The visual system and visual attention Dr Lubna
The visual system and visual attention Dr Lubna Ahmed Keele University, UK 1
Lecture overview o Attention – the basics o Visual system o Visual attention 2
Sensory input – limited in capacity
Some terminology o Perception Vs Attention
Physics Psychology Light Vision Sound Hearing Chemical Concentration Smell and Taste Heat & Pressure Touch, Pain and Body Senses Gravity & Acceleration Balance and Vestibular Senses Physical World Phenomenal World
What is attention? o State: the difference between attended and unattended processing o Process: the selection of relevant information o Attention facilitates perception and any further processing (e. g. memory) for relevant stimuli o Attention prevents perception and any further processing (e. g. memory) for irrelevant stimuli 6
Why do we need attention? o Limited capacity o Not all sensory information can be processed in detail o Attention is the ‘filtering system’ that funnels our resources
Visual attention o o o The visual system Automatic (bottom-up/ exogenous) Vs controlled (top-down/endogenous) Overt Vs covert attention Spotlight metaphors Selective attention
Receptor cells – specialised neurones which respond to a particular physical property of environmental stimuli
Anatomy of the eye
The electromagnetic spectrum 1 nm = 10 -9 m = 1/1, 000, 000 metres
The retina Pigmented layer Receptor cells Horizontal cells (H) Bipolar cells (B) Amacrine cells (A) n ns a Tr tio c du Retinal ganglion (RG) cells light
Transduction The conversion or translation of environmental physical energy into neural activity LIGHT ELECTRICAL IMPULSE Nerve cell firing Activated firing Resting discharge Impulses per second
The visual pathway Primary visual cortex or Striate cortex
The optic pathway is crossed Retinotopic map
RG Cells RFs Retina LGN Cortex Simple Cell RFs Complex Cell RFs Hyper-complex Cell RFs
Functional streams within geniculate-striate pathway o Based on lesion experiments in monkeys, Mishkin and Ungerleider (1982) proposed that primates have two geniculate-striate streams: Identification of objects, termed ‘what’ stream n Spatial location of objects and visual guidance of movements, termed ‘where’ stream’. n 1/30/2022 17
Higher visual processing Location, speed of movement Features, such as colour etc.
Spatial Attention o “What” & “where” pathways
Spatial Attention o Damage to this pathway (typically to the rightparietal cortex) can lead to Spatial Neglect
The visual system o A lot of what we know about functional localisation learnt from: n n n o Animal lesion studies Patients Neuroimaging Outstanding question – where in the visual system does attention take affect? n n Early vs late filtering If late, what happens to unattended information?
Visual attention o o o The visual system Automatic (bottom-up/ exogenous) Vs controlled (top-down/endogenous) Overt Vs covert attention Spotlight metaphors Selective attention
Automatic (bottom-up/ exogenous) Vs controlled (top-down/endogenous) o o o Cherry 1950’s Dichotic listening tasks People were able to shadow info from attended ear but recalled little from unattended ear
Automatic (bottom-up/ exogenous) Vs controlled (top-down/endogenous) o Visual attention o Gazzaley et al 2005 (f. MRI, EEG)
Automatic (bottom-up/ exogenous) Vs controlled (top-down/endogenous) o Visual attention o Gazzaley et al 2005 (f. MRI, EEG) Peak N 170 10 ms earlier for remember faces cf to ignore (P<0. 05)
o So we can direct attention in a task relevant manner…. . does this always work? o Are there situations or stimuli that ‘automatically’ capture our attention?
Automatic (bottom-up/ exogenous) Vs controlled (top-down/endogenous) o Cocktail party effect (Moray, 1959) n o A salient input like the participants name ‘garbs’ attention in the unattended ear Salient singleton capture n Jan Theeuwes et al, 2010.
Shiffrin and Schneider (1977) o Automatic processes (bottom up): n n o Controlled processes (top down): n n n o Fast. Suffer no capacity limitations Do not require attention or consciousness, can not be resisted (? ) Are of limited capacity Require attention Can be used flexibly in changing circumstances Attention : Attempts to coordinate bottom-up and top-down processes to guide behaviour in a task relevant manner. 28
Visual attention o o o The visual system Automatic (bottom-up/ exogenous) Vs controlled (top-down/endogenous) Overt Vs covert attention Spotlight metaphors Selective attention
What is Spatial Attention? o Deployment of attention across space o Essential for visual search Driving n Looking for your friends in a bar n During visual search n
Overt Attention
Overt Attention o Usually we move our eyes to a location in order to attend to it o When attention follows eye movements, it reflects overt attention o “Usually no object lying in the marginal portions of the field of vision catch our attention without at the same time ‘catching our eye’. . . ” (William James)
Covert Attention o Attending without moving the eyes
G L B T S D F U Z A F I Y Q W O Z M A H E I L P R
Covert Attention o Attention can be moved without moving the eyes o Although, attentional focus is clearer around fixation o William James described visual attention as having a focus, a margin, and a fringe
n
Visual attention o o o The visual system Automatic (bottom-up/ exogenous) Vs controlled (top-down/endogenous) Overt Vs covert attention Spotlight metaphors Selective attention
The nature of attention o The most common metaphor of attention is that it is like a spotlight o Allows us to attend to a particular region of the environment
The nature of attention o If attention is best at the focus, then the focus must be moved to attend to different items in the environment o Does n the spotlight move? If so, how?
