1-B-30 視覚的注意の時間的処理限界と順序知覚(2002年度 日本基礎心理学会第21回大会優秀発表賞)

The Japanese Psychonomic Society

NII-Electronic Library Service

The Japanese

t/t

'x.

Psychonomic Society

TheJmpaneseJburnalofRsyc'honomic Science

2003, VoL 22,No, 1,31-32

Prizewinner's

Summary1-B-30

Temporal

limits

of

visual

attention

on

order

_perception

Hiroyuki

TsuBoMi

and

Naoyuki

OsAKA

KIyotoUitiversit),*

When

four

letters

are repeatedly presented

by

the

RSVP

method with a gradually

increasing

duration of cxposure, an observer can

identify

the lettersbut misreports theirorder

(Holcombe,

Kanwisher, & Treisman, 2001). Holcombe et aL

(2001)

argued thattheorder illusioniscaused by

an observer failingto directtheirattention tothe arbitrary initialitem. Hewever, our findingsin

Experirnent1

of thepresentstudy indicatedthattheaccuracy of the order does not depcnd on the

ease of directingattention tothe$alient item, Inaddition, the erder

illusion

occurred when the

targetletterswere _presented ina deceleratingmanner. We

found

in

Experiment

2

thatan order

illusion

occurred when

the

observer attended

to

a

lctter

pre$ented very brieflyand was required to

report theorder of thesubsequent lctters_prcsented within 466 ms. These results suggest thatan

order illusioniscaused

by

the temporal

limits

of visual attention which also cause an attentional

blink,Key

words: visual attention, order _perception,attentional blink,RSVP

When four lettersare repeatedly _presented by the

RSVP

method with an exposure

duration

which is

gradually

increased,

it

is

more

diMcult

toperceive

their

order

than

whcn the lettersare _presented only

once

(Midstream

Order Deficit;Holcombe,

Kan-wishcr,

&

Treisrnan,2001X Holcornbe et aL

(2001)

argued thattoperceive theorder correctly attention needs to be directed to the arbitrary initialitem. When theitems are _presented once thereissaliency

between

the

initial

target

letter

and

the

_preceding

distractor

item,

therefore

it

attracts exogeneous

at-tention.However, there isnot a sa]icncy when the

lettersare _presented repeatedly, so an observer must

direct

endogeneous attention to

the

arbitrary

se-lected

letter.

Bccause

endogeneous attention

is

driven more slowly than exogeneous attention, an

order

illusion

results,

However,

in

allthe conditions where an order

illu-sion occurred, the targetletterswere presented

re-peatedly with deceleration.The

deceleration

may

be

the cause of the order illusion,We therefore first

examined which factorscause the order illusion.

Experiment

1

Method

Eight

naive university students

partici-* _{Department of Psychology,}

_Graduate

_School

_of

Letters, Kyoto University, Yeshida-honmachi,

Sakyo-ku, Kyoto 606-8501

pated. The stimuli were displayed inthecenter of a

gray-background computer monitor. Figure 1(a)isa

schematic

diagram

of

Experiment

1.

Each

trial

be-gan with a

fixation

display

for

357

ms, and thiswas

followed

by

a

display

of

20

black

letters,

at the same

location,without any interstimulus blanks, The

seven conditions which were used wcrc

the

combina-tionsof thefollowing:single or cycle; deceleratingor

not; and containing a salient item or not,

In

the

sing]e condition

the

firstletterwas _presented

for

17

ms and the exposure duration of subsequent letters

gradua]]y

increased

until the ninth serial _position.

After thc tenth

item,

the

letters

were presented

for

183 ms, Four lettersof Hiragana which were targets

were _presented once

between

the

tenth and

thir-teenth position,and the other

distractors

were

Ka-takana

letters.In thecycle condition

four

lettersof

Hiragana targets were repeated fivetimes, Inthe

decelerating

condition the exposure duration

gradu-ally

increased

as

in

the single condition.

In

thc

non-decelerating condition a]] of theitems were pre-sented ferthc same _periodof 183 ms each. Inthe

saliency condition the

letters

were ina set of four

which consisted of three

Hiragana

letters

and a sali-ent item which was either a Katakana

letter,

a red

colored Hiragana letter,or a digit. In the

non-saliency condition a]1 of

targets

were Hiragana

let-ters.

At

the end of a sequence, a cross sign was

The Japanese Psychonomic Society

NII-Electronic Library Service

The JapanesePsychonomic Society

32

The

Japanese

Journal

of

Psychonomic

Science

Vol.22,No, 1

(a)

a

:yOcniedeceler{

C singlc

Cms]

+"di"JT4its,r"tts,'+'-f;s.<ivtig'iat/:3iT Ei!bX?!a.k.

