乳児の顔認識に対する運動情報の効果(第23回大会 優秀発表賞抄録)

The Japanese Psychonomic Society

NII-Electronic Library Service

The JapanesePsychonomic Society

Thellapanese

_fournal

e.fPsychonomic Seience

20e5,VoL24,)"o,1,125-126

Summary

ofAwarded

PresentationIP75

The

effect

of

motion

information

on

infants'

recognition

of

unfamiliar

faces

Yumiko

OTsuKA,i･

2

So

KANAzAwA,3

Masami

K.

YAMAGucHi,i

Alice

J.

O'TooLE,4

and

Herv6

ABDi`

Chuo [iniversilyi,

jopan

Sociedy

for

ehe

l]},omotion

of

Science2,

Shuhutohu Ciniversit.y3.Universit.v

of

Texas at Dallas4

We

examined the role of motion iniormation on

infants'

recognition of unfamiliar faces.

Previous

studies suggested

that

motion

information

promotes infants'_perception

_(Kellman

&

Spclke, 1983; Otsuka

&

Yamaguchi,

2003),

and thereforewe theorized that motion information

should facilitate

infants'

_{face recognition, In}the _present study, we cornpared infants'recognltion memory

for

unfamiliar

faces

learnedinamoving or a static condition. Infantsaged

3-

to5-months

(N'--

24)w'erc familiarizedwith asmiling

female

face

eithcr inthemoving or static condition, After

familiarizution,infant$were testedusing a _pairof novel and

iamiliar

female

Caces,

We

feund

that

the infantsshowed a significant _preference for novel

faces

only in the moving condition. The

present results suggcst thatlearning from _thernoving condition _promotes

infants'

recognition of

unfamiliar faces.

Key

words: moving

face,

unfam"iar face,infants

Therc are two hypothcses about the possib]e role

of motion information

in

_thc

face

recognition: the supplemental

informatton

hypothesis and

the

repre-sentation enhancement hypot.hesis

(O'Toole,

Roark,

& Abdi, 2002). The supplemental

inforrnation

hy-pothesis positsthatmotion information contributes

to

face

recognition through supplying additional

identity

information

to

the

invariant

structure of

theface

_(e,g.

facialmovements characteristic _tothe

individualsuch as the way of srnJling).

This

hypo-thesispredicts that tnotion information weuld be

more effectivc for familiar

faces

than unfarniliar

faces.

This

is

because

some experience with a face

may be needed tolearnwhich movements are char-acteristic te the individuaL The representatien

en-hancement hypothesis _positsthat motion

informa-tion contributes to

face

recognition by facilitating

the perception of facialstructure.

According

tothis

hypothesis,

theeffect of motion

information

depends

o.n. the perception ef faccs rather than the prior ex-1Departmcnt

of Psychology,

Chuo

University,742L

1,

Higashinakano,

Hachioji-shi,

Tokyo

192-0393

This research was supported by RISTEX

Japan

Science

Technology Agency, and a

Grant-in-Aid

forscientific research

(15500172)

from

JSPS,

and

Chuo

University

Joint

Research Grant,

Copyright 2005.

perience with a face.Thus, the representation

en-hancement

hypothesis _predicts that the effect of

motion information would be found forboth

famil-iarand unfarniliar

faces.

As a matter of fact,studies with adult participants reveal a

fundarnental

differ-ence intheeffect of rnetlon for

familiar

and

unfamil-iar faces

_{O'Toole

et al.

2002),

Although

a

facili-tativeeffect of motion isconsistently found

for

_the

recognition of familiarfaces,thiseffect

is

lessclear

forthe recognition of unfamiliar

faces,

Aim of

the

_present study was to examine _therole

of motion information on infants'recognition of

un-familiar

faces.Several _previous studies have

sug-gested that rnotien

information

_prometes

_infants'

perception

(Kellman

&

Spelke, 1983;

Otsuka

&

Ya-maguchL 2003>.Therefore we theorizedthatmotion

information should

facilitate

infants'

face

recogni-tioneven

for

thefacesunfamiliar tothe infants,We

compared

infants'

recognition memory for

unfamil-iar

faces

which were learned ina moving or astatic

condition.