Spatial Cuing Paradigms o Subjects asked to make speeded reaction as soon as they detect onset of a light (Posner et al. , 1980) o Provided with a pre-cue: Central: Arrow pointing left or right, indicating where the target will appear n Peripheral: Presented outside fixation; one of the target locations would flash n
Spatial Cuing Paradigm o Attentional spotlight moves to the area of space indicated o Sometimes n They pointed in the wrong direction o Sometimes n the cues were invalid the cues were neutral Arrows pointing in both directions
Results
Exogenous and Endogenous Orienting o Endogenous n i. e. the arrow must be interpreted and acted upon o Exogenous n – controlled by intentions – automatically captures attention i. e. the flash
Exogenous Orienting o Many real world events trigger exogenous orienting If someone were to leave the lecture n The person will capture your attention exogenously n Eyes likely to move to them (overt attention) n Covert attention can also move like this n
Exogenous Cuing Paradigm o Cue (box flash) does not always predict target location o Performance is faster at cued location, but only for short SOAs (50 -250 ms) Fixation Cue Time SOA Targe t Posner & Cohen (1984)
Results
Results Facilitation
Results o When target appears more than 300 ms after cue, the benefit turns into a cost n i. e. RTs are slower o This cost is linked to Inhibition of Return (Posner & Cohen, 1984)
Results Inhibition
Results Fixation Cue Time SOA Target
Inhibition of Return (IOR) o. A cost returning to a recently attended location which was deemed task-irrelevant o What are the advantages of IOR? Foraging behaviour in evolutionary terms n System that biases orienting to novel locations n o Klein IOR (2000) provides an excellent review on
Three components of attention o Posner and Petersen (1990) n n n Disengagement of attention from a given visual stimulus Shifting of attention from one target stimulus to another Engaging or locking attention on a new visual stimulus 54
Focusing the Spotlight o We have shown that the spotlight can move o But what is the nature of the spotlight? Is the spotlight initially diffuse, becoming focussed with time? n Is it of a fixed-focus? n o Evidence from the Eriksen flanker task (Eriksen & Eriksen, 1974)
Eriksen Flanker Task o Presented with a string of letters o Identify the central letter (is it an ‘S’ or an ‘H’) Eriksen & Eriksen (1974)
SSSSS
Eriksen Flanker Task o Presented with a string of 5 letters o Identify the central letter (is it an ‘S’ or an ‘H’) o Central letter can be “flanked” by congruent letters or incongruent letters Congruent = HHHHH n Incongruent = SSHSS n Eriksen & Eriksen (1974)
Eriksen Flanker Task HHSHH
Factors affecting visual attention o External factors n n n o Perpetual load Cognitive Load Cultural environment Internal factors n n n Individual differences (intelligence, working memory) Personality types (anxiety levels) Depression / stress
Flanker task (Eriksen, 1974) n Spot light / zoom lens model of spatial attention Distractor Effects o congruent HHHHHHH Incongruent HHHSHHH attended distance Congruency Effect
o o Selective attention starts in a diffuse state then contracts to a focused state. Time effects this process n n o Z Eriksen & St James 1986: increasing SOA led to less influence from incompatible distracters Perceptual load Effects this process n n n X Lavie 1995 & Rees 1997 (V 5/MT) Handy, Soltani & Mangun 2001 La. Berge (1991) High perceptual load Z F T X X Low perceptual load Z X
Perceptual Load Theory o o o Everyone has limited attentional capacity The amount of attentional capacity allocated to the main task depends on its perceptual load, which is determined by “the number of units in the display and the nature of processing required for each unit” (Lavie & Tsal, 1994, p. 185) “Any spare capacity beyond that taken by the highpriority relevant stimuli is automatically allocated to the irrelevant stimuli” (Lavie, 1995, p. 452). Thus, the total available attentional capacity is always allocated to processing 63
High perceptual load Z F T X X o Low perceptual load Z X Mean target identification time as a function of distractor type (neutral vs. incompatible) and perceptual load (low vs. high). Based 64 on data in Lavie (1995).
