Attention, Effort, and
Resource Allocation
September 18, 2008
What Is Attention?
• “Every one knows what attention is. It is the
taking possession by the mind, in a clear and
vivid form of one out of what seem several
simultaneously possible objects or trains of
thought. Focalization, concentration of
consciousness are of its essence. It implies
withdrawal from some things in order to deal
with others…” (James, 1890, p. 403)
•
http://viscog.beckman.uiuc.edu/djs_lab/demos.html
Attention is...
• An attitude of mind
• Relative proportion of activated traces to all
memory traces – active workspace
• Some form of energy or desire involving will
or effort
• A filter
• An allocation of resources
• A spotlight, a selective attenuator
Characteristics
• General agreement regarding two
characteristics outlined by James:
– Bottom up (sensory) attention driven by
environmental events (stimulus-driven;
exogenous)
– Top down (volitional) attention to both
external and internal stimuli (goal-driven;
endogenous)
• May be thought of as automatic vs.
controlled, respectively.
Automatic Vs. Controlled
Processing Distinction
• Automatic processes
– no capacity limitation; fast
– do not require attention, effort
– difficult to modify once learned
• Controlled processes
– limited in capacity; slow
– require attentional resources, effort
– can be used flexibly in changing circumstances
Ventral right frontoparietal network and target detection
(detection of low-frequency targets; these regions not
active in the antipation period)
Stimulus-driven or
bottom-up attention
Corbetta & Shulman, Nat Rev NSci, 2002
Dorsal frontoparietal network for top-down control
of visual attention (these regions active in the anticipation period)
Corbetta & Shulman, Nat Rev NSci, 2002
Functional
Components of
Attention
What gains access to
WM?
Regulation of channel
signal strength
Automatic filtering for
behavioral importance
Knudsen, Ann Rev NSci, 2007
Purposes of Attention
•
•
•
•
•
•
•
Cope with inherent capacity limitation of brain
Facilitate stimulus detection
Facilitate stimulus perception
Facilitate thinking
James
Facilitate memory
Recruit relevant processors
Prepare for action
Varieties of Attention
• Focused attention (processing only one
input at a time)
– Orienting to sensory events
– Detecting signals for focal (conscious) processing
• Divided attention (processing multiple
inputs according to nature of inputs and
goals)
• Maintaining a vigilant state
Varieties of Attention
Dimensions of Attention
• Focality (detection vs. selective attention)
• Duration (brief vs. sustained attention)
• Input channel
– Visual attention
• Spatial
• Object-based
– Auditory attention
• What, where
• Cognitive effect (facilitation vs. inhibition)
Why have an attentional system? Processing economy
-brain as a limited-capacity processor; that which is selected should
be consistent with goals and expectations
Early Visual Attention
• Interface between Attention and STM:
Attentional blink
http://www.rit.edu/~gssp400/Blink/blinkinstr.html
Early Visual Attention (Cont’d)
• Interface between Attention and STM:
Repetition blindness
Early Visual Attention
• AB and RB can be doubly dissociated (Chun)
– Making targets different from distractors alleviates
AB but not RB
– Enhancing episodic distinctiveness of the two
targets eliminates RB but not AB
• May represent different things
– AB represents bottleneck in attentional processing
– RB reflects failure of token individuation
http://psych.hanover.edu/JavaTest/Cognition/Cognition/attentionalblink_instructions.html
Selective Attention
• Selective processing of some information but
not others
• Filter vs. capacity models
• Important research paradigms
– Dichotic listening
http://www.humnet.ucla.edu/humnet/linguistics/p
eople/schuh/lx001/Dichotic/dichotic.html
– Shadowing
Dear Linguist:
I am hoping you might be able to provide some information for me regarding dichotic listening.
I work in a call center in Vancouver, Canada. The position I work in requires that we monitor 2 separate
phone lines at the same time with 2 headsets (one on each ear) for 8 hour shifts. We monitor the content of
messages being left by callers, and subsequently accept or refuse their recorded greeting/message.