.ega...

_±

7...T

astt7r i-ft7lra-[:7(-tt 7rasrtx

sS)/I S333se K31oo/-leJF--- /9a- :sJ

+7Se-t]AhS(S[:-"7]40Vtx

(b)

]/

g;:f:,,

[

+s,-"it$,T"lttsLTJyt ±sL(asttssc-r;x withKatakana+7rbit71i-ft7lriE-[:71bt:)(asf:x withcolor +S.r-tifi.ttstt.$･(hr:.S.(bttfi.resf:x witbdigit +21E+trk21bf:2K-[:2r-t:2(asf:x Figure 1.

_{a)

A targetletters Experiment 2. 4ee3.,"W2euUltsUo

schematic diagram of the conditions of Experiment

for each condition of

Experiment

1.

(c)

The

mean

presented to

indicate

thatthe four targetsequence should

be

verbally reported. The observer was per-mitted toreport any letteras the

first,

except

in

the

single condition,

ResultsandDiscussion

The results of

Experi-ment 1are shown inFigure 1(b).A one-way ANOVA

and

Tukey's

HSD

of the order accuracy

demon-strated thattheorder accuracy was smaller when the

target was

_presented

with the deceleration

condi-tions

_(D-G)

than the other conditions

(A-C).

There

was little

difference

ofthe

identification

accuracy in any condition.

The

results suggest that the order

accuracy does not depend on the ease of directing

attention tothe salient item,and

that

theorder

illu-sion may

be

caused

by

the temporal

limits

ef visual

attention. These

_proposats

are supported by a report

of Raymond, Shapiro,

&

Arnell

(1992)

who noted that

directing

attention to a

briefly

_presented

itern

in-duced an

inability

toprocess an

item

which followed

it

within 500 ms

(attentional

blink}.InExperiment 2, we exarnined whether directingattention toa

letter

which was presented

briefiy

might cause an order

illusion,

Experiment

2

Method

Eightnewna)'veuniversitystudentspar-ticipated.The method was identicaltothat of

Ex-periment 1,except as follows.Allof thestimuli were

Hiragana letters,with Target

1

colored _green,

Tar-gets

2-5

colored red, and the

distracter

colored

brown. The firstfive letterswere distractors

pre-sented with a _gradually

increasing

exposure

dura-tionof

17,

33,

33,

50,

and

83

ms,

Tl

was presented

for

100

rns

between

theeighth and tenthserial

posi-(c)t4

es3b&2:

=Ilk,.D.r":,}[dm"fie-iltin

ol:

:Ti

_LE.",O:ge,,}

orkr

Ue A

･H

[:E fi G 1234SG7

Condhion Tl-TlLag

1,

(b)

The

mean number of correct

number of correct T2-T5 lettersin

tion,and was followed

by

distracters

for

183ms.

The

number of

distractors

before

T2

ranged

between

O

_(Lag

1)

and 6

(Lag

7). T2-T5 were _presented

consecutively for183 ms, after which

two

distractors

and a cross sign were

displayed,

The

ebservers were

required toreport

Tl

and thesequence of T2-T5 in

the

Tl-attended

condition, and only T2-T5 in_the

Tl-ignored condition.

Results

and

Discussion

The

results of

Experi-ment 2are shown inFigure 1(c).A two-way ANOVA

<condition

×T1-T2 Lag) of the T2-T5 order accuracy

indicated

that

the

Tl-ignored

condition

led

to

a

bet-terorder accuracy than theTl-attended, These

dif-ferences

were much

larger

in

Lag 1-3 and became

smaller as the Lag increased, There was

little

dif-fcrence inthe identificationaccuracy

for

all of the

conditions.

Thcse

results suggest thatan order

illu-sion occurred when theobserver attended toa briefiy

presented

letter

and was required

to

report the

se-quence of the subsequent

letters

presented within

466 ms.

Consequently,

we concluded that theorder

Musion

is

caused by the temporal limitsof visual

attention, which

is

a

factor

common with attentional

blink.

References

Holcombe, A.O. Kanwisher,

N.G.

&

Triesman,

A.

2001

The

midstream order

deficit,

llarcaptinn

&

Rsychophysics,

_63,

_322-329,

Raymond,

J.

E.

Shapire,

K.L.

&

Arne]1,

K.M. 1992

Temporary

suppression of visual _precessing inan

RSVP task:An attentional blink?

fournat

of

Ex-pen'mental

Psychology:Hdeman R?rceptionand

PeF

formance,

18,

849-860.