -1/4 turn; biologically impossible (shorter) -is the brain going to still prefer the shorter path ev en if it’s impossible? Or is it going to take into account that one path is biologically impossible and take the longer but the more realistic interpretation? -Both are correct  depends on how long the blank frame ( inter - frame interval -IFI ) is in between. -If the IFI is very tiny/short  people perceive the shorter pa th even though it’s impossible -If the IFI is longer (have enough time between the frames)  people see the longer, realistic path -depends how much time there is between th e two frames, the interval between the two frames  if there’s not enough time, the brain takes the default, short path but if you give more time, that is factored and taken into account and have more realistic path. -Picture of a lady (Shiffrar & Freyd 1990) -one possibility is that the arm is going through the arm – impossible (shorter) -one is that it goes under the arm – possible (longer) -Same as the boy picture  if the IFI is short, you see the impossible and if you increase the IFI, then you sta rt to see the physically possible interpretation even if involves longer trajectory of motion. LIMA16 Motion Perception Recall : measuring motion requires determining : -speed -direction -how the brain deals with motion and if neurons in the brain want to measure motion and it requires two variables  speed and direction -contrast borders or edges are more realistic  if you have a complex cell in v1 and it has a receptive field and a cont rast border that sweeps across the receptive field, how does that cell/brain know which direction the contrast border was moving in? How doweestimatespeedanddirectionwhenallwecansee ofmotionis whatever is falling through the small aperture? (wh y are we talking about small aperture?) -small aperture: each neuron of a receptor field can only see a tiny part of the image  that particular neuron can only see whatever part of the image that is falling in its aperture Aperture Problem Recall: comple x cells in V1 are D irectionally Selective (DS)  small receptive fields (each neuron is only able to see a tiny part of the image , not the whole ) -which direction is that straight line border moving in?  Can we tell which direction that straight line border is moving in (under the assumption that only thing you can see is whatever that falls inside the receptive field )  if you see the endpoint you can see which direction, but if everything outside is cov ered up and whatever yousee is insidetheaperture, can you stilltellwhat directionit is moving in? -if all you can see is an edge moving through a receptive field, you can’t tell which direction. -all 5 are consistent -aperture motion (the moti on of the straight edge viewed through the aperture) is ambiguous (no way you can tell what the true motion is) -MT needs inputs from V1 in order to take motion interpretations -at the level of v1, you cannot solve this problem  you can solve it at a higher up in the MT LIMA17 Aperture Problem (picture of circles) -perceived motion would be the same for A, B, and C (because the motion inside the aperture looks exactly the same even though the actual motion of the bar is different) -but the actual motion is differe nt (going straight, up and down) Aperture Problem (moving image of red and green) -one is going horizontal and one is going vertical -the green: what you can see inside the aperture -the red: what falls outside the aperture -in this case, everything is moving from left to right and in other case, everything is moving from top to bottom -but if you just focus just on the white circle with green part , there is no difference at all as far as the bars motion is concerned inside that aperture. a. Barber -pole Illusion -the pole is rotating (horizontal motion) but we see it as the stripes are going up (vertical motion) -where is there an unambiguous signal for the direction of motion? -we can either look at the middle of the pole or the edges of the pole -if we look at the middle of the pole, we don’t get the unambiguous signal  we face the aperture problem  Motion in the center is ambiguous (not hepful) -if we look at the edge, the T -junction is moving up  motion at the edge is unambiguous and it’s going to tell us that it’s going up , predict the direction (vertical)  this is why we see the barber pole as it is moving up Role of Terminators / Shape ofAperture -if you look at terminator (t -junction) of the circle  it’s going down obliquely -if you look at the horizontal line or the “Z”  they’re moving horizontally -if you look at the vertical line or the “Z”  they’re moving vertically -even though motion is ambiguous, if you track these terminators, you can LIMA18 predict what motion direction a human observer is going to perceive. Motion Integration -two separate motions -one is going horizontally to the left -one is going vertically down -what would happen if we put those two together  diagonal to the left 1. Component motion : you see two independent motions (you can see one through the other  like a window frame through a window frame ) 2. Coherent motion : the two motion signals are combined into a single global motion (you perceive a SINGLE direction of motion) When you do see coherent motion, how can we predict (b ased on the two original motion direction ) what the direction of the coherent motion would be? -Vector Sum: if you have two vectors (Aand B), you can get the sum (A+B) -horizontal motion is faster than the vertic al motion  it’s going to be diagonal but it’s going to be closer to be horizontal Chopstick Illusion -you can see either as tied together or independent Lecture 23-Motion Perception and the Brain Review: 1. Motion Integration-two separate motions -superimpose the two -possibility 1: when you integrate those two motions into a single motion (Coherent motion – when you integrate the two motions) -possibility 2: two separate motions  like window panels, you can see one through the other (Component Motion – you don’t integrate motion into a single motion) 2. Chopstick Illusion: -the two displays are identical in every way, but the one has an occlude that makes it look like it is tied to each other into a single object (Coherent motion – the 2 bars move together) LIMA19 -without the occlusion, you see a component motion – the 2 bars move independently This is reason why motion integration is important (the example of two bars – one looks like an X going up and down, one looks like two bars going horizontally past each other) -Why do the occluders matter?  