Orienting of Attention: Spatial Cognition Article

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/233813483
Orienting of attention
Article · January 1980
CITATIONS
6,456
READS
6,272
1 author:
Michael Posner
University of Oregon
468 PUBLICATIONS120,006 CITATIONS
SEE PROFILE
All content following this page was uploaded by Michael Posner on 30 November 2016.
The user has requested enhancement of the downloaded file.

Quatte y Jowad of Experimcntal
Pqchology (r98o) 32, 3-25
ORIENTING OF ATTENTION-
MTCHAEL I. POSNER
Univerity of Oregon
Bartlett viewed thinLing as a high level skiU exhibiting bauistic properties that he
called its "point of no retum", This paper cxploles one aspeci of cognition
through thc usc of a eimple model tark in which human subjicts are asked to
commit attention to a porition in viaual apacc other than 6xation.' This instruction
is orecuted by orienting r covcrt (attcntional) mechanism that seems
sufficiently
time locked to cxtemal cvcnts tlrat it! trajcctory can b€ traced across the visual 6eld
in terms of momentary changes in the emciency of detecting 6timuli. A com-
parison of results obtain€d with alert monkeys, brain injured and normal human
subjects 6hows the r€lationship of this covert rystem to saccadiC eye movements and
to various biain ststems controlling petception and motion. In eccordance with
Bartlett's insight, the possibility is erplored that sinilar principles apply to
orienting of attention toward 8cnsoqlinpltnnd orienting to the aerunticstr:ucrur€s
used in thinkins.
Introduction
Sir Fredcric
Bartlett wrotc
a book,
Thinkbg, during the
last
part
of his life (Bartlett,
r958). It is not as
widely known
as
his earlier
work, Remanbning
(Bartlcit, r93z),
butjt had
a strong impact
on me, perhaps
because it was among
the irst psyct
ology
books I read. Bartlett's
theme
was
as simple as it was powerful. Thinking iJa
skill and should be studied
with the techniques
that had proved
successful
i; the
study of other skilled behaviour. In particular, I was struck with Bartlett's
metaphor
that thinking like swinging a bat, has a ,,point of no return". Once
committcd
in a particular
direction,
thought
is ballistic in that it cannot he altered.
It nray be hard to understand
why this idea should haye
been so exciting to
someone reading the psychological
literaturc in r959. In retrospect,
ihat
captured the imagination
must have
becn the idea that a hidden psychological
process
like the formation of a thought might be rendered
sufficiently
concretc
to
measure. Twenty years
later, when
psychologists
routinely measure the speed of
rotation of visual images (Cooper
and Shcpard,
1973) or the time needed
to scan
the next item of an internally stored
list (Sternberg,
1969),
it is hard to reinstate
the excitcment
that thc prospect
of such research
could
have engendered
in at lcast
one
reader
of Bartlett's
book.
During the last few
years of rcscarch
on human cognition,
there
has grown
up a
numbcr of similar views irf how the human ncrvoui system is organizcd
in ihe
performance
of species-specific
human behaviour
such
as reading(LaBcrgc
and
t Text of the Seventh Sir Faederick
Bsrtlctt Lccturc gil|en at a mecting of the Experimental
Psychology
Socicty in OdorJ, 5 July t979.
oo33-555x/8o/orooo3
+ 23 goz.oo/o @ r98o The Experimental
Psychology
Society

