Cognitive, Attention, WEEK 3

Question 1

What is attention?

Answer 1

• Attention is the way in which we select one thing to be aware of out of what may be a large number of things to be aware of at a given time
• When looking at how attention is addressed across cultures, (e.g. English: pay attention, Hungarian: dedicate att, German: gift att) there is a theme as though attention is a resource of ours which we distribute into our environment > we are giving our attention, we are giving our resource out into the world
• Attention is important in the world as if we didn’t have it our conscious exp of the world would be overloaded because there are so many things happening in the world that we could not consciously experience all of it
• Attention reduces this overload into a serial stream of consciousness from what is happening in the world (e.g. in where’s wally? There are lots of things to look at but attention narrows our focus)
• As attention reduces the overload, att determines what we perceive

Question 2

Change blindness

Answer 2

• Change blindness is when observers fail to recognise large changes to objects or scenes when the change coincides w/ a brief visual disruption (Simons & Levin, 1997)

Question 3

What is inattentional blindness?

Simons & Chabris, 1999

Answer 3

• We overestimate how much of the world we are actually aware of > we can miss very salient things in the world > we regularly miss important events in our EV because we are focussing on something else
• E.G. the Gorilla study (Simons & Chabris, 1999) > ppts are shown a video of a black team and white team throwing a ball + had to count how many times one team passed the ball > in the video a person in a gorilla suit walks through > ppt tend to not have noticed the gorilla due to attention on the team
• Two video styles: transparent is where the teams and the unexpected event were filmed separately while the opaque is where it was all filmed together (more realistic, avoiding bumping into each other etc)
• Hard condition had two tasks as they had to keep 2 separate numbers in their mind + look for them

Question 4

Gorilla study (Simons & Chabris, 1999) results:

Answer 4

• Results show that inattentional blindness can easily be induced as although 70% of people saw the monkey in the opaque easy task, 30% did not, which means inattentional blindness had been induces
• IB is more easily induced in transparent conditions than opaque > makes sense as in the transparent condition, it would be possible for the gorilla to walk through other people while in opaque, people have to move out of the way > may make it harder to miss the gorilla
• IB is dependant on the difficulty of the task > the more the primary task needs your attention, the less likely to see the salient event > attention is a limited resource > In the hard condition, attention has been distributed more widely as they have 2 tasks, so less attention is available to notice the gorilla while in the easy task, more attention is available

Question 5

Central Capacity Theory: Kahneman, 1973

Answer 5

• Idea is we have a single central capacity aka single pool of resources (attention, central executive) that we distribute among various tasks > if we have more than one task to complete, we get more costs to some of those tasks as attention gets used up by other tasks (as there is a single pool shared between competing tasks > this resource is limited)

Question 6

Evidence supporting central capacity theory

Answer 6

• Evidence support central capacity comes from experiments using a car simulator > first do one task where you need to drive safely + brake quickly when seeing brake light or dual task condition where you have to do the same task but talk on the phone at the same time
  Results: ERP’s
• The amplitude of the ERP is much smaller in the dual task condition than in the other condition > this is a measure of how much the brain is processing the brake light, meaning in the dual task condition, the brake lights are processed much less.
• They have less neural resources dedicated to them (processing brake lights) when they are doing another task at the same time (on the phone) in comparison to when they are doing one thing + using all neural resources for one task.
• Attention is limited + distribute across tasks > due to this, when we do more than one thing at one time, each of those things we are doing will suffer + less processing of stimuli under those circumstances.

Question 7

What is the attentional blink?

Answer 7

• Attentional blink is the idea we can make something invisible if we show it very quickly after showing something important to them > discovered by Jane Raymond and Kim Shapiro

Question 8

How to test the attentional blink

Answer 8

• Have to present stimuli very quickly at around 10 Hz (a Hz is how many times something happens per second so present 10 images per second to make one invisible)
• Ppt has to do two tasks > they have to look out for 2 targets + report if they had seen it (first target is called T1 and second called T2) > have to remember something about targets like what they saw or if they saw it or not
• There also needs to be distractors/”masks” > items irrelevant to the task + happen between targets
• We can move when T2 is shown relative to T1 and this helps us measure how much time is left for processing of T2 based on its position relative to T1
• These results show the position of T2 relative to T1 so how long after T2 was shown > in the dual task where ppt had to look for T1 + T2, we can see when they are shown T2 very soon after T1, accuracy was low and there is a big dip around 270ms > this is the attentional blink
• Only around 50% of ppt report seeing T2 even if it happened > means we can make T2 invisible half the time if ppts are performing a task with a target that happened around 300ms earlier

Question 9

300ms in the attentional blink: why is it special?