There is a fairly large staff (approx. 60 employees) here that monitors the phone lines 24 hours a day, 7 days
a week. The content of the messages is often rude and unsavory - we monitor a singles meeting place chat
connector and callers can get fairly explicit - but this is not the issue that concerns me. The position did
involve listening to single lines before a company expansion over a year ago when it changed to double-line
monitoring. It seems that this double listening is affecting people negatively. Having done this for 8 months
now, I know that I have been experiencing psychological trauma (depression, confusion, irritability, chronic
headaches etc.), as have a number of my co-workers.
My question is this: Is this type of dual listening harmful in any way,or furthermore, have any studies been
done in regards to the effects this type of extended exposure to dual listening may have on the human brain or
psyche?
I have just started looking into this - your name came up in a web search for dichotic listening - and I would
greatly appreciate any information or reference that you could pass on.
Thanks.
-Name, Occupation
Cherry (1953)
• Interested in attentional “popout” of relevant
information (“cocktail party effect”)
• Two messages, same voice, both ears: subjects
could invariably separate messages, but with
difficulty; uses physical/source characteristics
• Dichotic listening with shadowing
– Recalls little if any content from other ear
– Often doesn’t recall language
– Can recognize it as speech/nonspeech, and can
recognize male-female
Broadbent (1958)
• Influential paper on focused (selective)
attention; felt by many to be a critical
‘cornerstone’ paper in cognitive psychology
• Influenced by Cherry’s shadowing results
• Used dichotic listening, and found a strong
tendency to report digits ‘by ear’, thus
reflecting a tendency to select based on
perceptual/physical characteristics of the
input
Broadbent’s Filter Theory
Stimulus #1
Stimulus #2
Limited capacity
STM
Sensory register/buffer
Selection Filter
Problems with Broadbent’s Theory: Recall
across ears can be organized (automatically?)
Problems With Broadbent’s Filter Model
• Sometimes attention doesn’t follow input
source: Gray & Wedderburn (1960)
fan
rage
tic
out
tas
ous
• Can demonstrate that “unattended”
information is processed phonologically or
semantically: Corteen & Wood (1972)
– EDR’s to shock-associated words in an unattended
channel – city name study
Early Vs. Late Selection
•
“Early” theories (Broadbent, Treisman)
– selection takes place well before extensive (e.g., semantic) analysis
takes place
– contradicted by studies showing semantic effects in unattended ear
•
“Late” theories (Deutsch & Deutsch)
– extensive analysis of stimuli takes place before selection
– raises issue of benefits of selective attention
– shadowing delayed by presentation of a synonym in the other ear;
recall biased by semantic interpretation (e.g., bank-river, vs. bankmoney)
Broadbent
Sensory register
Selection Filter
STM
Stim#1
Stim#2
Treisman
Sensory register
Attenuator
STM
Stim#1
Stim#2
Deutsch & Deutsch
Sensory register
Stim#1
Stim#2
STM
A Hybrid: Perceptual Load Theory
• 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
• Early selection occurs when load is high; late
selection dominates when load is low
Evidence for Perceptual Load Theory
rtxpqk
__x___
•
Mean target identification time as a function of distractor type (neutral
[N] vs. incompatible [Z]) and perceptual load (low vs. high). Based on
data in Lavie (1995). Nature of distractor has more effect on time when
perceptual load is low.