if you have occluders, then you don’t treat the edges as the end of the black bars  they are perceived as continuing (we don’t know where they end)  but without the occluders, it is perceived as the whole object Aperture problem MOTION PERCEPTIONAND THE BRAIN -complex cells that respond to motion in V1 have small receptive fields (each neuron can only see its aperture)  you’re going to face this huge ambiguity (you can’t tell if the motion is one direction or another direction) -we as human observers don’t face this ambiguity (we only see motion in one particular direction)  the motion we see -v1 by itself is not able to solve the aperture problem (motion is ambiguous because of the small receptive field  so they can’t determine uniquely which is the right motion) Example: two diamond shapes with these arrows are telling you which direction it’s moving. (they are moving in horizontal, opposite directions) -If you’re a v1 neuron, you have a small receptive field and one of the little circle is your receptive field  it won’t be able to tell a clear pronouncement because of the ambiguity(itmightinferthatthedarksquareis goingdiagonallydown,whichiswrong, because it’s going straight to the left) -if our brain does not integrate information across those neurons by themselves, not a single one of them will give a correct direction of motion. -aperture problem is only a problem is it’s a straight edge  if it’s a corner, you don’t have a problem (no aperture problem) -thus V1 cells CANNOT determine object motion EDGE MOTION vs OBJECT MOTION -these local motions signals must be integrated across multiple v1 cells in order to perceive unambiguous object motion -ex: two diamonds  the green arrow is showing you the actual direction, but if you have a receptive field placed in the red circle, it might look like it is moving diagonally upward/downward. LIMA20 -what if there was a higher level area of brain that was able to receive signal from both of the twov1neuronanditsreceptivefields(thetwocircles)Wouldthathelpresolvetheambiguity? YES -the signal that you’re getting is that one is going diagonally upward and one is going diagonally downward  then you can constrain what that direction that must be IF YOU COMBINE the two information -this integration is done inAREAMT -ex: the previous two gray diamonds  if you want to combine the two receptive fields from the dark gray diamond, it would be helpful  if you combine the one receptive fields from each dark and light gray receptive field, it WON’T be helpful and it would make it more ambiguous -if two edges are separate objects it is not a good idea to combine them -which edges belong to the SAME object and which edges belong to different objects. -if you integrate two motion signals from the same object, it is helpful to see which direction it is going -in order to know which motions to integrate  the brain has to know which edges belong to the same object -focusing on T-junction would not work either (it looks like it’s going up and down) AREAMT -motion blindness (can’t see smooth motion) had damage from a stroke in the MTarea  some evidence that MT is the motion area in the brain -all cells in MT exhibit directional selectivity (cells only respond to motion in one particular direction – each cell might have its own preferred direction; it’s not that they don’t respond completely if you change direction completely, but they will respond weakly because it’s not their preferred direction) -much larger receptive fields than v1 cells  this is significant in this context because you are able to view a larger view of image (see corners, edges) -an MT cell receives inputs from multiple v1 cells -cartoon ex: single empty neuron with a large RF (right circle) – it’s receiving multiple inputs from multiple v1 cells  this neuron MT cell is in position to seeing all the ambiguous small receptive fields in the V1 cells -Neurons in v1 analyze the LOCALmotions of contours/edges  whereas neurons in MT respond to the GLOBALmotions of objects. -what happens if nothing in the world is moving and if my eyes are sweeping left to LIMA21 right, what happens to the images on the retina?  SENSITIVITYTO MOTION When we wanted to study the eyes sensitivity to light, what psychophysical tool did we use? How did we go about determining the eyes’sensitivity to light? -Absolutethresholdforseeinglight:minimumamountoflightneededtoreliablyseesomething -in order to measure sensitivity to light  we measure their absolute threshold (if I have light that is physically present, but if the light is so dim, then it is below threshold)  when you increase signal in some level, people will just start to see it. -once I have theAT, how does that tell me about sensitivity  Sensitivity = 1/AT -lowAT = high sensitivity; highAT=low sensitivity Motion Threshold -we want theAT  it’s not feasible to say the minimum speed detected  Ex: three cirlces -A: No correlations – all dots are movingin random direction (like dust particles in air) -B: 50% Correlation – half of the dots have the same motion direction and the other half is going in random directions. -C: 100% Correlation – all of the dots are moving in the same direction. How should we measure threshold?  What is the minimum amount of correlated motion that you need before people will see correlated motion in one direction -Task of the subject: 2AFC (2 Alternative Forced Choice)  only two choices: is the motion going up or going down? (If the correlation is 0
More Less

Related notes for 01:830:301

Log In


Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.