Answer 9

• There seems to be something special about 300ms as we see a dip in graph but it then goes back up, also the accuracy was quite good at even 90ms so it is not just about speed > if we show T2 immediately after T1, then they are able to see it, if it happens 300ms later, they are worse and by 600ms this is much better
• This dip where we show T2 300ms after T1 is our attentional blink

Question 10

Event-related potential: N400

Answer 10

• Whenever we see a stimulus which has meaning, 400ms later, we get a negative ERP (N400) > N400 reflects processes to do with semantic processing (understanding the meaning)
• N400 can be used as a marker to show someone is processing the meaning of something even if they cannot tell us about it (brain activity can show what processes people are carrying out even if they don’t consciously know)
  Luck et Al, 1996: classic AB study but also looked at activity of N400 at the same time > can see the usual dip in A but in B where we look at N400, we can tell that the N400 does not care about the time differences even at lag 3 (300ms)
• Amplitude of N400 is the same across all dual task conditions > N400 doesn’t blink, only the person’s conscious experience + attention blinks > from this we can determine that the T2’s which ppt do not consciously remember or experience are still processed up to point of meaning (semantic processing)
• Things which happen around us, even if we are not paying attention to them or have a conscious experience, they are still processed up to point of meaning w/o us being consciously aware that it happened

Question 11

Interference Theory (Shapiro et Al., 1994)
(explanation of AB)

Answer 11

• Argues that AB happens because, we have a lot of distractor items between T1 + T2 and we have a limited amount of visual short term memory (temporal buffer) > your brain is trying to process all of these items rapidly and interference theory argues AB happens due to competition
• There is competition to retrieve whether T1 and T2 happened or not, to ignore the distractors, what were the items + which of them are T1 + T2 > competition from distractors cause the ppts to not be as capable as reporting accurately if they saw T2.

Question 12

Evidence for Interference theory

Answer 12

• If it is true that AB happens due to interference from other items, then AB should get worse if amount of items increase as more items need to go through the temporal buffer + can interfere w/ T1+T2
  Isaak et Al. (1999) reports that AB does increase w/ increasing numbers of distractors > as temporal buffer fills up, magnitude of AB increases and ability to detect T2 decreases as the buffer is overwhelmed by the large amount of intervening items.

Question 13

Unified model
(explanation of AB)

Answer 13

• Could have a unified theory that combines other theories on AB, finding similarities across them which may help describe the phenomenon itself
• Unified theory suggests because there is a mask following T1, you need to increase your attention to T1 to be able to process it > due to all the intervening distractors before and between T1, a lot of attention has to be paid to T1 to ignore the other items > this leaves less attention to process T2
• The amount of attention left to process T2 changes over time > after a certain amount of time, you no longer need to pay as much attention to T1 because you have completed processing it + attention can return and process T2
• T2 occurring at a long lag has more attention available for it to be processed than a T2 at a short lag
• Attention is a limited resource which gets used up to process T1 and deal w/ the distractors so it is not available to process things which happen very quickly afterwards > after 700ms, all the attention processing of T1 is done + have attention for processing T2

Question 14

The Cocktail Party Problem

theory of attention

Answer 14

Classic way of studying how we focus attention and when in the perceptual stream attention selection occurs
Cherry (1953) > Idea is in a cocktail party, or a loud you try listen to just one person (if you are in a party and there is a lot of sources of sound, you are still able to converse w/ the person next to you)
This is studied in dichotic listening tasks > dichotic meaning both ears so sounds are going into both ears simultaneously > then we ask the ppt to listen to one stream of speech + ignore the other stream
- 1/3 of ppts report hearing their name in the unattended channel (the speech you are not paying attention to) > ppts notice this even though they aren’t paying attention
- This suggests there may be a level of processing which happens of stimuli at an unconscious level > even though ppts didn’t pay attention to one stream of speech, they still heard their name which they wouldn’t be able to do if there wasn’t some level of processing in the brain
- Suggests there is an important difference between processing of stimuli when you pay attention to it vs processing of stimuli when you don’t pay attention
- The studies also show that paying attention to one voice over another becomes easier if the voices are physically different > supports bottom-up processing > if physical features of stimulus in the real world is similar, it is harder to ignore one over the other > bottom-up processing is to do w/ the stimulus which changes the way we experience things or pay attention