A Capacity Model
Kahneman (1973)
• Supplements previous bottom-up
analysis with a consideration of topdown influences
• Emphasizes concept of processing
resources
• Attention and mental effort are strongly
correlated
• Arousal can work to increase processing
resources
Kahneman’s
Model
Understanding the Effects of Attention is Important
When there are Limitations on Processing
• Resource-limited processes
– Processes dependent on the availability of resources that
can be devoted to task solution
– Applying more effort or processing resources increases task
performance
• If output not available until task is finished, then
devoting more resources decreases RT
• If output continuously available, then performance level
increases
• Data-limited processes
– Processes dependent upon the quality of data input, rather
than upon resource allocation
• Applying more resources may have little effect on
performance
• Most processes have both resource- and data-limited
components
Present
Absent
“Yes”
True Positive
False Positive
“No”
False Negative
True Negative
Basic Signal Detection Framework
Focused Visual Attention
• The Attentional “Spotlight” model
– items within a small portion of the field can be seen clearly
– Posner’s ‘covert’ shifting of the spotlight
– Problem: proximity not always facilitative –e.g. Driver &
Baylis, 1989 – common movement
• Zoom-lens model (Eriksen & Yeh, 1985)
–
–
–
–
Magnification inversely proportional to FOV
‘magnification’ can be increased or decreased
Grouping processes affect spatial extent of attention
Problem: attention can be object-based; objects outside the
zoom can be processed or even inhibited
Evidence for Spotlight
• Location-specific facilitation of
attention
– RT’s faster when object appears in cued
location
– search for target letter and report
orientation of “U”; done better when two
letters are close together
T

vs.
T

Focused Visual Attention
• The Attentional “Spotlight” model
– items within a small portion of the field can be seen clearly
– Posner’s ‘covert’ shifting of the spotlight
– Problem: proximity not always facilitative –e.g. Driver &
Baylis, 1989 – common movement overrides proximity
• Zoom-lens model (Eriksen & Yeh, 1985)
–
–
–
–
–
Magnification inversely proportional to FOV
‘magnification’ can be increased or decreased
Grouping processes affect spatial extent of attention
Split attention (Awh & Pashler)
Problem: attention can be object-based
Experiments Demonstrating Split Attention
(a problem for the zoom-lens model)
• Awh and Pashler
(2000).
• (a) Shaded areas
indicate the cued
locations and the near
and far locations are
not cued; (b) probability
of target detection at
valid (left or right) and
invalid (near or far)
locations.
•
Based on information in Awh
and Pashler (2000).
Object-Based Attention
Behrmann, et al (1998) JEP: HPP, 24, 1011-1036
Inhibition in Attention
Inhibition in Attention (Cont’d)
Negative Priming Paradigm
Prime
Probe
Task = name red drawing
Attended repeat
Unrelated
Unattended repeat
Attention as Facilitation
A
Preattentive Processing
Further Cognitive Processing
Stimulus 1
Stimulus 2
ATTENTION
Attention as Inhibition
B
Preattentive Processing
Further Cognitive Processing
Stimulus 1
Stimulus 2
ATTENTION
Inhibition of Return (IOR)
• “A reduced perceptual priority for information in a
region that recently enjoyed a higher priority”
(Samuel & Kat, 2003, p. 897)
• A bias favouring novel locations and objects
• Posner and Cohen (1984)
– IOR due to inhibition of perceptual and/or attentional
processes
• Taylor and Klein (1998)
– IOR due to inhibition of motor processes
• Prime and Ward (2004) ERP study
– IOR is a perceptual phenomenon not in motor response
Models of Visual Search
• Feature integration model
– rapid initial parallel process not dependent on attention
– subsequent attention-dependent serial processes in which
features are combined to form objects
– attention provides “glue” binding features together into an
object; can only do this one object at a time
– feature combination affected by stored knowledge
– without focused attention, features combined randomly,
producing ‘illusory conjunctions’
Display Size Effect
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
T
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
Find the “T”
L
L
L
L
T
L
Feature Search & Feature Singletons
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
O
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
Find the “O”
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
T
O
O
O
O
O
O
O
O
O
Find the “T”
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
Treisman & Gelade (1980)
L L L T L L T L L
L T L L T L L T L
L L L L T L L L T
L L T L T L T L L
L T L L L T L L L
L L L L T L L T L
L T T L L L L L L
L L L T L L L L T
T T L T T L L L L
L L T L L L T L L
L T L T L T L T L
L T L T L L L T T
L L L T L T T L L
Find the “Green T”
S L R M T N
X A Y
W S B X N
I
F M I
Q W A S Z
X L K P
M B E W E
R U
I N
S L R M T N
X A Y
W S
F M I
B X N
I
Q W A S Z
X L K P
M B E W E
R U
S L R M T N
U O W L
I N
X A Y
X P D G
H
Find the “B lue Letter”
Feature Conjunctions & Illusory Conjunctions
Find the “$” (or
“is there a $?”
