2P107 2種の運動刺激間に見られる知覚的ずれのISIの導入による反転(2003年度 日本基礎心理学会第22回大会優秀発表賞)

The Japanese Psychonomic Society

NII-Electronic Library Service

The JapanesePsychonomic Society

TheJinPanesefournalof RsychonomicScience

2004,VoL23,No.1,1IS-114

Summary

ofAwarded

Presentation2PI07

A

reversal

of

illusory

offset

between

two

kinds

of

motion

stimuli

by

a

longer

ISIi>

Kazushi

MARuyA

and

Takao

SATo

Lbeiversity

of

71)hyo*

Motion-defined motion

<MDM)

isthe rnotlon of patterns defincd by localmovement direction,

Motion- and luminance-defined motions

(LDM)

are _perceived as misaligned whcn

thcy

arc

prc-scnted _physicallywith the same speed and in_phase

(the

_perceptualoffset phenomenon,

POP).

In

thisstudy, we examined theeifects of ISIs

(inter

stimulus intervals)thatwere introduced between

pattern motion frames of thc POP. When an ISIwas introduced toMDM, the estimated offset

decreased rapidly as

the

ISI

increased

from

O

to

80

ms.

Furthermore,

when

the

ISI

was

80

ms most

subjects reported offscts in_the reverse

direction,

These

result$

indicate

thatsimple differencesin

the transduction

latency

for

MDM and LDM cannot explain the POP.

Key words: vision, motion _perception.second-order rnotion, _perceptual c)ffset

Several

studies

have

reported _perceptualerrers

in

the _position of moving objects

(e.g.

the flashlag

¢

ffect,

Nijhawan,

I994).

We

have

reported a

percep-tual offset phenomenon

(POP)

as one such illusion

{Maruya

& Sato,2003).In thisillusion,_gratings

de-fined

by

different

attributes were _perceived

to

be

misaligned

when they were

_presented

_physically

with the sarne spced and in _phase

(Figure

la).By

using the same methods as our _previous study the

presentstudy examined theeffect of

introducing

an

ISI

on thjs

illusion

toexplore theorigin of the phe-nomenon, An ISIhas often been used as a _procedure

to

disturb

early motion systems which are

based

on motion energy

_(Adelson

& Bergen, 1985;Georgeson

& Harris, 1990; Braddick, 1974). By utilizing this

procedure, we investigated

how

the

POP

varied

when a critical change of early rnotion _processing

occurred,

Method

In

thisstudy, we used a

luminance-defined

motion

(LDM)

and a motion-defined motion

{MDM,

cf, Zan-ker,1993)as two typesof motjon stjmuli,

The

stimuli

* _{Department of Psychology, Faculty of Letters,}

University of Tokyo, 7-3-1 Hongo, Bunkyo-ku,

Tokyo, 113-Oe33

")

_For

_TS

_{the study was supported}

_by

_HFSP

_and

grants-in-aidfor scicntific research

from

the

Ministry

of

Education,

Science,

Sports

and

ture,

KM

was supported by

JSPS,

were two types of square wave _pattcrns.

Each

of

Lhem consisted of 1024 random

dots

in

a stimulus

field

ef 120X300 pixels, which corresponds toa vjs-ual angle of 4× 10 deg. Squarc-wave _patterns were

generated by modulating the movement directionof

dots

(up

or

down;

motion-defined motion) or their

luminance

_{(luminance-defined}

rnotion), The spatial

frequency of

the

modulation was

O.2

c/d for

both

the

MDM

and LDM. For LDM stimuli the dots were static. For MDM stimuli the dots wcrc shifted bv 8

arc min every 20ms. The MDM and LDM stimuli

were _generated by shifting thesesquare waves

pat-tern every

120

ms

by

a

30

deg

_phase angle.

The

MDM and LDM patterns were moved at the same

speed, and _phase misalignments between the two

stimuli were measured

by

using themethod of

con-stant stimulL The _{physicalphase} offset was varied in

eight steps

between

12 and

-l2

deg

_phase angle.

The MDM stimuli were presented 2.5

dcg

above a

fixation

point and the LDM stimuli were

_presented

2,5

deg

below.Inconditions where an ISIwas

intro-duced either the luminance- or motion-defined

pat-terns

disappeared

for

a given time.

This

was

achiev-ed

by

making the

luminance

of thedotsuniform

Cfor

LDM) or by rnaking the localmotion clirection

ran-dom

_(for

MDM). An ISIwas introduced

for

either the

LDM or MDM stimulL and varied infivesteps from O to,80ms.

