Spatiotemporal integration of object features in the stream/bounce event perception(Summary of Awarded Presentation at the 25th Annual Meeting)

The Japanese Psychonomic Society

NII-Electronic Library Service

The JapanesePsychonomic Society

TheJaPaneseJournalof Ps.1/chonomicScience

2007,Vo],25,No.2,273-274

Summary

ofAwarded

PresentationIP08

Spatiotemporal

integration

of

object

features

in

the

stream/bounce

event

_perception

Yousuke

KAwAcHi*,

Takahiro

KAwABE**,

and

Jiro

GyoBA*

DqPartment

_of

Ils),cholagy,7bhoku UitiversitJ,*,LlserScienceinstitute,

_ts'ushu

University** When two identicalvisual objects move across each other ina two-dimensional display,there

isa _perception of streaming through, or bouncing off each other, at the positionof their

coinci-dence. Inour _prcliminary observation, a switch of object

features

such as color

between

the two

objeets, biased the_perception towards

bouncing.

We

have

examined the characteristics oi

spati-oternporal

integration

of object

features

inthestreamfbounce evcnt, By changing theduration of

the object _presentation after a color switch at theircoincidence we explored the spatioternporal

propertiesofobiect color tosubstantially affect the stream or

bounce

event.

The

results revealed

that

100

ms after the color switch, the observers can utilize

the

color change

lnformation

to

determine the streamfbounce event, Additional]y,the percentage of

bouncing

percepts reached a

rnaximurn rate 150 ms after theco]or switch. A follow-upexperiment showed thattheresults were

not affectcd by thcspccd of theobjects. This result suggested thatthe _{post-coincidence}duration iscrucial forspatiotemporal integrationof theobject color

feature

in

event perception.

Key words: spatiotemporal integration,feature,stream/bounce event _perception

In the _perception of stream/bounce events,

changes

in

the

features

of moving objects after their

coincidence have been reported toinfiuencethe

per-cept of the streamlbounce display

<e.g.

Feldman

&

Tremoulet, 2006).However, itisunclear hew long it

takes for the object featurechange to affect visual event perception. We manipu]ated thepresentation

duration of a stimulus

(Watanabe

& Shimojo, 2001)

afterswitching objects' colors at theceincidence.

We

attempted to specify the spatiotemporal condiition

for the visual system to utilize the object feature

information effectively invisual event _perception.

Experiment

la

and

b

Methods

Seven Tohoku university students including the

author

(YK)

participated

in

Experiment

la, In

Ex-periment lb,seven new naive students _{participated.}

A white fixationcross was _presented at thecenter of a CRT display. The display background wa$

black. The two moving objects, subtending O.5

de-grees,were colored red and green and thecolors were

subjectively isoluminant toeach other, Thc objects

Department of Psychology, Graduate School of

Arts and Letters,Tohoku University,27-1 Kawa-uchi, Aoba-ku, Sendai 980-8576

appeared separately, 4.2degrees frorneach other, and

O.9

degree$

above thewhite

fixation

cross, They

moved toward each other with a vclocity of 5,7

deg s-i, coincided and then continued as

they

moved

away from each other. After the coincidence the

colored objects either moved in_thc same

direction

(no

switch condition) or in the opposite dire¢tion

(switch

condition}, InExperiment 1bthe two objects

were

isochronous,

so that the object coLor was

con-stant

before

and after thecoincidence

_{control

condi-tion).The duration was varied after the coincidence

(Post

coincidence duration

(PCD)

of O,53.2,106.4,

159.6.212,8,319.2.or 372.4ms). An occluder was

presented at

O,9

degrees

above the

fixation

cross,

in

order to make the

instant

of thecolor switch

invisi-bleforthe _{participants.}

The participantswere asked to

judge

whether the

objects streamed through or bounced offeach other

at theceincidenee. If_partieipantswere unsure about

their _perception they could _pressthe "unsure" _key.

