P-2A-24 錯覚的同期に基づく知覚的群化 : 空間的相互作用と同期性の関係について(2001年度 日本基礎心理学会第20回大会優秀発表賞)

The Japanese Psychonomic Society

NII-Electronic Library Service

TheJapanesePsychonomic Society

1]-.e

fiipanese

.Sou"mt

of

kychonomic

Science

2002,

VoL

ZI,

No.1,

5]

･52

Prizewinner's

SummaryP-2A-24

NTT

Perceptual

_grouping

based

on

illusory

synchrony

:

On

the

relationship

between

synchrony

_perception

and

spatial

interactions

Isamu

MoToyosHi',""

Commztnication

Science

Labomton'es,

NTT

Corporation'

and

ISPS

Research

Fellozv"'

Synchrony

is

an

important

factor

for

binding

visual

inforrnatien

across

space.

It

is

also

one

of

the

basic

attributes

in

_time

_perception,

To

understand

the

mechanism

underlying

synchrony

perception

I

explored

visual

grouping

based

on

the

`synchreny

of orientation'

between

spatially separated

patterns

whose orientations changed

_periodically

in

time.

The

resurts showed

that

the

central

element

tended

to

be

_grouped

not with

the

element

whose

orientation

cha4ged

in

a

physical

synchrony

but

with

the

element

whose

orientation

changed

s]ightly

earlier.

Moreover,

this

tendency

became

more

profound

as

the

spatial

distance

between

elements

increased,

Thisillusory

synchrony was explained

by

a

_proposal

that

synchrony

_perception

is

mediated

by

spatiaHnterac-tions

between

feature

detectors,

which essentially

involve

mutual

delays

of

propagating

signals.

t

t

t

'Key

words :vision, synchrony, spatial

interaction,

_propagation

Synchrony

i's

ap

important

factor

for

binding

visual

'

information

across space.

It

is

also

one

of

the

basic

'

attributes of

_time

_perception.

IIere

I

report

a

perce-ptual

greuping

based

on an

illusory

synchrony of orientations.

This

effect could

lead

a

proposal

that

synchrony

between

visual stimuli

is

mediated

by

spatial-interaction signals which

_propagate

between

'

feature

detectors

with a

finite

velocity,

'

Methods

The

stimulus was a

horizontal

array

ef

three

local

patterns

whose

orientatiens

were alternated

in

time

(Figure

1a),

Each

pattern

was cemposed of

overlap-ping

vertical and

horizontal

Gabor

patches

<3

c/deg)

.

Their

relative

contrasts were modulat ¢

d

out

of

phase

around

the

average of

O.4

in

a

triangular

waveform

(temporal

frequency,

1.5Hz),

The

mean

luminance

'

of

the

di

splay

was

52

cdlm2

The

orientation of

the

left

'

er

right

e]ement

in

the

array

alternated with a

phase

difference

to

the

others

of

A

ip.

The

subjects were

'

_NTT

_{Communication}

_Sci6nce

_{Laboratories,}

_3-1

Morinosato

Wakamiya,

Atsugi,

Kanagawa,

243-O198

asked

to

fixate

on

the

center of

the

array and

to

indicate

which element

(left

or right> changecl

its

orientation

synchronously with

the

central element.

The

sttmuli were

_presented

until

_the

subject

respond-ed.

No

motion was seen

between

the

elements,

The

percentage

ef correct

_grouping

was measured

for

various

A

di

and

for

various

distanees

between

ele-ments

(Inter-Element

Distances,

IEDs),

Results

Figure1b

shows

the

percentage

of correct

_grouping

as a

function

of

the

_phase

difference

ri

di.

The

different

symbels

represent

_the

results

obtained

for

different

IEDs.

The

correct

response

decreases

when

the

absolute

phasie

difference

is

srnall, and

its

range

becomes

wider as

the

IED

increases.

This

indicates

a

decline

in

the

tempora]

acuity with an

increased

spatial

distance

between

the

elements.

A

more

strik-ing

feature

is

that

as

the

IED

increases

the

shape of

the

_{psychornetric}

function

becornes

asyrnrnetrical

with respect

to

a

_phase

difference

of

O

msec,

There

is

'

therefore

a

greater

decrease

of

the

correct response at

'

The Japanese Psychonomic Society

NII-Electronic Library Service

TheJapanesePsychonomic Society

52

The

Japanese

Journal'of

PsychonomicScience

Vol.

