Chapter 4 Attention PDF

Chapter 4 Attention Sept. 10th Attention Basic aspects of attention Controlled or focused on task/stimuli Select task relevant stimuli Or switch tasks/activities Can be “captured”... E.g., by sudden, loud noise Can help filtering of information Selective Attent...

Chapter 4 Attention Sept. 10th Attention Basic aspects of attention Controlled or focused on task/stimuli Select task relevant stimuli Or switch tasks/activities Can be “captured”... E.g., by sudden, loud noise Can help filtering of information Selective Attention Attention able to select only a subset of available info Some info unable to be attended due to resource limitations Attention “allows in” info relevant to accomplishing a task Filter out irrelevant information E.g., able to listen to conversation in a crowded, noisy room (Cocktail party effect) How is info selected? Dichotic Listening Task Dichotic listening task Different messages played on each ear Shadow (repeat) message from one ear (attended) Ignore message on other ear (unattended) Afterwards asked... What was the message in attended/unattended ear? Dichotic Listening Task From the unattended ear... Able to remember... Basic stimuli characteristics Speech, tone Types of sounds (male/female speakers) Couldn’t remember... verbal content of message Change in language Specific words (e.g., from word list)  Broadbent’s filter model Filter applied after an initial sensory stage Sensory Sensory Short- And before meaning processing informati Filter term memory on memory e.g., info from environment Filter out info based on physical aspects E.g., pitch, loudness, location After filter, process meaning (i.e., language) Some sensory information does not reach stage for meaning process Blocked out Not processed at level of perception Broadbent’s filter model During dichotic listening task Info in unattended ear... Sensory informati Sensory Filter Short- term memory Filtered out on memory No meaning is processed e.g., info from environment Attended ear info passes filter Meaning is processed Attention dependent on physical cues Attention models Broadbent’s model Early selection model Filter of info at early stages of processing Bottleneck in the initial stages of attention  Broadbent’s filter model Are all unattended stimuli blocked out? If unattended message doesn’t pass Sensory Sensory Short- informati Filter term filter... on memory memory Then we should be unaware of the message e.g., info from environment Moray (1959) “subjectively ‘important’ messages” Can break through filter Dichotic listening task E.g., Name embedded within unattended message ~33% detected name 4 out of 12 participants  Treisman’s Attenuation model Sensory informati According to the Attenuation model on Attended message passes through Sensory memory Unattended messages not completely blocked Become attenuated (“weakened”) Attenuator Can still be processed for meaning Dictionar y Unit Short- term memory Treisman’s Attenuation model Sensory Attenuator informati on Can process info at... Physical level Sensory Pitch, loudness memory Linguistic level Language, Syllables, words Attenuator Semantic level Meaning Weakens info based on relevance to task Dictionar E.g., weakened info from unattended ear, not y Unit important to completing shadowing task Short- term memory Treisman’s Attenuation model Sensory Dictionary Unit informati on Includes words learned Thresholds Sensory Low threshold memory More likely to attend to Important info (e.g., our names) Attenuator High threshold E.g., uncommon words Dictionar Less likely to be attended y Unit E.g., our name in unattended ear Able to perceive it as it’s easier to meet its activation threshold Short- term memory LATE Selection model Late processing Meaning & sensory/physical info is processed Then it can be selected for attention Sensory Physical & Short- Sensory informati on memory Semantic Filter term memory processing  Load Theory Is N present in the circle of letters? Load Theory Processing capacity is limited Perceptual Load High Load Low Load Increases based on... Amount of information to process Complexity/difficulty of task  Load Theory Is N present in the circle of letters? Load Theory High perceptual load No/reduced resources to process distractors High Load Low Load All resources taken up by task Distractor does not get processed Perceptual Perceptual Capacity Capacity Low perceptual load Resources Resources available to process all stimuli available Including distractor for All distractor Distractor gets processed resources used by Resources task used by low load Load Theory Low load Which bar is green? More likely to be aware of square Resources available to process it High Load task Which bar is larger? Less likely to notice square due to lack of resources Load Theory Load theory can sometimes use early or late selection If perceptual load is high Distraction would not be processed into awareness No perceptual resources to devote to distraction Early selection If perceptual load is low Distraction would be processed Must