The Japanese Psychonomic Society

NII-Electronic Library Service

The JapanesePsychonomic Society '

114 The Japane$e

Journal

of Psych

'onom]cScience

VoL

23,No. 1

a

b

-;:. mm///lt

I

Phy$icalconfiguration

r.,.aog.s

g･2g,g,eg

.2-ag･f,

,<

±

_x(iii

i

,ajagi,,,

･

tt･ag/tt

-'t"aj/:tt'ttttsuSt

smm

. : : :

.

:

.

muttt.... :

lPerception

×

_"/"

XT

MDM

LDM :Z::::T;X:;ILMDM+LDMI e 2o 4o 6e ISI

(ms)

Figurel.

_<a)

A _phenomenon of

effset.

(b)

Amounts of _perceptual

function

of the ISL

eo perceptual

offset as a

offset direction

by

using the 2-AFC method. The

experimcnt was conducted

in

sessions with a fixed

ISL

Each block had 160 trials,and within

b]ocks,

eight offset conditions werc _presented 20 times ina

randomizcd order. The subjeets performed 2 blocks

for

every shift condition.

Results

and

discussion

We

calculated

the

rate of the responses

for

LDM

being ahead of

MDM

as a function of physical offset

in each condition.

After

thaL the null _peint forthe

apparent offset was calculated

fer

these

data

with a

probitanalysi$,

The

mcans of the_perceptualoffset of

fivesubiects are plotted as a

function

ofISIinFigure

lb.When an ISIwas introduced

for

the

LDM

stimuli

theoffset was virtually constant. Inconditions where

an ISIwas introduced

for

MDM

stimulL the offset

rapidly decreased as the ISIincreased

from

O

to 60

ms.

When

the

ISI

was 80ms, al] of the subjects

perceived an

Musory

offset

in

the reverse direction

to

that

perceived intheother conditions. In normal

conditions when the ISI was Oms, the

LDM

was

perceived to

be

before

the

MDM.

However, the LDM

was _perceived tebeafter the

MDM

when the ISIwas

80

ms.

This decrea$e and revcrsal of _perceptual offset

in-dicates

that a simp]e difference

in

transduction

la-tency

between

MDM

and LDM cannot explain all

aspects of the

POP,

Frame onsets of _pattern motion occurred at thesame timc regardless of

the

lengthof

the

ISI,

and the differenceinthe latency shou!d be

constant

for

the

ISI

conditions. Ifalatcncy differencc

was a cause of the

POR

the

ISI

should not

have

infiuenced

the POP. The current results are obvi-ously

inconsistent

with

this

_prediction.

One exp]anation

for

thc reduction and reversal ef

the

POP

isthat a critica] change rnay occur to a

position processing of moving objects through

intro-duction of an ISL

ln

our pilotstudy, MDM was

per-ceived

to

be

fasterthan

LDM

for

thesame conditions

we used

in

the _present investigation,

According

to

this,ifthe_perceived speed differencewas integrated

in

some temporal window, the

MDM

should be beforc

the

LDM.

The

currcnt results were consistent with

this _prediction.

We

reported

that

such a perceived

speed differencedidnot explain the

POP,

However, it

is

possible thatacriticalchange occurred totheflow

of information inposjtionprocessing by the

intro-duction of an ISI.The ISIcou]d then cause anabsence

of

tundamental

frequency-based

information

and

in-fiuencehigher-order

information.

References

Adelson,

E,

H.

&

Bergen,

J,

1985.

Spatiotemporal

ergy models

for

the _perception ef motion.

"mrnal

of

the

_QPtical

Society

ofAmerica.

A,

2,284 299,

Braddick,

O.

I974,A short-range process

in

apparent

motion.

Vision

Research,

14,519-527,

Gcorgeson,

M.

A,,

&

Harris,M.G.1990.The temporal range of motion sensing and motion perception,

VisionResearch,

30,

615-619,

Maruya,

K.,& Sato,T.2003.

Perccptual

offset

tween first-and sccond-order motion

rnal

of

Vision,

3,

8e3.

Nijhawan, R.1994.Motion extrapolation ineatching, NLiture,370,256-257,

Zanker,

j.

M 1993.Thetu rnotion:

A

_paradoxical

mulus toexplore higher order motion extraction.

Vision

Research,33,533-569.