Lecture 6+8: Attention Flashcards
What is attention?
- in the research world, attention has been defined many ways.
- “Everyone knows what attention is …. It implies withdrawal
from some things in order to deal effectively with others, and is a condition which has a real opposite in the confused, dazed, scatterbrained state which…is called distraction” - William James (1890) - Attention is best understood in terms of what it does rather
than what it is. - It requires withdrawal from one thing and the focus on another.
- Attention lets us focus on something, that is the thing it lets us do.
What happens when we do not have attention?
- Spatial (unilateral) neglect → attentional disorder
- Damage to the parietal lobes → the parietal lobe has regions that help direct attention
- Results in an inability to attend to information in space contralateral to the brain damage. Cannot direct attention to spacial information on the contralateral side of the brain damage.
Spatial neglect
- Inattention to information in ‘contra-lesional’ space
- Often following right hemisphere damage
- Right hemisphere is specialized for spatial processing - Attentional deficit presents across sensory modalities
(not just vision)
- Is not due to impairment in sensory processing
- defecit due to higher order attentional processing to a certain area
- If you show something to that unattended area, the areas will active but will not spread to other areass.
What are the deficits that come with spatial neglect?
- Left side of the world is out of awareness
- They read only words on the right side
- They eat from one side of the plate
- They can only describe half of imaginations and memories
- They will dress only one side of their body - Awareness doesn’t resolve the condition
- This attention deficit spans all modalities, all sensory information (vision, touch, smell)
- They are not aware that they are neglecting that side. If you point it out they will notice → will still not fix the problem.
- The locus of the defecit is about the person, the observer center.
What tests can we use to determine if someone has spatial neglect?
- Object drawing → will only draw half the object
- Line cancellation→ will only cancel half of the lines
Attentional Processing in the brain
- Spatial neglect indicates that attention is a brain mechanism
- Attention is not some specific area in the brain, it will engage a network of brain regions.
- A distributed network of prefrontal cortex (PFC) and parietal cortical regions (they work together)
- Directs processing for attended-to task
Top-down attention vs Bottom-up attention
- Different aspects involved in different forms of attention. For example, top down and bottom up attention recruit diffferent regions of the network.
- Intraparietal sulcus + FEF in preparing to attention t something (top-down attention) → voluntary top down attention.
- Temporo-parietal junction (TPJ) and VFC in bottom-up attentional orienting → capture your attention automatically = something in your external world grabbing your attention.
Types of attention
They all interact with one another.
Arousal
- Attention when you are very alert
- Optimal level for cognitive function
- you want some arousal so that you are engaged
- too much arousal = stress = impairs cognitive function and impairs you.
Top down attention
- goals and expectations determine attention
- when you chose what you want to focus on
Types of top down attention
-
Sustained attention
* Maintain focus on one input for a long period of time
* Vigilance
* Focus on a particular task for a long period of time -
Divided attention
* Shifting attentional focus between tasks
* Multi-tasking
* Rapidly shifting between two things -
Selective attention
* Focus on one input and ignore other information
* Attend to certain things and block out the rest.
Why do we have selective attention?
- Required because of limited resources
- You must prioritize what to process to act effectively
- What you attend to will depend on a given goal. Focus on what is necessary and filter everything else out. What is priority will change based on your task and goals (priority is dynamic).
- Spatial-based (focus attention on a particular space) versus feature-based (focus on a particular thing) attention
Experiment to measure spatial and feature attention (flanker task)
Give them a cue that will direct their attention/make them focus on a particular space.
* Stimuli can match the cue (congruent) or mismatch (incongruent)
* You are faster at detecting images when it is congruent.
Change blindness
- The failure to detect changes in stimuli in an attended
- It is due to the limits of selective attention (a failure of selective attention).
- Prevalent in every day life
zone. Even when we chose to attend to something, we are not processing everything that is there. Some things are missed. - Continuity errors in film: people do not notice differences across scenes
How do you measure change blindness?
