1-B-15 3次元物体における運動の予期 : 追従眼球運動の効果(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

TheJtiPaneseJournatofPs)'chonomic Science

2003,Vol,22,No.1,25--26

Prizewinner's

Summary1-B-15

Predictingfuture

position

of

three-dimensional

moving

Effect

of

tracking

eye

movements

objects

TakahiroKAwABE

and

Kayo

K),ushu

UitiversiCy*MIuRA

We

investigated

the effects of tracking eye movements en

_predicting

a

future

_positionof a

three-dimensional moving object. Each stimulus movie was composed of fourobject frames. The

participants

judged

whether the

fourth

object

frame

contained an object which had been changed an equivalent amount totheobject

in

thesecond and thirdobiect

frames,

The

results showed

that

tracking

eye movements are required toaccurately _predictthefuture

position

ofa moving object

although cast shadows are also neccssary forthe _predictionof the three-dimensional rotation. The results therefore

imply

theexistence of a

labeling

mechanism which

is

mediated

by

tracking

eye

movements.

Key

words: prediction ofmotion,

tracking

eye movements, cast shadow

Introduction

This study examined whether or not tracking eye

movements affect the _performance of

_predicting

the

future _position of a moving obiect. Ithas been

suggested thatthe

future

positionof a moving object

is

underestimated

(Finke

&

ShyL

1988},

In

thisstudy

we hypothesized thateye movements would give an

additional cue forthe _predictionbecause tracking

eye rnovements often co-occur with a change of

im-age frame, and consequently the visual system

is

likely

to

associate

the

eye movcments with the

repre-sentation which istobe accumulated.

Method

Stimuli Figure la shows an example of the

stim-uli used

in

Experirnent

1.

A

frame

containing

the

object

(an

object

frame)

was _presented

for

250ms,

and thiswas followed by an interval

for

250ms,

Each stimulus movie was composed of four object

frames

and fourintervals.

In

each of the second and thirdobject

frames

the object was rotated

by

25

degrees

either clockwisc or countercleckwise about itsaxis. Inthefourth object

frarne,

the object was rotated

by

either

19,

21,

23,25,

* _{Department of Psychology, Graduate School ef}

Human-Environmental

Studies,Kyushu sity,

6-19-1

Hakozaki,

Higashi-ku,

Fukuoka

812-8581

27, 29,or 31 degrees.

The

relative rotation angles

were therefore:

-6,

-4,

or

--2

degrees,

indicating

an

undershot of the rotation; O degrees; 2,4,or

6

de-grees,

indicating

an overshoot of therotation. Halfof

the stimulus movies had a cast shadow

but

theother

halfdid not.

Figure

1b

shows an example of

the

stimuli used in

Experiment 2.

The

duration

and presentation order

of theframes were thesame as

in

Experiment 1.The

object was a sphere and itmoved totheright or

left

by

a visual angle of 1,43 degrees in each of the second and thirdebject frames. Inthe fourth object

frame, the object was _presented as moved

by

either

1.00,1.144,1,287,1,43,1,564,1,716,or 1.859degrees

(equivalent

to

either 21, 24, 27, 30, 33, 36, or 39

p[xels).The relative rnovement

distance

was either

-9,

-6,

-3,

_{O,3,6,or 9 pixels,respectively.}

Procedure

The participants indicated with

as-signcd keys their

judgment

of whether the fourth

object

frarne

contained an object which

had

been

rotated an equivalent amount

(Experimcnt

1) or

moved an equivalent amount

(Experiment

2)tothe

object

in

thesecond and thirdobject frames. Inthe

"fixation

condition" the

_participants

were

encour-aged to look at a fixation_point presented at the

center of the display. In the "tracking

condition:',

they were required totracka thin arm indicatedby

The Japanese Psychonomic Society

NII-Electronic Library Service

TheJapanesePsychonomic Society

26

The Japanese

Journal

(a)

-(b)

ensm of Psychonomic S d.5 1O.5 D-,5

-1-1,5

-2-2,5

cience VoL 22,No. 1

Figure

1.

Bar:95% confidefice intervel

oWITHOUTSHADOW aWITHSHAOOW 1 o

-1

-2

-3

4

FOCATIeN TRACwaNG FIxaTION TRACKIrc

VIEWINGCONDITION

(c)

(d)

See

text

fer

details.

Results

and

Discussion

The

response

biases

of theparticipantswere

com-puted

from

the weighted rnean

for

each condition

and are shown

in

Figures

lcand ld. The error bars are the95% confidence intervals,The resu]ts show-ed that

in

the "tracking

cendition" of

Experirnent

1

(Figure

lc} the predictions were accurate for the

objects with cast shadows. In Experiment 2 the

predictions

for

the

objects

both

with, and without,

cast shadows were accurate

in

the

"tracking

condi-tion"

(Figure

ld).

These results show that tracking eye movements

can p]ay a critical role

for

prediction of the

future

position of a moving object. In Experiment 2,the

predictions thatwere made with tracking eye

move-ments were accurate.

We

propose

that

this

is

be-cause the tracking cye rnovements labeledthe repre-sentation of thcmeving objects and thestored repre-sentation became more accurate,

Moreover, we have shown inExperiment 1_thata

cast shadow

is

also required to predict the

future

positionof a three-dimensional rnoving object, The

rr]otion ef

the

three-dirnensional

objects

in

Experi-ment 1

did

not seem to

be

projectedequivalently on

theretina, that

is,

image

compression occurred near

the edge of objects, Therefore we _propose

that

foran

accurate _predictiona three-dimensional cue, such as

cast shadow, was required _toaccurately perceive the

distance

traveled

by

the moving objects.

On

the ether hand, thesphere used inExperiment 2didnot

cause distortionof

the

projected

image.

Therefore

only tracking eye movements were required foran

accurate _prediction.

References

Finke,

R.

A.

&

Shyi,

G.

C,-W,

1988 Mental

tionand representational momentum

for

complex

implied

metions,

fournal

of

Experimental