Question 15

Johnsrude et Al. (2013)

Answer 15

• Dichotic listening task > 2 speakers in each ear, one is a target to pay attention to while one is a distractor
• Target-to-masker ratio (dB) refers to how much louder is the target than the mask > correct responses refers to how often do ppts detect correct things the target says
• 3 conditions: Circle is where you know the target and you have to listen to them over the unknown masker. Square is where the masker is the familiar target + should be ignored > unfamiliar person is target. Triangle is novel where you don’t know either person (not familiar)
• We are better at detecting what someone is saying if that voice is of a person you know > supports top-down as we have experience hearing this person so we are better at hearing what they say even if someone else is talking
• Squares are higher than triangles so you can better detect the things someone you don’t know is saying if the person you ignore is familiar than if you don’t know either speaker > better at paying attention to something if the thing you ignore is something you have experience with
• This supports top-down processing because we are better at paying attention and ignoring certain things when it is something which we have experience with

Question 16

Attention as early selection: Broadbent’s (1958) Theory

Answer 16

• Argues we select the thing we are paying attention to at the earliest point in your visual or any sensory processing stream
• Broadbent’s filter theory has the idea that inputs from the environment enter our sensory register + there is a selective filter early on (attention). This filter selects a small amount of aspects of the input entering the sensory system to get processed further > only the things selected in this filter are then processed to the level of meaning > only process meanings of what you are paying attention to
• The sensory inputs are filtered based on physical characteristics > filtering prevents overloading of the limited capacity mechanism > inputs remaining in the buffer after filter are available for later semantic processing
• This idea works for basic bottom up processing idea from Cherry > unattended stimuli undergo limited processing before being filtered out
• Accounts for findings from dichotic listening tasks > filters select an input depending on it’s most prominent physical characteristic

Question 17

Evidence against Broadbent

Answer 17

• At least some inputs from the unattended stream are processed semantically (e.g. hearing your name in a conversation you weren’t paying attention to)
• Stimuli that people don’t report ever consciously experiencing can still change their behaviour > e.g. blindsight or T2, we may not consciously remember seeing T2, but our brain processes seeing it
• Suggests that attention doesn’t occur early because some things we don’t pay attention to still get processed to the level of semantics

Question 18

Attention as late selection: Deutsch & Deutsch (1967)

Answer 18

• Idea that everything entering our sensory system gets processed up to the point of meaning and then attention selects what is relevant for the response + processed to the highest level for your task in the world + produce an appropriate response

Question 19

Evidence against Deutsch & Deutsch

Answer 19

• Comparison of ERP’s of when people are paying attention to a stimulus + aren’t paying attention, activity is very different around 100ms after the stimulus appears > this is too quick to be a late selection, after 100ms, info has not been processed to level of meaning
• Suggests if ERP is different between attended and unattended stimuli, this suggests that attention doesn’t happen just at a late stage > these ERP’s support Broadbent

Question 20

Attention as flexible selection: Treisman’s (1960) Leaky Filter

Answer 20

• Idea is that if your processing stream reaches capacity of what it could potentially process in your sensory world, then attention applies > so if the stream reaches capacity early on then attention also applies early on
• Processing stream reaches capacity early when there are too many things to process in depth at an early stage causing attention to have to apply early on
• If processing stream reaches capacity later, perhaps if there aren’t many things to process at once, then attention also applies later
• This means attention can be flexible depending on the type of task the ppt is completing

Question 21

Covert attention: The Posner Cueing Paradigm

Answer 21

• there seems to be a flexible influence of both bottom-up and top-down processing on when attention selection happens in processing stream
• The paradigm supports this: sighted people can look at things they are not directly looking at > can pay attention to things in your periphery > this is an aspect of covert attention
• The task looking at covert attention (w/o moving eyes) works by doing one endogenous trial and one exogenous trial.