Flicker technique paradigm
* Two highly similar visual images (e.g., scenes) are presented with an interstimulus “mask” (grey mask to disrupt selective attention)
* Sometimes there are small changes in the images (e.g., color change, removal of window of a building)
* When asked if the two images are the same, or what changed, people are often inaccurate
- people tend to not see the differences unless they are very big differences.
Theories of selective attention
- We filter out information when processing information
- Selective attention is lieke a bottle neck → we process info along one path and somewhere along the path there is a filter that only allows you to process some of it.
- Theories on how this works: Early selection models Attenuator model : Late selection model : Load theory
Broadbent’s early selection filter model
- You filter information at the level of perception, before information is processed for meaning (semantic analysis)
- Attented to information that meets your goal will be pulled forward to make its way to long term memory
- Early filter = filter out info early on when we are processing it → filter out distractions and irrelevant info before it is processed for meaning.
- We select info based on perception.
- Selects information for further processing at the sensory level (spatial location, frequency of sound)
- Attended information is processed for meaning, enters awareness and leads to a response
- Information not selected by the filter decays
Dichotic Listening tasks
- Present two simultaneous messages to each ear
- Participants are better able to recall information ear by ear than the simultaneous message
* In the ex. better remembering 2,5,6 and 8,4,1 than 2,8; 5,4; 6,1 - Information is selected for attention, at perception
Results:
* Asked to recall the digits either ear by ear or pair by pair.
* People are much better at remembering digits ear by ear.
* Remembering pair by pair forces you to move the filter from right ear to left ear to right ear which causes you to lose information.
Shadowing Task
Participants are given two messages: one to each ear. They are asked to pay attention to only one message and repeat it.
* People do not remember the content of an unattended message, but they might notice some sensory features. The meaning is not being processed but they can tell you somethinng about sensory or perceptual characterisitics.
- Can detect a new noise; gender of the speaker
* Evidence that unattended information is not processed for meaning
Problem with early selection filter models
“Name” experiment
- In certain situations, un-attended information can “break through” to our awareness
- At a party, you can attend to one conversation, yet hear your name if spoken in a non-attended-to conversation. Your name will draw your attention away from your current conversation.
- Did an experiment where they added your name in the message going to the unattented ear. Found that, 33% of the participants would hear their own name. This showed that some info might be processed.
- At a party, you can attend to one conversation, yet hear your name if spoken in a non-attended-to conversation. Your name will draw your attention away from your current conversation.
Problems with early selection filter models
- Participants presented with a word (e.g., apple) paired with an electric shock. Some words were paired with an electric shock.
- Next, they did the shadowing task with the ‘shocked’ word in
the unattended ear - Participants had increased skin conductance when the
‘shocked’ word was presented in the unattended ear. The words that has a shock will have meaning.- Sweat more = suggests that they did process meaning of the unattend words.
Treisman’s attenuator model
- An early filter dials down the influence of unattended material
- Some aspects of unattended material to be processed for meaning
- We do filter select information by the physical characteristics or the perceptual level
- all enters the sensory buffer and then is filtered out, and then you attend to it.
- The filter is not an all or none. It just turns the volume down of anything you are not paying attention to. This is what allows a shock or your name to slip through the filter and be processed for meaning.
Late selection filter models
- We process input to the level of the meaning, and then select what we want to process further
- Suggests that we selected our attention to something.
- When we do select attention to something we’re going to perceive both attended and the unattended information. We’re going to perceive it. We’re going to do some sort of analysis on it. We’re going to attach meaning to it, and only at that level, are we going to then attend to what we want to versus not.
- The filter is later in the processing pipeline.
Controlled vs automatic tasks
-
Controlled tasks: Those that require effort and voluntary top down attention
- Stroop Test: naming the color of the ‘ink’ (engages top down processing)
- Automatic tasks: Those that are highly familiar and well practiced and do not require voluntary top down attention
- Stroop Test: reading color names * does require access meaning
Stroop Task
- When it is congruentt, it is very easy to name the ink colour.