select the “task” to filter out the distraction Late selection  Load Theory High Load Low Load (e.g., 26485) (e.g., 01234) Load Theory Working memory/cognitive load WM Capacity WM Capacity is high... E.g., memorizing (random) 5-digit Resources available to numbers inhibit Distraction of the faces would All distractor interfere with holding numbers in resources used by Resources WM task, used by Reduced resources to inhibit nothing left WM low distractor to suppress load More difficult to use top-down distractor control for goal-oriented behavior When WM load is low... E.g., easy digit series (01234) Have enough resources to inhibit https://www.youtube.com/watch?v=IGQmdoK_ZfY  Inattentional blindness Inattentional blindness Failure to perceive an object without attention For example, can be unaware of an object even if we fix our eyes on it For example, Gorillas in Our Mist (Simons & Chabris, 1999) video of two “teams” passing a basketball count the ball passes After a while, a gorilla or man with an umbrella go by 46% of participants didn’t notice them pass by Attending to one activity made them fail to perceive the gorilla/umbrella man  Inattentional blindness (Simons, 2010) Even if familiar with the previous study Some participants unable to perceive unexpected events Being aware of inattentional blindness does not necessarily help avoid it  Driving & Load Theory (Murphy & Greene, 2016) Use driving simulator for gap perception task Image from Murphy & Greene Gap perception (2017) Can car pass through gap between parked cars If gap too narrow, go around cars (on the right) Low perceptual load Gap very wide or very narrow High perceptual load Gap a little wide or a little narrow for car to pass by In a few trials, a pedestrian or animal appears on the left (critical stimulus, CS) After trial asked if they noticed anything different Driving & Load Theory (Murphy & Greene, 2016) Under low perceptual load More likely to be aware of CS, than under high perceptual load High (visual) perceptual load increased chances of inattentional blindness i.e., not processing other stimuli due to demands of task Driving & Load Theory (Murphy & Greene, 2017) How could WM/Cognitive load play a role while driving? (in addition to perceptual load) Driving & Load Theory (Murphy & Greene, 2017) Driving Simulator Drive through several towns Cognitive Load Task Directions on billboard at town entrance Low load: “Take the M9” High load: “Take the M25 or the R167” Remember directions and take route after passing town Driving & Load Theory (Murphy & Greene, 2017) Perceptual Load Task Find red Mercedes among parked cars Low load: white cars, non-Mercedes cars High load: red cars, including Mercedes of different colors Distractors Billboards of a red Mercedes Told to ignore all billboards Congruent: on side of target car Incongruent: opposite side of target car Driving & Load Theory (Murphy & Greene, 2017) Unexpected stimuli Auditory screeching breaks horn beeping simple tone Visual “swerving vehicle in the oncoming lane” “pedestrian leaving the footpath” Billboards with celebrity faces After each car search, asked if they detected anything unusual  Driving & Load Theory (Murphy & Greene, 2017) Results – Response Time (RT) High & Low Cognitive load High Perceptual load took more time to search for car Decreased distractor interference (i.e., reduced processing) ”used up” perceptual capacity, so difference in Cognitive load didn’t matter High Cognitive Load Distractor interfered under Low Perceptual load Driving & Load Theory (Murphy & Greene, 2017) Results - Awareness of unexpected stimuli Auditory Low perceptual load: 100% High perceptual load: 62.5% Visual Low perceptual load: 92.5% High perceptual load: 30% More likely to detect unexpected stimuli under low perceptual load Had enough remaining resources to detect stimuli High perceptual load Inattentional blindness & deafness for unexpected stimuli Controlled processes Practice of controlled processes Deliberate/conscious use May be “slow” & effortful to perform Takes longer to perform Resource limited Can use serial processing (step-by-step) Practice helps automatization E.g., Reading Slow and effortful at first Then, faster, smoother Finally, automatic But, in other conditions (noisy roommate)... It can become effortful Automaticity Automatic Tasks Can be effortless May use minimal or no mental resources Can be done with no awareness or intention Practice can help tasks become automatic For example, reading In a Stroop task Name the font color of the words Mismatching of font color and color word Causes interference Automatic response of reading the word disrupts naming the font color Attention Attentional resources are limited Can only attend to a certain amount of info at any given time Having to think and generate an answer disrupts driving performance Not simply the acts of talking or listening Divided attention (DA) Strayer et al., (2003) What are other ways DA affects driving? Placed participants in driving simulator Driving only vs Driving & phone convo (hands-free) cell phone conversations Convo topics participants found interesting Had to follow a pace car Plus, different driving conditions Low or High density  Divided attention (DA) Strayer et al., (2003) When driving & talking (dual task)... Impaired (slower) reaction for breaking (high-traffic) Increased distance following distance Both traffic conditions Can be trying to compensate for talking on the phone Increase in perceptual load (# of cars in traffic) decreased break onset Even though the cars were in another lane, didn’t interact with participant’s car Divided attention (DA) (Strayer et al., 2003) Memory & DA Driving only vs driving & phone convo (DA) Drive by several billboards Track eye movements while driving Finally, given recognition test for billboards Did you see this billboard? Yes or No Memory & DA (Strayer et al., 2003) Results: Recognition of billboards Driving only > Driving & convo Avg. amount of fixation time (in msec.)... Driving only (1,122) ≈ Driving & convo (1,009) If they saw (fixated) the billboard, what’s the probability of recognizing a billboard?.50 (Driving) vs.24 (DA)  Memory & DA (Strayer et al., 2003) Active engagement in phone convo Shift in attention from context of driving to conversation More realistic driving task & conversations of interest (for participants) Both tasks compete for limited attentional resources If tasks they require more resources than available... Results in devoting less attention to each, impairing performance In DA condition Despite fixating on billboards unable to create durable memory Inattentional blindness Driving & DA (Cooper & Strayer, 2008) Practice driving and talking on the phone... In highway or city scenario For multiple days On final day drive in different scenario does practice transfer to another scenario? Only Driving (Single task, ST) Driving & convo (Dual task, DT)  Driving & DA (Cooper & Strayer, 2008) Collisions More collisions under DA for both scenarios (Day 1) Improvement on Day 4 (no difference b/w conditions) But, no transfer of improvement for DA Similar to Day 1 Break RT No improvement of practice DA took longer to apply breaks  Driving & DA (Cooper & Strayer, 2008) Overall, unable to eliminate costs of DA with practice Failure of transfer when driving in new scenario Practice unable to overcome limitations of bottleneck of attentional resources Didn’t decrease resources needed for driving or talking tasks Driving might be too dynamic a task to ”practice away”  Feature integration theory (FIT) Illusory conjunctions (Treisman, 1986) Quick presentation of stimuli 200 msec 5 8 Participants asked about numbers & shapes 29% said... Combinations of shapes and colors not presented E.g., green triangle More often than feature errors (13%) Features not presented E.g., a purple shape or a diamond Feature integration theory (FIT) Binding of features disrupted Limited or divided attention E.g., if focused on both numbers & shapes 5 8 Feature integration theory (FIT) Two stage process 1. Preattentive stage Stage prior to attention Automatic processing detect features from visual field Features are “unbound” Or independent Not associated with particular a stimulus Unaware of process Feature integration theory (FIT) Preattentive stage “Pop out” features Some features are important in early visual processing For example... Orientation Color Movement Feature integration theory (FIT) Participants were more likely to report seeing a triangle after seeing (Treisman, 1986)... Option b: Has features of line in correct orientation Plus, closure (circle) Option a: “seldom” saw triangle Features: components of triangle Option c: Same as a Diagonal line in different orientation Option d: Closure, but lines w/different orientation Feature integration theory (FIT) Preattentive stage Even though consciously we detect a triangle It is detected first as independent features i.e., lines & closure Feature integration theory (FIT) 2.Focused attention stage Attention helps bind features of stimuli E.g., features of triangle & purple into purple triangle Combination of independent features It is a conscious process “Disruptions” or divided attention can lead to... Binding problems (e.g., Illusory conjunctions) Feature integration theory (FIT) Focused attention stage E.g., visual search Target: red O Target: white rectangle with a given orientation Feature integration theory (FIT) Focused attention stage E.g., visual search Target: red O Does not “pop out” Shares features with distractors Conjunction of two features Circle & red More distractors with shared features, longer to find target Faster search if items share only one feature Feature integration theory (FIT) Focused attention stage serial search Search item by item Bind item’s features of letter & color Then, decide if it’s target  Feature integration theory (FIT) How to reduce illusory conjunctions? Treisman (1986) E.g., What was the middle shape? If told random pairings E.g., you’re about to see “an orange triangle, a blue ellipse and a black ring” More likely to have illusory