M. I. POSNIiR
Samuels,
1974; Posner,
1978). 'l'he
idca
of a limited
capacity
attcntional
systcm
has becn a central
feature
of these
views. Although some
have arsued ihat a
skills
approach is antithetical
to the study
of inteinar attentional
i,echanisms
(Neisser,
1976),
most work on skill has
also
assumcd,
with Bartlett,
the importance
of mechanisms
of limited capacity (Broadbent,
ry77).
__-Currently,
the study of spatial
attention
in alert ruonkeys
(Mountcastle,
r97g;
Wurtz
and
Mohler, r976;.Robinson,
Goldberg
and
St4nton,
r97g), brain
injuicd
persons
(Weiskrantz,
Warrington,
Sanders
and
Marshall,
r974)
and normal
subjects
(Posner,
1978, Chapter
7) seems
to me to be a most promising
model system
for
relating an imponant component
of complcx
humarr
cognition
to studies
of the
neural
systems
underlying performance. while orientinglo stimuri
in visual
space
is a restricted
sensc of attention, I believe
that its study
is capabte
of providing us
both with important
testa
of the adcquary
of general
models of human "oqniiinn
and with new insights
into the role of attention
in more complex
humen aitivitv.
Accordingly,
this
paper
will be devoted to a discussion
ofthe rcsults
of exoerime.ris
in human
spatial attention
and a comparison
of them with animal approaches
to
the same
topic. If there should
emerge
satisfactory
convergence
between
human
performance
and physiological
approaches
with this simple
model system,
I believe
that the pgr.chological
metlods used
to explorc
attertion
in more
comrlcxlasks:will
rcceive
added support. In addition,
studics
of human
p"rf,,r-ancc may help
investigators
of neural
systems
toward the needed
integration
of their stuiies of
separate anatomical
structures.
Orintiag
I will use the term orienting
to mean
the aligning
of attention
,rvith
a source
of
sensory
input or an internal semantic
structure stored
in memory. The term
orienting
has been
closely tied to a reflex (Sokolov,
1963),
the operation
of which is
indcxed by a variety of autonomic,
CNS and overt chanees. The idea of an
orienting rcflex
is related
1o
the
mental
operation
oforienting
as I use
it. However,
the orienting reflex does not distinguish
between
aligning of attention and the
resulting perception
of a stlmulus.
Detcct;ng
- I distin_guish
orienting from another cognitive
act that I call tletecting. By
detecting I will mean
that a stimulus
has reached
a level
of the nervous
sysrem
ar
which it is now possible
for the subject
to report
its presence
by arbitrary
,".pon.""
that the expcrimenter
may assign. These
may be verbal (,,i scc it',)-or manual
(pressing
a kcy). I)etccting
mcans
to be
aware
or conscious ofthc stimulus. 'I.he
distinction bctwccn orienting and detecting allows onc to cxplore thc pro-
position
that some
responses
(e.g.
saccadic
cye
movements)
may bc available
io a
stimulus
before it has been
detected
in the
sense
used here. This distinction
makes
it reasonable
that a normal
subject
may movc
his
eyes toward
a stimulus,
but not be
able
to otherwise
report it, or that a brain damaged
subject
might have impairments
in <ictecting an
event
to which he can orient (Weiskrantz,
Warrington,
Sanders
and
Marshall, r974).

ORIENTINC OF ATITNTION <
Locus
of control
It is also
important to make a distinction between
external
and central
control
over
orienting. If oricnting
to memory
and to external
stimuls
events
is to hat/e a
common
base'
it is clear that we must be able to orient attention
in the absence
of
an external
stimulus. . Srmilarly,
movements
of the eyes
can either
be driven by
strmulus
rnput or result
from a search
plan internal
to the organism.
Ooert and coaert
orienting
Finally, it is important to distinguish
between
overt changes
in orienti+g
that
can
\c observed
in head
and ey€ movem€nts,
and the purely
iovert oricnting that
may be achieved
by the central
mechanism
alone. In order to make
this distinc_
tion, one must
be able
to measure
covert
orienting
by means
other
than observation
of overt head
and
eye movements. With humarisubjects
it is possible
to manipu,
late the direction of attentio-n
by an instruction,
ly changing
ihe probability of a
target
event,
or by the use
of appropriate
overt movements. To measure
whether
orienting
occurs,
changes
in the efficiency
of detecting
events
that occur
at various
spatial positions are examined. Some
variant of riental chronometry
(posner,
r,978).such
as
rcaction
time (Posner,
Nissen
and Ogden,
r97g),
threshold
detection
(liTrDCto-ttE), evoked
potential_
amplitude
{Iffiris and
Hillyard, 1977)
or changes
in firing ratcs.of
single
cells
(Mountcastle,
r976)
can
be used
as
a deiirrl
dent measure
of processing
eflicicncy.
..
It is important to keep in mind thi dcfinitions
of orienting
and detecting
and the
distinction
between
external and
central
control
as
we revieJexperimental
evidence.
Evidencc
is examined
in four major sections. The first establishes
the ability of
subjccts
to shift attention
around
the visual field in accordance
with instructions.
8y measuring both the facilitatory and inhibitory efiects of odenting on the
efficiency
of detection,
it is possible
to examine
ihe relationship
of tf,e covert
attentional
mechanism
to the fine structure of the retina. The sccond
section
supports
the idea of analogue
movements
of attention
across
the visual
field that
are measurcd by time locking of attention shifts to external cucs. The third
section examincs
the relationship
bet\,veen
movcmcnts of attention :rnd overt
changes
of eye
position. In this section
it is possible
to compare
our results
with
those
arising_from
single
cell recording
and to distinguish
betwecn
theories
out_
lining the relationship
between
percepiual
and moveirent systcms. The fourth
section deals
with thc crucial role of peripheral
stimuli in iontrolling attention
movements. The.concluding
portion of the paper
examines
the impiications
of
our results
on spatial
attcntion
for more
complex
performance.
rt is not
obvi<,us
rh* ,h,f,hs:j;;l 11"""H",""r*s anything
rnore
than
the
movemcnt
of thc cycs
to positions
in the visual
6eld. Ceriainly,
no one
would
drspute
the close
connection
between
movements
of our eyes
and
slifts of attention.
Nonethelcss,
thcre has always
been
speculation
th"t oie c."r. shift attention in-
oependent
ot eye movements.
For
example,
Wundt (r9rz, p. zo) commented
on

