2-B-33 感情誘発刺激を用いた注意の瞬き現象モデルの検討(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 JapanesePsychonomic Society

77ieYkiPanesefourvzal of PsychenemicScience

2003,VoL22,No.1,5a-54

Prizewinner's

Sumrnary2-B-33

Attentional

blink

with

negative

stimuli

Serial

or

_parallel

_processing

?

Ken

KiHARA,*

Naoyuki

OsAKA,*

and

Yoshiaki

NAKAJiMA**･

i

K)roto

Uitiversily*

and

Osaha

Uleiiversity**

In

a rapid serial visual _presentation

(RSVP}

_{ta$k,the processing} of a _preceding

target

(Tl}

interrupts

the

response toa subsequent target

_(T2).

This _phenomenon

is

thought toindicatethe

temporal

limitations

of attention and isnamed attentional blink

(AB).

In _previous studies

two

conflicting medels

have

been

_{proposed for theAB phenomenon: theserial processing model and}

theparallelprocessing model,

The

presentstudy examined which could bemore appropriate,

We

used negativity bias,which isa _phenomenon that negative stimuli are more rapidly _processed

because they attract more attention than neutral stimuli. When T2 was negative the

AB

deficit

decreased

(Experiment

1).However, when Tl was negative the AB deficitincreased

(Experiment

2>.Moreover, when

both

Tl

and

T2

were negative the AB deficitsobserved

in

Experiment 1and

Experiment

2canceled each other

(Experirnent

3),

It

was therefQreconsidered that

the

AB was

due

toattentional

interference

_quring

parallel processing of RSVP

items.

Key words: attentional blink,negativity

bias,

_parallel

_processing

model

Raymond,

Shapiro

and

Arnel]

(1992)

demonstrated

atemporal limitationof human attention by using a

rapid serial visual _presentation

(RSVP)

task.

When

people were required to _process a target

(Tl)

this

taskinterrupted theresponse toa subsequent target

(T2)

that

appeared within approximately 500 ms of T

1.

The

investigators

named this_phenomenon

"atten-tional"blink

_(AB}

because

they

found

thatthe

failure

todetect T2 was caused not by sensory factor$but

by attentional

factors.

The models of AB which

have

been

proposed

in

_previous studies can

be

divided

into 2 _groups: whether

failure

_torecognize T2

oc-curred as a result of

delay

and extinction of T2

representation caused by serially _processing of

RSVP

items;

or whether failuretorecognize

T2

oc-curred through interference

during

the_parallel proc-essing of _the

items.

In the present study we tested an

AB

task

for

which

both

of the _present rnodels would _predicta

differentoutcome. To be more specific,

both

of the

present models _predict that the AB deficitwill

de-crease when T2 contains items which attract

atten-tion. This isbecause

the

serial processing model

* _{Graduate School of Letters,}

_Kyoto

_University,

Yoshida-honmachi,

Sakyo-ku,

Kyoto

606-8501

** _{Graduate School of}

_Human

_Sciences,

_Osaka

versity, Yamada-eka _{1-2,Suita,Osaka 565-0871} i

_Now

_atthe

_Graduate

_School

_of

_Hurnan

_Sciences,

Waseda

University.

predicts the saliency

items

would show a lasting

representation, whereas the _{parallel processing}

model predicts that thesesaliency

items

would tend toovercome interferencefrom other

items.

However,

both

models would

differ

in

the _predictionof an AB

deficit

when the Tl items attract attention. The seria] _processing model would _predict that the

AB

deficit

would

decrease

because

_the saliency items

could

be

_{processed quickly} and thereforeT2 would also

be

processed prornptly. Incontrast,

the

_parallel

processingmodel would _predictthat T1 would cause

a strong interferencewith theneutral T2 and make

the

AB

deficit

increase.

We

used negative words a$ attention-attracting

items

because

many studies of other

investigators

have _proposed that negative stimuli attract _greater

attention and are _processed more rapidly.

This

phe-nomenon

is

called negativity bias

(e.g.

Cacioppo

&

Gardner, 1999). In addition,

Ogawa

and

Suzuki

{2001)

have

shown thata negativity

bias

appears in

the

AB.

Therefore,

we consider that negativity

bias

is

an effective

_phenomenon

with which to verify an

AB

model, InExperiment

1

we used negative words for T2 toreconfirm _the appearance of a negativity bias in the

AB.