- When the colour of the ink and the name does not match it is slower.
The stroop test is evidence for what model?
- It is evidence for the late filter model.
- Activate both at the level of meaning which caused interference
- For the interference effect to occur on the Stroop task, you must process the written color name (unattended information) for the meaning.
- Automatic information is processed at the level of meaning because if it was not then it would not interfere.
Removing automatic Processing
Study (hyptnosis)
- Hypnotized English-speaking participants to think color
names were meaningless (in a language they did not speak) - Removes the meaning of the words (i.e., color) thus automatic processing and also removes the Stroop interference effect. They stop processinng the words as meaning.
Load Theory and Flanker task
Participants are asked to search the circular display for the letter X and N. They need to press seperate buttons for each.
* There are 2 conditions:
- low load: only one type of non target distractors (o)
- high load: all of the non target stimuli in the cirlce were distractors –> resembled the targets.
- Then they asked people to do this task with letters flanking the display. They can be proccessed and distract people from that given task.
- Sometimes the flanker was the other target. Ex: when the X was in the display, the N was a flanker (incompatible).
- Sometimes the flanker was nothing (neutral)
- And sometimes there was no flanker (compatible)
- The idea is that if you are looking at the display and you see and X and it is flanked by an N. If you are processing the N, the other target you are looking for, you are going to be slower to detect the X.
- Slower for incompatible trials then neutral trials.
- The flanker affect only occurs under the low load condition, because you are more open to the distraction. In the low load, it is easy so your attentional processes are seing what else is in the world so you get distracted and succome to ther flanker effect.
How do we measure inattentional blindness?
Has a design to assure that we are measuring inattentional blindness and nothing else.
* Participants focus on a task in a space
* An unexpected target is presented in that space. Something new is introduced into that given space that they are already attending to.
Experiment
* Crosses with vertical and horizontal ‘arms’ of different lengths are presented very quickly
* A mask
* Participants asked to determine which cross arm is longer (task they have to do to assure that they are focused on space)
* Critical trials = a small black square is included
* Participants were later asked if they saw this black square, and many said ”No”. They didnt see this new stimuli introduced.
Can the meaning of information that people are inattentionally blind to can still influence or affect their behaviour?
- Same experiment but they presented words into the quadrant.
- Participants are more likely to complete the word stem with the what was presented during a period of “inatttentional blindness” (e.g., armpit)
→ asked to complete a word stem completion task, given the stem of a word and asked to complete it with any word that comes to mind. Most often would fill woth words from the word stem completion task.
→ they are not aware they are processing it. - Things that are in your innatentional zone that you are unaware off can affect your behaviour.
Posner cuing task
- Fixation display: Fixate on the center a screen
-
Cue Display: After a time period, A ‘space’ cue directs attention to an area (blue square; left or right)
3.** Target Display:** time interval and then reaction time to detect the target is measure.
On its own you find that when the cue is in the same place as the target, people respond faster and can detect that belly to different locaion.
What happens when you varry the time between the cue and target?
Posner Cuing Task
Duration between the cue and target is the SOA: stimuli onset asynchrony.
* Short time interval (SOA) < 200 m
* Long time interval (SOA) ~ 300 ms
Different results for short and long SOAs:
* Short SOA = people are rly good at detecting the target (faster on valid than non-valid) → spatial attentional cue is helpind detect the target).
* Long SOA = people are less good at detecting the valid trial than the invalid (spatial attentional cue is inhibiting detecting the target).
Inhibition of return (IOR)
Posner Cuing task
Inhibition of return (IOR): attention is inhibited from going to a recently attended space after a long duration between space cue and target (SOA)
* Cue = “pay attention here”, you dont see a target there so you shift your attention to somewhere else to try and find target
* For you to go back is harder because you have already marked that spot where the cue is as “searched”. Therefore, you will inhibit returning to that space.