21,

No.

1

a

_di

_to

_di

+A

di

_coVn"t",ast

b`co

C

SI

su

E

-4--IM

"

'

--

-..

-

i,,,

i.,.

/

-:

':T-l...

H

.

pa.

1 m

-'

'-`

aoo

-:oo

D loo 2oo

re

Oo

--"

1 2 3

'/'

.s

-

AO(msec)

a-

IED{deg)

Figure1,

(a)

Selected

snapshots

of

the

stimuli used

in

the

experiment, and

the

tirne

course

of

orientation

alternation

(the

contrast

difference

between

vertical

and

horizontal

components),

(b)

The

percentage

of

correct

grouping

as

a

function

of

the

temporal-phase

difference

between

the

target

and

the

others.

Different

symbols

represent

the

results

for

different

IEDs.

(c)

The

difference

between

the

_positive

and

negative

phase-difference

thresholds

that

gave

a

75%

correct

respense,

_plotted

as a

function

of

the

IED.

IED,

the

correct

response

decreases

to

even

less

than

O.5

at

approximately

A

ip

of

-40

msec.

Similar

results were obtained

for

the

two

naive

subjects.

The

asymmetry

indicates

a

tendency

that

the

subjects

'

grouped

not with

the

element whose orientation

chan-ged

in

a

_physical

synchrony

(A

ip

==O.e _msec)

but

_with

the

element

whose

orientation

changed slightly

ear-lier

than

the

central element.

The

amount of

the

as}rmmetry was

analyzed

by

comparing

the

_positive

and

negative

threshold

phase

difference

that

_gave

a correct response of

75%.

Figure

1

c

shows

the

difference

between

the

positive

and negative

phase

thresholds

plotted

against

the

IED.

It

is

clear

that

the

amount

of

the

asymmetry

increase's

_{proportionally}

with

the

IED.

'

/t

Discussion

'

'

Why

did

the

orientatien synchrony-basecl

_grouping

exhibit such a

tempora]

asymmetrY?

one

simple explanatien

is

the

difference

in

latency

to

each

local

element, which can

be

affected

by

many

factors

such

as

the

stimulus

eccentricity and attention

<Hikosaka

et al.,

1993).

Given

that

the

subjects always

fixated

and

paid

attention

to

the

central

element,

it

is

likely

that

neural

signals

for

the

peripheral

elements were

delayed

relative

to

that

for

the

central

element.

This

might

have

produced

the

observed

temporal

asym-metry

_;

a

_physically

advanced signal appeared

simul-taneously.

However,

this

_proposal

does

not

explain

why

the

temporal

asymmetry

and

the

temporal

acuity

both

depended

on

the

spatial

separation

between

elements.

Moreover

this

notion

is

isomorphic

and

by

itself

reveals nothing about

how

the

visual

system

determines

the

synchrony.

'

I

propose

a simpler explanation:

the

synchrony of

visual

_pattern

information

is

mediated

by

spatial

interactions

between

local

feature

detectors.

It

has

been

suggested

that

the

segmentation

and

_grouping

of

patterns

is

partialiy

subserved

by

interactions

between

cortical neurons with receptive

fields

which

are

Iocated

at

different

positions

(Gilbert

&

Wiesel,

1985).

Note

that

such

interactions

inevitably

involve

mutual

delays

of signals

between

neural

ttnits

due

to

the

finite

velocity

of

signal

propagation.

These

delays

indeed

_produce

asymmetric neural

responses,

depending

on whether one element

has

an

advanced

or

lagged

temporal

_phase.

Moreover

_given

the

delays

are

_proportional

to

the

distance

between

neural units,

it

is

natural

that

the

amount

of

the

temporal

asyrn-metry

increases

proportionally

with

the

distance

between

elements.

The

decrine

in

temporal

acuity

with

the

IED

can

be

also explained only

by

assuming a random

fluctuation

of

the

_propagation

velocity,

which

is

plausible

in

a

biological

system.

References

'

Gilbert,

C,D.,

&

Wiesel,

T.N.

1985

Intrinsic

nectivity and receptive

field

properties

in

visual

cortex.

VZsioge

Research,

25,

365'374

Hikosaka,

O.,

Miyauchi,

S,,

&

Shimojo,

S.

1993

Focal

visual attention

_produces

illusory

temporal

order

and motion sensation.

Vision

Resea7ch,

33,

1219T

1240