In

Experiment

2 we used negative words

for

T1 toexamine which was amore

appropri-ate AB model. Experiment

3,

using negative words

The Japanese Psychonomic Society

NII-Electronic Library Service

The JapanesePsychonomic Society

54 The

_Japanese

_Journal

of

Psychonomic

Science

VoL

22,

No.

1

Table 1

Mean

detection

rate of T2 when Tlwas

identified

correctly,

Lag

1

Lag 3 Lag 7

NeutralNegative

NeutralNegative

NeutralNegative

Experimentl

M

SD

Experiment2 M

SD

Experiment3

M

SD

,31a(.26).36a(.25),45a,

b(.29)

.50b,

,(.26>.20b(.17).36aC20)

.41ab<.29).34aC24),52b(,26)

.52c(.25).45aC31).62cC23)

,83d(.25).74c(.22).74d(.22)

.83d(.20).74eC16).83dC15>

Note.

The

mean of an

proposition deduced from Experiments 1and 2,

Method

Partieipants There were 21,

14,

and

25

partici-pants who

took

part

in

Experiments 1,2,and 3,

respectively,

Stimulus A listof negative and neutral two-letter

Kanji

words were selccted

from

Gotoh

and

Ohta

(2001).

All

of the words were controlled forvisual

complexity and

familiarity.

Procedure The RSVP consisted of Tl, T2, and

distractors.

All of the words subtended a visual angle of

2,50

in

width and

1.250

in

height,

A white

display was uscd forTl and the other words were

displaycd

as b]ack against a _gray

field.

Each

word

was presented

for

80

ms with an interstimu]us

inter-val ef20 ms,

The

number of stirnuli occurring before

T1 ranged from 8to12,

in

a random order.

A

display

of

T2

occurred as the first, third or seventh item after

T1,

and each

T2

pesition was termed

Lag

1,

Lag

3,

or Lag 7,respectively. The participantswere asked to

identify the white target words and to detect the

biack

targetwords

that

appeared at thebeginning of

the trial

for

2,OOOms.

In

Experiment

1,neutral and

negative T2 were presented.

In

Experiment 2,

neu-traland negative Tl were _presented.

In

Experiment

3,

both

targetswere neutral or negative. Inall of the

experiments the

distractor

words were neutraL

Results

and

Discussion

The

dependent

variab]e was the

detection

rate of

T2, based only on trials

in

which Tl was identified

correctly. To testtheeffect of the negative stirnuli. a

T2

Lag

(Lags

1

vs.

3

vs..7)XAffectivity

(neutral

vs.

negative) within-subject

ANOVA

was carried out on

experiment with a differentsubscript

differs

significantly atp<.05.

each expenment.

The

T2 detection rates and the

results of the _post

hoc

analyses are _presented in

Table 1.Experiments

1

indicated

thatthenegativity

bias

affected the AB, The results of Experiment

2

corresponded with _predictionsbased on the

parallel

processing model.

Therefore,

it

was suggested

that

RSVP iternswere _processed

in

_paralleland that an

AB

occurs

due

to

an interferencebetween

T1

and T2,

Experiment

3

indicated

thatthe AB deficitsshown

in

Experiments 1 and 2 canceled each other, which

$eerned to be a reasonab]e result.

From

the total results

it

was suggested thatthe AB was due toan attentional

interference

during

_{parallelprocessing}ef

RSVP items.

There are several studies supporting the serial

model of the

AB,

howcver,

and we believe these

findings do not conflict with our results, Itis

as-sumed

that

emotional processing

differs

from

identi-cal _processingand thattheseare _performed

in

paral-lel.Therefore, with the assumption thatemotional

processes interferewith each other

but

identical

processes are serial, the result of an

investigation

using emotional

items,

as used

in

our study, could

imply thatitems are processed

in

paraLleL

References

Cacioppo,

J.

T.

&

Gardner,

W, L, 1999 Emotion.

nual Review

of

Psycholog),,

50,

191-214,

Gotoh,

F.,

&

Ohta,

N.

2001Affective

valence of compound kanjiwords, 71sukuba

Rsychological

search,

23,

45-52

(in

Japanese),

Ogawa,

T. & SuzukL N, 2001

On

the saliency of

negative stimuli:

Evidence

from attentional blink.

Manuscript

subrnitted for _publication,

Raymond

J.

E. Shapiro,

K,

L.

&

Arnell,K.M. 1992

Ternporary suppression of visual _processing

in

an

RSVP task:

An

attentional

blink?

Jbumal

of

Perimentat

RsycholQgy: llbemanPlercePtionand

formance,

18,849-860.