Question 22

The Posner Cueing Paradigm: Exogenous & Endogenous trials

Answer 22

• Endogenous trial: ppt has to stare in the centre of the screen, in the centre an arrow appears pointing right or left. Stimulus will also appear on either left or right side of the screen, in valid trials the arrow will point in a direction and the stimuli appears in the same direction Invalid trials are the opposite > top-down + goal driven because ppt chooses to use arrow direction to dictate where to look
• Ppts job is to press a button (L/R) depending on where they saw the stimulus
• Exogenous trial: Ppt has to stare in the centre but this time a X will flash on either side of the screen and after the stimulus will appear on either side of the screen > can be valid or invalid > ppt has to ignore X and press button of stimuli location > bottom-up as X doesn’t signify anything + attention gets pulled to stim

Question 23

The Posner Cueing Paradigm: Results

SOA = time between two stimuli

Answer 23

• Ppts RT’s are much faster in the valid trials compared to invalid trials (v RT= 430ms, inv RT= 460ms > consistent at short SOA (100ms) + long SOA (800s))
• Faster in an end trial when cue is valid as you covertly pay attention to one side + detect things happening in that side of space faster > slower in inv trials as limited att resource is being used for the other side of space
• Ppts RT’s are faster in valid trials than invalid ones if the SOA is short (100ms). But at longer SOAs (800ms) we are faster at invalid trials than valid ones
• Bc at short SOA, X drags our attention to one side + if the stimuli is presented quickly after X, our attention is still there so we detect it quickly
• BUT, at long SOA, our attention starts going back to the centre after 100ms of looking at the side where X was, while our att goes back to the centre we are quicker at detecting if something happens on the other side of space so faster in invalid trials
• The different results indicate two different att systems

Question 24

Where is the endogenous and exogenous system in the brain?

Answer 24

• Top-down, endogenous system is in more of the dorsal areas of the parietal lobe > area is controlled by peoples expectations + intentions + involved when central cues are presented
• The bottom-up, exogenous system is involved in the ventral regions of the parietal lobe + some of the ventral regions of the frontal lobe > Automatically shifts att, involved when uniform peripheral cues are presented, usually when stimulus is salient.

Question 25

Attention as a spotlight (Posner)

Answer 25

• Posner argues if top-down att is a limited resource, it may be comparable to a spotlight > i.e. we are looking in the centre of something, an arrow appears and our spotlight shifts to the area
  This brings up questions such as > what processing do things outside of the spotlight receive? If we are in charge of our endogenous attention system then how do we choose to distribute our attention to different things?

Question 26

Visual search paradigm

Answer 26

• Feature search paradigm is where the targets have a unique feature which is not shared by other items > target is different in a scene by one feature thus it “pops out” at us > e.g. the red A is surrounded by other A’s but because it is red, it pops out + is salient
• Conjunction search paradigm: is where’s wally? Target has no unique feature which is not shared by other items in the scene which makes visual search more difficult > requires combination of features to identify a target > e.g. it is more difficult to find the red A because the letter H now shares its red feature and there are other blue A’s > the combination of these two factors make the target unique + this combo is what the ppt looks for

Question 27

Results of visual search paradigms

Answer 27

• For feature search, regardless of how many distractors, the RT to spot the target is the same across conditions > is why it “pops out”, can’t help but immediately notice the A
• For conjunction search, the amount of distractors do matter as RT’s increase significantly when there are more distractors. RT’s for conjunction search is longer than for feature search even when there are few distractors
• Can be explained by feature integration theory > can do feature search quickly because you do it in the early parallel stage regardless of amount of distractors while conjunction search needs top-down attention which is slow because you need to pay attention to different features to perceive the object + serially check each object to see it is the target > hence why it takes longer in many distractor conditions

Question 28

Feature Integration Theory: Treisman (1988,1992)

Answer 28

• Argues there is parallel processing before attention is applied > in parallel processing, perceptual features are individually processed so if a target has a unique feature compared to it’s distractors then this feature “pops out” and you don’t need to pay attention
• If the target is conjunction of several features, then it cannot “pop out” because you cannot detect a combination of features in early parallel processing (as it isn’t one feature but many combined)
• The only way to find the target is to pay attention to each item individually + search them serially (look at 1, combine its features + decide if that item is what you are looking for then look at 2)
• If an object shares features with other objects, then it cannot be detected from a single perceptual feature and spatial attention is needed to search all candidates serially
• Essentially, object is only an object if you pay att to it > early visual processing is only about individual features, but to bind these features and perceive conjunctions which we do in everyday life, we need to pay attention to them
  Object > pre-attentive processing of object (if it pops out, don’t need att) > focused att to bind featues > perception of object