Find the triangle (or
“is there a triangle?”
Similarity vs. Feature Conjunctions
Q
@
A
?
B
*
%
E
P
^
$
*
@
W
!
M
*
>
Q
K
&
O
S
G
B
W
E
&
J
P
/
A
C
#
!
W
+
N
%
+
F
+
Q
(
Z
X
N
*
Q
(
T
K
T
)
H
A
S
O
S
B
K
I
Y
&
S
Y
W
F
^
T
Y
T
Z
J
C
Z
Q
{
[
B
I
@
U
K
P
%
D
K
R
C
H
#
X
O
G
K
C
U
\
A
M
L
D
X
S
K
U
D
=
V
*
)
B
N
S
>
}
r
@
a
?
c
*
x
f
q
^
@
#
@
s
y
b
!
#
p
l
!
s
t
h
c
p
Find the “R”
t
n
r
<
]
b
d
#
“
x
?
o
*
+
k
+
q
“
a
%
m
*
s
)
u
j
t
)
l
e
t
p
t
c
l
j
z
/
w
y
w
f
@
u
z
^
a
k
%
a
b
+
[
l
R
(
v
l
q
=
e
l
s
d
o
#
s
g
p
l
d
v
}
b
n
m
e
t
s
i
u
e
+
w
+
*
c
=
\
<
}
Key Takeaway Points
• Generally agreed that two processes are involved in
visual search (parallel & serial), though recent
neuroimaging data suggests otherwise – substantial
overlap in brain areas involved
• Different visual features are processed independently
• Speed of visual search depends on set size and
similarity of targets to distractors
• Perceptually grouped objects will be selected or
rejected together; grouping probably takes place
prior to attentional ‘enhancement’
Posner’s paradigm
Posner’s paradigm
Posner’s paradigm
Posner’s paradigm
•
Posner’s paradigm
Posner’s paradigm
Posner’s paradigm
Posner’s paradigm
•
Posner’s components of
attention
• Three components involved in visual
attention:
– Disengage attention from a given stimulus
– Shift attention from one stimulus to
another
– Engage attention on a new stimulus
Processing Components of
Attention (Posner & Rothbart)
• Alerting (NE)
• Orienting (ACh)
• Executive (DA)
Dual Task Performance
• Relevant to processing capacity
• Interference methodology a useful tool to
determine whether two tasks share resources
• What determines degree of interference?
– Task similarity
– Task difficulty
– Practice/expertise
• Sensitivity (d') to auditory and visual signals as a function
of concurrent imagery modality (auditory vs. visual).
Adapted from Segal and Fusella (1970).
Multiple-resource Theories
• Wickens (1984).
A proposed dimensional structure of human
processing resources. From “Processing resources in attention” by Wickens,
C.D. in Varieties of Attention edited by R. Parasuraman and D.R. Davies ©
1984 by Academic Press. Reproduced by permission of Elsevier.
Norman & Shallice
• Three levels of functioning:
– Fully automatic processes, controlled
by well-learned schemas
– Partially automatic processing,
controlled by contention scheduling
– Deliberate control by a conscious,
supervisory attentional system
Summary: Important
Concepts
• Limited-capacity for informationprocessing (information bottleneck)
leads to selective attention
• Attentional acts take time and effort
• Attentional control re: goals and plans
• Automatic vs. Controlled processing
• Attention and consciousness