M. t. POSNER
the ability to separate
the line of fixation from the line of attcntion. Natural
language refers
to the ability to look out of the corner
of our eyes,
and athletic
coaches
instruct
their players
to do so in order to confuse their opposition.
Many experimental studies
using
methods
of mental
chronomet'ry
lGrindley and
Townsend,-r968;lVleTens,
r956; Mowrer, t9+r ; Shiffrin
and
Gard.,"r, ,972; w"re
not successful in showing
this ability,
at
least in empty
visual
fields. More iecently,
successful
rcports of attention
shifts in the absence
of eye
movements
have
bein
frequent (Eriksen
and Hofrman, 1973; Klein, 1979;
posner,
Nissen
and Ogden,
r9?8; Shaw, 1978;
Von Voorhis
and Hillyard, 1977;Wurtz and
Mohler, r976j.
Wc (Posner,
Nissen and Ogden, r.978)
sought
to determine
*h.thei .esporr"e"
to clear above
threshold luminance
increments
in dark fields would occui morc
quickly when subjecte
knew where the stimulus
would occur
than when they did
not. Wc used
differences
in rcaction
time to a stimulus
at expected
and
unexpected
positions
in the viaual field as a measurc
of the cfficicncy
of detection
due to turning
ettention
toward
thc expected position. To insure
that the difiercnccs
in rcactio'
time did not depend
upon_shifting
one's
eycs,
we
monitored
eye
rnovements
by use
of EOG. We used
only those
trials in which the eyes remained
fixated. In order
to elimimte "_":T tTpgnT preparations es a contributor, we used either a single
key (simple RT) which the rubject pressed
regardless
of where the stimulus
occurred
or made the response
choicc
unrelated
to stimulus position. Figure r
illustrates
the sequence of event8 within a trial. The subject
was
presentei
with
a
plus
sign or an arrow pointing
to thc right or [eft. If the plus
sien
was
Dresented.
the detection
stimulus was
equally likely to occur
to the ieft or-rieht of fi""tion.
If an arrow
was presented,
the probability
was o.8
that
the
detection
itimulus would
D.t.ction
- .tl6ut!r
,oo-6d6i-{-,oio "$!!e!e$o ri,ins
+ Cu.
Fi
oor."'iion
votid c!.
I
_ t ,"i"i,"1.".
Ftounr r. Organizrtion
of individual tiiah ilr thc ccntral
cuc experiments.
occur on the indicatcd
side (valid) and o,2 that it rryould
occur on the other side
(invalid). One can
then examine
both the benefis from knowing
rvhere
in space
the stimulus
will occur, and the cost when it occurs
at a positioi other than the
expected position.