* Adaptative, it helps us search our environment efficiently so that we do not continue to search/focus in one location.
Visal Search Tasks
name the 2 different search tasks?
Feature Search
* Search for an object that is different from the distractors based on one feature (search for “that” feature to find the object).
* Can use bottom-up and automatic processing (quick and automatically processed)
Conjunction Search
* Search for an object that is different from the distractors across many features.
* Requires top-down attention. Need to engage attentional processes.
*Requires you to hold the multiple feautures in your memory to guide your search.
The pop out effect
- The time needed to find a target that is different by one feature from distractors is independent of the number of distractors (set size)
- A pop out effect exists for features processed automatically in the visual cortex
- Only works for features that are processed by the modules in the visual cortex.
it would take the same time because it is a feature search. Pop out effect!
Will not get a pop out effect if the object shares a lot of similar features as the distractors.
Conjunction searches and set size
- What you usually find is that people take longer to do this, in general, compared to feature searches because you need to engage these top down attentional processes to unite multiple features to guide your search.
- Features that you are looking at do overlap with the distractor in some way.
- Takes more time to find a target with more distractors.
- Reaction time to find the object will increase as the number of distractor items in your visual field increases.
- Reaction time is dependent on set size.
- Shows interplay between attention and memory - you need to keep tract of the different features you are looking for. + perception
Embodied theories of attention
Eye tracking tools
Eye movements = determining where eyes are fixating to gatther innfo and how much people move between these fixations.
* The idea that we can use eye movement to measure attention assumes that attention is embodied (mental activities are linked with physical organs).
* Eye movements detect visual attentional goals
* Overt visual attention: attending to something with your eye movements (what eye tracking measures)
* Covert visual attention: attending to something without eye movements (usually prior to overt attention). Harder to measure.
- Eye movement as people viewed the image freely vs when you asked them to estimate the whealth of the people vs the age of the people.
- Attending to specific target information depending on their goals.
Cultural differencess in visual attention
- Measured eye movements as two groups (Western vs East Asian) attended to images with a central object (tiger vs airplane) and background (mountain vs forest).
- They wanted to see if people from different cultures would focus more on a central visual entity or be more likely to process an image wholestically.
- Western students were more lokely to fixate on the central object (more eye fixation to that)
- East asian students spend more time looking at the background.
- mapping onto some of the cultural differences that you might see with a focus on more individual and more holistic.
Sustained Attention: Vigilance decrement
Sustained attention is very hard to hold up and decreases over time (vigilance decrement)
Study:
Students watched one-hour videos of lectures
* At various points, asked if they were mind wandering (not paying attention). Mind wandering = symptom of breaking sustained attention
* Tested on lecture material
* more mind wondering in the second half of lectured
*Participants scored worst on material from second half of the lecture (vigilance was lower = experiencing vigilance decrement)
Two interpretations of this:
* Overload theory: increase attentional demands with time when you are doing a task (you become overloaded). You need more and more attentional processing over the time of the lecture.
* Underload theory: cause boredom overtime → lose focus → mind-wandering → divided attention
Example of a task switching experiment
- Participants were shown a letter and digit in a quadrant, whichever quadrant it was in, indicated what task they had to do. One task: decide if letter is vowel or consonant. Otther task: decided if number is even or odd.
- Right after people switched the task, their reaction time to make the decision increased. Performance = decrease.
MEasuring attentional capture of faces
- Task design: A ‘signal’ (circle) is superimposed on different types of visual stimuli (faces, objects, nothing)
- Go Trial: When the signal is green, indicate if the vertical line
is on the left or right - No-go Trial: When the signal is red, press a task neutral button
- Upright face will capture your attention because we process faces all the time whcih might have an affect on your performance.
results of go, no go task?
Faces below pull attentional resources away from the task that people are supposed to be doing.