Method

Participants

Twenty-fourhea]thy3-te5-month-olds

(mean

age=-'-128.6

days,

ranging

from

81 to 162

The Japanese Psychonomic Society

NII-Electronic Library Service

TheJapanesePsychonomic Society

126 The

_Japanese

_Journal

of Psychonomic Scjence Vol.

(a)MovinglStatic

(b}Static

Figurel.

Examples

of _the familiarization

stimuli

(a),

and the

test

stimuli

(b).

Tablc 1

Prcfercnce

for

nevel face mean _preference

scores

in

percentages, standard

deviation,

and t

scores

_{vs.

chance)

Static

cendition 24,No. 1

Moving

condition

M

63.12 45.00

SD

14.21

11.37

t

3./.?.*.

i,52

EP

<

.O1,

two-tailedtest.

days) _participatedinthisexperiment. An additional

11 infants were excluded from the ana]ysis due to

fussiness

(3>

or side

bias

greaterthan

90

%

(6>,

or

due

to looking times in the familiarizationtrialsthat

were less

than

20s

(2).

Apparatus

All

oi thestimuli were

displayed

on a

21-inch CRT monitor.

The

infant's

viewing distance

was approximate]y 40cm. A CCD camera

just

below

the monitor screen was used to videotape the

behav-ior

of

infant

throughout the experiment.

Stimuli All of the stimuli were produced from

two video recordings, which were taken inthe Uni-versity of

Texas

at

Dallas.

The

familiarization

stim-uli consisted of a smiHng

female

face

seen either

in

a moving or astatic condition. A moving stimulus was composed of 33

frames,

which were shown at a rate

of

25

frames

per second.

A

static stimulus was

com-posed of thelast

frame

of the moving stimulus

(Fig-ure la).The size of the familiarizationstimuli was

220

on cach sidc,

Test

stirnuli consisted of a static

female iacewith aneutral expTession

_CFigure

1b)for

all infants.Hair was excluded so that only the

inter-nal

features

were visible.

The

sizeof the teststimuli

was 16" x lg",

Procedure First,the

infants

were familiarized

with a smiling female face which was

displayed

in

either a moving or static condition en the center of

theCRT rnonitor. The fami]iarizationtrialwas

com-posed of 2 _{presentations,}each of 15 seconds, Aftcr

familiarization,

infants

were testedusing one nove]

and one familiar faces which were shown side by

side. The testtrialswere composed of 2

presenta-tions,each of 10 seconds

duration.

in which the

position of the novel and the familiar

face

was

re-versed, The

familiarization

facedirfered

between

the

infantsso that the novel and farniliarrelationship of

the _teststimuli was reversed,

Infants'

looking _times

for

each stimulus were measured

based

on the video recordings, Only the infanVs looking behavior was visible

in

the

video.

Results

and

Discussion

The totallooking time during 2 familiarizatien

trialsdid not differbetween the moving and static conditions

(t<22)=O.53,

P=.6,

two-tailed).We

calcu-lateda _preference seore for each infant,This was

done by dividing the infanVs looking time

to

the

novel

female

face

during

2 test trials

by

the total

looking time over thesetesttrials,and then

multiply-ing

this

ratio by 100.Two-tailed one-sample t-tests

(vs.

a chance

level

ef 50%) revealed significant

pref-erence in the rnoving condition

_(t(11)=3.2,

_P<,Ol)

but not in the static condition,

(t(ll)=1.52,

P=.16).

These results suggest that infantscould learn

unfa-miliar

face

only

in

themoving condition.

The

present resu]ts suggest thatlearningfrom themoving

condi-tion _prornotes infants' recognition of unfamiliar

faces,

In

centrast

to

the studies with adult

partici-pants,we

found

facilitative

effect of motion

informa-tionfor

the

recognition of unfamiliar facesininfants.

Our

finding

suggests that motien informatjon

con-tributes to the recognition of unfamiliar

faces

in

young infants who are

in

the course of _perceptual

development.

References

Kellman, P,

J,

&

Spelke,E.S.1983 CQgnitive

ogy, 15(4):483-524,

0'Toole,A.

J.

Roark, D,A,

&

AbdL

H.

20e2 Trends in

CQgnitive

Science,

6{6):

261-266.

Otsuka,

Y.

&

Yamaguchi,

M.

K.

2003

fournal

of