conjunctions If told familiar objects “a carrot, a lake and a tire” Made less mistakes Spot the bicycle Feature integration theory (FIT) Treisman (1986) Expectations can influence visual search Top-down processing Spot the bicycle Easier if you know where to look i.e., use prior knowledge Selective Attention Top-down selective attention Goal-oriented attention Requires effort to select & maintain Intentional focus i.e., focus on info that will help complete a task A process under our control  Goal-oriented attention (Theeuwes et al., 1998) Task: See grey circles w/8s  Goal-oriented attention (Theeuwes et al., 1998) Task: See grey circles w/8s Then, 8s  letters And, Grey circles  red Except one When circles change color, move your eyes to remaining grey circle (target). Does remaining grey circle have a “C”? In some trials, another red circle appears i.e., a distractor Track to check if eye movement was influenced by distraction Goal-oriented attention (Theeuwes et al., 1998) No distractor Eye movement directly to target Target Distractor consistently Distractor disrupted eye movements towards target First direct at distraction, then target Goal-oriented attention (Theeuwes et al., 1998) Goal-oriented action (eye movement to target) disrupted Target by stimulus-driven distraction Distractor Even if the distraction is task irrelevant Plan & action of eye movement disrupted Participants asked if distractor impacted their eye movement Responded they were sure distractor stimulus didn’t affect their eye movement  Goal-oriented attention (Theeuwes et al., 1998) Top-down (voluntary) action disrupted by... bottom-up (involuntary) attention capture Target Distractor Why might attention be susceptible to this disruption? May have goal/intention in mind, but might need to attend to unexpected event/stimulus Attentional capture Unintentional shift of attention Automatic Stimulus may “pop out” Attentional capture Bottom-up selective attention Stimulus-driven Stimulus “captures” attention Can happen despite intentions i.e., involuntary How much can automatic capture influence goal- oriented attention?  Goal-oriented attention (Theeuwes et al., 2000) Study task Shown nine grey diamonds w/circled 8s One diamond changed to red (at Change to red: 700 different times) ms 50, 100, 150, 200, 250, or 300 Location of distractor ms before last All diamonds & 8s replaced by display circled letters (all grey) Except target, a grey diamond Is the target a “C”? Ignore distractor  Goal-oriented attention (Theeuwes et al., 2000) The longer the time between distractor presentation and target presentation... Faster reaction time for target 50 ms & 100 ms slower than no distractor, and all other times No distractor = 150, 200, 250, 350 ms  Goal-oriented attention (Theeuwes et al., 2000) Still get disruption from task irrelevant stimulus But, if distractor shown with enough time before... Top-down or goal-oriented attention can override the bottom-up distraction Top-down control needs time to avoid interference from attentional capture i.e., distraction from irrelevant stimulus  Attention & Learning (Della Libera & Chelazzi, 2009) Training sessions across 3 days Target figure (same color as cue) Judge whether it is the same or different from figure on the right Target & distractor superimposed Received high (€.10) or low (€.01) reward after trial  Attention & Learning (Della Libera & Chelazzi, 2009) Target & distractors could be typically associated with high or low rewards Then, test session (5 days later) Same judgement task No rewards  Attention & Learning (Della Libera & Chelazzi, 2009) Results Regardless of item (reward) history Target items response was equal But Distractors responses influenced by history Interference from (training) Targets associated with high Item History rewards Easier to ignore (training) Targets & Distractors associated with low rewards  Attention & Learning (Della Libera & Chelazzi, 2009) Selection of stimulus with history of reward... Acted as a distraction (at test) Even after several days since creation of association But if ignoring stimulus was associated with reward... Easier to ignore later Item History Overall, selection or suppression of stimulus influenced by reward history Attention & Threats (Nissens et al., 2016) Fixate on target among distractors Target had unique shape (e.g., circle among diamonds) Trial Types Threat: shock if look at distractor E.g., if red diamond is present Safe: no shock E.g., if blue diamond is present May also be shocked if take too long to respond/fixate on target Attention & Threats (Nissens et al., 2016) Results More first saccades towards distractor under Threat trials Threat Trials Great proportion of saccades to distractor during the early/short latencies Attention & Threats (Nissens et al., 2016) Attentional bias towards threats can be automatic and involuntary Attention captured even when threat is not salient Can occur during early stages of visual selection Voluntary attention (controlled top-down process) takes time to influence visual selection