Disruption of visual search by the misbinding illusion(Summary of Awarded Presentation at the 31st Annual Meeting)

(1)

The Japanese Psychonomic Society

NII-Electronic Library Service

TheJapanesePsychonomic Society

TheJapaneseJburnatofPsychonomicScience

2D13,

VeL32,No. 1,143-144

Summary

of

Awarded

Presentation

DisruP

Megumi

SuzuKi

tion

ofvisual

search

by

the

misbinding

illusion

a*,

Jeremy

WoLFEb,

Tbdd

S. HoRowiTzb,

and

Yasuki

NoGucHi

"Kbbe

university

bHarvard

Medical

School

In

cluttered environments, attentive searching

is

necessary

to

find

targets

defined

by

a coajunction of multiple

features

(e.g.,

color and

shape),

Thus

it

has

classically

been

assumed

that

the

allocation of attention causes

binding,

In

contrast,

here

we show a

_phenomenon

where

the

causality

between

attention and

binding

is

reyersed;

the

visual

precessing

of

feature

binding

affects spatial shifts of attention.

Subjects

searched

for

targets

with a

combination

of

particular

colors and orientations

(e,g,,

red, vertieal

bars)that

were embedded

in

_peripheral

areas

of

a

search array.

When

acentral area consisted of

bars

with regular color-orientation

_pairings,

these

regular

_pairings

automatically

changed

perceptual

bindings

in

the

_peripheral

area,re$ulting

in

target

omissien.

These

results

indicate

that

the

visu-alsystem

_performs

a rapid

extrapolation

of

binding

information

spreading

frorn

center

to

_peripheral

areas, which

throws

offattention.

Our

results

thus

_provided

evidence supporting

the

hypothesis

that

binding

affbcts attention.

Key

words:

_preattentive

_process,

feature

binding,

misbinding

illusion,

visual

search

Mounting

evidence

indicates

that

visual

features

such as

shape and celQr are

_processed

in

separate regions

in

the

brain.

Hew

our

brain

binds

the

features

associated with

the

same

eb-ject

is

an

important

_problem.

A

dominant

theory

(Treisman

&

Gelade,

1980)

argues that

the

integration

of

features

requires attention,

IYpical

examples

are

findings

in

visual search

para-digms,

When

targets

can

be

distinguished

frorn

distracters

by

a

simple

feature,

these

targets

are readily apparent,

When

the

targets

are

defined

by

aceajunction of rnultiple

features,

how-ever, an attentive search

is

necessary

to

find

them.

This

indi-cates

that

attention

deployment

results

in

the

integration

of

different

features

of an object,

In

contrast,

there

are

also

_{psychophysical}

data

suggesting

an early mechanism

for

feature

binding

that

does

not require

at-tention

(e.g.,

the

McCollough

effect),

It

remains unclear

how

this

early

binding

mechanism

is

the

visual search

findings

above.

In

this

study we addressed

this

issue

using

the

misbinding

illusion

(Wli,

Kanai,

&

Shimojo,

2004),

a

phenom-enon representing

the

early

binding

mechanism,

In

this

illu-sion, central and

_peripheral

_portions

of a

transparent

motion

field

combine color and rnotion

in

opposite ways.

In

observers

of such stimuli,

perceptual

color-motion

bindings

in

the

periphery

change to match

the

bindings

in

the

center,

resu!t-ing

in

erroneous celer-metion

pairings

in

the

periphery

The

*

_{Corresponding}

authon

Department

of

PsychologF

Grad-uate

School

of

Humanities,

Kobe

University)

1-1

Rokke-dai-cho,

Nada,

Kobe

657-8501,

Iapan.

E-rnail:g-bt-bs-88@vahoo,co.jp

Cepy

current

study

first

shews

that

this

illusion

also

takes

_place

in

color-orientatien

bindings

(Exp.

1).

The

illusion

is

then

applied

to

avisual search

_paradigm

(Exp.

2).

Experiment

1 Methods.

Eleven

and nine subjects

participated

in

Exper-iments

1a

and

lb,

respectively

In

each

trial,

we

_presented

sub-jects

with an array of

864

stimuli

(27

deg

×

18 deg;

cell size of

the array:

O.75 deg

×

O,75

deg)

in

which

four

types

of

bars

(red-horizontal,

red-vertical,

_{green-horizontal,}

and

green-ver-tical)

were

intermixed.

Each

array consisted of

three

fields:

a

central

field

(13.5

deg

×

18 deg,

432 items)

and

two

_peripheral

fields

(Ieft

and right surreunds,

6.75 deg

×

18 deg,

216 items

for

each),

In

the

left

and right surrounds,

the

array was

com-posed

of equal numbers of red

(R)

and

_green

(G)

bars

(108

for

each).

Half

of

the

_green

bars

were

horizontal

(H)

and other

halfwere

vertical

(V).

Fer

the

_peripheral

red

bars,

the

ratio of

H

to

V

changed across

trials

from

O

:

100 to

100

:

O. The

cen-figuration

of

the

central

field

was chosen

frem

four

condi-tions,

as

fo11ows,

In

the

"No

Stimulus"

_(or

NS)

condition,

there

was no

bar

in

the

central

field

(peripheral

bars

only),

In

the

"Random"

condition, equal numbers of

the

four

types of

bars

(R-H,

R-Y

G-H,

and

G-V)

were

intermixed

in

the

central

field.

These

two

conditions served as controls.

In

the

"Dou-ble"

conditions, allred

bars

were of one orientation and all

green

bars

were of

the

other orientation.