More perceptual load evidence No-Go trials: depends either on detecting the presence of any blue shape (low load) Lavie (2005) or on discrimination of conjunctions of colour and shape (high load; e. g. target responses are made only if there is a blue square and a red circle). 65
Load theory of selective attention o Resolution to the early vs late selection debate o If relevant processing exhausts capacity (high perceptual load), then irrelevant distractors will not be perceived: early selection o If relevant processing does not exhaust capacity (low perceptual load), then irrelevant distractors will be perceived: late selection 66
Load theory of selective attention o Early selection is bottom-up: involves passive exclusion of irrelevant distractors, depending on stimulus characteristics o Late selection is top-down: involves active rejection of irrelevant information to maintain goal-directed behaviour 67
WMC and selective attention o n n o Lavie 1995 & Rees 1997 (V 5/MT) Handy, Soltani & Mangun 2001 La. Berge (1991) The process can be controlled in a top-down manner (e. g. via instructions): n n n o Eriksen & St James 1986: increasing SOA led to less influence from incompatible distracters X Perceptual load affects this process n n n o Z Time affects this process (Belopolsky, 2007 &2010) RSVP (focused) Vs Global shape (diffuse attention) Singleton capture reduced in focused condition External Cognitive load also affects this process High perceptual load Z F T X X Low perceptual load Z X
Effects of cognitive load on distractor processing • When distractors are processed (e. g. , under low perceptual load), active control is needed to minimize distractor interference. • Imposing cognitive load will impair this active control, leading to greater distractor interference, compared to when cognitive control is available for selective attention. 69
Effects of cognitive load on distractor processing Low memory load High memory load 526917 4 Z X X Z 4? 8? 70
Effects of cognitive load on distractor processing Low memory load 526917 4 WM available to ignore distractor: weak interference High memory load WM not available to ignore distractor: strong interference Z X X Z 4? 8? 71
Q: Is the central name a politician or a musician? De Fockert et al, 2001 Face distractor present 0 1234 or 0 3124 2 David Bowie Mick Jagger 2 Face distractor absent 0 1234 or 0 3124 John Major Elton John 2 72
Q: Is the central name a politician or a musician? De Fockert et al, 2001 Face distractor present 0 1234 or 0 3124 2 David Bowie Mick Jagger 2 73
Working memory load x Face distractor De Fockert et al, 2001 C B Face present > Face absent low memory load high memory load R L • fusiform gyrus (BA 19/37) • lingual gyrus (BA 18) • inferior occipital gyrus (BA 18) Height threshold p <. 001 uncorrected 74
The role of working memory load in selective attention o Working memory is needed to retain relevant information in visual processing. o Reducing the availability of working memory for selective attention results in greater distractor processing (because the distinction between stimuli of high and low priority is weakened). 75
o o Distractor Effects …a complex spotlight Linear gradient model The influence of peripheral distractors is inversely proportional to the distance from the focus of attention (C. W. Eriksen & St James, 1986) attended distan ce o Centre-surround inhibition or ‘Mexican hat’ profile n Hopf et al pnas 2005 (MEG study) n Muller et al 2005
Centre-surround inhibition or ‘Mexican hat’ profile Task search for pop out red ‘C’ amongst distractor Blues and respond to direction of C. Targets always 8º from fixation point. Probe (white circle) presented at varying distances from target. PD 0= probe at target location PD 1 -4= 1. 35, 2. 7, 4. 05, 5. 4º Level of Processing of probe as a function of Probe-target distance used to observe surround-inhib model. Level of probe processing Hopf et al pnas 2005 (MEG study)
Centre-surround inhibition or ‘Mexican hat’ profile Muller et al 2005 Used flanker task. Cong/incong & neutral distractors. Letters= E, F, X, O. Letters 4º from fixation. Letters presented in all squares simultaneously for 150 ms. Task to identify letter in first square as: X or 0 (in perceptual low load) F or E (in perceptual high load) One compatible/incompatible letter presented in another square, Neutral letter in the others. Target-Distractor distances: D 1 -D 4=1. 3, 2. 5, 4. 7, 6. 5º Found ‘hat’ distribution. Hat*Load: did find predicted ‘shrinking’ with Perceptual load
Visual attention o o o The visual system Automatic (bottom-up/ exogenous) Vs controlled (top-down/endogenous) Overt Vs covert attention Spotlight metaphors Selective attention
- Slides: 80