Question 29

Support for FIT: Illusory conjunctions

Answer 29

• need attention to be able to bind features + process conjunctions between them
• Illusory conjunctions show that sometimes people experience shapes + objects which are not really there and are the combination of other shapes and objects
• Illusory conjunction tasks have different shapes of different colours shown which have to be named by the ppt > the shapes flash for a certain amount of time then is blocked using a mask > ppt has to recall 2 numbers + one shape
• Results usually show ppts recall number but then confuse different features of other shapes to make up a shape which is not there (e.g. if there is a blue triangle and yellow circle, but because the screen flashes quickly you may perceive there being a yellow triangle even though there is not) > bound different features
• happens because we do not have enough attentional resources for that because the ppt is looking at the numbers on the screen + is on the screen for a short time > this combination uses up attention + causes conjunctions of things which aren’t there > need attention to process conjunctions between features.

Question 30

When do illusory conjunctions occur?

Answer 30

• When focused attention is absent
• When relevant stored knowledge is absent
• When spatial attention is diverted > looking for other things, attention is diverted
• When display is presented in peripheral vision
  Illusory conjunctions are not guessing! > people report what they feel like they saw
• can occur with high confidence
• do not occur under all circumstances that decrease performance
• Illusory conjunctions happen across space but not time! > if there is a blue triangle + yellow circle which appear very quickly one after the other, ppt do not experience illusory conjunctions > only happens if the two shapes are close together in the same location + space

Question 31

Evidence against FIT: negative priming tasks

Answer 31

• ppt shown primes, e.g. one would be a red dog with a blue hammer overlapped + then had to name the probe (dog)
• Ppt is looking for the red aspect of the prime + features in it, ignoring blue
• primes are shown before probes > attended repetition is where the same stimulus is shown as a prime + probe, attended semantic is where the stimulus shares a similar meaning but looks different (cat + dog are both animals)
• Ignored repetition is where you are to ignore the dog in the prime because it is blue + focus on the trumpet + ignored semantic is where you ignore the cat but focus on the guitar
• Faster RT for attended repetition + slowest for ignored repetition (blue dog)
  slowest RT for ignored repetition
• Slower to identify that a dog is a dog in the probe if just before you are presented with a dog or cat (which you weren’t supposed to pay attention to) > blue cat + blue dog interfere with how quick people can name the red dog > but people weren’t paying attention to the blue cat + dog, if you aren’t paying attention to this then you shouldn’t even know they are cats or dog under FIT, only meant to process the trumpet + guitar
• slows people down even more than the control where the table + guitar having nothing to do with the probe > suggests if we are slower to reacting to the probe after seeing the blue dog + cat, that must mean we processed them on some level > not paying att but still processed its meaning
• you can process combined features without paying attention to them and they can subsequently influence your behaviour.

Question 32

Strengths and weaknesses of FIT

Answer 32

+ An important contribution to explaining what happens within the attentional spotlight
+ Influenced thinking on a variety of topics from early sensory encoding to later attentional control
- Doesn’t explain why the similarity of distractors is influential
- Doesn’t explain how sometimes we can process things to the level of meaning w/o paying attention
- Patients w/ damage to their attentional systems do not seem to have individual damage to their conjunction or single feature targets > some patients are not good at feature search and conjunction search > but if FIT was correct, then patients would be bad at just conjunction search if feature search doesn’t require attention

Question 33

Guided Search Theory/Dual Path Model (Wolfe, 1998)

Answer 33

• Suggests simultaneous mix of parallel+ serial strategies for visual search > combines top-down + bottom-up
• Our top-down knowledge helps us narrow down what parts of the world we should be paying attention to when doing certain tasks + prior knowledge can make search more efficient (guided search)
• Argues we use a mix of parallel and serial processing from early on > we pay attention to what looks most likely to help the relevant task on the activation map > rules out certain aspects so attention can be concentrated (serial aspect)
• Suggests some kind of feed-forward process where we continuously update our activation map indicating which areas of the visual field are most promising + use serial, slow processing on these aspects
• Bottom-up effect > tells us what areas to search. Top-down effect > narrows down what areas are important > these combined helps us solve our visual search more efficiently.