Thus

the

central

field

in

the

Double-1

condition consisted of

R-H

and

G-V

bars,

while

that

in

the

Deuble-2

condition consisted of

R-V

and

(2)

The Japanese Psychonomic Society

NII-Electronic Library Service

TheJapanesePsychonomic Society

144 The

Japanese

Journal

of

Psychonomic

Science

Vbl.

32,

No.

1 G-Hbars,

Subjects

estimated

the

ratio of

R-H

te

R-V

bars

in

the

pe-ripheral

fields.

They

_pressed

one

key

to

_indicate

that

the

num-ber

of

_peripheral

R-H

bars

was

_greater

than

the

number of

R-V

_bais,

and

_pressed

another

to

indicate

the

reverse.

They

were asked

to

ignore

stimuli

in

the

central

field

and

to

attend

te

the

surround

(left

or right, assigned randomly

for

each

sub-ject).

The

actual

_percentage

of

_peripheral

R-H

bars

(%R-H)

was varied randomly across

trials

from

O

to

100%,

and we measured

the

_probability

of

subjects reporting

R-H

bars

as

greater

in

number

(%

`H") as a

function

of

%R-H.

Misbind-ing

illusiens

should

be

observed as changes

in

the

_point

ef subjective equa!ity

(PSE)

between

R-H

and

R-V

Thus,

Dou-ble-1

sheuld

bias

the

apparent orientation of

_peripheral

red

bar$

toward

horizontal,

while

Double-2

should

bias

the

appar-entorientationofperipheralredbarstowardvertical.

Every

trial

began

with

fixation

for

1

s,

fo11ewed

by

a

200

ms

presentation

of

the

array

After

400

ms, subjects were

prompt-ed

for

a

key

response.

Experiment

1a

involved

the

two

control

conditions

_(NS

and

Random),

Subjects

_performed

four

ses-sions of

90 trials

ea[h.

In

Experiment

lb,

arl

four

cenditions were randomly

intermixed.

Each

_participant

completed six sessions of

90 trials

each,

Results.

In

Experiment

la,

no significant

difference

in

PSE

was ebserved

between

NS

and

Randorn

(p=.64),

which

indicated

that

the

bars

in

the

central

fieLd

themselves

did

not

induce

any

_perceptual

biases.

In

Experirnent

lb,

PSEs

in

the

two control conditions

(NS

and

Random)

again

did

not

differ

{p=.14).

In

contrast, we obseryed a significant

decrease

ofthe

PSE

in

Double-1

cempared

to

Random

(p=.O1),

anda

signifi-cant

increase

in

Double-2

compared

to

Random

(p=.03).

These

results

are similar

to

those

observed

in

color-motion misbinding

(Wu,

Kanai,

&

Shimojo,

_2004),

_indicating

that

the

misbinding

illusion

can also

take

_place

between

color and

eri-entationinformation.

Experiment2

Methods.

In

Experiment

2,

we applied

the

color-erienta-tion

misbinding

illusion

to

a visual search

_paradigrn.

Eleven

subjects

_participated

_in

_Experiment

2,

As

in

Experiment

1,

the

array consisted of

four

_types

of

bars

and was subdivided

into

three

fields,

Each

field

had

an equal number of red and

_green

bars,

and

the

H

:

V

ratio ofthe

_peripheral

_green

bars

was

fixed

at

50

/

50. The

targets

of

visual

search were

either

R-V

or

R-H

bars

_(determined

randemly

fbr

each subject).

These

target

bars

were

_placed

in

either

the

left

or right surround.

In

the

target

field,

a number of

target

bars

could

be

2,

12,

22,

38,

54,

or

70,

which corresponded

to

respective

target

per-centages(%

Thrgets)

of

2,

11,

20,

35,

50,

or

6S%ofthe

108 to-tal

red

bars

in

a

target

field.

There

were

no

targets

in

the

cen-tralateral

field,

Subjects

reported whether

the

left

or right surround contained

the

target

bars,

ignertng

the

stimuli

in

the

central

field,

We

compared

task

accuracy

for

the

two

types

of central

fields

(Double

and

NS).

In

both

conditions,

target

con-junctions

were absent

from

the

central

field.

Each

participant

performed

six session$

of

96 trials

each.

Fer

each

type

of central

field,

changes

in

task

accuracy as a

function

of

the

%Tlargets

were

fitted

by

a sigmoid

function.

We

compared

the

75%threshold

of

the

function

between

Double

_(misbinding)

and

NS

(no

misbinding) conditions.

Results.

We

observedasignificant

increase

in

the

75%

threshold

in

Double

compared

to

NS

(p=.O07).

These

results were censistent with

the

view

that

the

central

displa7

in

the

Double

trials

induced

illusory

bindings

in

the

_peripheral

tar-gets,

which

drew

attention away

from

subjects and resulted

in

the

failure

to

detect

targets.

Discussion

In

contrast

to

a

_previeus

view

that

an allocation ofattention causes an

integration

of

features

at an attended

_position,

we

tested

the

hypothesis

that

the

visual

_precessing

of

feature

binding

affects

the

allocation of attentien

during

visual search.

SpecificallM

we used

the

misbinding

illusion

as an example of early

feature

binding

(ipdependent

ofattention) and examined

how

the

illusion

interacted

with attention

in

visual search.

Results

revealed

that

the

misbinding

induced

by

the

central

stimuli

changed

_perceptual

color-orientation

binding

in

the

peripheral

stimuli

(Exp.

1),

reducing

the

saliency

of

targets

in

the

_peripheral

field

_(Exp.

2),

These

results

_provide

evidence

supporting

the

hypothesis

that

binding

affects attention.

References

Ireisman,

A,

M.,

&

Gelade,

G,

_(1980).

A

feature-integration

theory

ofattention.