Attention Flashcards
Overt vs covert attention
Overt = externally observable –> turning head and eyes towards stimulus (behaviourism)
Covert = not directly observable –> can move attention without moving eyes
- von Helmholtz (1896) - dark room, electric spark to illuminate text + fixate in centre –> could read the place where covert attention was paid
Auditory attention
- Cherry (1953) - dichotic listening - different stimuli in each ear, could focus on + shadow one attended to (covert because not moving ears or head)
- Triesman & Geffen (1967) - when hear ‘tap’ tap - respond reliably if in attended ear BUT not if in unattended ear
Filter model (auditory attention)
Filter model (Broadbent 1958) - early stages = high capacity –> passes through bottleneck; late stages = lower capacity –> information has been filtered out
CRITICISMS:
- Triesman (1960) - can switch sides if attended message changes to irrelevant side - attention filter is not absolute –> filter attenuation model (filter doesn’t ignore information, it attenuates it) - all info weakly activated higher level processing
- Cherry (1953) - own name enters awareness, even if in ignored ear (cocktail party effect)
- Corteen + Wood (1972) - city names previously paired with shock elicit electrodermal responses even if in ignored ear
Attention as a spotlight
- endogenous cues
- exogenous cues
- how wide is spotlight?
Posner et al., 1980:
- attention moves around scene highlighting important parts, disengages from one area + shifts to engage another
Endogenous cues = shift is internally driven
- cue arrow (80% accuracy) before target –> invalid cue = cueing cost (RT slower), valid cue = cueing benefit (RT faster)
Exogenous cues = shift is stimulus driven - draws attention reflexively
- bright flash in periphery (50% accuracy) - not predictive of target
- valid cue = faster RT
- if increased delay (cue-target interval) - benefit for valid cues reverse - there is inhibition of return (inhibitory lag)
How wide? - Eriksen + Eriksen (1975) - flanker paradigm
- incongruent flankers - RT slower; congruent - RT faster
- greater separation, RT faster
- spotlight ~1 degree separation
Unilateral neglect
Damage to posterior parietal cortex (usually R)
- inattention to side of space opposite lesion
- Wilson, Cockburn + Halligan (1987) - fail to cross out lines/stars on L of paper
- Bisiach + Luzatti (1978) - neglect of remembered scenes (miss out L on one side but if from other POV remember it because it’s on right - Milan square)
- Posner et al., (1984) - response to valid cues the same as controls BUT much worse for invalid - difficulty switching to new targets (Styles, 1997)
- Driver + Halligan (1991) - compare objects + spot the difference –> worse on L than R and ALSO worse if difference was on the L of objects on the R
Unilateral extinction
Damage to posterior parietal cortex
- detect stimuli presented to R OR L if alone BUT not if bilaterally presented stimuli - if R+L together, miss L out
- Attention drawn away from LHS
- Moray (1975) - even healthy participants miss brief targets if they have to monitor two streams simultaneously
- can’t attend to 2 targets simultaneously - due to attention (Duncan, 1980)
- greater impairment if seen as separate object as opposed to merely obscured (Mattingley, Davis + Driverr, 1997) - patient VR
Balint’s syndrome
Balint (1909)
- bilateral parietal damage
- fixity of gaze
- problem shifting attention
- simultagnosia - can’t perceive multiple objects at once
- Findlay + Gilchrist (2003) - FIT - they have attention impairment so struggle to bind separate features together - attention needed to combine features
- if attention was a spotlight - they’d be able to see multiple OBJECTS at one
How does attention alter perception?
- Hawkins et al., (1990) - barely-visible targets followed by mask, pointer around fixation to cue (valid cue - 0.93 sensitivity; neutral - 0.69; invalid - 0.46)
- Pack, Carney + Klein (2013) - low contrast X/O discrimination - valid cue increased % correct, low contrast + invalid = chance, low contrast + valid = 66.6% correct
Feature integration theory
Triesman + Gelade (1980)
- Stage 1 - features all processed in parallel (make map for each feature)
- Stage 2 - attention combines all features in spotlight to create a master map
- visual search: parallel (pops out, orientation/size/colour), series (shift attention around, if present usually only need to search around 50% of it) - fo multiple features
What does attention select?
- objects vs spotlight
Egly, Driver + Rafal (1994) - cueing paradigm: 2 types of invalid cue - same object + other end VS different object + same end (same distance)
- faster if same object –> attention selects whole object
Richard et al., (2008) - flanker effects, same vs separate objects
- no evidence of interference from incongruent flankers if on different object
Multiple object tracking paradigm
- Pylyshyn + Storm (1988) - track >1 moving objects + ignore distractors
- Pylyshyn et al., (2008) –> some dots presented on moving objects - if on tracking object, able to detect
How much can attention select?
Enumeration
- Jevons (1871) - can count 4 objects in parallel (at same speed per set size -1, 2,3, 4)
- Trick + Pylyshyn (1994) - count sets of 1-8 objects: 1-4 RT is same, 4+ RT increases linearly (series processing)
Visual STM - briefly present object, remove, re-present with change
- Luck + Vogel (1997) - 1-3/4 pretty accurate, 4+ many more errors
MOT task
- 4+ ability drops off rapidly
Early selection
- arguments for
- arguments against
Early selection = attention selects after basic sensory processing happens
- high level processing of objects outside attended region is greatly attenuated
- attention drawn by salient/unique features only
FOR:
- Triesman (1964); Triesman + Gelade (1980) - attention can select efficiently based on simple features - shown by parallel search in visual search paradigm
BUT - complex stimuli can be processed in parallel too
- Nakayama + Silverman (1986) - depth + colour –> can search for this combo efficiently in parallel (pre-attentively)
- Macleod et al., (1991) - motion X form - finding moving X among static X and O –> shape + movement but still parallel
Late selection
-arguments for
Late selection = attention selects after basic sensory and high-level processing happen –> all objects in visual field undergo high-level processing
- Yarbus (1967) - presented paintings of a scene + eye movement patterns recorded –> they were related to task at hand
- Henderson + Hayes (2017) - even first saccades driven by meaning not salience
- Shiffin + Gardner (1972) - 4 letter-like stimuli, mask before + after presented –> successive condition (2 then 2 = less demanding) + simultaneous condition (4 at once) had same performance - at least 4 items processed at high level
BUT: limited number
- Duncan (1980) - selecting objects to control actions –> find digit among 4 letters (successive or simultaneous) - could be one in vertical pair AND one in horizontal (successive has faster RT)
Attention and saccades
If one is inhibiting distractor (like when given cue), saccades curve away; if one is not, saccades curve towards (distracts attention)
- Walker et al., (2006) - valid cue reverses curve direction –> successful inhibition
- Shepherd et al., (1986) - arrow shows where to make saccade to (but target on other side 80% time) –> cue-target lag (70ms - too short for saccades - covert attention) –> RT increased if lag 140ms because had to move eyes opposite way to direct attention