In Experiment la there were two color switch

(no

switch or switch) conditions and seven PCD

copdi-tions,InExperiment lb therewere seven PCD

condi-tion$only. There were 20 trial6foreach condition.

Results and Discussion

The resu]ts are shown inthe upper _partof

Figure

The Japanese Psychonomic Society

NII-Electronic Library Service

The JapanesePsychonomicSociety

274 The

_Japanese

_Journai

of Psychonomic Science VoL25, No. 2

100fi8 9ov 80Fa 7o-v 60ctwa soe 4ozV 30z= 202 io o 90gv 80'tu 70-Y 60: 50e 4oZa 3oz= 202 ie o O tOO 200 ₃₀₀ PCD

(ms)

400

o loe 2oo 3oo 4oo

PCD

(ms)

Figure 1. The mean _percentage of bouncing percept

in

Experiments

la,

b

and 2. The

error

bars

denote the mcan standard errors

of mean

_<N=7).

1. Because more than ten _percent oi the responses wcre 'unsure'

in

the

Oms

PCD

condition

in

both

Experiments la and b,we excludcd thiscondition

from

the statist.ical analysis. A two-way ANOVA revealed that the main effects of both PCD and color

switch were significant

(F(5,30)=2.62,

p<O.05;

F(1,6)=41.76,

p<O.OOI,

respectively).

There

was

also a significant interaction between thc factors

(F(5,

30)=-,8.94,P<O.OOI). A _post-hoc analysis indi-cated thatinthePCD condition$ from

100

ms to

350

ms, the _percentage of bounce _percepts was

sig-nificanUy higher inthc switch condition than

in

the no switch condition

(all

_P<.O.05).

For

the switch condition, the percentage of bounce _percepts was

higherjnPCD 150 ms to

350

ms _than PCD

50

ms

{all

P<O.05),

For Experiment

lb

a one-w･ay

ANOVA

revealed _{that thc} main effect of

PCD

was not

sig-nificant

(F(5,

30)=O.99,p=O.442).

From theresults of

Experiments laand b,

it

was apparent thatacritical

duration of approximately

150ms

is

required

for

observers to utilize object color information

effec-tively for_perceivingthe streamfbounce event.

Experiment

2

Method

Seven

Tohoku university students participatcd

in

thisexperiment.

Except forthefollowing,thesti'muli were identical

to

that

of

Experiment

1.

In

Experiment

2

there

were

threeconditions

(no

switch condition, switch

condi-tion,and contrel condition) in a within-subject

de-sign, Inaddition, _the spced of theobject motion was

doubled. The motion

distance

was also doub]ed,

Results

and Discussion

The results are shown inthe lower _partof Figure 1.

As in

Experirnent

1,

regardless of the speed of the

object motion, when the

PCD

was above 100ms the

differencebetween theswitch and control condition, and between the no switch and control condition,

reached significant levels

(P<O,05).

Moreover,

for

the switch condition thepercentage of

bouncing

per-cept was higherinthePCD [rom 150 to 350 rns than

in the PCD 50ms

_{a]l

_p<O,05).

Meanwhile, inthe

contrel condition, the _percentage of

bouncing

per-cepts was

higher

in

the PCD 200 to350 ms than in

the

PCD

50ms

(all

_P<

O.05).

General

Discussion

Thc results revealed that

150

ms was required

for

observers to consistently

judge

a perception of

streaming or bouncing

(Experiment

1). Moreover,

the temporal characterTstTc of the effect of color

switch was invariant even when the motion speed

was doubled

(Experiment

2).

A

critical duration

thereforeseems to

be

necessary

for

the

spatiotempo-ral integrationof object featurestoform the repre-$entation of thestreamlbounce event.

Referenees

Feldtnan,

J.

&

Tremoulet

2e06 Individuation of

vis-ual objects over tirne.

CQgnition,

99,135-161,

Watanabe,

K, & Shimojo, S,2001 Postcoincidence trajectoryduration affects motion cvcnt

percep-tion.PercePtion

&

Psychophysics,63,16-28.